framework,version,device,op_name,kernel_source,gemm_dtype,m,n,k,latency
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,65536,16384,6.364296468098958
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,65536,12288,9.215521240234375
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,65536,16384,11.52805684407552
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,65536,12288,4.453387959798177
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,65536,10240,4.127893320719401
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,65536,10240,8.042812601725261
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,65536,8192,6.145487976074219
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,65536,8192,2.660447947184245
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,65536,8192,1.9477973937988282
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,65536,7168,5.031881713867188
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,65536,7168,2.385689544677734
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,65536,10240,2.4512255350748697
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,65536,6144,4.368965148925781
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,65536,6144,2.278307088216146
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,65536,12288,3.0193471272786456
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,65536,5120,3.4225824991861975
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,65536,5120,1.6605738321940104
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,65536,6144,1.3965802510579428
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,65536,4096,2.6403284708658856
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,65536,4096,1.4575584411621094
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,65536,7168,1.5909088134765625
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,65536,3584,2.1838058471679687
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,65536,3584,1.303217061360677
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,65536,5120,1.1459349314371745
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,65536,3072,2.0740853627522786
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,65536,3072,1.168768056233724
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,65536,16384,3.9546686808268228
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,65536,2560,1.4099732716878255
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,65536,2560,1.0718196868896483
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,65536,3584,0.8373834609985351
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,65536,2048,1.229210662841797
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,65536,2048,0.7242687861124675
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,65536,3072,0.7402751922607422
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,65536,1536,0.9088693618774414
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,65536,1536,0.6402293523152669
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,65536,4096,0.9248490651448569
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,65536,1024,0.6169653574625651
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,65536,1024,0.4165514628092448
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,65536,2560,0.6379733403523763
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,65536,768,0.47673600514729814
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,65536,768,0.3213856061299642
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,65536,1024,0.3512778600056966
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,65536,512,0.3612640062967936
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,65536,512,0.2539466698964437
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,65536,768,0.2996629397074381
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,65536,256,0.23111999829610191
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,65536,256,0.20681920051574706
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,65536,256,0.2080608050028483
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,65536,128,0.1937130610148112
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,65536,128,0.18331093788146974
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,65536,2048,0.5440693537394206
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,65536,64,0.16623892784118652
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,65536,64,0.18472213745117189
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,65536,32,0.16997440656026203
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,65536,32,0.18552746772766113
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,65536,128,0.18400106430053711
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,65536,1536,0.4568309466044108
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,65536,512,0.26416640281677245
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,16384,16384,2.7646283467610675
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,16384,65536,6.867780049641927
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,16384,16384,1.7168853759765625
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,16384,65536,11.778009033203125
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,16384,12288,2.2327435811360674
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,16384,12288,1.3295360565185548
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,16384,10240,1.6989312489827473
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,16384,10240,1.3783285776774088
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,16384,10240,0.6727711995442708
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,16384,8192,1.2003392537434896
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,16384,8192,1.0463210423787435
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,16384,12288,0.7899274826049805
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,16384,7168,1.099127451578776
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,16384,7168,0.7990901311238606
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,16384,16384,1.0488661448160808
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,16384,6144,0.9752768198649088
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,16384,6144,0.6659264246622721
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,16384,8192,0.5442933400472005
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,16384,5120,0.7442464192708333
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,16384,5120,0.5326208114624024
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,16384,7168,0.44638932545979815
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,16384,4096,0.5730197270711263
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,16384,4096,0.4496373176574707
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,16384,5120,0.3397311846415202
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,16384,3584,0.5196629206339518
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,16384,3584,0.3353930791219076
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,16384,6144,0.3814880053202311
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,16384,3072,0.436293347676595
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,16384,3072,0.2967285474141439
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,16384,4096,0.2720106760660807
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,16384,2560,0.38356266021728513
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,16384,2560,0.23159359296162924
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,16384,2560,0.18692800203959148
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,16384,2048,0.2995423952738444
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,16384,2048,0.20751892725626625
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,16384,3584,0.24938559532165527
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,16384,1536,0.23918293317159017
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,16384,1536,0.15411200523376464
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,16384,3072,0.2115818659464518
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,16384,1024,0.15920000076293944
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,16384,1024,0.11868373552958171
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,16384,1024,0.09996480147043864
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,16384,768,0.12448320388793946
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,16384,768,0.1029311974843343
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,16384,1536,0.13077332973480224
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,16384,512,0.09388480186462403
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,16384,512,0.0848586638768514
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,16384,768,0.08570240338643392
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,16384,256,0.060088535149892174
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,16384,256,0.07116053104400635
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,16384,512,0.07653120358784994
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,16384,128,0.05322986841201782
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,16384,128,0.06430720090866089
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,16384,128,0.057684266567230226
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,16384,64,0.047866666316986085
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,16384,256,0.06406186819076538
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,16384,64,0.06309546629587809
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,16384,32,0.045627733071645096
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,16384,32,0.06259306669235229
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,16384,2048,0.1565834681193034
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,12288,65536,5.406477864583334
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,12288,16384,2.181169128417969
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,12288,65536,8.6654052734375
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,12288,16384,1.4225716908772787
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,12288,12288,1.4328821818033854
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,12288,12288,1.0604000091552734
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,12288,12288,0.5988778432210287
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,12288,10240,1.2644927978515625
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,12288,10240,0.9666175842285156
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,12288,16384,0.7942997614542644
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,12288,8192,1.0321088155110678
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,12288,8192,0.7546229044596354
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,12288,10240,0.49503361384073896
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,12288,7168,0.8066218694051107
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,12288,7168,0.6200181325276692
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,12288,8192,0.3814026514689127
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,12288,6144,0.6662623723347981
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,12288,6144,0.6133567810058593
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,12288,7168,0.35346240997314454
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,12288,5120,0.5755125045776367
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,12288,5120,0.4056394577026367
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,12288,6144,0.2945941289265951
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,12288,4096,0.447218132019043
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,12288,4096,0.30971412658691405
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,12288,5120,0.2680490811665853
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,12288,3584,0.41141440073649094
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,12288,3584,0.27214186986287436
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,12288,4096,0.2135541280110677
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,12288,3072,0.3319775899251302
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,12288,3072,0.23954453468322753
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,12288,3584,0.1961898644765218
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,12288,2560,0.2847370783487956
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,12288,2560,0.21682346661885582
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,12288,3072,0.16918932596842448
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,12288,2048,0.23062613805135093
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,12288,2048,0.1584895928700765
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,12288,2560,0.1500127951304118
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,12288,1536,0.1743765354156494
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,12288,1536,0.12751893202463788
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,12288,1536,0.10426560242970784
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,12288,1024,0.12473920186360676
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,12288,1024,0.09906880060831705
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,12288,2048,0.12437866528828938
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,12288,768,0.0942624012629191
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,12288,768,0.08490880330403647
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,12288,1024,0.07985280354817709
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,12288,512,0.07141653696695963
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,12288,512,0.06981013615926107
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,12288,768,0.06786453723907471
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,12288,256,0.04615893363952637
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,12288,256,0.06241706609725952
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,12288,512,0.06125333309173584
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,12288,128,0.038157868385314944
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,12288,256,0.0521717349688212
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,12288,128,0.053529600302378334
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,12288,64,0.03537919918696086
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,12288,64,0.05412906805674235
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,12288,32,0.03821440140406291
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,12288,32,0.053532799084981285
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,12288,128,0.04686506589253743
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,10240,65536,4.578451029459635
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,10240,65536,7.394614156087239
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,10240,16384,1.7135306040445966
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,10240,16384,1.2095882415771484
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,10240,16384,0.6465002695719402
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,10240,12288,1.2352383931477866
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,10240,12288,0.9432320276896158
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,16384,65536,5.737407938639323
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,10240,10240,1.0782602945963542
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,10240,10240,0.8711456298828125
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,10240,12288,0.4844789187113444
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,10240,8192,0.7846015930175781
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,10240,8192,0.6230986913045247
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,10240,10240,0.40250666936238605
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,10240,7168,0.6807658513387044
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,10240,7168,0.5524821599324544
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,10240,8192,0.33293654123942057
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,10240,6144,0.573855972290039
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,10240,6144,0.4277183850606282
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,10240,7168,0.2997941335042318
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,10240,5120,0.4926602681477864
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,12288,65536,3.858598327636719
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,10240,5120,0.42735573450724285
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,10240,4096,0.3864213307698568
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,10240,4096,0.27682558695475257
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,10240,4096,0.18366187413533527
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,10240,3584,0.3348842620849609
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,10240,3584,0.23579840660095214
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,10240,5120,0.22786453564961753
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,10240,3072,0.29546346664428713
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,10240,3072,0.2038815975189209
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,10240,6144,0.2591615994771322
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,10240,2560,0.2349781354268392
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,10240,2560,0.1724789301554362
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,10240,3584,0.16583253542582194
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,10240,2048,0.1953173319498698
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,10240,2048,0.14218239784240722
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,10240,3072,0.14669119517008464
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,10240,1536,0.14756906827290853
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,10240,1536,0.11501333713531495
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,10240,2560,0.12863679726918537
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,10240,1024,0.10375573635101318
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,10240,1024,0.08389546871185302
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,10240,1536,0.08900907039642333
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,10240,768,0.08190186818440756
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,10240,768,0.07411413192749024
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,10240,1024,0.0689237356185913
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,10240,512,0.059556265672047935
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,10240,512,0.06282240152359009
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,10240,768,0.05890986522038778
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,10240,256,0.04047466516494751
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,10240,2048,0.10652799606323242
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,10240,256,0.053547732035319005
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,10240,128,0.03146986762682597
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,10240,128,0.04734293222427368
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,10240,512,0.05283840099970499
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,10240,64,0.02818560004234314
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,10240,128,0.04081706603368123
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,10240,64,0.04553920030593872
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,10240,32,0.04509653250376384
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,10240,256,0.045628801981608076
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,8192,65536,3.9063402811686196
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,8192,65536,5.525870768229167
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,8192,16384,1.4034591674804688
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,8192,16384,1.1095349629720053
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,8192,16384,0.5320469220479329
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,8192,12288,0.875495465596517
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,8192,12288,0.828005345662435
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,8192,12288,0.397051747639974
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,8192,10240,0.7691392262776693
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,8192,10240,0.7392117182413738
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,10240,65536,2.9549110412597654
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,8192,8192,0.7860309600830078
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,8192,8192,0.5002346674601237
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,8192,10240,0.3438613255818685
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,8192,7168,0.5342549641927083
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,8192,7168,0.4286944071451823
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,8192,8192,0.2698303858439127
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,8192,6144,0.43680534362792967
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,8192,6144,0.3630602518717448
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,8192,7168,0.2500256061553955
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,8192,5120,0.3732768058776855
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,8192,5120,0.33015359242757164
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,8192,6144,0.21701332728068032
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,8192,4096,0.3072319984436035
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,8192,4096,0.23338774045308433
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,8192,4096,0.15453119277954103
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,8192,3584,0.26382932662963865
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,8192,3584,0.21261226336161293
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,8192,5120,0.19349013964335124
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,8192,3072,0.22743679682413737
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,8192,3072,0.1822015921274821
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,8192,3584,0.14001173973083497
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,8192,2560,0.18456320762634276
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,8192,2560,0.15401172637939453
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,8192,3072,0.12487040360768635
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,8192,2048,0.1546496073404948
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,8192,2048,0.12234453360239665
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,10240,32,0.02950826684633891
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,8192,1536,0.12289493083953858
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,8192,1536,0.1014677365620931
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,8192,1536,0.07550826867421469
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,8192,1024,0.08542079925537109
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,8192,1024,0.07693333625793457
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,8192,2048,0.0915445327758789
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,8192,768,0.0659168004989624
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,8192,768,0.06781333287556966
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,8192,2560,0.1076917330423991
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,8192,512,0.047203199068705244
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,8192,512,0.05536853472391764
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,8192,512,0.04491413434346517
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,8192,256,0.03182186683019002
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,8192,256,0.04724586804707845
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,8192,768,0.049858132998148605
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,8192,128,0.029294933875401812
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,8192,128,0.04285973310470581
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,8192,1024,0.058303999900817874
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,8192,128,0.03662079970041911
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,8192,64,0.024380799134572348
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,8192,256,0.04018773237864177
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,8192,32,0.024875734249750772
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,8192,64,0.04069439967473348
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,8192,32,0.039308798313140866
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,7168,65536,3.502572886149088
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,7168,65536,5.041527303059896
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,7168,16384,0.9975210825602213
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,7168,16384,1.1105578104654947
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,7168,12288,0.7778090794881185
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,7168,12288,0.7814058939615885
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,7168,12288,0.34730027516682943
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,7168,10240,0.6638197580973307
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,7168,10240,0.7456170399983724
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,7168,16384,0.4558709462483724
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,7168,8192,0.5312277475992839
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,7168,8192,0.47846186955769854
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,8192,65536,2.3956672668457033
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,7168,7168,0.497760009765625
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,7168,7168,0.45595839818318684
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,7168,10240,0.29093545277913413
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,7168,6144,0.4019359906514485
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,7168,6144,0.33918825785319007
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,7168,8192,0.2409056027730306
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,7168,5120,0.32127145131429036
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,7168,5120,0.28778454462687175
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,7168,7168,0.22021013895670571
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,7168,4096,0.2588874657948812
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,7168,4096,0.21998079617818198
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,7168,6144,0.19804372787475585
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,7168,3584,0.23311146100362143
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,7168,3584,0.19583999315897624
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,7168,5120,0.17004799842834473
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,7168,3072,0.19791146914164226
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,7168,3072,0.1669823964436849
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,7168,3584,0.12714773019154865
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,7168,2560,0.17431999842325846
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,7168,2560,0.1427135944366455
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,7168,4096,0.13898879686991375
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,7168,2048,0.13307946523030598
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,7168,2048,0.11263039906819661
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,7168,3072,0.10933972994486492
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,7168,1536,0.10290239651997883
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,7168,1536,0.09150613149007161
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,7168,2560,0.09673386414845785
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,7168,1024,0.07454720338185629
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,7168,1024,0.06844267050425211
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,7168,2048,0.08128213087717692
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,7168,768,0.05713813304901123
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,7168,768,0.06378560066223145
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,7168,1024,0.0521727999051412
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,7168,512,0.04280106623967488
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,7168,512,0.05059093236923218
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,7168,1536,0.06748906771341959
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,7168,256,0.02826026678085327
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,7168,256,0.04579413334528605
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,7168,768,0.044921600818634035
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,7168,128,0.025887999931971235
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,7168,256,0.036212265491485596
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,7168,128,0.03999040126800537
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,7168,64,0.021491199731826782
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,7168,64,0.03767786820729573
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,7168,32,0.021917865673700968
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,7168,32,0.03714346488316854
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,7168,128,0.033353598912556966
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,7168,512,0.04023146629333496
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,6144,65536,3.345680999755859
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,6144,65536,4.557231140136719
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,6144,16384,0.878105608622233
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,6144,16384,1.0306314468383788
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,6144,12288,0.6874687830607097
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,6144,12288,0.7245215733846029
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,6144,12288,0.30968106587727867
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,6144,10240,0.5850133260091146
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,6144,10240,0.638481076558431
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,6144,16384,0.4237813313802083
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,6144,8192,0.45581973393758135
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,6144,8192,0.4298090616861979
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,6144,10240,0.2760053316752116
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,6144,7168,0.4035509427388509
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,6144,7168,0.3660661379496256
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,6144,8192,0.22442879676818847
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,6144,6144,0.35154987970987955
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,6144,6144,0.309059206644694
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,6144,7168,0.20488106409708656
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,6144,5120,0.29625492095947265
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,6144,5120,0.2579594612121582
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,6144,6144,0.17991679509480793
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,6144,4096,0.22387199401855468
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,6144,4096,0.20969279607137045
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,6144,5120,0.16003093719482422
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,6144,3584,0.1994442621866862
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,6144,3584,0.17665173212687174
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,6144,4096,0.12894826730092365
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,6144,3072,0.17294507026672362
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,6144,3072,0.15486613909403485
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,6144,3584,0.1169706662495931
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,6144,2560,0.14820586840311686
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,6144,2560,0.13153493404388428
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,6144,3072,0.10150720278422039
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,6144,2048,0.11649599870045979
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,6144,2048,0.10372479756673177
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,6144,2560,0.09014933109283448
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,6144,1536,0.09148693084716797
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,6144,1536,0.08625919818878174
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,6144,2048,0.07437333265940348
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,6144,1024,0.06118186712265015
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,6144,1024,0.06304959853490194
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,6144,1024,0.04827306667963664
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,6144,768,0.04811413288116455
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,6144,768,0.057956266403198245
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,6144,1536,0.06275946696599324
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,6144,512,0.0359658678372701
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,6144,512,0.05531200170516968
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,6144,768,0.04122346639633179
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,6144,256,0.030935466289520264
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,6144,256,0.04745386838912964
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,6144,512,0.03784106572469075
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,6144,128,0.02012373407681783
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,6144,256,0.03424853483835856
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,6144,128,0.035588268438975015
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,6144,64,0.018131200472513834
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,6144,64,0.033573333422342935
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,6144,32,0.018462934096654258
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,6144,32,0.03299626708030701
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,6144,128,0.031496532758076984
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,5120,65536,3.0070782979329427
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,5120,65536,3.495071919759115
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,5120,16384,0.9967936197916666
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,5120,16384,0.8931680043538412
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,7168,65536,2.033781305948893
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,5120,12288,0.6560672124226887
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,5120,12288,0.6036191940307617
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,5120,16384,0.3468053181966146
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,5120,10240,0.49475733439127606
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,5120,10240,0.4719498634338379
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,5120,12288,0.27949441274007164
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,5120,8192,0.39102293650309244
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,5120,8192,0.39439147313435874
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,5120,8192,0.19636480013529461
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,5120,7168,0.35374720891316735
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,5120,7168,0.33003412882486977
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,5120,10240,0.23188799222310386
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,5120,6144,0.2962026596069336
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,5120,6144,0.2838047981262207
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,5120,6144,0.15723306337992352
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,5120,5120,0.24025492668151854
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,5120,5120,0.23186453183492026
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,5120,7168,0.1771018664042155
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,5120,4096,0.19351360003153484
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,5120,4096,0.18292053540547687
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,6144,65536,1.6911936442057294
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,5120,3584,0.17084800402323405
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,5120,3584,0.17508586247762042
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,5120,4096,0.11307413578033447
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,5120,3072,0.15017174084981283
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,5120,3072,0.14086613655090333
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,5120,5120,0.14022293090820312
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,5120,2560,0.12415146827697754
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,5120,2560,0.11326186656951905
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,5120,3584,0.10204586982727051
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,5120,2048,0.10186879634857178
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,5120,2048,0.0941759983698527
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,5120,3072,0.08997440338134766
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,5120,1536,0.07642239729563395
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,5120,1536,0.07705492973327636
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,5120,2560,0.0786240021387736
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,5120,1024,0.05297600030899048
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,5120,1024,0.05993173519770304
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,5120,1024,0.04213759899139404
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,5120,768,0.04113279978434245
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,5120,768,0.054612267017364505
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,5120,1536,0.055359999338785805
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,5120,512,0.029782400528589888
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,5120,512,0.04441279967625936
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,5120,768,0.03671893278757731
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,5120,256,0.02093120018641154
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,5120,256,0.03791466554005941
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,5120,512,0.03305173317591349
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,5120,2048,0.06557333469390869
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,5120,128,0.01773866613705953
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,5120,128,0.03476586739222209
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,5120,64,0.016566399733225504
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,5120,256,0.030871466795603437
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,5120,64,0.033446399370829265
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,5120,32,0.017025067408879598
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,5120,128,0.02844053308169047
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,5120,32,0.03269973397254944
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,4096,65536,2.8853429158528643
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,4096,65536,2.7584510803222657
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,4096,16384,0.7641866683959961
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,4096,16384,0.6959370930989583
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,4096,12288,0.4699808120727539
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,4096,12288,0.5307498613993327
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,4096,12288,0.23221227327982583
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,4096,10240,0.40531625747680666
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,4096,10240,0.42268905639648435
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,4096,16384,0.2997866630554199
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,4096,8192,0.31581226984659827
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,4096,8192,0.3376405398050944
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,4096,10240,0.20241173108418783
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,4096,7168,0.28231681187947594
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,4096,7168,0.2935978571573893
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,4096,8192,0.16256853739420574
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,4096,6144,0.23928960164388022
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,4096,6144,0.2529845396677653
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,4096,7168,0.14973759651184082
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,4096,5120,0.2000725269317627
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,4096,5120,0.2080832004547119
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,4096,6144,0.13204906781514486
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,4096,4096,0.1519978682200114
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,4096,4096,0.16163093249003094
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,4096,5120,0.1198581298192342
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,4096,3584,0.1396447976430257
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,4096,3584,0.14645973841349286
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,4096,4096,0.0955637296040853
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,4096,3072,0.11843093236287434
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,4096,3072,0.12359573046366375
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,4096,3584,0.08750293254852295
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,4096,2560,0.10049599806467693
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,4096,2560,0.10184533596038818
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,4096,2560,0.06670506795247397
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,4096,2048,0.08171839714050293
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,4096,2048,0.08849813143412272
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,4096,3072,0.07625173727671306
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,4096,1536,0.06200746695200602
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,4096,1536,0.07078293164571127
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,4096,1536,0.046646400292714434
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,4096,1024,0.04285226662953694
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,4096,1024,0.0550442655881246
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,5120,65536,1.3381781260172525
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,4096,2048,0.057429333527882896
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,4096,768,0.03201813300450643
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,4096,768,0.04850879907608032
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,4096,512,0.024488532543182374
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,4096,512,0.040756265322367355
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,4096,1024,0.03594880104064942
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,4096,256,0.017101866006851197
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,4096,256,0.03345066706339518
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,4096,512,0.027970133225123088
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,4096,128,0.01411626636981964
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,4096,768,0.030482133229573567
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,4096,128,0.03025919993718465
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,4096,64,0.013651200135548911
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,4096,64,0.028842665751775104
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,4096,32,0.014493866761525472
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,4096,256,0.026791467269261675
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,4096,128,0.025676800807317095
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,4096,32,0.028230400880177815
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,3584,65536,2.2630335489908857
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,3584,65536,2.503980763753255
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,3584,16384,0.5897792180379231
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,3584,16384,0.6393152236938476
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,3584,12288,0.3899477322896322
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,3584,12288,0.4730496088663737
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,3584,12288,0.20614080429077147
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,3584,10240,0.33106346130371095
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,3584,10240,0.40519040425618486
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,3584,16384,0.26988906860351564
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,3584,8192,0.27637866338094075
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,3584,8192,0.32193066279093424
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,3584,10240,0.1799839973449707
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,3584,7168,0.24123199780782065
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,3584,7168,0.2780277252197266
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,3584,8192,0.14726932843526203
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,3584,6144,0.2042367935180664
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,3584,6144,0.23803413709004723
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,3584,7168,0.13232853412628173
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,3584,5120,0.1648853302001953
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,3584,5120,0.19216747283935548
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,3584,6144,0.12015039920806884
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,3584,4096,0.13474666277567546
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,3584,4096,0.1504405339558919
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,3584,5120,0.10712959766387939
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,3584,3584,0.11798186302185058
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,3584,3584,0.13375040690104167
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,4096,65536,1.1401685078938804
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,3584,3072,0.10677119890848796
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,3584,3072,0.13384853998819987
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,3584,4096,0.08706773122151693
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,3584,2560,0.08642133076985678
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,3584,2560,0.09830186367034913
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,3584,3584,0.07815893491109213
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,3584,2048,0.06901866594950358
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,3584,2048,0.08191573619842529
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,3584,3072,0.06929066975911459
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,3584,1536,0.054545064767201744
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,3584,1536,0.06356159845987955
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,3584,2560,0.06042133172353109
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,3584,1024,0.0371829350789388
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,3584,1024,0.05118826627731323
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,3584,2048,0.05095680157343546
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,3584,768,0.02762986620267232
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,3584,768,0.04424639940261841
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,3584,1024,0.03282026648521423
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,3584,512,0.0211136003335317
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,3584,512,0.037085866928100585
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,3584,1536,0.042771200338999435
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,3584,256,0.015482667088508605
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,3584,256,0.03190400004386902
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,3584,768,0.02802133361498515
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,3584,512,0.025864533583323163
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,3584,128,0.014044800400733947
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,3584,256,0.02550400098164876
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,3584,128,0.028657066822052
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,3584,64,0.012019200126330058
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,3584,64,0.026902399460474652
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,3584,32,0.012087466319402058
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,3584,128,0.023910399278004965
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,3584,32,0.0262442668279012
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,3072,65536,1.9557781219482422
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,3072,65536,2.367665100097656
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,3072,16384,0.4624373435974121
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,3072,16384,0.6116320292154949
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,3072,12288,0.3572821299235026
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,3072,12288,0.4698645273844401
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,3072,12288,0.2070517381032308
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,3072,10240,0.3192053476969401
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,3072,10240,0.3872096061706543
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,3072,16384,0.2655285358428955
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,3072,8192,0.24434879620869956
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,3072,8192,0.30392319361368814
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,3072,10240,0.1800117333730062
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,3072,7168,0.21802666982014976
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,3072,7168,0.27454185485839844
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,3072,8192,0.14615146319071454
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,3072,6144,0.18106239636739094
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,3072,6144,0.22830293973286947
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,3072,7168,0.13263893127441406
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,3072,5120,0.15935254096984863
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,3072,5120,0.19270720481872558
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,3072,6144,0.11901546319325765
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,3072,4096,0.12373973528544109
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,3072,4096,0.14854399363199872
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,3072,5120,0.10675626595815021
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,3072,3584,0.11086933612823487
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,3072,3584,0.12573226292928058
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,3072,4096,0.08631253242492676
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,3072,3072,0.09374293486277262
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,3072,3072,0.11024959882100423
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,3072,3584,0.07812159856160482
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,3072,2560,0.08022399743398032
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,3072,2560,0.0916320006052653
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,3072,3072,0.06942719618479411
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,3072,2048,0.06397333145141601
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,3072,2048,0.07854932943979899
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,3072,2560,0.06036800146102905
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,3072,1536,0.04783466657002767
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,3072,1536,0.061953067779541016
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,3584,65536,0.9974655787150064
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,3072,1536,0.0420906662940979
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,3072,1024,0.03194560011227925
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,3072,1024,0.04802026748657227
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,3072,768,0.026604799429575603
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,3072,768,0.043816534678141277
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,3072,2048,0.05018453200658163
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,3072,512,0.018984532356262206
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,3072,512,0.035469865798950194
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,3072,768,0.0276309331258138
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,3072,256,0.014148267110188803
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,3072,512,0.025516800085703534
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,3072,256,0.03107946713765462
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,3072,1024,0.03123626708984375
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,3072,128,0.012481066584587096
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,3072,128,0.026452267169952394
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,3072,64,0.010420266787211101
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,3072,64,0.02485760052998861
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,3072,128,0.023695999383926393
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,3072,256,0.024681599934895833
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,3072,32,0.011105066537857056
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,3072,32,0.026577067375183106
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,2560,65536,1.591696039835612
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,2560,16384,0.49574826558430984
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,2560,65536,2.2173642476399738
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,2560,16384,0.6910005569458008
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,2560,12288,0.2924799919128418
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,2560,12288,0.4191786766052246
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,2560,12288,0.18304212888081867
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,2560,10240,0.2649322668711344
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,2560,10240,0.3619455973307292
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,2560,16384,0.235259739557902
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,2560,8192,0.20634454091389975
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,2560,8192,0.2908543904622396
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,2560,10240,0.16011519432067872
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,2560,7168,0.1850389321645101
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,2560,7168,0.2602997303009033
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,2560,8192,0.1302346706390381
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,2560,6144,0.15986773173014324
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,2560,6144,0.21877973874409995
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,2560,7168,0.11828479766845704
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,2560,5120,0.13446292877197266
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,2560,5120,0.18522666295369467
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,2560,6144,0.10732053120930989
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,2560,4096,0.10955413182576497
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,2560,4096,0.14442453384399415
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,2560,5120,0.09654719829559326
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,2560,3584,0.09591360092163086
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,2560,3584,0.1256597359975179
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,2560,4096,0.07747306823730468
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,2560,3072,0.08325973351796469
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,2560,3072,0.10534506638844807
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,2560,3584,0.07056533495585124
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,2560,2560,0.07010560035705567
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,2560,2560,0.09216746489206949
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,2560,3072,0.06172693173090617
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,2560,2048,0.05714026689529419
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,2560,2048,0.07621119817097982
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,2560,2560,0.05394986470540365
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,2560,1536,0.044251732031504315
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,2560,1536,0.06030826568603516
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,3072,65536,0.9437599817911784
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,2560,1024,0.02882560094197591
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,2560,1024,0.04488319953282674
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,2560,2048,0.0459114670753479
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,2560,768,0.023600000143051147
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,2560,1024,0.028391466538111372
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,2560,768,0.03956480026245117
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,2560,512,0.017046399911244712
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,2560,512,0.03346773386001587
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,2560,1536,0.03699520031611125
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,2560,256,0.012567466497421265
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,2560,256,0.028888533512751263
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,2560,256,0.023129600286483764
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,2560,128,0.00976213316122691
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,2560,512,0.022746666272481283
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,2560,768,0.024294400215148927
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,2560,64,0.008366933465003968
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,2560,128,0.025784534215927125
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,2560,32,0.00836906631787618
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,2560,64,0.02272319992383321
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,2560,32,0.022593067089716593
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,2560,128,0.02221440076828003
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,2048,65536,1.2770751953125
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,2048,16384,0.385644785563151
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,2048,65536,2.053475189208984
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,2048,16384,0.5638453165690105
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,2048,12288,0.2448202610015869
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,2048,12288,0.40456212361653643
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,2048,12288,0.15703786214192708
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,2048,10240,0.20332585970560707
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,2048,10240,0.3375061353047689
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,2048,16384,0.20466133753458657
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,2048,8192,0.17357014020284017
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,2048,8192,0.26688960393269856
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,2048,10240,0.1381269296010335
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,2048,7168,0.14514026641845704
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,2048,7168,0.23041280110677084
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,2048,8192,0.11354133288065593
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,2048,6144,0.12955626646677654
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,2048,6144,0.2008906682332357
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,2048,7168,0.10279786586761475
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,2048,5120,0.10318079789479573
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,2048,5120,0.16667200724283854
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,2048,6144,0.09485440254211426
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,2048,4096,0.08343146642049154
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,2048,4096,0.1291530688603719
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,2048,5120,0.08513387044270834
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,2048,3584,0.07194560368855794
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,2048,3584,0.11277759869893392
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,2048,4096,0.06932266553243002
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,2048,3072,0.06405653158823649
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,2048,3072,0.096998397509257
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,2048,3584,0.06285119851430257
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,2048,2560,0.051481600602467856
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,2048,2560,0.08046080271402994
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,2048,3072,0.05503253142038981
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,2048,2048,0.04299946626027425
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,2048,2048,0.06714346408843994
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,2560,65536,0.8627498626708985
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,2048,1536,0.03224746584892273
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,2048,1536,0.053471998373667395
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,2048,2560,0.047670400142669676
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,2048,1024,0.021893332401911415
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,2048,1024,0.0400490681330363
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,2048,1536,0.03252266645431519
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,2048,768,0.018105600277582803
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,2048,768,0.037095467249552413
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,2048,2048,0.039920000235239665
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,2048,512,0.013036800424257913
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,2048,512,0.03028693397839864
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,2048,1024,0.02517333428064982
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,2048,256,0.010021332899729412
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,2048,256,0.026613332827885944
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,2048,512,0.020588799317677816
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,2048,768,0.021473066012064616
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,2048,128,0.00872320036093394
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,2048,128,0.02452053427696228
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,2048,64,0.007952000200748443
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,2048,256,0.02106346686681112
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,2048,128,0.02018666664759318
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,2048,32,0.007896533111731212
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,2048,32,0.022036266326904298
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,1536,65536,0.9407914479573568
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,1536,65536,1.8981695810953778
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,1536,16384,0.24053014119466148
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,1536,16384,0.48924585978190105
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,1536,16384,0.17313812573750814
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,1536,12288,0.18360959688822429
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,1536,12288,0.37021013895670574
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,1536,12288,0.1357653299967448
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,1536,10240,0.1591093381245931
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,1536,10240,0.31553386052449545
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,1536,10240,0.12022293408711751
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,1536,8192,0.13197866280873616
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,1536,8192,0.2529930591583252
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,2048,65536,0.7249994913736979
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,1536,7168,0.11005653540293377
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,1536,7168,0.21743466059366862
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,1536,8192,0.09780266284942626
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,1536,7168,0.09003413518269857
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,1536,6144,0.19122986793518065
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,1536,6144,0.0823199987411499
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,1536,5120,0.08015573024749756
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,1536,5120,0.15768639246622723
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,1536,5120,0.0756341298421224
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,1536,4096,0.06398080190022787
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,1536,4096,0.12048106988271076
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,1536,4096,0.06023039817810059
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,1536,3584,0.05599573453267416
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,1536,3584,0.1053930679957072
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,1536,3584,0.05532373189926147
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,1536,3072,0.04847253163655599
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,1536,3072,0.08976960182189941
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,1536,3072,0.04821759859720866
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,1536,2560,0.04075839916865031
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,1536,2560,0.07503999869028727
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,1536,2560,0.04092693328857422
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,1536,2048,0.032596266269683837
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,1536,65536,0.6190879821777344
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,1536,2048,0.059779198964436855
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,1536,1536,0.026147200663884478
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,1536,1536,0.047532800833384195
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,1536,1536,0.02847893238067627
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,1536,1024,0.017935999234517417
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,1536,1024,0.03914239803949992
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,1536,2048,0.034016001224517825
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,1536,768,0.014995200435320535
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,1536,768,0.035734399159749346
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,1536,1024,0.02187839945157369
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,1536,512,0.012437333663304646
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,1536,768,0.018568533658981323
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,1536,512,0.030859732627868654
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,1536,256,0.009900800387064616
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,1536,256,0.02683093349138896
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,1536,512,0.01811306675275167
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,1536,128,0.008362666765848795
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,1536,256,0.019387733936309815
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,1536,128,0.025077333052953083
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,1536,64,0.008330666522185007
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,1536,128,0.018915200233459474
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,1536,32,0.007100800176461537
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,1536,64,0.02072426676750183
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,1536,32,0.02193386753400167
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,1024,65536,0.5996693293253581
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,1024,65536,1.7232341766357422
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,1024,16384,0.1565301259358724
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,1024,16384,0.4442303975423177
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,1024,16384,0.1414197285970052
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,1024,12288,0.12220053672790528
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,1024,12288,0.33665173848470054
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,1024,12288,0.1113770643870036
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,1024,10240,0.10613333384195964
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,1024,10240,0.2853482564290365
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,1024,10240,0.09996159871419272
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,1024,8192,0.08684053421020507
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,1024,8192,0.2281343936920166
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,1024,8192,0.08157760302225749
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,1024,7168,0.07603946526845297
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,1024,7168,0.20240319569905602
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,1024,7168,0.07573973337809245
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,1024,6144,0.06488960186640422
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,1024,6144,0.17471893628438312
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,1024,6144,0.07022613684336344
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,1024,5120,0.05410879850387573
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,1024,5120,0.14147839546203614
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,1024,5120,0.06585386594136557
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,1024,4096,0.04348906675974528
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,1024,4096,0.10965332984924317
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,1024,4096,0.05168106555938721
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,1024,3584,0.038523733615875244
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,1024,3584,0.09273706277211508
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,1024,65536,0.49605973561604816
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,1024,3072,0.03237333297729492
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,1024,3072,0.078220796585083
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,1024,3584,0.045679998397827146
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,1024,2560,0.027964800596237183
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,1024,2560,0.06577173471450806
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,1024,3072,0.04039573272069295
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,1024,2048,0.022836265961329143
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,1024,2048,0.053574399153391516
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,1024,2560,0.03489813407262166
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,1024,1536,0.018462934096654258
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,1024,1536,0.04489920139312744
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,1024,2048,0.028382933139801024
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,1024,1024,0.012919466694196066
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,1024,1024,0.03493333260218302
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,1024,1536,0.023939200242360435
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,1024,768,0.010981333255767823
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,1024,1024,0.018206934134165444
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,1024,768,0.03097493251164754
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,1024,512,0.008429867029190064
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,1024,512,0.02725546757380168
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,1024,768,0.015723733107248943
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,1024,256,0.006302933394908905
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,1024,512,0.01530346671740214
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,1024,256,0.023242666323979696
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,1024,256,0.017795199155807497
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,1024,128,0.005888000130653381
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,1024,128,0.022344533602396646
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,1024,128,0.017369600137074788
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,1024,64,0.005063466727733612
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,1024,64,0.020537600914637247
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,1024,32,0.005123200019200643
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,1024,32,0.01869866649309794
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,768,65536,0.4819967905680339
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,768,65536,1.6485621134440105
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,768,16384,0.12187093098958332
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,768,16384,0.4248799959818522
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,768,16384,0.14049386978149414
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,768,12288,0.09194560050964355
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,768,12288,0.3233269373575846
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,768,12288,0.1117024024327596
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,768,10240,0.08650240103403727
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,768,10240,0.27621119817097983
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,768,10240,0.09949973424275717
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,768,8192,0.06922346750895182
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,768,8192,0.2225887934366862
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,768,8192,0.08131840229034423
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,768,7168,0.061262933413187656
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,768,7168,0.19401706059773763
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,768,7168,0.07511786619822183
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,768,6144,0.052900266647338864
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,768,6144,0.16865599950154622
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,768,6144,0.06977386474609375
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,768,5120,0.04411413272221883
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,768,5120,0.13757120768229167
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,768,5120,0.06517546574274699
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,768,4096,0.03508586486180623
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,768,4096,0.10494080384572346
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,768,65536,0.4959690729777019
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,768,3584,0.030385067065556843
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,768,4096,0.05064853429794312
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,768,3584,0.08834239641825357
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,768,3072,0.026011733214060466
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,768,3072,0.07464106877644858
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,768,3584,0.04541866779327393
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,768,2560,0.02342933416366577
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,768,3072,0.03939733505249023
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,768,2560,0.05985920031865438
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,768,2048,0.01837013363838196
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,768,2048,0.04944213231404622
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,768,2560,0.03405653238296509
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,768,1536,0.015135999520619711
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,768,2048,0.028564266363779706
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,768,1536,0.04129066864649455
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,768,1024,0.010987733801205952
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,768,1024,0.032876799503962204
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,768,1536,0.023942399024963378
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,768,1024,0.017745065689086913
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,768,768,0.00934933324654897
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,768,768,0.030931200583775836
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,768,512,0.007088000078996022
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,768,512,0.025568000475565594
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,768,768,0.015713066856066386
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,768,256,0.006316799918810527
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,768,512,0.014859732985496522
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,768,256,0.02355413238207499
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,768,128,0.005492266515890757
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,768,128,0.021042132377624513
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,768,256,0.01736853321393331
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,768,64,0.005083733300367991
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,768,128,0.016988799969355265
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,768,64,0.018556799491246542
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,768,32,0.005131733417510986
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,768,32,0.019643733898798622
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,512,65536,0.3448298772176107
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,512,65536,1.5771626790364583
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,512,16384,0.08641706307729086
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,512,16384,0.40576852162679033
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,512,16384,0.10914666652679443
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,512,12288,0.06581120093663534
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,512,12288,0.3089994748433431
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,512,12288,0.08740586439768473
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,512,10240,0.05556480089823405
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,512,10240,0.2600032011667887
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,512,10240,0.07948693434397379
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,512,8192,0.04532693227132161
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,512,8192,0.2093536059061686
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,512,8192,0.06546773513158163
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,512,7168,0.03992960055669149
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,512,7168,0.18529280026753742
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,512,7168,0.06129706700642904
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,512,6144,0.034883201122283936
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,512,6144,0.15877013206481932
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,512,6144,0.05814400116602579
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,512,5120,0.029028266668319702
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,512,5120,0.12983787059783936
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,512,65536,0.3811392148335775
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,512,4096,0.02294506629308065
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,512,4096,0.09843413035074869
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,512,5120,0.05413013299306234
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,512,3584,0.020654932657877604
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,512,3584,0.08203413486480712
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,512,4096,0.04123626550038655
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,512,3584,0.03671679894129436
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,512,3072,0.018322134017944337
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,512,3072,0.06610986789067587
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,512,3072,0.03179840048154195
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,512,2560,0.015853866934776306
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,512,2560,0.05718400080998739
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,512,2048,0.01358080009619395
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,512,2048,0.04607573350270589
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,512,2560,0.027585067351659137
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,512,1536,0.010856533050537109
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,512,1536,0.03745493491490682
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,512,1536,0.019398399194081626
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,512,2048,0.023135999838511147
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,512,1024,0.008358400066693623
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,512,1024,0.03153066635131836
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,512,768,0.007143466671307881
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,512,768,0.028805333375930785
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,512,1024,0.015306666493415833
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,512,512,0.0058794667323430385
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,512,512,0.024845866362253825
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,512,768,0.013271466890970866
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,512,256,0.004647466540336609
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,512,256,0.021191465854644775
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,512,512,0.013649066289265951
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,512,128,0.004197333256403605
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,512,256,0.016131200393040977
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,512,128,0.01882986625035604
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,512,64,0.004208000004291534
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,512,128,0.01575040022532145
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,512,64,0.019435733556747437
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,512,32,0.0042133331298828125
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,512,32,0.018478933970133463
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,256,65536,0.21124480565388998
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,256,65536,1.5149951934814454
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,256,16384,0.05559999942779541
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,256,16384,0.39131094614664713
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,256,16384,0.10776853561401367
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,256,12288,0.04178666671117147
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,256,12288,0.2965504010518392
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,256,12288,0.08594559828440348
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,256,10240,0.035531731446584065
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,256,10240,0.2496799945831299
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,256,10240,0.0787221352259318
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,256,8192,0.02948586742083232
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,256,8192,0.20101760228474935
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,256,8192,0.06407466729482016
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,256,7168,0.026088533798853557
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,256,65536,0.37111892700195315
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,256,7168,0.1765941301981608
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,256,6144,0.02211093306541443
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,256,6144,0.15060267448425294
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,256,7168,0.06076373259226481
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,256,5120,0.017347200711568197
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,256,5120,0.12341012954711914
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,256,6144,0.05662506818771362
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,256,4096,0.013375999530156455
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,256,4096,0.08899412949879965
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,256,5120,0.05332479874293009
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,256,3584,0.012195199728012085
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,256,4096,0.039044264952341715
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,256,3584,0.07600639661153158
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,256,3072,0.010809600353240967
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,256,3072,0.060481067498524985
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,256,3584,0.03538879950841268
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,256,2560,0.009473066528638203
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,256,3072,0.030892799297968548
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,256,2560,0.051575465997060144
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,256,2048,0.007978666822115581
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,256,2048,0.0413269321123759
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,256,2560,0.0267850657304128
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,256,1536,0.007640533149242401
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,256,1536,0.03543146848678589
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,256,2048,0.022666666905085245
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,256,1024,0.005459199845790863
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,256,1024,0.029079467058181763
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,256,1536,0.01904426614443461
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,256,768,0.004929066697756449
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,256,1024,0.01493013302485148
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,256,768,0.027214932441711425
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,256,512,0.004271999994913737
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,256,512,0.022782933712005616
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,256,768,0.01320319970448812
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,256,256,0.0034677334129810332
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,256,256,0.020702934265136717
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,256,512,0.01363200048605601
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,256,128,0.0030432000756263735
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,256,128,0.01862506667772929
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,256,256,0.015692800283432007
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,256,64,0.003011200080315272
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,256,64,0.01665386656920115
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,256,128,0.01569386621316274
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,256,32,0.003186133255561193
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,256,32,0.01656426688035329
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,128,65536,0.1695199966430664
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,128,65536,1.5028213500976562
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,128,16384,0.045612800121307376
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,128,16384,0.3858314514160156
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,128,16384,0.11203093528747558
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,128,12288,0.033989334106445314
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,128,12288,0.2921333312988281
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,128,65536,0.3866645177205404
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,128,10240,0.02998080054918925
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,128,12288,0.09289173285166422
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,128,10240,0.2461141268412272
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,128,8192,0.024105600516001382
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,128,8192,0.19775466918945311
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,128,10240,0.08438826402028402
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,128,7168,0.021163733800252278
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,128,8192,0.06571520169576009
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,128,7168,0.17362772623697917
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,128,6144,0.01597119967142741
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,128,6144,0.14777812957763672
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,128,7168,0.061543468634287515
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,128,5120,0.0120170662800471
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,128,6144,0.06005866527557373
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,128,5120,0.11949333349863689
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,128,4096,0.009496532877286275
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,128,4096,0.08664639790852866
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,128,5120,0.0551797350247701
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,128,3584,0.008321066697438557
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,128,4096,0.04057493209838867
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,128,3584,0.07227733135223388
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,128,3072,0.007531733314196269
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,128,3072,0.058837334314982094
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,128,3584,0.03577599922815959
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,128,2560,0.006761600077152252
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,128,2560,0.047295999526977536
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,128,3072,0.03210986653963725
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,128,2048,0.005864533285299936
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,128,2048,0.039903998374938965
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,128,2560,0.027592533826828004
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,128,1536,0.005138133466243744
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,128,2048,0.02380266586939494
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,128,1536,0.03489919900894165
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,128,1024,0.004229333500067393
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,128,1024,0.028909866015116376
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,128,1536,0.01975040038426717
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,128,768,0.003790933390458425
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,128,1024,0.01568000018596649
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,128,768,0.02683200041453044
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,128,512,0.0034506666163603462
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,128,512,0.02304640014966329
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,128,768,0.013274666666984559
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,128,256,0.003028266628583272
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,128,512,0.013661866386731466
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,128,256,0.02069973349571228
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,128,128,0.0026229334374268847
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,128,128,0.018503467241923012
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,128,256,0.015718400478363037
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,128,64,0.002584533393383026
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,128,64,0.016619732975959776
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8192,128,128,0.015639467040697734
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,128,32,0.0025770666698614756
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,128,32,0.01804800033569336
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,64,65536,0.1643189271291097
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,64,16384,0.04635519981384277
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,64,16384,0.38507518768310545
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,64,12288,0.03405119975407918
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,64,12288,0.29146347045898435
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,64,10240,0.02953280011812846
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,64,65536,1.4991349538167318
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,64,10240,0.24418346087137857
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,64,8192,0.02347093423207601
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,64,7168,0.020651733875274657
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,64,8192,0.19654186566670734
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,64,6144,0.015364266435305276
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,64,7168,0.17278292973836262
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,64,5120,0.01018773317337036
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,64,6144,0.14717119534810383
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,64,4096,0.008168533444404602
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,64,5120,0.1186784029006958
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,64,3584,0.0070816000302632645
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,64,4096,0.08691093126932779
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,64,3072,0.006728533407052357
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,64,3584,0.07199359734853109
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,64,3072,0.056064001719156896
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,64,2560,0.006450133522351582
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,64,2048,0.005555200080076853
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,64,2560,0.04828266700108846
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,64,2048,0.03967039982477824
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,64,1536,0.005070933202902476
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,64,1024,0.004220800101757049
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,64,1536,0.0343722661336263
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,64,768,0.003806933263937632
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,64,1024,0.028060799837112425
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,64,512,0.0031850665807724
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,64,512,0.022745599349339805
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,64,768,0.02690453330675761
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,64,256,0.0026410666604836782
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,64,128,0.00258240004380544
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,64,256,0.02056106726328532
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,64,64,0.002198400100072225
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,64,128,0.01777600049972534
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,64,32,0.0025920001169045764
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,64,64,0.017217065890630087
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,64,32,0.014545067151387533
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,32,65536,0.16251306533813475
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,32,16384,0.04930026531219482
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,32,16384,0.38482348124186194
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,32,12288,0.03183359901110332
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,32,65536,1.4868863423665366
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,32,12288,0.29109226862589516
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,32,10240,0.027562665939331054
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,32,10240,0.24364906946818032
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,32,8192,0.022283732891082764
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,32,8192,0.19551893870035808
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,32,7168,0.018937599658966065
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,32,7168,0.17063147226969402
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,32,6144,0.013215999801953635
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,32,6144,0.1462293306986491
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,32,5120,0.009994666775067646
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,32,5120,0.11667839686075847
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,32,4096,0.00788800021012624
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,32,3584,0.007209600011507671
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,32,4096,0.08655253251393637
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,32,3072,0.006403199831644694
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,32,3584,0.07221120198567708
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,32,3072,0.05835200150807699
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,32,2560,0.005497600138187409
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,32,2048,0.00472320020198822
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,32,2560,0.04734933376312256
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,32,1536,0.004690133531888326
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,32,2048,0.04103786547978719
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,32,1024,0.0038634667793909705
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,32,1536,0.03480960130691528
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,32,768,0.003398400048414866
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,32,1024,0.02881706754366557
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,32,512,0.0031285333136717476
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,32,768,0.02650559941927592
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,32,256,0.0026496000587940215
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,32,512,0.023098667462666832
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,32,128,0.0026101333399613695
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,32,256,0.01957226594289144
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,32,64,0.002196266750494639
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,32,128,0.01857173244158427
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,32,32,0.0022026665508747103
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,32,64,0.016921599706014
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,32,32,0.0168938676516215
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,65536,16384,2.6959498087565104
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,65536,16384,5.651384480794271
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,65536,12288,4.210236867268881
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,65536,12288,1.9674463907877606
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8192,2048,64,0.023459200064341226
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,65536,10240,3.9478431701660157
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,65536,10240,2.0091978708902998
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8192,1536,6144,0.09916906356811524
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,65536,8192,2.706207021077474
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,65536,8192,1.3460426330566406
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,65536,10240,1.0971871693929036
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,65536,7168,2.575885772705078
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,65536,7168,1.2082164764404297
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,65536,8192,0.8651242574055991
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,65536,6144,1.778428777058919
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,65536,6144,1.085901896158854
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,65536,12288,1.2811103820800782
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,65536,5120,1.4606240590413413
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,65536,5120,1.001544507344564
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,65536,5120,0.5805930455525716
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,65536,4096,1.1822826385498046
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,65536,4096,0.7538464228312175
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,65536,6144,0.6673749287923176
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,65536,3584,1.0163413365681966
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,65536,3584,0.6138794581095378
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,65536,16384,1.7494336446126302
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,65536,3072,0.8985013326009115
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,65536,3072,0.5013397216796875
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,65536,7168,0.7895605087280273
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,65536,2560,0.7478047688802083
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,65536,2560,0.4410602569580078
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,65536,3072,0.38290026982625325
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,65536,2048,0.6011168162027996
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,65536,2048,0.4101141293843587
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,65536,3584,0.4324362754821777
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,65536,1536,0.46463893254597977
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,65536,1536,0.2890240033467611
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,65536,2560,0.33064746856689453
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,65536,1024,0.3224725405375163
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,65536,1024,0.19901760419209796
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,65536,4096,0.4675733248392741
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,65536,768,0.241266139348348
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,65536,768,0.17098132769266766
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,65536,1024,0.18009279568990072
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,65536,512,0.18688853581746417
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,65536,512,0.14434879620869953
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,65536,1536,0.2409663995107015
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,65536,256,0.1229354699452718
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,65536,256,0.11588799953460693
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,65536,512,0.13743893305460614
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,65536,128,0.10038613478342692
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,65536,128,0.1011082649230957
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,65536,768,0.15454187393188476
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,65536,64,0.08511359691619873
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,65536,64,0.10275946458180744
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,65536,32,0.08735040028889975
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,65536,32,0.10202986399332684
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,65536,128,0.09405439694722494
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,16384,65536,5.687898763020834
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,16384,65536,3.3006100972493493
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,65536,256,0.10868053436279297
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,16384,16384,1.3322463989257813
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,16384,16384,0.8257781346638998
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,16384,12288,0.8781802495320639
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,16384,12288,0.6527936299641927
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,65536,2048,0.28563305536905925
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,16384,10240,0.7634666442871094
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,16384,10240,0.5752096176147461
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,16384,10240,0.2989301363627116
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,16384,8192,0.6014303843180339
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,16384,8192,0.38999999364217125
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,16384,12288,0.37098986307779946
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,16384,7168,0.5325514793395996
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,16384,7168,0.36563520431518554
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,16384,16384,0.4605418523152669
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,16384,6144,0.4531231880187988
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,16384,6144,0.2820373217264811
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,16384,7168,0.22989865938822426
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,16384,5120,0.38296321233113606
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,16384,5120,0.2505269368489583
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,16384,8192,0.243776003519694
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,16384,4096,0.29818134307861327
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,16384,4096,0.19409173329671223
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,16384,6144,0.20069653193155923
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,16384,3584,0.26355733871459963
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,16384,3584,0.17965013186136883
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,16384,4096,0.1421120007832845
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,16384,3072,0.22418452898661295
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,16384,3072,0.14995306332906086
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,16384,5120,0.17442026138305664
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,16384,2560,0.200437339146932
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,16384,2560,0.13256746927897137
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,16384,3584,0.129149866104126
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,16384,2048,0.1578858693440755
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,16384,2048,0.10734612941741943
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,16384,3072,0.11343466440836589
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,16384,1536,0.11686293284098308
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,16384,1536,0.08894826571146647
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,16384,2560,0.09882346789042154
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,16384,1024,0.08287786642710368
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,16384,1024,0.06872959931691489
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,16384,1024,0.05456639925638834
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,16384,768,0.06725333531697592
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,16384,768,0.06261759996414185
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,16384,1536,0.0709824005762736
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,16384,512,0.04876586596171061
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,16384,512,0.05555839935938517
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,16384,768,0.04807680050532023
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,16384,2048,0.08471786975860596
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,16384,256,0.03314666748046875
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,16384,256,0.043982934951782224
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,16384,128,0.02889066735903422
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,16384,128,0.03924373388290405
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,16384,128,0.03291306694348653
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,16384,64,0.02437333265940348
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,16384,64,0.039341866970062256
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,16384,32,0.024771199623743693
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,16384,32,0.03898026545842488
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,16384,256,0.03540480136871338
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,16384,512,0.04362560113271077
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,12288,65536,2.723309834798177
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,12288,65536,4.288866170247396
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,12288,16384,1.0021856307983399
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,12288,16384,0.7926090876261394
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,12288,12288,0.6676170349121093
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,12288,12288,0.5601280212402344
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,12288,12288,0.29591999053955076
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,12288,10240,0.5879701614379883
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,12288,10240,0.49700158437093095
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,12288,16384,0.3816735903422038
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,12288,8192,0.474564266204834
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,12288,8192,0.3297375996907552
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,12288,10240,0.25253973007202146
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,12288,7168,0.40679146448771164
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,12288,7168,0.27669334411621094
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,12288,8192,0.20763093630472818
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,12288,6144,0.3526026725769043
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,12288,6144,0.24465600649515787
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,12288,7168,0.1857770601908366
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,12288,5120,0.29605865478515625
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,12288,5120,0.2082848072052002
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,12288,6144,0.16463039716084799
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,12288,4096,0.2326591968536377
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,12288,4096,0.16774613062540691
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,12288,5120,0.14209493001302081
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,12288,3584,0.1996938705444336
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,12288,3584,0.14360639254252117
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,12288,4096,0.11655786832173665
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,12288,3072,0.17469013532002767
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,12288,3072,0.1250997304916382
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,12288,3584,0.10484586556752522
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,12288,2560,0.14818453788757324
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,12288,2560,0.1071786642074585
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,12288,3072,0.09211093584696452
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,12288,2048,0.12264533042907715
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,12288,2048,0.08824960390726724
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,12288,2560,0.08094826539357504
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,12288,1536,0.09229653676350912
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,12288,1536,0.07727253437042236
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,12288,1536,0.057790935039520264
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,12288,1024,0.06398293177286783
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,12288,1024,0.05565866629282633
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,12288,2048,0.0693674643834432
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,12288,768,0.05060586531956991
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,12288,768,0.05551573435465494
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,12288,768,0.040106666088104245
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,12288,512,0.03650240103403728
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,12288,512,0.051395201683044435
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,12288,1024,0.044595201810201004
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,12288,256,0.031048534313837688
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,12288,256,0.0456714669863383
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,12288,256,0.029794132709503172
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,12288,128,0.019887999693552653
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,12288,128,0.03408533334732056
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,12288,128,0.02749226689338684
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,12288,64,0.017504000663757326
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,12288,64,0.033166933059692386
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,12288,512,0.036160000165303546
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,12288,32,0.018132267395655315
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,12288,32,0.03202880024909973
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,10240,65536,2.422804260253906
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,10240,65536,3.603844197591146
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,10240,16384,1.0019210815429687
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,10240,16384,0.7413493474324544
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,10240,16384,0.3311274528503418
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,10240,12288,0.6009344100952149
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,10240,12288,0.45455147425333664
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,16384,65536,2.0054773966471355
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,10240,10240,0.5346986770629882
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,10240,10240,0.40382080078125
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,10240,12288,0.23986666997273765
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,10240,8192,0.39514452616373696
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,10240,8192,0.2767082532246908
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,10240,10240,0.2115818659464518
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,10240,7168,0.3288106600443522
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,10240,7168,0.24782719612121581
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,10240,8192,0.17498453458150226
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,10240,6144,0.2936565399169922
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,10240,6144,0.21365440686543785
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,12288,65536,1.4457504272460937
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,10240,5120,0.2538826624552409
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,10240,5120,0.20484906832377114
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,10240,6144,0.13876907030741376
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,10240,4096,0.193341859181722
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,10240,4096,0.1461429278055827
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,10240,7168,0.15487039883931478
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,10240,3584,0.1599199930826823
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,10240,3584,0.12551146348317463
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,10240,5120,0.12138880093892415
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,10240,3072,0.14121707280476886
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,10240,3072,0.1088213364283244
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,10240,3584,0.08808746337890624
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,10240,2560,0.11943786938985188
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,10240,2560,0.09825173219045004
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,10240,4096,0.09891093571980794
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,10240,2048,0.10171093146006267
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,10240,2048,0.08151679833730062
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,10240,3072,0.07936426798502604
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,10240,1536,0.07646506627400716
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,10240,1536,0.06555413405100505
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,10240,2048,0.058856534957885745
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,10240,1024,0.05362133185068766
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,10240,1024,0.049394134680430093
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,10240,1536,0.049968000253041586
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,10240,768,0.041698133945465087
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,10240,768,0.04902720053990682
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,10240,2560,0.06868053277333577
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,10240,512,0.02927146752675374
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,10240,512,0.04121066729227702
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,10240,768,0.03501546780268351
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,10240,256,0.02108586629231771
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,10240,256,0.04309440056482951
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,10240,512,0.031380265951156616
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,10240,128,0.017385600010553996
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,10240,128,0.031572266419728594
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,10240,128,0.024327466885248818
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,10240,64,0.01569493313630422
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,10240,1024,0.038646399974823
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,10240,32,0.01716266671816508
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,10240,64,0.031221334139506025
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,10240,32,0.031880533695220946
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,10240,256,0.026778666178385417
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,8192,65536,2.0218143463134766
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,8192,65536,2.5137972513834637
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,8192,16384,0.6500703811645507
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,8192,16384,0.6250368118286133
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,8192,12288,0.6009962717692058
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,8192,12288,0.4224213282267253
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,8192,12288,0.20543039639790855
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,8192,10240,0.38811521530151366
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,8192,10240,0.2850165367126465
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,8192,16384,0.2808298746744792
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,8192,8192,0.3154560089111328
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,8192,8192,0.25975467363993326
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,8192,10240,0.17550400098164876
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,8192,7168,0.272270933787028
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,8192,7168,0.20839145978291831
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,8192,8192,0.14308907190958658
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,8192,6144,0.24333440462748207
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,8192,6144,0.17657920519510906
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,8192,7168,0.12650240262349446
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,8192,5120,0.1846175988515218
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,8192,5120,0.15204052925109862
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,10240,65536,1.2114805857340494
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,8192,4096,0.1571605364481608
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,8192,4096,0.1504458745320638
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,8192,6144,0.11365653673807781
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,8192,3584,0.13530559539794923
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,8192,3584,0.11009173393249512
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,8192,5120,0.09935572942097982
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,8192,3072,0.11550079981486003
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,8192,3584,0.07513706684112549
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,8192,3072,0.09230613708496094
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,8192,2560,0.09578346411387126
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,8192,2560,0.08578240076700847
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,8192,4096,0.08268266518910726
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,8192,2048,0.08135999838511149
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,8192,2048,0.07158400217692057
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,8192,2560,0.05744959910710653
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,8192,1536,0.06031680107116699
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,8192,1536,0.05995839834213257
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,8192,2048,0.04911359945933024
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,8192,1024,0.04258453448613485
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,8192,1024,0.04519786834716797
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,8192,3072,0.06607040166854858
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,8192,768,0.03373333215713501
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,8192,768,0.043347199757893876
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,8192,768,0.029949865738550824
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,8192,512,0.02471253275871277
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,8192,1024,0.03262186646461487
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,8192,512,0.03777066469192505
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,8192,256,0.01755519906679789
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,8192,256,0.03215466737747193
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,8192,256,0.023452800512313843
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,8192,128,0.016218666235605875
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,8192,128,0.03271680076917012
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,8192,128,0.022088533639907836
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,8192,1536,0.04102720022201538
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,8192,64,0.01740586757659912
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,8192,512,0.02632960081100464
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,8192,32,0.018161066373189292
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,8192,64,0.031778132915496825
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,8192,32,0.033293867111206056
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,7168,65536,1.8046133677164715
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,7168,65536,2.433721669514974
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,7168,16384,0.5098976135253906
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,7168,16384,0.610859743754069
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,7168,12288,0.49310719172159834
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,7168,12288,0.3419466654459635
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,7168,12288,0.18492906888326008
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,7168,10240,0.3432640075683594
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,7168,10240,0.3064480145772298
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,7168,16384,0.23829332987467447
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,7168,8192,0.27191146214803064
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,7168,8192,0.21023146311442056
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,7168,10240,0.15718612670898438
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,7168,7168,0.24027093251546225
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,7168,7168,0.1956714630126953
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,7168,8192,0.1250442663828532
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,7168,6144,0.19545706113179523
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,7168,6144,0.16539626121520995
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,7168,7168,0.11260586579640705
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,7168,5120,0.1692405382792155
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,7168,5120,0.13980906804402668
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,7168,6144,0.1005834658940633
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,7168,4096,0.13164587020874025
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,7168,4096,0.10788160165150959
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,7168,5120,0.09008320172627768
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,7168,3584,0.12295253276824951
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,7168,3584,0.10009173552195232
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,7168,4096,0.07286079724629721
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,7168,3072,0.10057919820149738
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,7168,3072,0.08839253584543863
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,7168,3584,0.06636586586634317
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,7168,2560,0.08833279609680175
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,7168,2560,0.0760416030883789
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,7168,3072,0.05877759853998819
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,7168,2048,0.0683199961980184
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,7168,2048,0.0632213314374288
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,7168,2560,0.05156266689300537
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,7168,1536,0.0535370667775472
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,7168,1536,0.05324693520863851
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,7168,2048,0.04408533175786336
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,7168,1024,0.03614399830500285
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,7168,1024,0.04193280140558879
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,7168,1536,0.037088000774383546
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,7168,768,0.029589333136876422
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,7168,768,0.040408531824747726
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,7168,768,0.026632533470789595
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,7168,512,0.021895466248194377
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,7168,512,0.0347264011700948
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,7168,1024,0.028971733649571736
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,7168,256,0.014890666802724203
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,7168,256,0.028922667105992634
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,7168,256,0.02062186598777771
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,7168,128,0.013686399658521017
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,7168,128,0.02752959926923116
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,7168,512,0.024578134218851723
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,7168,64,0.01204159955183665
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,7168,64,0.027565866708755493
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,7168,32,0.012169599533081055
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,7168,32,0.02635306715965271
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,7168,128,0.019357866048812865
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,6144,65536,1.9084351857503254
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,6144,65536,1.6938026428222657
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,6144,16384,0.4810976028442383
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,6144,16384,0.4364501317342122
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,8192,65536,1.0516682942708333
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,6144,12288,0.36371946334838867
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,6144,12288,0.39778881072998046
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,6144,12288,0.17041600545247396
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,6144,10240,0.31439040501912435
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,6144,10240,0.26167893409729004
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,6144,16384,0.21846292813618978
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,6144,8192,0.24278613726298012
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,6144,8192,0.2111423969268799
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,6144,10240,0.154366938273112
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,6144,7168,0.21634666124979654
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,6144,7168,0.18161066373189289
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,6144,8192,0.12457706928253173
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,6144,6144,0.18258879979451498
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,6144,6144,0.15430827140808107
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,6144,6144,0.10022079944610596
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,6144,5120,0.15072959264119465
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,6144,5120,0.12412693500518798
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,6144,7168,0.11265920003255207
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,6144,4096,0.12333973248799641
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,6144,4096,0.10266346931457519
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,6144,4096,0.0729535977045695
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,6144,3584,0.10838613510131836
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,6144,3584,0.09436586697896322
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,6144,5120,0.08821226755777994
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,6144,3072,0.09045120080312094
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,6144,3072,0.07994133631388346
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,6144,3584,0.06595093409220378
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,6144,2560,0.07805866400400797
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,6144,2560,0.07419946988423666
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,6144,3072,0.058132266998291014
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,6144,2048,0.06124053398768107
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,6144,2048,0.060236799716949466
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,7168,65536,0.9070250829060873
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,6144,1536,0.04673813184102376
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,6144,1536,0.06317333380381265
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,6144,2048,0.04207786719004313
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,6144,1024,0.03221653302510579
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,6144,1024,0.039842132727305093
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,6144,2560,0.05034026702245077
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,6144,768,0.025796266396840413
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,6144,768,0.03816320101420085
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,6144,1536,0.035743999481201175
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,6144,512,0.019215999046961467
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,6144,512,0.033192533254623416
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,6144,768,0.026289065678914387
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,6144,256,0.013726933797200521
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,6144,256,0.02766079902648926
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,6144,512,0.023449599742889404
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,6144,128,0.012423466642697651
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,6144,1024,0.027714133262634277
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,6144,128,0.025399466355641682
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,6144,64,0.010433066884676616
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,6144,64,0.025566933552424113
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,6144,32,0.010781866312026978
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,6144,32,0.024668800830841064
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,6144,256,0.020514132579167683
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,6144,128,0.018999467293421425
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,5120,65536,1.5876448313395182
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,5120,65536,1.520193099975586
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,5120,16384,0.4564341227213542
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,5120,16384,0.40889279047648114
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,5120,12288,0.2902346611022949
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,5120,12288,0.27946240107218423
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,5120,12288,0.1559541384379069
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,5120,10240,0.252129062016805
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,5120,10240,0.24364693959554037
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,5120,16384,0.19258880615234375
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,5120,8192,0.20306347211201986
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,5120,8192,0.18978026707967122
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,5120,10240,0.1349141279856364
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,5120,7168,0.18277866045633953
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,5120,7168,0.17004159291585286
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,5120,8192,0.11016213099161784
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,5120,6144,0.1568448066711426
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,5120,6144,0.13954453468322753
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,5120,7168,0.10010666847229004
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,5120,5120,0.12766506671905517
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,5120,5120,0.11822293599446614
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,5120,6144,0.08964373270670573
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,5120,4096,0.10353813171386719
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,5120,4096,0.09668800036112467
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,5120,5120,0.07946133613586426
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,5120,3584,0.09025279680887857
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,5120,3584,0.08897066911061605
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,5120,4096,0.06428586641947429
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,5120,3072,0.07668906847635905
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,5120,3072,0.07815039952596028
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,5120,3584,0.058151467641194665
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,5120,2560,0.06483199993769327
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,5120,2560,0.06991999944051107
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,5120,3072,0.05157119830449423
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,5120,2048,0.05290453433990479
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,5120,2560,0.04482346773147583
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,5120,2048,0.05389226675033569
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,5120,1536,0.040829865137736
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,5120,1536,0.04696960051854451
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,6144,65536,0.881165885925293
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,5120,1024,0.027410133679707842
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,5120,1024,0.03765226602554321
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,5120,2048,0.03789653380711873
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,5120,768,0.022061866521835328
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,5120,768,0.03540693521499634
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,5120,1536,0.03219733238220215
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,5120,512,0.0164682666460673
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,5120,512,0.031149866183598836
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,5120,768,0.023491199811299643
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,5120,256,0.012380799651145935
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,5120,1024,0.025268266598383587
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,5120,256,0.027048534154891966
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,5120,128,0.00960213343302409
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,5120,128,0.02289173404375712
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,5120,512,0.021336533625920615
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,5120,64,0.00834986666838328
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,5120,64,0.02648640076319377
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,5120,128,0.017778132359186807
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,5120,32,0.010815999905268351
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,5120,32,0.023628799120585124
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,4096,65536,1.2752447764078776
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,4096,65536,1.3529557545979818
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,4096,16384,0.3097514788309733
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,4096,16384,0.350432014465332
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,4096,16384,0.17011839548746746
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,4096,12288,0.23256853421529136
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,4096,12288,0.2663029352823893
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,4096,12288,0.1320629358291626
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,4096,10240,0.20543786684672036
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,4096,10240,0.21304213205973305
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,4096,10240,0.11413333415985108
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,4096,8192,0.1652565320332845
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,4096,8192,0.16354880332946778
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,5120,65536,0.7220021565755208
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,4096,7168,0.13685654004414877
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,4096,7168,0.14579413731892904
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,4096,8192,0.09314133326212565
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,4096,6144,0.12387306690216064
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,4096,6144,0.12223040262858073
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,4096,7168,0.0840277353922526
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,4096,5120,0.10016427040100098
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,4096,5120,0.10218026638031005
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,4096,5120,0.06685866514841715
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,4096,4096,0.0800160010655721
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,4096,4096,0.08669119675954183
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,4096,6144,0.0769589344660441
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,4096,3584,0.07181440194447836
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,4096,3584,0.07957653204600015
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,4096,3584,0.04979519844055176
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,4096,3072,0.06248960097630819
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,4096,3072,0.06912213166554769
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,4096,4096,0.05454933245976766
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,4096,2560,0.05084586540857951
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,4096,2560,0.06005866527557373
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,4096,2560,0.038389333089192706
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,4096,2048,0.041093333562215166
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,4096,2048,0.04969173272450765
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,4096,3072,0.04376000165939331
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,4096,1536,0.03270080089569092
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,4096,1536,0.04473280111948649
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,4096,1536,0.027688533067703247
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,4096,1024,0.022894932826360067
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,4096,1024,0.035018666585286455
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,4096,2048,0.03268373409907023
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,4096,768,0.017666133244832356
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,4096,768,0.03203946749369303
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,4096,1024,0.02193280061086019
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,4096,512,0.01306880017121633
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,4096,768,0.020207999149958293
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,4096,512,0.028998400767644244
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,4096,256,0.010006399949391682
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,4096,256,0.024733867247899374
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,4096,256,0.016961065928141277
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,4096,128,0.008376533786455791
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,4096,128,0.022677334149678548
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,4096,512,0.018962132930755615
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,4096,64,0.007918933530648549
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,4096,64,0.020887466271718343
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,4096,32,0.00796693315108617
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,4096,128,0.016129066546758018
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,4096,32,0.02061226765314738
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,3584,65536,1.1128672281901042
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,3584,65536,1.336664581298828
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,3584,16384,0.29895146687825525
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,3584,16384,0.3513717333475749
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,4096,65536,0.6091487884521485
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,3584,12288,0.23041386604309083
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,3584,12288,0.2668298721313477
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,3584,16384,0.13790720303853352
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,3584,10240,0.16668799718221028
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,3584,10240,0.19124693870544435
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,3584,10240,0.09400213559468587
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,3584,8192,0.1350613276163737
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,3584,8192,0.1465941270192464
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,3584,12288,0.10958293279012044
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,3584,7168,0.12491733233133953
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,3584,7168,0.13581226666768392
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,3584,8192,0.07704213460286459
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,3584,6144,0.10259093443552654
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,3584,6144,0.11062719821929931
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,3584,7168,0.07206186453501383
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,3584,5120,0.08665599822998046
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,3584,5120,0.09720426400502523
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,3584,6144,0.0639520009358724
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,3584,4096,0.06746346950531006
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,3584,5120,0.05918293396631876
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,3584,4096,0.07901546955108643
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,3584,3584,0.05677333275477091
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,3584,3584,0.06980053583780924
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,3584,3584,0.042233598232269284
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,3584,3072,0.05058026711146037
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,3584,3072,0.06285013357798258
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,3584,4096,0.04700266520182292
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,3584,2560,0.043432533740997314
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,3584,2560,0.05398186842600504
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,3584,3072,0.03700799942016601
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,3584,2048,0.03526293436686198
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,3584,2048,0.04649279912312825
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,3584,2560,0.03345599969228109
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,3584,1536,0.027276800076166792
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,3584,1536,0.04038293361663818
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,3584,1536,0.023989333709081014
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,3584,1024,0.0194432000319163
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,3584,1024,0.03265173236529033
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,3584,2048,0.027847466866175334
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,3584,768,0.014794666568438211
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,3584,768,0.03087573250134786
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,3584,1024,0.019284266233444213
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,3584,768,0.017742933829625447
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,3584,512,0.011308800180753071
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,3584,512,0.027040000756581622
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,3584,256,0.008393599589665731
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,3584,256,0.02391466697057088
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,3584,256,0.01648533344268799
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,3584,128,0.007101866602897644
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,3584,128,0.020850133895874024
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,3584,512,0.016572800278663636
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,3584,64,0.007125333448251088
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,3584,64,0.021293866634368896
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,3584,32,0.007495466868082683
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,3584,128,0.01607360045115153
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,3584,32,0.020159999529520668
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,3072,65536,0.9050954818725586
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,3072,65536,1.2097450256347657
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,3072,16384,0.2403722604115804
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,3072,16384,0.31158294677734377
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,5120,256,0.018556799491246542
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,3072,12288,0.19113066991170247
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,3072,12288,0.23846400578816734
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,3072,16384,0.13838079770406086
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,3072,10240,0.16196160316467284
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,3072,10240,0.18703360557556153
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,3072,10240,0.09477012952168783
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,3072,8192,0.12824853261311847
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,3072,8192,0.14814292589823405
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,3072,12288,0.10742613474527996
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,3072,7168,0.11293439865112305
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,3072,7168,0.12386986414591472
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,3584,65536,0.49933439890543624
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,3072,6144,0.09169812997182211
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,3072,6144,0.10808426539103191
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,3072,8192,0.07684906323750815
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,3072,5120,0.07932799657185872
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,3072,5120,0.09088426431020101
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,3072,7168,0.07065707047780355
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,3072,4096,0.06425813436508179
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,3072,4096,0.07731733322143555
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,3072,6144,0.06282453139623007
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,3072,3584,0.05513386726379395
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,3072,3584,0.07018453280131022
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,3072,4096,0.04597973426183065
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,3072,3072,0.04787520170211792
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,3072,3072,0.05818986495335897
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,3072,5120,0.05583360195159912
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,3072,2560,0.04004266659418742
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,3072,2560,0.053438933690389004
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,3072,3584,0.04207786719004313
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,3072,2048,0.03255680004755656
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,3072,2048,0.04484906593958537
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,3072,2560,0.03299946586290996
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,3072,1536,0.025259733200073242
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,3072,1536,0.03904106616973877
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,3072,3072,0.037378132343292236
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,3072,1024,0.01835626761118571
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,3072,1024,0.0324458658695221
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,3072,1024,0.018782933553059898
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,3072,768,0.014844800035158793
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,3072,768,0.031162667274475097
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,3072,1536,0.0240447998046875
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,3072,512,0.011872000495592753
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,3072,512,0.02746880054473877
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,3072,512,0.016568533579508462
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,3072,768,0.01773866613705953
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,3072,256,0.009546666344006857
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,3072,256,0.026410667101542155
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,3072,128,0.009402666489283245
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,3072,128,0.02342080076535543
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,3072,256,0.015148799618085226
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,3072,128,0.014526933431625366
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,3072,64,0.008846933643023174
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,3072,64,0.022660267353057862
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,3072,32,0.00909440020720164
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,3072,32,0.022525866826375328
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,2560,65536,0.8394378662109375
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,2560,65536,1.3281567891438804
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,2560,16384,0.21299413045247398
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,2560,16384,0.29943892161051433
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,2560,16384,0.13789119720458984
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,2560,12288,0.1601077397664388
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,2560,12288,0.22476479212443032
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,2560,12288,0.10722239812215169
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,2560,10240,0.13007040023803712
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,2560,10240,0.17934826215108235
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,3072,65536,0.49913921356201174
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,2560,8192,0.10278293291727703
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,2560,8192,0.13713280359903973
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,2560,10240,0.09456533590952555
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,2560,7168,0.09440639813741049
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,2560,7168,0.11981759866078694
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,2560,8192,0.07606613636016846
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,2560,6144,0.07703466415405273
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,2560,6144,0.1009717305501302
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,2560,7168,0.07016639709472657
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,2560,5120,0.06711253325144449
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,2560,5120,0.08653759956359863
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,2560,6144,0.06299839814503988
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,2560,4096,0.05089919964472452
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,2560,4096,0.06878613630930583
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,2560,5120,0.05490026473999023
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,2560,3584,0.046487466494242353
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,2560,3584,0.06370773315429687
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,2560,3584,0.0408458669980367
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,2560,3072,0.04025919834772746
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,2560,4096,0.044948267936706546
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,2560,3072,0.0544490655263265
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,2560,2560,0.032459733883539836
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,2560,2560,0.05069546699523926
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,2560,2560,0.03283733328183492
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,2560,2048,0.02723520000775655
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,2560,2048,0.04284160137176514
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,2560,3072,0.036391464869181316
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,2560,1536,0.020334933201471964
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,2560,1536,0.036771198113759354
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,2560,1536,0.023497599363327026
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,2560,1024,0.014292266964912415
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,2560,1024,0.029632000128428142
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,2560,2048,0.02764906684557597
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,2560,768,0.011895466844240825
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,2560,768,0.028819199403127032
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,2560,768,0.017332265774408974
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,2560,512,0.009283199906349182
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,2560,1024,0.018574933211008705
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,2560,512,0.025006934007008867
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,2560,256,0.0067775999506314594
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,2560,256,0.022757333517074586
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,2560,256,0.014493866761525472
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,2560,128,0.006285866598288219
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,2560,512,0.016089600324630738
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,2560,128,0.020785067478815714
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,2560,64,0.00555733342965444
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,2560,64,0.018870399395624796
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,2560,128,0.01406613290309906
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,2560,32,0.005869866907596588
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,2560,32,0.018474666277567546
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,2048,65536,0.6163882573445638
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,2048,65536,1.0381429036458334
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,2048,16384,0.15459413528442384
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,2048,16384,0.27089068094889324
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,2048,16384,0.10485866864522297
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,2048,12288,0.1171114683151245
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,2048,12288,0.20360320409139
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,2560,65536,0.479859193166097
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,2048,10240,0.10652053356170654
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,2048,10240,0.16790186564127604
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,2048,12288,0.08283946514129639
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,2048,8192,0.08361173470815023
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,2048,8192,0.13044053713480633
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,2048,10240,0.07433173656463624
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,2048,7168,0.07772053082784017
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,2048,7168,0.11378986835479736
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,2048,8192,0.06046293179194132
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,2048,6144,0.06362666686375937
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,2048,6144,0.09612906773885091
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,2048,7168,0.056332798798878994
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,2048,5120,0.05441493193308512
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,2048,5120,0.07877439657847086
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,2048,5120,0.04410986502965291
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,2048,4096,0.04376426537831624
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,2048,4096,0.06593813498814902
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,2048,6144,0.050246401627858484
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,2048,3584,0.03727893431981404
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,2048,3584,0.05917973518371582
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,2048,3584,0.03327680031458537
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,2048,3072,0.03224640091260274
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,2048,3072,0.05151146650314331
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,2048,4096,0.03689066569010417
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,2048,2560,0.028442666927973433
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,2048,2560,0.046753064791361494
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,2048,3072,0.029306666056315107
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,2048,2048,0.022871466477711995
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,2048,2048,0.03925866683324178
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,2048,2560,0.026375466585159303
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,2048,2048,0.022286933660507203
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,2048,1536,0.018261333306630455
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,2048,1536,0.03502506812413533
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,2048,1024,0.013099732995033263
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,2048,1024,0.02903253237406413
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,2048,1024,0.014935466647148132
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,2048,768,0.010843732953071594
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,2048,768,0.028359466791152955
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,2048,1536,0.019798400004704793
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,2048,512,0.008769067128499348
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,2048,512,0.02512213389078776
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,2048,512,0.014122666915257773
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,2048,768,0.014924800395965577
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,2048,256,0.007134933272997539
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,2048,256,0.022667733828226726
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,2048,128,0.006735999882221222
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,2048,128,0.020619734128316244
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,2048,256,0.013249066472053529
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,2048,64,0.006252799928188324
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,2048,128,0.012809600432713827
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,2048,32,0.0063285330931345625
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,2048,64,0.019870932896931967
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,2048,32,0.0185589333375295
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,1536,65536,0.4824192047119141
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,1536,65536,0.9547786712646484
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,1536,16384,0.12010239760080973
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,1536,16384,0.25022613207499184
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,1536,16384,0.10451200008392333
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,1536,12288,0.09127999941507975
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,1536,12288,0.18427947362263997
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,2048,65536,0.37970240910847985
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,1536,10240,0.08861119747161865
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,1536,10240,0.15778452555338543
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,1536,12288,0.08313279946645101
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,1536,8192,0.0701311985651652
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,1536,8192,0.11925653616587321
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,1536,8192,0.06024640003840128
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,1536,7168,0.060635733604431155
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,1536,7168,0.10290773709615071
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,1536,10240,0.0741365353266398
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,1536,6144,0.052934400240580236
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,1536,6144,0.08888320128122965
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,1536,6144,0.04962666829427083
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,1536,5120,0.043668266137441
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,1536,7168,0.055461335182189944
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,1536,5120,0.07282346884409586
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,1536,4096,0.03338773250579834
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,1536,4096,0.06065920193990072
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,1536,4096,0.03624000151952107
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,1536,3584,0.03057066599527995
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,1536,5120,0.043543465932210285
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,1536,3584,0.054774399598439535
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,1536,3072,0.02653119961420695
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,1536,3072,0.04762986501057943
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,1536,3072,0.02967039942741394
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,1536,3584,0.032959999640782674
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,1536,2560,0.022547199328740438
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,1536,2560,0.04309653441111247
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,1536,2048,0.01768640081087748
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,1536,2048,0.037034666538238524
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,1536,2560,0.02608213424682617
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,1536,2048,0.02233920097351074
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,1536,1536,0.01504746675491333
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,1536,1536,0.03295466701189677
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,1536,1024,0.010968533158302308
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,1536,1024,0.028126933177312213
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,1536,1024,0.014959999918937683
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,1536,768,0.008938666184743245
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,1536,1536,0.019346133867899577
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,1536,768,0.026794666051864625
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,1536,512,0.007558399935563405
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,1536,512,0.022962133089701332
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,1536,768,0.014514133334159851
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,1536,256,0.005898666878541311
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,1536,512,0.013639466961224875
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,1536,256,0.02241920034090678
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,1536,128,0.005456000069777171
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,1536,128,0.018823466698328652
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,1536,256,0.01320319970448812
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,1536,64,0.00506986677646637
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,1536,64,0.018407466014226277
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,1536,128,0.012875733772913614
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,1536,32,0.005066666503747304
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,1536,32,0.017832533518473307
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,1024,65536,0.324232546488444
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,1024,65536,0.8820661544799805
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,1024,16384,0.08438613414764404
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,1024,16384,0.229476261138916
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,1024,16384,0.07512746651967367
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,1024,12288,0.06537813345591227
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,1024,12288,0.17122559547424315
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,1536,65536,0.37472426096598305
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,1024,10240,0.0541536013285319
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,1024,10240,0.14048852920532226
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,1024,12288,0.06006186803181966
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,1024,8192,0.04503466685612996
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,1024,8192,0.1111967960993449
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,1024,10240,0.0550495982170105
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,1024,7168,0.03782186508178711
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,1024,7168,0.09303253491719564
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,1024,8192,0.04369920094807943
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,1024,6144,0.03384213447570801
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,1024,6144,0.07996906439463297
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,1024,7168,0.04110720157623291
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,1024,5120,0.028588799635569255
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,1024,5120,0.06375253200531006
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,1024,6144,0.03749973376592
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,1024,4096,0.02358186642328898
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,1024,4096,0.05288639863332113
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,1024,5120,0.03314346671104431
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,1024,3584,0.021221333742141725
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,1024,3584,0.04903573195139567
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,1024,3584,0.02523946762084961
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,1024,3072,0.018569600582122803
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,1024,3072,0.04297600189844768
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,1024,4096,0.027151999870936076
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,1024,2560,0.016057599584261575
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,1024,2560,0.039179734388987225
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,1024,2560,0.019859200716018675
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,1024,3072,0.022258132696151733
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,1024,2048,0.013285332918167114
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,1024,2048,0.03458026647567749
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,1024,1536,0.010829866925875346
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,1024,1536,0.030884265899658203
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,1024,1536,0.014910933375358582
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,1024,1024,0.008376533786455791
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,1024,2048,0.017449599504470826
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,1024,1024,0.02608533302942912
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,1024,768,0.007216000060240428
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,1024,768,0.02472533384958903
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,1024,1024,0.01204159955183665
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,1024,768,0.012388267119725545
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,1024,512,0.005914666752020518
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,1024,512,0.02239039937655131
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,1024,256,0.004694400231043497
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,1024,256,0.018989866971969603
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,1024,512,0.012082133690516155
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,1024,128,0.004265599946180979
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,1024,256,0.011582932869593303
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,1024,128,0.018677333990732826
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,1024,64,0.004177066683769226
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,1024,128,0.011281067132949829
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,1024,64,0.017860267559687296
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,1024,32,0.004223999877770742
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,1024,32,0.016613333423932394
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,768,65536,0.26219199498494467
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,768,65536,0.8546719868977866
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,768,16384,0.0664693315823873
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,768,16384,0.2196298599243164
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,1024,65536,0.25623359680175783
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,768,12288,0.0483456015586853
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,768,12288,0.16398720741271972
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,768,16384,0.07430400053660074
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,768,10240,0.044055465857187906
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,768,10240,0.13421759605407715
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,768,12288,0.05951786835988363
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,768,8192,0.03354986508687337
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,768,8192,0.10328213373819987
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,768,8192,0.04280960162480672
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,768,7168,0.02923626701037089
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,768,7168,0.08726507027943929
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,768,7168,0.03967039982477824
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,768,6144,0.025142399470011394
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,768,6144,0.0722815990447998
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,768,6144,0.034969600041707356
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,768,5120,0.021705599625905354
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,768,5120,0.06000746488571167
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,768,5120,0.032308266560236616
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,768,4096,0.01780479947725932
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,768,4096,0.04962773323059082
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,768,4096,0.02686186631520589
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,768,3584,0.016056533654530844
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,768,3584,0.045201067129770914
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,768,3584,0.02436800003051758
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,768,3072,0.013936000068982443
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,768,3072,0.0408458669980367
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,768,3072,0.021899733940760294
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,768,2560,0.012095999717712403
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,768,2560,0.037058134873708085
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,768,2560,0.019822933276494346
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,768,2048,0.01018773317337036
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,768,2048,0.03280426661173503
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,768,65536,0.25181973775227867
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,768,1536,0.008398933211962382
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,768,2048,0.017026132345199584
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,768,1536,0.028757333755493164
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,768,1024,0.006684799989064534
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,768,1024,0.02480106751124064
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,768,1536,0.014494933684666953
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,768,768,0.005995733539263407
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,768,768,0.02400533358256022
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,768,768,0.012404266993204753
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,768,512,0.005108266572157542
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,768,512,0.021580799420674642
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,768,512,0.011963733037312825
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,768,256,0.004379733403523763
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,768,256,0.01884373426437378
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,768,256,0.011693867047627766
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,768,128,0.003509333233038584
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,768,128,0.016673066218694053
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,768,128,0.011240532994270325
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,768,64,0.0033952000240484873
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,768,64,0.019050665696461997
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,768,32,0.004673066735267639
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,768,32,0.017829332749048868
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,512,65536,0.18335787455240887
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,512,65536,0.8127146402994792
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,3072,2048,0.027506132920583088
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,512,16384,0.04945493141810099
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,512,16384,0.2115327994028727
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,512,65536,0.24869972864786782
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,512,12288,0.03720106681187947
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,512,12288,0.15859413146972656
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,512,12288,0.05871253410975138
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,512,16384,0.07435626983642578
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,512,10240,0.03073173364003499
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,512,10240,0.129367462793986
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,512,8192,0.024362667401631673
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,512,8192,0.09714986483256022
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,512,8192,0.04235733350118001
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,512,7168,0.021514666080474854
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,512,10240,0.05359679857889811
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,512,7168,0.08129920164744059
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,512,6144,0.01801066597302755
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,512,6144,0.06597226858139038
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,512,7168,0.038007465998331706
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,512,5120,0.015737600127855935
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,512,6144,0.035412267843882246
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,512,5120,0.057163735230763756
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,512,4096,0.012995200355847678
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,512,4096,0.04520639975865682
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,512,5120,0.032092799743016556
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,512,3584,0.011844266454378765
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,512,4096,0.026334933439890545
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,512,3584,0.04296640157699585
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,512,3072,0.0108842670917511
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,512,3072,0.037836798032124835
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,512,3584,0.024344533681869507
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,512,2560,0.009265066186587016
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,512,3072,0.021975467602411904
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,512,2560,0.035086933771769205
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,512,2048,0.00790293316046397
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,512,2048,0.03090133269627889
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,512,2560,0.019030400117238364
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,512,1536,0.007109333574771881
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,512,2048,0.016911999384562174
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,512,1536,0.028683733940124512
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,512,1024,0.005527466535568237
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,512,1024,0.023306665817896526
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,512,1536,0.014058666427930197
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,512,768,0.0046965335806210835
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,512,1024,0.012070399522781373
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,512,768,0.022938666741053264
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,512,512,0.004228266576925913
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,512,512,0.020721065998077392
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,512,768,0.012020267049471537
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,512,256,0.0035061334570248926
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,512,256,0.018643200397491455
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,512,256,0.011583999792734782
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,512,128,0.0030752000709374744
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,512,128,0.017014400164286295
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,512,128,0.01123199959595998
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,512,64,0.003032533327738444
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,512,64,0.017845332622528076
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,512,32,0.003078400095303853
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,512,32,0.014727466305096946
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,256,65536,0.11212800343831379
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,256,65536,0.7681280136108398
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,256,65536,0.2500895977020264
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,256,16384,0.037213865915934244
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,256,16384,0.20534720420837402
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,256,16384,0.07354133129119873
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,256,12288,0.027857067187627156
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,256,12288,0.15545280774434406
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,256,12288,0.05899093151092529
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,256,10240,0.021633066733678184
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,256,10240,0.12394026915232341
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,256,10240,0.05220906734466553
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,256,8192,0.018167465925216675
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,256,8192,0.09317119916280112
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,256,8192,0.04120426575342814
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,256,7168,0.01529813309510549
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,256,7168,0.07650346755981445
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,256,7168,0.03784639835357666
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,256,6144,0.013725866874059042
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,256,6144,0.06110506852467855
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,256,6144,0.03460906744003296
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,256,5120,0.012051199873288471
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,256,5120,0.05076266527175903
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,256,5120,0.03172159989674886
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,256,4096,0.00993386705716451
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,256,4096,0.04312533140182495
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,256,4096,0.026293333371480303
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,256,3584,0.009238400061925252
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,256,3584,0.04112639824549357
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,256,3584,0.024053333202997844
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,256,3072,0.008351999521255492
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,256,3072,0.036992001533508304
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,256,3072,0.021460266908009847
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,256,2560,0.007480533421039581
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,256,2560,0.03471893469492594
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,256,2560,0.0192522664864858
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,256,2048,0.006301866471767425
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,256,2048,0.030381866296132404
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,256,2048,0.016590933005015053
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,256,1536,0.005524266759554545
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,256,1536,0.027029333511988322
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,256,1536,0.014502400159835815
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,256,1024,0.004664533336957296
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,256,1024,0.023084799448649086
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,256,1024,0.011703466375668842
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,256,768,0.00421013335386912
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,256,768,0.022808533906936646
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,256,768,0.011703466375668842
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,256,512,0.003500800083080927
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,256,512,0.020722132921218873
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,256,512,0.0116074671347936
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,256,256,0.003028266628583272
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,256,256,0.01867199937502543
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,256,256,0.0116757333278656
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,256,128,0.00262719988822937
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,256,128,0.016777600844701132
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,256,128,0.010870400071144103
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,256,64,0.002570666621128718
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,256,64,0.017001599073410034
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,256,32,0.0026367999613285064
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,256,32,0.01797013282775879
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,128,65536,0.09309759934743246
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,128,65536,0.7638495763142903
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,128,65536,0.24958826700846354
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,128,16384,0.0285045325756073
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,128,16384,0.20130453109741211
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,128,16384,0.07369386355082194
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,128,12288,0.020804266134897866
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,128,12288,0.15122666358947753
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,128,12288,0.05868373314539591
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,128,10240,0.015793066223462424
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,128,10240,0.12178239822387696
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,128,10240,0.052243200937906895
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,128,8192,0.013090133666992188
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,128,8192,0.0893461306889852
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,128,8192,0.04128533204396566
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,128,7168,0.011462400356928509
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,128,7168,0.07377386887868245
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,128,7168,0.037427198886871335
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,128,6144,0.010429867108662923
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,128,6144,0.06011306842168172
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,128,6144,0.034229334195454916
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,128,5120,0.0092031995455424
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,128,5120,0.048987734317779544
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,128,5120,0.03171520034472148
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,128,4096,0.007625600198904674
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,128,4096,0.0417525331179301
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,128,4096,0.025988266865412397
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,128,3584,0.007132799923419952
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,128,3584,0.039690665404001874
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,128,3584,0.023950932423273723
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,128,3072,0.006298666695753734
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,128,3072,0.034962133566538496
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,128,3072,0.02144533395767212
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,128,2560,0.005845333139101664
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,128,2560,0.03295893271764119
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,128,2560,0.018978132804234823
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,128,2048,0.005092266698678335
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,128,2048,0.02903146743774414
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,128,2048,0.016684800386428833
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,128,1536,0.004625066618124644
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,128,1536,0.026807467142740887
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,128,1536,0.014130133390426635
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,128,1024,0.003788800040880839
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,128,1024,0.022742400566736855
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,128,1024,0.011592533191045125
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,128,768,0.003369600077470144
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,128,768,0.022730666399002075
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,128,768,0.011966933806737263
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,128,512,0.003065599997838338
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,128,512,0.020794665813446044
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,128,512,0.01162559986114502
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,128,256,0.0026367999613285064
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,128,256,0.01853013237317403
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,128,256,0.011192533373832702
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,128,128,0.0025600001215934755
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,128,128,0.017257599035898845
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,128,128,0.011158399780591329
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,128,64,0.0024288001159826914
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,128,64,0.017333332697550455
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,128,32,0.002609066665172577
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,128,32,0.015586133797963461
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,64,65536,0.08828372955322265
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,64,65536,0.7611765543619792
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,64,16384,0.025649066766103106
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,64,16384,0.19922560056050617
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,64,12288,0.01730666756629944
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,64,12288,0.14910933176676433
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,64,10240,0.014443733294804893
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,64,10240,0.12129706541697186
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,64,8192,0.012552533547083536
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,64,8192,0.08921919663747152
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,64,7168,0.010994133353233338
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,64,7168,0.07189013163248698
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,64,6144,0.009592533111572266
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,64,6144,0.05773546695709229
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,64,5120,0.008407466610272725
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,64,5120,0.047914667924245195
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,64,4096,0.007158400118350982
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,64,4096,0.041383465131123856
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,64,3584,0.00636053333679835
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,64,3584,0.03909759918848674
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,64,3072,0.005955199897289276
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,64,3072,0.03507200082143148
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,64,2560,0.005471999943256378
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,64,2560,0.03299093246459961
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,64,2048,0.004730666677157084
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,64,2048,0.02900159955024719
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,64,1536,0.004262400170167288
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,64,1536,0.026922667026519777
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,64,1024,0.003492266684770584
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,64,1024,0.022755199670791627
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,64,768,0.0030623999734719592
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,64,768,0.02267199953397115
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,64,512,0.0030080000559488933
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,64,512,0.020593067010243736
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,64,256,0.002644266684850057
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,64,256,0.018266665935516357
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,64,128,0.002540799975395203
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,64,128,0.01667520006497701
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,64,64,0.0022751999398072558
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,64,64,0.016518400112787882
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,64,32,0.0022698665658632917
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,64,32,0.014545067151387533
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,32,65536,0.08627306620279948
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,32,65536,0.7548949559529622
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,32,16384,0.025843199094136553
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,32,16384,0.19794774055480957
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,32,12288,0.01566506624221802
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,32,12288,0.14891626040140787
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,32,10240,0.011178666353225708
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,32,10240,0.12154239813486736
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,32,8192,0.009634133179982502
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,32,8192,0.08941973050435384
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,32,7168,0.008861866593360902
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,32,7168,0.07293546994527181
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,32,6144,0.007964799801508587
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,32,6144,0.058422398567199704
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,32,5120,0.011024000247319539
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,32,5120,0.0491104006767273
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,32,4096,0.009317333499590557
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,32,4096,0.041152000427246094
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,32,3584,0.008612266182899475
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,32,3584,0.03910826841990153
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,32,3072,0.0075989335775375364
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,32,3072,0.034970664978027345
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,32,2560,0.0067007998625437425
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,32,2560,0.03298773368199666
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,32,2048,0.004665599763393402
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,32,2048,0.028919466336568195
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,32,1536,0.004242133100827535
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,32,1536,0.026652799050013228
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,32,1024,0.003499733408292135
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,32,1024,0.022835199038187662
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,32,768,0.003081600119670232
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,32,768,0.022663466135660806
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,32,512,0.0029653333127498626
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,32,512,0.0208512008190155
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,32,256,0.002644266684850057
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,32,256,0.018548266092936198
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,32,128,0.0022709332406520844
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,32,128,0.016635732849438985
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,32,64,0.0022154666483402254
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,32,64,0.014843733112017313
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4096,32,32,0.0022111999491850535
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4096,32,32,0.014833066860834757
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,65536,16384,2.458489481608073
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,65536,16384,1.6145013173421223
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,768,10240,0.05391039848327637
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,65536,12288,2.008713658650716
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,65536,12288,1.2644266764322916
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,768,1024,0.012007466952006022
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,65536,10240,1.4840746561686198
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,65536,10240,0.9689109166463217
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4096,512,512,0.012001066406567892
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,65536,8192,1.2200244903564452
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,65536,8192,0.8185962677001953
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,65536,10240,0.5422538757324219
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,65536,7168,1.0372693379720053
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,65536,7168,0.6618538538614909
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,65536,8192,0.4489386558532715
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,65536,6144,0.8457269032796224
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,65536,6144,0.5068895975748698
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,65536,12288,0.6561130523681641
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,65536,5120,0.7352437337239583
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,65536,5120,0.44263680775960285
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,65536,16384,0.842901357014974
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,65536,4096,0.5739775975545247
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,65536,4096,0.3472309430440267
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,65536,6144,0.3485578536987305
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,65536,3584,0.5044885317484538
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,65536,3584,0.29630934397379555
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,65536,7168,0.3912757237752279
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,65536,3072,0.45121173858642577
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,65536,3072,0.28506558736165366
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,65536,5120,0.3055829366048177
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,65536,2560,0.38313706715901696
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,65536,2560,0.21952853202819825
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,65536,4096,0.250544007619222
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,65536,2048,0.3128085454305013
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,65536,2048,0.17490986188252766
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,65536,3584,0.21460159619649252
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,65536,1536,0.22510827382405602
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,65536,1536,0.14102293650309244
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,65536,3072,0.20175573031107583
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,65536,1024,0.1613525390625
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,65536,1024,0.10852266947428386
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,65536,2560,0.17290453910827636
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,65536,768,0.1271114667256673
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,65536,768,0.09545173645019531
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,65536,1536,0.1240447998046875
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,65536,512,0.08994452953338623
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,65536,512,0.07576426664988199
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,65536,768,0.081822935740153
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,65536,256,0.06362666686375937
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,65536,256,0.0642848014831543
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,65536,1024,0.09498666922251384
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,65536,128,0.052729598681131994
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,65536,128,0.06310293277104696
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,65536,128,0.05067733526229858
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,65536,64,0.049968000253041586
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,65536,64,0.06215893427530924
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,65536,32,0.04853013356526693
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,65536,32,0.061692798137664796
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,65536,2048,0.1474122683207194
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,65536,256,0.05764586528142294
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,16384,65536,1.679093297322591
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,16384,65536,2.6721033732096355
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,16384,16384,0.6796543757120769
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,16384,16384,0.4993141492207845
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,16384,12288,0.4921066602071126
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,16384,12288,0.317137082417806
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,65536,512,0.07230400244394938
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,16384,10240,0.38549760182698567
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,16384,10240,0.2365503946940104
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,16384,10240,0.15732372601826985
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,16384,8192,0.3039999961853027
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,16384,8192,0.19016532897949218
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,16384,12288,0.1891306718190511
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,16384,7168,0.2666133403778076
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,16384,7168,0.17691307067871093
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,16384,16384,0.23097920417785645
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,16384,6144,0.2258240063985189
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,16384,6144,0.14966079394022624
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,16384,8192,0.12853866418202717
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,16384,5120,0.18301013310750325
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,16384,5120,0.13362986246744793
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,16384,7168,0.11872213681538898
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,16384,4096,0.1519477367401123
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,16384,4096,0.10331520239512126
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,16384,5120,0.08917439778645833
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,16384,3584,0.1287722667058309
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,16384,3584,0.09528640111287436
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,16384,6144,0.1032853364944458
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,16384,3072,0.11757866541544597
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,16384,3072,0.0840170701344808
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,16384,4096,0.07532373269399008
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,16384,2560,0.10019520123799641
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,16384,2560,0.0719754695892334
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,16384,2560,0.05325440168380737
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,16384,2048,0.08158186276753744
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,16384,2048,0.06049280166625977
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,16384,3584,0.06730026404062907
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,16384,1536,0.06035733222961426
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,16384,1536,0.05533653497695923
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,16384,3072,0.06123520135879516
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,16384,1024,0.043501865863800046
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,16384,1024,0.04528426726659139
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,16384,1024,0.031496532758076984
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,16384,768,0.033796266714731855
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,16384,768,0.03902399937311808
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,16384,1536,0.03836053212483724
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,16384,512,0.024035199483235677
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,16384,512,0.034732798735300704
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,16384,768,0.027765333652496338
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,16384,256,0.01728640000025431
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,16384,256,0.0309173325697581
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,16384,512,0.023957333962122598
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,16384,128,0.01569919983545939
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,16384,128,0.03195733428001404
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,16384,256,0.021040000518163047
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,16384,64,0.017407999436060587
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,16384,128,0.01895786722501119
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,16384,32,0.018172800540924072
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,16384,64,0.03306986689567566
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,16384,32,0.031880533695220946
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,16384,2048,0.04620586633682251
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,12288,65536,1.3656895955403647
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,12288,65536,2.056217575073242
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,12288,16384,0.46828053792317703
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,12288,16384,0.36771411895751954
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,12288,12288,0.3558453241984049
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,12288,12288,0.250053326288859
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,12288,12288,0.16401920318603516
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,12288,10240,0.3141269365946452
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,12288,10240,0.20892906188964844
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,12288,16384,0.20924480756123862
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,12288,8192,0.24352854092915854
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,12288,8192,0.17033920288085938
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,12288,10240,0.140227206548055
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,12288,7168,0.2159050623575846
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,12288,7168,0.14588479995727538
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,12288,8192,0.11262933413187663
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,12288,6144,0.18561174074808756
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,12288,6144,0.12918293476104736
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,12288,7168,0.10257813135782878
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,12288,5120,0.1554207960764567
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,12288,5120,0.11194346745808918
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,12288,6144,0.08998719851175943
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,12288,4096,0.12085119883219402
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,12288,4096,0.08410773277282715
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,12288,5120,0.07950826485951742
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,12288,3584,0.10514026482899982
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,12288,3584,0.0810378630956014
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,12288,4096,0.06451626618703207
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,12288,3072,0.091102933883667
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,12288,3072,0.06979733308156331
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,12288,3584,0.05892159938812256
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,12288,2560,0.07876160144805908
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,12288,2560,0.061647999286651614
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,12288,3072,0.05266986687978109
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,12288,2048,0.061437865098317467
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,12288,2048,0.0530240019162496
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,12288,2560,0.046299731731414794
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,12288,1536,0.04795946677525838
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,12288,1536,0.047490131855010984
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,12288,2048,0.03946453332901001
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,12288,1024,0.032789333661397295
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,12288,1024,0.039179734388987225
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,12288,1536,0.03393919865290324
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,12288,768,0.026295467217763262
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,12288,768,0.03521386782328288
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,12288,1024,0.027268266677856444
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,12288,512,0.019058134158452353
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,12288,512,0.03141546646753947
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,12288,768,0.02432533303896586
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,12288,256,0.014127999544143677
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,12288,256,0.026650667190551758
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,12288,256,0.01816320021947225
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,12288,128,0.012471466263135274
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,12288,128,0.02518293261528015
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,12288,512,0.02145919998486837
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,12288,128,0.016910932461420693
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,12288,64,0.010449066758155823
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,12288,32,0.010829866925875346
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,12288,64,0.02542826731999715
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,12288,32,0.023255467414855957
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,10240,65536,1.1799306233723958
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,10240,65536,1.46343994140625
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,10240,16384,0.45219628016153973
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,10240,16384,0.3508586565653483
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,10240,16384,0.18178240458170575
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,10240,12288,0.300599479675293
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,10240,12288,0.21828479766845704
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,16384,65536,0.9201514561971029
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,10240,10240,0.26496639251708987
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,10240,10240,0.21185812950134278
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,10240,12288,0.14004480044047038
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,10240,8192,0.21206186612447103
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,10240,8192,0.14618773460388185
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,10240,10240,0.1147274653116862
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,10240,7168,0.19591466585795086
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,10240,7168,0.138701868057251
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,10240,8192,0.09817492961883545
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,10240,6144,0.16245120366414387
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,10240,6144,0.11852906545003254
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,12288,65536,0.770748774210612
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,10240,5120,0.1313098669052124
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,10240,5120,0.09998719692230225
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,10240,7168,0.08896746635437011
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,10240,4096,0.10608212947845459
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,10240,4096,0.08390080134073893
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,10240,6144,0.078165332476298
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,10240,3584,0.09356693426767984
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,10240,3584,0.07995413144429525
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,10240,5120,0.06899413267771402
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,10240,3072,0.081605331103007
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,10240,3072,0.06670719782511393
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,10240,3584,0.05168639818827311
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,10240,2560,0.06842986742655435
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,10240,2560,0.05952320098876953
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,10240,4096,0.05678079922993978
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,10240,2048,0.05672640005747477
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,10240,2048,0.05055573383967081
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,10240,3072,0.046133331457773846
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,10240,1536,0.04324373404184977
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,10240,1536,0.044921600818634035
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,10240,2048,0.03502080043156942
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,10240,1024,0.030689066648483275
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,10240,2560,0.04056959946950277
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,10240,1024,0.03747520049413045
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,10240,768,0.0242741326491038
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,10240,768,0.0335647980372111
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,10240,1024,0.02432960073153178
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,10240,512,0.01761386593182882
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,10240,512,0.02909653385480245
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,10240,1536,0.02960639993349711
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,10240,256,0.012562132875124612
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,10240,256,0.02489173412322998
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,10240,256,0.01616213321685791
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,10240,128,0.01090666651725769
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,10240,128,0.02291839917500814
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,10240,128,0.015337600310643514
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,10240,64,0.009594666957855224
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,10240,64,0.022768000761667885
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,10240,32,0.009879466891288758
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,10240,32,0.02253440022468567
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,10240,512,0.019356799125671387
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,8192,65536,1.2197376251220704
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,8192,65536,1.0107285181681314
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,8192,16384,0.316810671488444
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,8192,16384,0.2617589314778646
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,8192,16384,0.1501322587331136
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,8192,12288,0.235098663965861
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,8192,12288,0.18789440790812176
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,8192,12288,0.11572053432464599
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,8192,10240,0.20266772905985514
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,8192,10240,0.15700693130493165
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,10240,65536,0.6295413335164388
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,8192,8192,0.16174826622009278
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,8192,8192,0.1266208012898763
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,8192,10240,0.10021653175354003
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,8192,7168,0.14578240712483723
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,8192,7168,0.12033066749572754
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,8192,8192,0.08069226741790772
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,8192,6144,0.12252586682637531
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,8192,6144,0.10220373471577962
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,8192,7168,0.07408959865570068
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,8192,5120,0.10328319867451985
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,8192,5120,0.090066130956014
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,8192,6144,0.06511360009511312
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,8192,4096,0.08107413450876871
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,8192,4096,0.065993599096934
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,8192,4096,0.04799893299738566
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,8192,3584,0.0713141361872355
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,8192,3584,0.06170879999796549
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,8192,5120,0.05820586681365967
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,8192,3072,0.06315306822458902
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,8192,3072,0.05594559907913208
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,8192,3584,0.043689600626627606
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,8192,2560,0.05192853212356567
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,8192,2560,0.049621331691741946
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,8192,3072,0.03954133192698161
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,8192,2048,0.04261759916941325
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,8192,2048,0.04305493434270223
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,8192,2560,0.03410559892654419
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,8192,1536,0.03185813426971436
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,8192,1536,0.040166401863098146
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,8192,2048,0.030169600248336793
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,8192,1024,0.023150932788848878
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,8192,1024,0.034993068377176924
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,8192,1024,0.020666666825612388
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,10240,768,0.022243199745814006
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,8192,768,0.019078399737675986
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,8192,768,0.031724800666173295
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,8192,512,0.015244799852371215
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,8192,512,0.029412267605463664
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,8192,1536,0.02521066665649414
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,8192,256,0.01123520036538442
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,8192,256,0.02532800038655599
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,8192,512,0.016714666287104288
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,8192,128,0.009506133198738099
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,8192,128,0.02407360076904297
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,8192,256,0.014526933431625366
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,8192,64,0.009551999966303508
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,8192,768,0.019074134031931558
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,8192,32,0.009988266229629516
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,8192,64,0.022732800245285033
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,8192,32,0.022778666019439696
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,8192,128,0.014084266622861228
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,7168,65536,1.0383605321248373
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,7168,65536,0.9078794479370117
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,7168,16384,0.315939203898112
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,7168,16384,0.21791572570800782
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,7168,12288,0.1969589392344157
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,7168,12288,0.17934080759684246
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,7168,12288,0.09248106479644776
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,7168,10240,0.17211947441101075
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,7168,10240,0.14010880788167318
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,7168,16384,0.11714773178100586
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,7168,8192,0.13542505900065105
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,7168,8192,0.11248213450113934
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,7168,10240,0.08022293249766031
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,7168,7168,0.11866346995035808
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,7168,7168,0.10332159996032715
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,8192,65536,0.5334314982096354
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,7168,6144,0.10167786280314128
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,7168,6144,0.08616106510162354
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,7168,8192,0.06542719999949137
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,7168,5120,0.08401386737823487
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,7168,5120,0.07441173394521078
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,7168,7168,0.06073493162790934
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,7168,4096,0.06867093245188395
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,7168,4096,0.06449919939041138
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,7168,6144,0.054720000425974524
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,7168,3584,0.06120213270187378
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,7168,3584,0.0608842651049296
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,7168,5120,0.04712959925333659
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,7168,3072,0.05116906563440958
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,7168,3072,0.052985600630442296
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,7168,3584,0.03618026574452718
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,7168,2560,0.04465493361155192
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,7168,2560,0.04770239988962809
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,7168,4096,0.04023466507593791
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,7168,2048,0.03591786623001099
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,7168,2048,0.04113599856694539
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,7168,3072,0.03262506723403931
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,7168,1536,0.027085866530736285
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,7168,1536,0.03673386573791504
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,7168,2560,0.029435733954111736
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,7168,1024,0.019874133666356406
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,7168,1024,0.029857067267100017
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,7168,1536,0.02148800094922384
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,7168,768,0.015213867028554281
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,7168,768,0.028663466374079388
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,7168,2048,0.025127466519673663
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,7168,512,0.01179093321164449
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,7168,512,0.024692267179489136
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,7168,1024,0.017734400431315103
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,7168,256,0.00853653351465861
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,7168,256,0.02246933380762736
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,7168,768,0.01611840029557546
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,7168,256,0.014424533645311991
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,7168,128,0.007495466868082683
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,7168,128,0.01985599994659424
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,7168,512,0.014493866761525472
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,7168,64,0.006668800115585327
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,7168,32,0.006737066805362702
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,7168,64,0.020819199085235596
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,7168,32,0.01856106718381246
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,7168,128,0.013300266861915589
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,6144,65536,0.9581941604614258
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,6144,65536,0.8639647801717123
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,6144,16384,0.28425172170003254
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,6144,16384,0.21349973678588868
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,6144,12288,0.18130879402160643
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,6144,12288,0.1580565293629964
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,6144,12288,0.09210986296335856
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,6144,10240,0.16029225985209147
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,6144,10240,0.1366655985514323
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,6144,16384,0.11771946748097736
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,6144,8192,0.12292693456013996
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,6144,8192,0.10442240238189697
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,6144,10240,0.0803541342417399
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,6144,7168,0.11224533716837566
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,6144,7168,0.1002570629119873
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,6144,8192,0.06616426706314087
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,6144,6144,0.09493227005004883
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,6144,6144,0.0818399985631307
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,6144,7168,0.059852798779805504
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,6144,5120,0.08007360299428304
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,6144,5120,0.07592639923095704
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,6144,6144,0.05390186707178751
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,6144,4096,0.06340906620025635
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,6144,4096,0.05956159830093384
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,6144,5120,0.04696000019709269
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,6144,3584,0.05621546506881714
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,6144,3584,0.05650773445765177
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,6144,4096,0.039800532658894855
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,6144,3072,0.048128000895182294
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,6144,3072,0.050964268048604336
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,6144,3584,0.03609493176142375
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,6144,2560,0.04089920123418172
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,6144,2560,0.044225064913431804
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,6144,3072,0.032682667175928756
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,6144,2048,0.0329749325911204
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,6144,2048,0.040009601910909014
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,6144,2560,0.028493867317835493
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,6144,1536,0.025949867566426595
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,6144,1536,0.03627200126647949
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,6144,1536,0.02133973240852356
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,6144,1024,0.018338133891423546
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,6144,1024,0.030938667058944703
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,6144,2048,0.024949334065119424
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,6144,768,0.014789332946141561
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,6144,768,0.02868693272272746
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,6144,1024,0.01737173398335775
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,6144,512,0.012131200234095255
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,6144,512,0.026025599241256712
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,6144,768,0.016132266322771708
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,6144,256,0.009604266285896302
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,6144,256,0.02429973284403483
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,6144,256,0.012429866194725036
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,6144,128,0.009550933043162029
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,6144,128,0.022690133253733317
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,6144,512,0.014114133516947427
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,6144,64,0.008712533116340637
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,6144,64,0.02192853291829427
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,6144,128,0.013006933530171714
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,6144,32,0.008995200196901958
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,6144,32,0.021320533752441407
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,5120,65536,0.8383541107177734
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,5120,65536,0.8397408167521159
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,5120,16384,0.21869014104207357
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,5120,16384,0.19264853795369466
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,7168,65536,0.42779305775960286
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,5120,12288,0.16319360733032226
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,5120,12288,0.14912427266438802
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,5120,16384,0.11807893117268879
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,5120,10240,0.12909226417541503
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,5120,10240,0.11763947010040283
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,5120,12288,0.0918783982594808
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,5120,8192,0.10314666430155436
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,5120,8192,0.092795729637146
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,5120,10240,0.08011199633280436
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,5120,7168,0.09065279960632325
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,5120,8192,0.06575146516164145
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,5120,7168,0.09078933397928873
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,5120,6144,0.07733120123545328
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,5120,6144,0.07584746678670248
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,5120,6144,0.05345813433329264
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,5120,5120,0.06432640155156454
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,5120,5120,0.06422933340072631
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,5120,7168,0.05925120115280151
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,5120,4096,0.052805332342783604
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,5120,4096,0.05403733253479004
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,5120,5120,0.04657919804255168
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,5120,3584,0.045621331532796225
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,5120,3584,0.05184746583302816
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,5120,4096,0.03916693528493245
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,5120,3072,0.04014506737391154
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,5120,3072,0.045737600326538085
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,5120,3584,0.03588800032933553
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,5120,2560,0.03318079908688863
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,5120,2560,0.04154239892959595
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,5120,3072,0.03196053306261699
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,5120,2048,0.027790933847427368
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,5120,2048,0.036738133430480956
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,6144,65536,0.42828054428100587
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,5120,1536,0.02059946656227112
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,5120,1536,0.032945066690444946
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,5120,2560,0.02800319989522298
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,5120,2048,0.02444480061531067
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,5120,1024,0.015204266707102457
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,5120,1024,0.028515199820200603
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,5120,768,0.012145066261291504
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,5120,768,0.026745599508285523
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,5120,1536,0.02095359961191813
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,5120,512,0.00918933351834615
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,5120,512,0.022997333606084188
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,5120,512,0.013838932911554972
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,5120,256,0.007090133428573608
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,5120,256,0.020658133427302043
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,5120,768,0.015629866719245912
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,5120,128,0.0063274666666984555
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,5120,128,0.018636800845464072
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,5120,1024,0.017543466885884602
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,5120,256,0.011988266309102377
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,5120,128,0.01200213332970937
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,5120,32,0.005901866654555003
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,5120,64,0.019321600596110024
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,5120,32,0.018659200270970663
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,4096,65536,0.5892960230509441
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,4096,65536,0.6935818354288737
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,4096,16384,0.15704533259073894
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,4096,16384,0.17057706514994303
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,4096,16384,0.08678613503774008
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,4096,12288,0.12202880382537842
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,4096,12288,0.1295039971669515
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,4096,12288,0.06877120335896811
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,4096,10240,0.10916053454081218
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,4096,10240,0.10541333357493084
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,4096,10240,0.06016106605529785
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,4096,8192,0.08723626931508383
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,4096,8192,0.09019947052001953
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,5120,65536,0.41005652745564775
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,4096,7168,0.07416000366210937
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,4096,7168,0.08092373212178548
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,4096,8192,0.04908373355865479
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,4096,6144,0.06543253262837728
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,4096,6144,0.06785066922505698
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,4096,7168,0.04552746613820394
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,4096,5120,0.053686400254567467
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,4096,5120,0.06001493136088053
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,4096,5120,0.03704640070597331
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,4096,4096,0.0435914675394694
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,4096,4096,0.051565865675608315
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,4096,6144,0.04105493227640788
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,4096,3584,0.038610132535298665
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,4096,3584,0.04805013338724772
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,4096,3584,0.02809813419977824
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,4096,3072,0.03370026747385661
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,4096,4096,0.030479999383290608
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,4096,3072,0.04307093222935994
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,4096,2560,0.027426133553187054
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,4096,2560,0.03957120180130005
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,4096,2560,0.022396800915400188
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,4096,2048,0.023966934283574423
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,4096,2048,0.034942932923634845
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,4096,3072,0.025256532430648803
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,4096,1536,0.01933120091756185
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,4096,1536,0.03171306649843852
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,4096,2048,0.019638399283091225
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,4096,1024,0.01390720009803772
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,4096,1024,0.026771199703216553
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,4096,1536,0.016951467593510947
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,4096,768,0.011655466755231221
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,4096,768,0.024716800451278685
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,4096,1024,0.014477866888046264
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,4096,512,0.00844266712665558
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,4096,768,0.012827733159065246
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,4096,512,0.022669865687688192
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,4096,256,0.006679466863473256
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,4096,256,0.021410133441289267
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,4096,256,0.011236266295115153
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,4096,128,0.006270933151245117
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,4096,128,0.01895466645558675
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,4096,512,0.011992533008257549
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,4096,64,0.006311466793219249
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,4096,64,0.019268266359965005
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,4096,128,0.010426666339238484
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,4096,32,0.006306133170922597
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,4096,32,0.018567466735839845
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,3584,65536,0.5975039800008137
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,3584,65536,0.6881408055623373
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,3584,16384,0.14882879257202147
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,3584,16384,0.1629813353220622
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,4096,65536,0.3082378705342611
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,3584,12288,0.11289280255635578
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,3584,12288,0.12452586491902669
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,3584,16384,0.08730986913045248
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,3584,10240,0.10718719959259033
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,3584,10240,0.1100490649541219
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,3584,12288,0.06791253089904785
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,3584,8192,0.08303679625193278
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,3584,8192,0.08812053203582763
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,3584,10240,0.06003413200378418
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,3584,7168,0.07519466876983642
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,3584,7168,0.0782912015914917
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,3584,7168,0.04490880171457927
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,3584,6144,0.06394559939702352
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,3584,6144,0.06966826915740967
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,3584,8192,0.04863893191019694
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,3584,5120,0.05179200172424316
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,3584,5120,0.05952426592508951
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,3584,5120,0.03533013264338176
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,3584,4096,0.04150079886118571
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,3584,4096,0.04952853520711263
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,3584,6144,0.04033066829045613
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,3584,3584,0.038320000966389975
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,3584,3584,0.048148266474405926
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,3584,3584,0.028104533751805622
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,3584,3072,0.03293866713841756
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,3584,3072,0.043649065494537356
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,3584,4096,0.0297760009765625
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,3584,2560,0.027413332462310792
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,3584,2560,0.039188265800476074
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,3584,3072,0.02478826642036438
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,3584,2048,0.02339199980099996
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,3584,2048,0.03413653373718262
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,3584,2560,0.021946666638056438
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,3584,1536,0.0179967999458313
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,3584,1536,0.030906667311986286
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,3584,1536,0.016964266697565712
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,3584,1024,0.013727999726931252
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,3584,1024,0.026754132906595868
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,3584,2048,0.019770665963490804
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,3584,768,0.010442666212717692
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,3584,768,0.02478613257408142
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,3584,1024,0.0141184002161026
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,3584,512,0.008387200037638346
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,3584,512,0.02226346731185913
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,3584,768,0.012822399536768595
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,3584,256,0.005921066800753275
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,3584,256,0.020407466093699138
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,3584,256,0.010790399710337321
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,3584,128,0.005565866827964783
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,3584,512,0.011596799890200297
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,3584,128,0.01821333368619283
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,3584,64,0.005013333261013031
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,3584,64,0.018381865819295247
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,3584,32,0.00508480022350947
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,3584,128,0.010422399640083313
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,3584,32,0.016691199938456216
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,3072,65536,0.4512309392293294
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,3072,65536,0.6044330596923828
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,3072,16384,0.11822400093078614
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,3072,16384,0.147433598836263
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,3072,16384,0.08761493364969888
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,3072,12288,0.09209173520406087
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,3072,12288,0.11381759643554687
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,5120,64,0.005470933516820272
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,3072,10240,0.08247573375701904
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,3072,10240,0.09194666544596354
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,3072,10240,0.05889706611633301
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,3072,8192,0.06572159926096598
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,3072,8192,0.07805013656616211
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,3072,12288,0.06958719889322916
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,3072,7168,0.057689599196116124
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,3072,7168,0.07011840343475342
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,3584,65536,0.30792214075724283
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,3072,6144,0.05004160006841024
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,3072,6144,0.0616703987121582
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,3072,8192,0.049269334475199385
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,3072,5120,0.040312532583872476
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,3072,5120,0.05337493419647217
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,3072,6144,0.04034239848454793
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,3072,7168,0.04509653250376384
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,3072,4096,0.03356373310089111
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,3072,4096,0.04531306823094686
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,3072,3584,0.02899199922879537
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,3072,3584,0.043150933583577473
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,3072,3584,0.028042666117350262
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,3072,3072,0.025389866034189863
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,3072,3072,0.038998401165008544
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,3072,5120,0.03548053503036499
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,3072,2560,0.021809067328770956
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,3072,2560,0.0351744016011556
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,3072,4096,0.03031253417332967
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,3072,2048,0.01818986733754476
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,3072,2048,0.032282666365305586
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,3072,2560,0.022259199619293214
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,3072,1536,0.014229333400726319
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,3072,1536,0.029814400275548297
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,3072,3072,0.02485333283742269
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,3072,1024,0.0109525332848231
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,3072,1024,0.025254400571187337
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,3072,2048,0.019299199183781944
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,3072,768,0.008957866827646892
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,3072,768,0.024222934246063234
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,3072,768,0.012922666470209756
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,3072,512,0.007192533214886982
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,3072,512,0.022668800751368203
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,3072,1024,0.013751467068990072
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,3072,256,0.006358399987220764
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,3072,256,0.02023893396059672
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,3072,1536,0.01689173380533854
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,3072,128,0.005895466605822245
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,3072,128,0.020196266969045005
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,3072,256,0.010856533050537109
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,3072,64,0.005459199845790863
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,3072,128,0.010357333223025005
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,3072,512,0.011593600114186604
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,3072,64,0.018461867173512777
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,3072,32,0.018931200106938682
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,2560,65536,0.4261472066243489
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,2560,65536,0.5770666758219402
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,2560,16384,0.10479040145874023
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,2560,16384,0.14139307339986165
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,2560,16384,0.08666453361511231
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,2560,12288,0.08304746945699057
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,2560,12288,0.10680320262908935
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,2560,12288,0.06805226802825928
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,2560,10240,0.070413867632548
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,2560,10240,0.08485226631164551
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,3072,65536,0.3034058570861816
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,2560,8192,0.05310826698939005
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,2560,8192,0.0703648010889689
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,2560,10240,0.059918932120005285
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,2560,7168,0.048503466447194415
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,2560,7168,0.06345280011494955
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,2560,8192,0.048070398966471355
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,2560,6144,0.03937919934590657
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,2560,6144,0.0546666661898295
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,2560,7168,0.044123733043670656
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,2560,5120,0.03467306693394979
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,2560,5120,0.05002346833546957
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,2560,6144,0.03946453332901001
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,2560,4096,0.026862933238347368
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,2560,4096,0.04149760007858276
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,2560,5120,0.035369598865509035
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,2560,3584,0.024785067637761435
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,2560,3584,0.04055253267288208
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,2560,4096,0.029793065786361695
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,2560,3072,0.020702934265136717
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,2560,3072,0.03666666746139526
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,2560,3584,0.027214932441711425
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,2560,2560,0.018517333269119262
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,2560,2560,0.033181866010030106
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,2560,3072,0.024399999777475992
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,2560,2048,0.014748799800872802
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,2560,2048,0.02913600007692973
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,2560,2560,0.021926399072011313
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,2560,1536,0.012339199582735699
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,2560,1536,0.028888533512751263
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,2560,1536,0.016567466656366985
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,2560,1024,0.009309867024421692
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,2560,1024,0.02476053237915039
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,2560,2048,0.01904426614443461
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,2560,768,0.007573333382606506
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,2560,768,0.023015467325846355
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,2560,1024,0.014008532961209616
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,2560,512,0.006764799853165944
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,2560,512,0.022056533892949422
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,2560,768,0.012458667159080505
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,2560,256,0.005505066613356272
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,2560,256,0.01925440033276876
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,2560,256,0.010454400380452474
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,2560,128,0.004705066482226053
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,2560,512,0.011577600240707397
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,2560,128,0.017522132396697997
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,2560,128,0.010422399640083313
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,2560,64,0.004636799792448679
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,2560,64,0.01673813263575236
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,2560,32,0.004712533454100291
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,2560,32,0.01664746701717377
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,2048,65536,0.33159465789794923
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,2048,65536,0.5481738408406576
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,2048,16384,0.0822762648264567
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,2048,16384,0.13110400040944417
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,2048,16384,0.05650773445765177
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,2048,12288,0.06234560012817383
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,2048,12288,0.09443093140920003
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,2560,65536,0.30808000564575194
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,2048,10240,0.05395306746164957
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,2048,10240,0.08216959635416667
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,2048,10240,0.03943253358205159
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,2048,8192,0.04354879856109619
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,2048,8192,0.0639903982480367
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,2048,12288,0.04533226490020752
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,2048,7168,0.03736639817555745
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,2048,7168,0.058874666690826416
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,2048,7168,0.030638933181762695
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,2048,6144,0.03244480093320211
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,2048,6144,0.05307519833246867
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,2048,8192,0.033191466331481935
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,2048,5120,0.028495999177296956
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,2048,5120,0.04724586804707845
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,2048,5120,0.025038933753967284
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,2048,4096,0.02375146746635437
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,2048,4096,0.040727468331654866
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,2048,6144,0.027602134148279826
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,2048,3584,0.020804266134897866
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,2048,3584,0.03820586601893107
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,2048,3584,0.019831466674804687
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,2048,3072,0.018234666188557944
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,2048,3072,0.035027201970418295
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,2048,4096,0.021284266312917074
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,2048,2560,0.016337066888809204
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,2048,2560,0.03130773305892944
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,2048,2560,0.016135467092196147
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,2048,2048,0.013567999998728434
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,2048,2048,0.028913066784540815
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,2048,3072,0.018206934134165444
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,2048,1536,0.01112000048160553
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,2048,1536,0.026786132653554277
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,2048,1536,0.01249066690603892
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,2048,1024,0.008360532919565837
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,2048,1024,0.02378666599591573
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,2048,2048,0.014122666915257773
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,2048,768,0.007145600020885467
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,2048,768,0.022959999243418374
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,2048,1024,0.011215999722480774
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,2048,512,0.005913599828879039
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,2048,512,0.02062293291091919
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,2048,768,0.010417067011197408
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,2048,256,0.004702933132648468
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,2048,256,0.019003732999165853
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,2048,256,0.00960106650988261
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,2048,128,0.00425600012143453
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,2048,512,0.009946667154630025
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,2048,128,0.017143466075261436
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,2048,64,0.004228266576925913
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,2048,64,0.016761600971221924
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,2048,128,0.009541333715120951
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,2048,32,0.004192000130812327
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,2048,32,0.016495999693870545
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,1536,65536,0.25625173250834143
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,1536,65536,0.506334940592448
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,1536,16384,0.06686399777730306
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,1536,16384,0.11899733543395996
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,1536,16384,0.05528000195821127
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,1536,12288,0.05109440088272095
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,1536,12288,0.08813013235727946
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,1536,12288,0.043816534678141277
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,1536,10240,0.0428117314974467
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,1536,10240,0.07367359797159831
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,2048,65536,0.1908768018086751
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,1536,8192,0.03455253442128499
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,1536,8192,0.06028480132420858
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,1536,10240,0.03904213507970174
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,1536,7168,0.029643734296162922
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,1536,7168,0.05353600184122721
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,1536,8192,0.032323199510574344
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,1536,6144,0.026764800151189167
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,1536,6144,0.04742613236109416
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,1536,7168,0.029356799523035687
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,1536,5120,0.021824000279108684
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,1536,5120,0.04295253356297811
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,1536,6144,0.026798933744430542
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,1536,4096,0.01874986688296
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,1536,4096,0.03732159932454427
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,1536,5120,0.023989333709081014
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,1536,3584,0.016459733247756958
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,1536,3584,0.03496106863021851
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,1536,4096,0.0213045338789622
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,1536,3072,0.014833066860834757
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,1536,3072,0.0332096000512441
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,1536,3584,0.019033600886662803
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,1536,3072,0.017833600441614784
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,1536,2560,0.01262933313846588
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,1536,2560,0.030823467175165813
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,1536,2560,0.015432533621788026
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,1536,2048,0.01081706682840983
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,1536,2048,0.02749333381652832
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,1536,1536,0.008937600255012512
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,1536,1536,0.025753599405288697
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,1536,1536,0.01209386686484019
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,1536,1024,0.0070816000302632645
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,1536,1024,0.022835199038187662
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,1536,2048,0.014127999544143677
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,1536,768,0.0063498665889104204
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,1536,768,0.02218986749649048
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,1536,768,0.010519466797510783
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,1536,512,0.00553706685702006
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,1536,512,0.021112533410390218
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,1536,512,0.00997759997844696
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,1536,256,0.004586666822433472
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,1536,256,0.017646932601928712
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,1536,256,0.009593600034713745
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,1536,128,0.003870933254559835
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,1536,128,0.01727573275566101
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,1536,128,0.00909546713034312
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,1536,64,0.0036864000062147772
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,1536,64,0.01660480002562205
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,1536,32,0.00383146678407987
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,1536,32,0.016247466206550598
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,1024,65536,0.18356374104817708
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,1024,65536,0.4604757308959961
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,1536,65536,0.18226879437764484
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,1024,16384,0.04660586516062419
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,1024,16384,0.1099669297536214
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,1024,16384,0.05537386735280355
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,1024,12288,0.03556693394978841
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,1024,12288,0.0797322670618693
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,1024,12288,0.04238826831181844
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,1024,10240,0.029531733194986982
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,1024,10240,0.06627626816431681
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,1024,10240,0.038148268063863115
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,1024,8192,0.023995733261108397
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,1024,8192,0.05366186698277792
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,1024,8192,0.03125333388646444
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,1024,7168,0.021496532360712688
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,1024,7168,0.05026026566823324
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,1024,65536,0.1815253257751465
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,1024,6144,0.018888533115386963
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,1024,6144,0.04373546838760376
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,1024,7168,0.02864426573117574
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,1024,5120,0.01606826682885488
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,1024,5120,0.0391103982925415
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,1024,5120,0.023485867182413737
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,1024,4096,0.013639466961224875
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,1024,6144,0.026016000906626386
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,1024,4096,0.03511573473612468
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,1024,3584,0.012134400010108948
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,1024,3584,0.032935466368993124
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,1024,3584,0.018953599532445273
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,1024,3072,0.011026133100191753
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,1024,3072,0.030880000193913775
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,1024,4096,0.020245333512624107
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,1024,2560,0.00941439966360728
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,1024,2560,0.027670399347941084
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,1024,3072,0.017366399367650352
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,1024,2560,0.01532799998919169
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,1024,2048,0.007868800063927968
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,1024,2048,0.025047467152277632
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,1024,1536,0.0067552000284194945
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,1024,1536,0.02479893366495768
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,1024,1536,0.012044800321261089
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,1024,1024,0.005509333312511444
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,1024,2048,0.013636266191800436
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,1024,1024,0.02206826607386271
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,1024,768,0.004980266590913137
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,1024,768,0.020719999074935914
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,1024,1024,0.01076586643854777
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,1024,768,0.010339200496673584
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,1024,512,0.004216533402601878
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,1024,512,0.01925440033276876
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,1024,256,0.0034495999415715536
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,1024,256,0.017119999726613364
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,1024,512,0.009935999910036724
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,1024,128,0.0034143999218940735
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,1024,256,0.009181867043177288
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,1024,128,0.015560533603032431
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,1024,64,0.0029663999875386557
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,1024,64,0.014712533354759217
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,1024,128,0.009124267101287841
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,1024,32,0.003031466652949651
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,1024,32,0.015000533064206442
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,768,65536,0.16110827128092448
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,768,65536,0.43354132970174153
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,768,16384,0.04242453177769979
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,768,16384,0.1076639970143636
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,768,16384,0.05472853183746338
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,768,12288,0.03171520034472148
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,768,12288,0.07740693092346192
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,768,12288,0.04236266613006592
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,768,10240,0.02680319945017497
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,768,10240,0.06360640128453574
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,768,10240,0.037110400199890134
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,768,8192,0.021935999393463135
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,768,8192,0.05327359835306803
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,768,8192,0.031351466973622635
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,768,7168,0.019437867403030395
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,768,7168,0.04743146498998006
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,768,7168,0.02858453392982483
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,768,6144,0.017122133572896322
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,768,6144,0.043229866027832034
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,768,6144,0.026133332649866742
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,768,5120,0.014789332946141561
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,768,5120,0.03906986713409424
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,768,65536,0.18086719512939453
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,768,4096,0.01218986709912618
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,768,4096,0.03357760111490886
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,768,5120,0.023528534173965453
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,768,3584,0.011441066861152649
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,768,3584,0.035131732622782394
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,768,3584,0.018205867211023966
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,768,3072,0.01035093367099762
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,768,4096,0.02020373344421387
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,768,3072,0.03097066680590312
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,768,2560,0.009168000022570292
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,768,2560,0.028988800446192425
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,768,2560,0.01529706617196401
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,768,3072,0.01700586676597595
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,768,2048,0.007964799801508587
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,768,2048,0.02638079921404521
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,768,1536,0.006735999882221222
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,768,1536,0.024485333760579427
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,768,1536,0.011975466211636861
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,768,2048,0.013692800203959146
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,768,1024,0.005514666438102722
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,768,1024,0.022257065773010253
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,768,768,0.005129600067933401
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,768,768,0.020910932620366415
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,768,1024,0.010404266913731893
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,768,512,0.0042698666453361515
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,768,768,0.010371200243631999
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,768,512,0.018759467204411826
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,768,256,0.003818666686614355
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,768,256,0.016555733482042947
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,768,512,0.009590400258700053
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,768,128,0.003433600068092346
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,768,256,0.009116799632708231
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,768,128,0.016521599888801575
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,768,64,0.0033813332517941795
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,768,128,0.009103999535242716
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,768,32,0.0034272000193595886
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,768,64,0.01498133341471354
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,768,32,0.014514133334159851
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,512,65536,0.10094079971313477
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,512,65536,0.4231914520263672
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,512,16384,0.03236266573270162
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,512,16384,0.10060799916585286
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,512,16384,0.0546666661898295
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,512,12288,0.02342080076535543
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,512,12288,0.06979520320892334
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,3072,32,0.005510400235652924
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,512,10240,0.020138667027155558
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,512,10240,0.057358932495117185
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,512,12288,0.04211946725845337
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,512,8192,0.016966400543848674
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,512,8192,0.047602132956186934
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,512,10240,0.036218667030334474
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,512,7168,0.014913066228230795
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,512,7168,0.043859199682871504
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,512,7168,0.028264532486597698
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,512,6144,0.013340800007184347
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,512,6144,0.03922986586888631
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,512,8192,0.030927999814351397
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,512,5120,0.011383466919263204
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,512,5120,0.03587520122528076
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,512,6144,0.025630933046340943
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,512,5120,0.023526400327682495
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,512,4096,0.009707732995351156
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,512,4096,0.03089066743850708
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,512,3584,0.008938666184743245
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,512,3584,0.030713599920272828
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,512,3584,0.0186463991800944
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,512,4096,0.020240000883738198
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,512,3072,0.008074666559696197
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,512,3072,0.0288810670375824
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,512,2560,0.0071370666225751235
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,512,2560,0.026825600862503053
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,512,3072,0.016899200280507405
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,512,2048,0.006306133170922597
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,512,2560,0.014907733599344889
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,512,2048,0.023331199089686075
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,512,1536,0.005493333439032236
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,512,1536,0.02432533303896586
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,512,2048,0.013766400019327798
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,512,1536,0.012124799688657125
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,512,1024,0.004624000191688538
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,512,1024,0.021610667308171592
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,512,768,0.0038965334494908653
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,512,768,0.019245866934458414
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,512,1024,0.010449066758155823
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,512,512,0.0034847999612490333
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,512,512,0.01919040083885193
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,512,768,0.0100000003973643
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,512,512,0.009554133812586466
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,512,256,0.0030613332986831666
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,512,256,0.01665600041548411
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,512,256,0.009149866302808125
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,512,128,0.002595199892918269
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,512,128,0.01622293293476105
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,512,128,0.009155199925104777
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,512,64,0.0026047999660174055
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,512,64,0.014469333489735923
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,512,32,0.002632533262173335
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,512,32,0.01443839967250824
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,256,65536,0.060646398862202966
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,256,65536,0.39812161127726237
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,256,16384,0.018117332458496095
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,256,16384,0.09874880313873291
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,512,65536,0.18068159421284993
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,256,12288,0.017114667097727458
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,256,12288,0.06562453508377075
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,256,16384,0.053352534770965576
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,256,10240,0.015372799833615622
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,256,10240,0.053216000398000084
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,256,12288,0.04083306789398193
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,256,8192,0.012379733721415202
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,256,8192,0.04538453420003255
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,256,10240,0.036245334148406985
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,256,7168,0.011272533734639486
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,256,7168,0.04214186668395996
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,256,8192,0.030501333872477214
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,256,6144,0.010014933347702027
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,256,6144,0.03720959822336833
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,256,7168,0.028650667270024615
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,256,5120,0.008775466680526733
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,256,5120,0.03491306702295939
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,256,6144,0.025616000096003216
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,256,4096,0.007540266712506611
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,256,5120,0.023129600286483764
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,256,4096,0.030850134293238324
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,256,3584,0.006949333349863689
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,256,3584,0.03049066662788391
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,256,4096,0.02023680011431376
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,256,3072,0.00629013329744339
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,256,3584,0.01817493240038554
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,256,3072,0.027613866329193115
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,256,2560,0.005898666878541311
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,256,2560,0.025523199637730913
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,256,3072,0.017002665996551515
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,256,2048,0.005096533397833506
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,256,2560,0.015251200397809347
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,256,2048,0.023015467325846355
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,256,1536,0.004637866715590159
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,256,1536,0.022728532552719116
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,256,2048,0.013537066181500754
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,256,1024,0.0037685332198937735
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,256,1536,0.011988266309102377
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,256,1024,0.02177066604296366
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,256,768,0.00344106654326121
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,256,768,0.019621332486470543
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,256,1024,0.010771200060844421
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,256,512,0.003010133405526479
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,256,512,0.018594133853912353
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,256,768,0.010007466872533162
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,256,256,0.0026335999369621276
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,256,256,0.016594133774439492
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,256,512,0.009665066997210186
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,256,128,0.0025546667476495106
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,256,256,0.009276800354321798
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,256,128,0.014906666676203408
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,256,64,0.002204799900452296
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,256,65536,0.1813759962717692
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,256,128,0.008772266904513042
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,256,64,0.014888532956441245
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,256,32,0.002253866692384084
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,256,32,0.015192533532778421
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,128,65536,0.05079040129979452
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,128,16384,0.012056533495585125
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,128,16384,0.09243840376536051
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,128,65536,0.3923349380493164
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,128,12288,0.012442666292190551
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,128,12288,0.06348906755447388
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,128,16384,0.05303680102030436
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,128,12288,0.04110506772994995
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,128,10240,0.011292800307273865
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,128,10240,0.05328960021336874
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,128,8192,0.012012799580891926
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,128,8192,0.043144532044728595
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,128,10240,0.03619946638743083
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,128,7168,0.010835199554761251
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,128,7168,0.04120426575342814
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,128,8192,0.03049280047416687
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,128,6144,0.009517866373062133
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,128,6144,0.03654400110244751
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,128,7168,0.02802346746126811
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,128,5120,0.008359466989835102
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,128,5120,0.03492799997329712
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,128,6144,0.02595626711845398
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,128,4096,0.007073066631952922
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,128,4096,0.029088000456492107
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,128,5120,0.023092265923817953
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,128,3584,0.006723199784755707
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,128,4096,0.020169599850972494
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,128,3584,0.02883946696917216
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,128,65536,0.1813802719116211
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,128,3584,0.01857173244158427
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,128,3072,0.00595413347085317
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,128,3072,0.027261867125829058
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,128,2560,0.005409066875775656
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,128,2560,0.024734934171040855
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,128,3072,0.01699519952138265
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,128,2048,0.004785066843032837
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,128,2048,0.024277333418528238
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,128,2560,0.015337600310643514
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,128,1536,0.004165333261092504
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,128,1536,0.022667733828226726
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,128,2048,0.013630933562914529
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,128,1024,0.0035018667578697203
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,128,1536,0.012016000350316365
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,128,1024,0.020660267273585
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,128,768,0.003403733422358831
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,128,768,0.019475199778874717
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,128,1024,0.01034773290157318
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,128,512,0.0029674666623274487
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,128,512,0.01807360053062439
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,128,768,0.009992532928784688
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,128,256,0.002712533374627431
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,128,256,0.01656426688035329
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,128,512,0.009516800443331402
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,128,128,0.0022261333962281544
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,128,256,0.009133866429328919
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,128,128,0.008712533116340637
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,128,64,0.002284800012906392
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,128,32,0.0025727999707063037
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,128,64,0.014439466595649719
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,128,32,0.017192532618840538
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,64,65536,0.04543360074361165
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,64,16384,0.014849066734313965
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,64,16384,0.09317973454793295
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,64,12288,0.011997866630554199
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,64,65536,0.3908768018086751
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,64,10240,0.010482133428255717
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,64,12288,0.06160853306452433
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,64,8192,0.009122133255004883
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,64,10240,0.05208213329315185
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,64,7168,0.008795733253161114
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,64,8192,0.04410986502965291
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,64,6144,0.007937066753705342
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,64,7168,0.041169067223866776
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,64,5120,0.00749120016892751
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,64,6144,0.036773331960042316
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,64,4096,0.006677333513895671
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,64,5120,0.03471786578496297
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,64,3584,0.0063274666666984555
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,64,4096,0.02983786662419637
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,64,3072,0.005523199836413065
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,64,3584,0.029190399249394733
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,64,2560,0.0050687998533248905
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,64,3072,0.026826665798823042
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,64,2048,0.004661333560943603
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,64,2560,0.02673706610997518
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,64,1536,0.0042133331298828125
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,64,2048,0.02272426684697469
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,64,1024,0.003409066547950109
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,64,1536,0.022832000255584718
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,64,768,0.003033600002527237
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,64,1024,0.020615466435750327
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,64,512,0.0029696000119050344
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,64,768,0.020460800329844157
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,64,256,0.0026367999613285064
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,64,512,0.018557866414388023
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,64,128,0.0022079999248186748
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,64,256,0.016673066218694053
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,64,64,0.0022272000710169475
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,64,128,0.014521599809328715
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,64,32,0.002163200080394745
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,64,64,0.014595199624697366
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,64,32,0.015246933698654175
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,32,65536,0.045060265064239505
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,32,16384,0.010150399804115296
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,32,16384,0.09263359705607097
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,32,65536,0.3873994509379069
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,32,12288,0.00820479989051819
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,32,10240,0.007526400188604991
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,32,12288,0.06473173300425211
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,32,8192,0.011005866527557372
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,32,10240,0.05152320067087809
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,32,7168,0.010161067048708599
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,32,8192,0.044437332948048906
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,32,6144,0.009310932954152425
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,32,7168,0.04119573434193929
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,32,5120,0.008831999699274699
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,32,6144,0.035572266578674315
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,32,4096,0.007914666831493378
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,32,5120,0.03460906744003296
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,32,3584,0.006265600025653839
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,32,4096,0.029317333300908403
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,32,3072,0.005494399865468343
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,32,3584,0.0290613333384196
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,32,2560,0.005082666873931885
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,32,3072,0.026973867416381837
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,32,2048,0.0042346666256586705
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,32,2560,0.02542399962743123
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,32,2048,0.022759467363357544
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,32,1536,0.0038719999293486277
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,32,1536,0.02304640014966329
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,32,1024,0.003369600077470144
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,32,1024,0.020808533827463786
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,32,768,0.002997333308060964
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,32,768,0.019708800315856933
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,32,512,0.0026186667382717133
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,32,512,0.018642133474349974
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,32,256,0.0025685332715511323
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,32,256,0.016476800044377647
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,32,128,0.002251733342806498
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,32,128,0.015219199657440185
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,32,64,0.0021226666867733
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,32,64,0.014735999703407287
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2048,32,32,0.002131200085083644
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,32,32,0.014539733529090881
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,65536,16384,1.1983744303385415
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,65536,16384,0.683359972635905
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,65536,12288,1.0008874893188477
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,65536,12288,0.6649770736694336
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2048,1536,1024,0.01090773344039917
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,65536,10240,0.7468714396158854
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,65536,10240,0.43717120488484695
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2048,128,128,0.01591146687666575
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,65536,8192,0.6457941055297851
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,65536,8192,0.41533972422281906
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,65536,12288,0.34737812678019203
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,65536,7168,0.5381792068481446
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,65536,7168,0.358900260925293
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,65536,10240,0.2936319986979167
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,65536,6144,0.44903999964396163
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,65536,6144,0.26853758494059243
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,65536,8192,0.24669119517008462
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,65536,5120,0.38994452158610027
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,65536,5120,0.21821333567301432
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,65536,16384,0.46367467244466143
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,65536,4096,0.3064576148986816
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,65536,4096,0.1786677360534668
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,65536,6144,0.18656214078267414
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,65536,3584,0.2640501340230306
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,65536,3584,0.16484373410542805
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,65536,5120,0.16400000254313152
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,65536,3072,0.23716373443603517
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,65536,3072,0.14470507303873698
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,65536,7168,0.21805119514465332
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,65536,2560,0.19662507375081378
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,65536,2560,0.1201738675435384
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,65536,3584,0.11874773502349853
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,65536,2048,0.15434667269388835
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,65536,2048,0.09746987024943034
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,65536,3072,0.10691946347554523
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,65536,1536,0.1281973361968994
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,65536,1536,0.08273599942525228
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,65536,2560,0.09318079948425292
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,65536,1024,0.08707520167032877
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,65536,1024,0.06593706607818603
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,65536,4096,0.13456746737162273
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,65536,768,0.06856746673583984
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,65536,768,0.058305064837137856
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,65536,768,0.04533333381017049
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,65536,512,0.05000213384628296
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,65536,512,0.05020693143208822
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,65536,1024,0.05269866784413656
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,65536,256,0.03368853330612183
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,65536,256,0.047005867958068846
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,65536,256,0.0324181338151296
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,65536,128,0.02955840031305949
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,65536,128,0.03996373414993286
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,65536,512,0.04016960064570109
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,65536,64,0.027433600028355914
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,65536,64,0.03987413247426351
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,65536,32,0.027012266715367633
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,65536,32,0.040956799189249674
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,65536,128,0.0288917342821757
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,16384,65536,0.8478015899658203
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,16384,65536,1.2598837534586589
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,16384,16384,0.37060906092325846
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,16384,16384,0.22025386492411295
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,65536,1536,0.06666773160298665
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,16384,12288,0.24553386370340982
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,16384,12288,0.1697717348734538
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,65536,2048,0.07888960043589274
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,16384,10240,0.20667947133382164
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,16384,10240,0.14340799649556477
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,16384,10240,0.09296212991078695
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,16384,8192,0.16655893325805665
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,16384,8192,0.11077120304107665
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,16384,12288,0.11049599647521972
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,16384,7168,0.1466741402943929
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,16384,7168,0.10430933634440105
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,16384,16384,0.1440351963043213
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,16384,6144,0.1275552034378052
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,16384,6144,0.0951136032740275
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,16384,8192,0.07607680161794027
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,16384,5120,0.10394666989644367
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,16384,5120,0.08027946949005127
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,16384,7168,0.0700704018274943
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,16384,4096,0.09326079686482748
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,16384,4096,0.06425493160883586
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,16384,4096,0.045612800121307376
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,16384,3584,0.08090026378631592
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,16384,3584,0.06346133152643839
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,16384,5120,0.054058667023976645
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,16384,3072,0.06680959860483805
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,16384,3072,0.05234346787134806
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,16384,6144,0.06231679916381836
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,16384,2560,0.05796480178833008
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,16384,2560,0.04956479867299397
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,16384,2560,0.033470932642618814
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,16384,2048,0.04730666478474935
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,16384,2048,0.04317973454793294
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,16384,3584,0.04152213335037232
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,16384,1536,0.034663466612497966
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,16384,1536,0.037646933396657305
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,16384,3072,0.03708373308181763
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,16384,1024,0.02334400018056234
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,16384,1024,0.033139199018478394
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,16384,1024,0.020259199539820354
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,16384,768,0.019179733594258626
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,16384,768,0.029009066025416058
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,16384,1536,0.025167999664942424
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,16384,512,0.01556373337904612
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,16384,512,0.027050666014353436
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,16384,768,0.01781546672185262
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,16384,256,0.011632000406583149
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,16384,256,0.024412800868352257
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,16384,256,0.013260799646377563
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,16384,128,0.009199999769528707
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,16384,128,0.022653865814208984
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,16384,128,0.01246399978796641
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,16384,64,0.009591466188430786
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,16384,512,0.01575040022532145
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,16384,64,0.022689066330591836
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,16384,32,0.009998933474222819
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,16384,32,0.022593067089716593
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,12288,65536,0.7038431803385417
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,12288,65536,1.014251708984375
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,16384,2048,0.02890133261680603
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,12288,16384,0.26845760345458985
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,12288,16384,0.18021012941996256
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,12288,12288,0.19889492988586427
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,12288,12288,0.12988266944885254
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,12288,12288,0.08697066307067872
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,12288,10240,0.16014186541239422
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,12288,10240,0.11863466898600261
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,12288,16384,0.11135359605153401
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,12288,8192,0.12606720129648846
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,12288,8192,0.09224747021993002
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,12288,10240,0.07490666707356772
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,12288,7168,0.11300053596496581
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,12288,7168,0.0822101354598999
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,12288,8192,0.06066879828770956
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,12288,6144,0.0976202646891276
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,12288,6144,0.07579733530680338
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,12288,7168,0.05601919889450073
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,12288,5120,0.0808288017908732
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,12288,5120,0.06567253271738688
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,12288,6144,0.049211732546488446
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,12288,4096,0.06716586748758951
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,12288,4096,0.05215679804484049
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,12288,5120,0.04421439965565999
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,12288,3584,0.05760960181554159
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,12288,3584,0.05290133158365885
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,12288,4096,0.036448001861572266
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,12288,3072,0.05036160151163736
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,12288,3072,0.04768213431040446
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,12288,3584,0.03359359900156657
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,12288,2560,0.041866668065388996
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,12288,2560,0.04328320026397705
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,12288,3072,0.030125866333643597
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,12288,2048,0.034808532396952314
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,12288,2048,0.03788906733194987
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,12288,2560,0.027636265754699706
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,12288,1536,0.026732800404230754
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,12288,1536,0.03392639954884847
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,12288,1536,0.020364799102147422
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,12288,1024,0.018910932540893554
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,12288,1024,0.029767467578252153
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,12288,2048,0.023755733172098795
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,12288,768,0.015415466825167336
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,12288,768,0.027005867163340254
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,12288,768,0.015129599968592325
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,12288,512,0.01242026686668396
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,12288,512,0.024915200471878052
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,12288,1024,0.016552533706029257
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,12288,256,0.009664000074068705
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,12288,256,0.023449599742889404
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,12288,256,0.01158186693986257
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,12288,128,0.009268266956011455
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,12288,128,0.022087466716766358
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,12288,128,0.01246506671110789
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,12288,64,0.008776533603668212
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,12288,64,0.021065600713094077
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,12288,32,0.00883840024471283
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,12288,32,0.023149865865707397
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,12288,512,0.01325866679350535
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,10240,65536,0.8477951685587565
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,10240,65536,0.6437536239624023
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,10240,16384,0.23143572807312013
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,10240,16384,0.160696538289388
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,10240,16384,0.1120469331741333
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,10240,12288,0.17423572540283203
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,10240,12288,0.11928533713022868
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,10240,12288,0.0861728032430013
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,10240,10240,0.14541865984598795
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,10240,10240,0.10452799797058106
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,16384,65536,0.5173642794291179
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,10240,8192,0.1120479981104533
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,10240,8192,0.08163413206736246
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,10240,10240,0.0741546630859375
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,10240,7168,0.1032960017522176
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,10240,7168,0.07986453374226889
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,10240,8192,0.06125226815541586
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,10240,6144,0.08724479675292969
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,10240,6144,0.0689301331837972
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,12288,65536,0.4075167973836263
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,10240,5120,0.07002027034759521
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,10240,5120,0.05929493506749471
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,10240,7168,0.05561813513437906
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,10240,4096,0.05815253257751465
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,10240,4096,0.05137493213017782
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,10240,5120,0.042762664953867595
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,10240,3584,0.04877440134684245
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,10240,3584,0.04733653465906779
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,10240,6144,0.04953386783599854
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,10240,3072,0.0434005339940389
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,10240,3072,0.04466773271560669
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,10240,3584,0.03320106665293376
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,10240,2560,0.03566933472951253
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,10240,2560,0.03994880119959514
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,10240,3072,0.030100266138712566
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,10240,2048,0.03023359974225362
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,10240,2048,0.03524479866027832
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,10240,4096,0.03623893260955811
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,10240,1536,0.02379306753476461
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,10240,1536,0.033590400218963624
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,10240,2560,0.02686400016148885
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,10240,1024,0.016939733425776163
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,10240,1536,0.02067199945449829
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,10240,1024,0.026758400599161784
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,10240,768,0.013296000162760415
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,10240,768,0.025183999538421632
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,10240,2048,0.023710934321085612
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,10240,512,0.010311466455459595
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,10240,512,0.02181866765022278
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,10240,768,0.014071466525395713
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,10240,256,0.007909333209196727
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,10240,256,0.021497599283854165
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,10240,1024,0.01616426706314087
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,10240,128,0.006739200154940288
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,10240,512,0.013229866822560629
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,10240,128,0.01919999917348226
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,10240,64,0.006301866471767425
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,10240,64,0.018669867515563966
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,10240,32,0.006301866471767425
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,10240,32,0.01917866667111715
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,10240,128,0.010423466563224792
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,8192,65536,0.6437706629435221
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,8192,65536,0.5106783866882324
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,8192,16384,0.15721707344055175
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,8192,16384,0.1339135964711507
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,8192,16384,0.08014506498972575
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,8192,12288,0.1222431977589925
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,8192,12288,0.10291199684143067
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,8192,12288,0.06363840103149414
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,8192,10240,0.10857813358306885
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,8192,10240,0.09286613464355468
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,10240,65536,0.39949013392130533
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,8192,8192,0.08502293427785237
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,8192,8192,0.07256853580474854
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,8192,10240,0.05502506494522095
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,8192,7168,0.0765397310256958
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,8192,7168,0.07085653146107992
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,8192,8192,0.04504213333129883
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,8192,6144,0.06521173318227133
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,8192,6144,0.06143146753311157
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,8192,6144,0.037298134962717694
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,8192,5120,0.05426880121231079
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,8192,5120,0.05134826501210531
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,8192,7168,0.04146666526794433
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,8192,4096,0.044385067621866864
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,8192,4096,0.04601066509882609
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,8192,4096,0.028257066011428834
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,8192,3584,0.039637335141499835
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,8192,5120,0.03294186592102051
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,8192,3584,0.04335466623306274
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,8192,3072,0.03446400165557861
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,8192,3072,0.04063253402709961
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,8192,3072,0.023176532983779908
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,8192,2560,0.029475200176239013
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,8192,2560,0.037222401301066084
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,8192,3584,0.026045866807301837
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,8192,2048,0.024280534187952677
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,8192,2048,0.03308373292287191
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,8192,2048,0.01867093245188395
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,8192,1536,0.019299199183781944
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,8192,1536,0.02954240043958028
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,8192,2560,0.021090133984883627
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,8192,1024,0.014826666315396628
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,8192,1024,0.02531733314196269
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,8192,1536,0.01656426688035329
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,8192,768,0.011011200149854024
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,8192,768,0.025282132625579833
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,8192,1024,0.01304533382256826
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,8192,512,0.009224533041318258
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,8192,512,0.022808533906936646
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,8192,512,0.010920533537864685
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,8192,256,0.007869866490364075
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,8192,256,0.022115200757980347
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,8192,768,0.012088533242543538
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,8192,128,0.005946666498978933
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,8192,128,0.020054399967193604
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,8192,128,0.00960106650988261
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,8192,64,0.006461866696675618
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,8192,64,0.018684800465901694
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,8192,32,0.0067114666104316715
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,8192,256,0.009987200299898785
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,10240,256,0.010849066575368245
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,7168,65536,0.5362666447957356
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,7168,65536,0.4614261309305827
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,7168,16384,0.13334506352742512
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,7168,16384,0.11241707007090251
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,7168,16384,0.07801173528035482
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,7168,12288,0.1005247990290324
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,7168,12288,0.09011627038319905
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,8192,65536,0.28865172068277994
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,7168,10240,0.08832000096638998
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,7168,10240,0.08558826446533203
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,7168,12288,0.06194666624069214
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,7168,8192,0.06836586793263753
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,7168,8192,0.0685856024424235
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,7168,10240,0.05387733379999796
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,7168,7168,0.06313173373540243
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,7168,7168,0.06565119822820029
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,7168,7168,0.04069226582845052
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,7168,6144,0.061826133728027345
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,7168,6144,0.05834559996922811
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,7168,8192,0.04419306516647339
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,7168,5120,0.04957333405812581
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,7168,5120,0.049975466728210446
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,7168,5120,0.03254186709721883
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,7168,4096,0.03961919943491618
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,7168,4096,0.04255466858545939
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,7168,6144,0.03629013299942017
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,7168,3584,0.03589013417561849
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,7168,3584,0.04118826786677043
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,7168,3584,0.025572266181310015
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,7168,3072,0.031038933992385866
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,7168,3072,0.038123734792073566
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,7168,4096,0.02768426736195882
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,7168,2560,0.02608426610628764
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,7168,2560,0.034780800342559814
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,7168,2560,0.02097919980684916
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,7168,2048,0.021112533410390218
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,7168,2048,0.030896000067392987
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,7168,3072,0.022871466477711995
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,7168,1536,0.017413334051767985
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,7168,1536,0.029029333591461183
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,7168,1536,0.01606826682885488
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,7168,1024,0.012682666381200155
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,7168,1024,0.024199465910593666
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,7168,2048,0.018281600872675576
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,7168,768,0.00978559950987498
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,7168,768,0.02279040018717448
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,7168,768,0.011692800124486287
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,7168,512,0.007572266459465027
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,7168,512,0.020521599054336547
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,7168,1024,0.012902399897575379
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,7168,256,0.005846400062243144
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,7168,256,0.01854506731033325
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,7168,256,0.009976533055305482
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,7168,128,0.005115733544031779
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,7168,128,0.016809600591659545
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,7168,512,0.010418132940928141
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,7168,64,0.0046634669105211895
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,7168,64,0.017059199015299478
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,7168,32,0.004642133414745331
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,7168,32,0.016646400094032288
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,7168,128,0.00918826659520467
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,6144,65536,0.4767168045043945
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,6144,65536,0.4450709342956543
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,6144,16384,0.1272885322570801
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,6144,16384,0.11231466929117839
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,6144,16384,0.07882560094197591
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,6144,12288,0.09903039932250976
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,6144,12288,0.08857386906941732
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,8192,32,0.019988266626993816
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,6144,10240,0.0836575984954834
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,6144,10240,0.07698667049407959
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,6144,10240,0.05329279899597168
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,6144,8192,0.06714560190836588
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,6144,8192,0.06396053234736124
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,6144,12288,0.06229973236719767
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,6144,7168,0.06032533248265585
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,6144,7168,0.059723734855651855
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,7168,65536,0.28519573211669924
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,6144,6144,0.0492298682530721
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,6144,6144,0.05064853429794312
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,6144,8192,0.044898132483164474
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,6144,5120,0.04169066747029622
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,6144,5120,0.045553068319956466
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,6144,7168,0.0404639999071757
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,6144,4096,0.03418026765187581
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,6144,4096,0.03949013153711955
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,6144,6144,0.03643626769383748
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,6144,3584,0.03014613389968872
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,6144,3584,0.03859093189239502
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,6144,5120,0.032553599278132124
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,6144,3072,0.026284799973169966
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,6144,3072,0.03605333169301351
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,6144,3584,0.025572266181310015
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,6144,2560,0.022374399503072104
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,6144,2560,0.03313279946645101
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,6144,4096,0.027739733457565308
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,6144,2048,0.018564265966415406
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,6144,2048,0.03060800035794576
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,6144,2048,0.018212266763051353
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,6144,1536,0.01495680014292399
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,6144,3072,0.022680532932281495
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,6144,1536,0.027324799696604413
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,6144,1024,0.01125973363717397
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,6144,2560,0.020424532890319824
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,6144,1024,0.025090134143829344
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,6144,768,0.009237333138783773
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,6144,768,0.02243306636810303
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,6144,768,0.011662933230400085
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,6144,512,0.007632000247637431
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,6144,512,0.022210133075714112
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,6144,1024,0.012880000472068786
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,6144,256,0.006658133367697399
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,6144,256,0.019661867618560792
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,6144,256,0.009595732887585957
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,6144,512,0.010763733585675558
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,6144,128,0.005916800101598104
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,6144,128,0.018497065703074137
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,6144,64,0.005480533341566721
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,6144,64,0.0187008003393809
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,6144,32,0.0055189331372578945
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,6144,1536,0.016165332992871602
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,6144,32,0.018619734048843383
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,6144,128,0.009139200051625569
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,5120,65536,0.4145119984944661
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,5120,65536,0.3978517214457194
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,5120,16384,0.10561280250549317
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,5120,16384,0.10324266751607258
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,5120,12288,0.08199893633524577
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,5120,12288,0.08171306451161703
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,5120,12288,0.06256426572799682
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,5120,10240,0.07146026293436686
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,5120,10240,0.06730026404062907
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,5120,16384,0.07995413144429525
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,5120,8192,0.05567786693572998
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,5120,8192,0.057384534676869714
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,5120,10240,0.05392746527989706
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,5120,7168,0.049634134769439696
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,5120,7168,0.053572265307108555
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,5120,8192,0.04489386479059855
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,5120,6144,0.04171200195948283
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,5120,6144,0.047622398535410566
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,5120,7168,0.040705064932505294
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,5120,5120,0.035895466804504395
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,5120,5120,0.042019200325012204
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,5120,6144,0.036559998989105225
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,5120,4096,0.02832319935162862
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,5120,4096,0.03717439969380697
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,5120,5120,0.03227413296699524
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,5120,3584,0.025587199131647746
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,5120,3584,0.03590826590855916
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,5120,4096,0.02758293350537618
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,5120,3072,0.02248106598854065
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,5120,3072,0.03515413204828898
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,5120,3584,0.025520000855127973
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,5120,2560,0.019157334168752035
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,5120,2560,0.031378134091695147
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,5120,3072,0.022741333643595377
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,5120,2048,0.01597760021686554
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,5120,2048,0.028947200377782183
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,5120,2560,0.020304000377655028
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,5120,1536,0.013154133160909017
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,5120,1536,0.02674773335456848
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,6144,65536,0.2822538693745931
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,5120,1024,0.009242666761080424
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,5120,1024,0.022657066583633423
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,5120,1536,0.015757866700490317
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,5120,768,0.008030933141708375
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,5120,768,0.02174293398857117
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,5120,2048,0.017777067422866822
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,5120,512,0.006694399813810985
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,5120,512,0.020457599560419717
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,5120,1024,0.012430933117866517
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,5120,256,0.0054666668176651
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,5120,256,0.018362667163213095
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,5120,768,0.011277866363525391
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,5120,128,0.0050335998336474095
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,5120,128,0.01660693287849426
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,5120,512,0.009995733698209126
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,5120,64,0.0046517332394917805
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,5120,256,0.009543466567993163
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,5120,64,0.016582399606704712
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,5120,32,0.0046304002404212955
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,5120,128,0.00918933351834615
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,5120,32,0.01657919983069102
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,4096,65536,0.33194986979166663
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,4096,65536,0.3483872095743815
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,4096,16384,0.08247466882069907
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,4096,16384,0.08660693168640136
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,4096,12288,0.061368532975514735
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,4096,12288,0.06961920261383056
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,4096,12288,0.03794346650441487
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,4096,10240,0.051312001546223965
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,4096,10240,0.06162773370742798
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,4096,16384,0.04873600006103516
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,4096,8192,0.0424949328104655
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,4096,8192,0.05068266789118449
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,4096,10240,0.033872000376383465
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,4096,7168,0.03653333187103271
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,4096,7168,0.04814293384552002
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,4096,8192,0.028511999050776164
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,4096,6144,0.03277653257052104
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,4096,6144,0.043169065316518145
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,4096,7168,0.026498132944107057
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,4096,5120,0.02699306607246399
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,4096,5120,0.03823466698328654
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,4096,6144,0.02447999914487203
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,4096,4096,0.022776534159978233
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,4096,4096,0.03489919900894165
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,4096,5120,0.021938133239746093
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,4096,3584,0.019938133160273232
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,4096,3584,0.03368639945983887
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,4096,4096,0.018696532646814982
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,4096,3072,0.017854932943979898
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,4096,3072,0.03181013266245524
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,4096,3584,0.017848533391952515
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,4096,2560,0.015372799833615622
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,4096,2560,0.028934399286905926
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,4096,3072,0.015801599621772765
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,4096,2048,0.01322773297627767
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,4096,2048,0.026987733443578084
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,4096,2560,0.014089600245157877
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,4096,1536,0.01051093339920044
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,4096,1536,0.024856533606847128
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,4096,2048,0.012918399771054587
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,4096,1024,0.008402132987976074
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,4096,1024,0.02265066703160604
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,4096,1536,0.0116074671347936
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,4096,768,0.007097599903742473
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,4096,768,0.020924800634384157
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,4096,768,0.009192533294359843
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,4096,512,0.00592853327592214
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,4096,1024,0.009588266412417095
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,4096,512,0.01856213410695394
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,4096,256,0.004717866579691568
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,4096,256,0.016557866334915163
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,4096,256,0.007955199976762136
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,4096,128,0.004211199780305227
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,4096,128,0.016506666938463845
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,4096,512,0.008753066261609394
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,4096,64,0.003857066730658213
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,4096,64,0.01648640036582947
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,4096,32,0.0038474666575590765
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,4096,128,0.007904000083605449
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,4096,32,0.016531200210253397
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,3584,65536,0.308185609181722
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,3584,65536,0.3694730758666992
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,3584,16384,0.08392106691996257
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,3584,16384,0.09089279969533284
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,5120,65536,0.2869290669759115
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,3584,12288,0.06411306858062744
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,3584,12288,0.07277546723683676
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,3584,16384,0.04772053162256877
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,3584,10240,0.053541334470113125
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,3584,10240,0.06373013257980346
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,3584,12288,0.03778986533482869
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,3584,8192,0.0434826652208964
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,3584,8192,0.050994133949279784
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,3584,8192,0.028262400627136232
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,3584,10240,0.033973332246144614
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,3584,7168,0.038660267988840744
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,3584,7168,0.048786131540934245
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,3584,6144,0.03303146759668986
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,3584,6144,0.04392106533050537
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,3584,6144,0.023891200621922813
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,3584,5120,0.028141866127649944
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,3584,5120,0.039246932665507
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,3584,7168,0.02621866663297017
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,3584,4096,0.023240532477696738
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,3584,4096,0.0347541332244873
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,3584,5120,0.021524266401926676
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,3584,3584,0.020730666319529214
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,3584,3584,0.03391893307367961
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,3584,4096,0.01862506667772929
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,3584,3072,0.018407466014226277
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,3584,3072,0.030933332443237305
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,4096,65536,0.16684266726175945
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,3584,2560,0.016301866372426352
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,3584,2560,0.03086400032043457
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,3584,3584,0.01737920045852661
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,3584,2048,0.013691733280817667
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,3584,3072,0.015726932883262636
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,3584,2048,0.02688213388125102
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,3584,2560,0.013749333222707114
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,3584,1536,0.01123413344224294
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,3584,1536,0.025209599733352663
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,3584,1024,0.00855573316415151
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,3584,1024,0.02291413346926371
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,3584,1536,0.011262933413187664
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,3584,768,0.0071712002158164975
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,3584,768,0.02057066758473714
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,3584,2048,0.012814933061599731
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,3584,512,0.00591893345117569
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,3584,512,0.018987733125686645
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,3584,1024,0.009618133306503296
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,3584,256,0.004659200211366018
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,3584,256,0.018780799706776936
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,3584,768,0.009134933352470398
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,3584,128,0.0038730666041374207
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,3584,128,0.016353066762288412
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,3584,256,0.008350933591524761
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,3584,64,0.0034815999368826545
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,3584,128,0.007980800171693166
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,3584,64,0.015754666924476624
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,3584,32,0.0038453333079814913
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,3584,512,0.008778666456540424
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,3584,32,0.01683733264605204
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,3072,65536,0.2601109345753988
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,3072,65536,0.33753280639648436
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,3072,16384,0.06665600140889485
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,3072,16384,0.07989866733551025
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,3072,12288,0.05018666585286459
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,3072,12288,0.060083198547363284
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,3072,12288,0.03760106563568115
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,3072,10240,0.04209173520406087
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,3072,10240,0.05596693356831869
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,3072,16384,0.04740906556447347
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,3072,8192,0.034429868062337235
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,3072,8192,0.045349331696828206
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,3072,10240,0.03259306748708089
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,3072,7168,0.030117332935333252
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,3072,7168,0.04421866734822591
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,3072,8192,0.028458666801452637
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,3072,6144,0.026308266321818034
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,3072,6144,0.040856532255808514
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,3072,7168,0.02571093241373698
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,3072,5120,0.022356265783309938
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,3072,5120,0.03739413420359294
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,3072,6144,0.023177599906921385
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,3072,4096,0.018604799111684164
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,3072,4096,0.03131519953409831
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,3072,5120,0.02114880084991455
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,3072,3584,0.01639359990755717
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,3072,3584,0.03110719919204712
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,3072,4096,0.01835626761118571
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,3072,3072,0.014876799782117209
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,3072,3072,0.029153066873550414
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,3072,3584,0.01662613352139791
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,3072,2560,0.012813867131868998
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,3072,2560,0.02765546639760335
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,3072,3072,0.015392000476519266
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,3072,2048,0.011032533645629884
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,3072,2048,0.025873066981633504
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,3584,65536,0.15881279309590657
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,3072,1536,0.008795733253161114
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,3072,1536,0.023432532946268715
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,3072,2560,0.013659733533859252
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,3072,1024,0.007068799932797749
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,3072,1024,0.020734934012095134
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,3072,1536,0.010826667149861652
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,3072,768,0.0063360000650088
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,3072,768,0.019091200828552247
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,3072,1024,0.009522133072217305
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,3072,512,0.0051136001944541935
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,3072,512,0.019065600633621217
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,3072,2048,0.012483200430870056
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,3072,256,0.00469760000705719
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,3072,256,0.0182805339495341
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,3072,512,0.00870293378829956
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,3072,128,0.00421013335386912
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,3072,256,0.007960533102353413
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,3072,768,0.008772266904513042
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,3072,128,0.016698666413625083
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,3072,64,0.003824000060558319
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,3072,32,0.003869866579771042
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,3072,64,0.01581760048866272
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,3072,32,0.01672853430112203
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,3072,128,0.00790826678276062
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,2560,65536,0.21719679832458497
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,2560,16384,0.056170666217803956
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,2560,65536,0.31175572077433267
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,2560,16384,0.07426986694335938
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,2560,12288,0.040728533267974855
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,2560,12288,0.05547626813252767
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,2560,12288,0.036722131570180255
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,2560,10240,0.03530559937159221
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,2560,10240,0.05178133249282837
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,2560,16384,0.046298666795094805
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,2560,8192,0.027382399638493853
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,2560,8192,0.04254293441772461
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,2560,10240,0.03261760075887044
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,2560,7168,0.025756800174713136
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,2560,7168,0.04147306680679321
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,2560,8192,0.02731200059254964
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,2560,6144,0.02135573426882426
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,2560,6144,0.036907732486724854
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,2560,7168,0.02521066665649414
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,2560,5120,0.018946133057276406
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,2560,5120,0.03370453516642253
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,2560,6144,0.02309760053952535
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,2560,4096,0.015255467096964518
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,2560,4096,0.02992960015932719
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,2560,4096,0.018173867464065553
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,2560,3584,0.013814399639765421
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,2560,3584,0.029509333769480388
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,2560,5120,0.02086720069249471
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,2560,3072,0.012780800461769104
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,2560,3072,0.026919466257095338
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,2560,3584,0.0165994664033254
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,2560,2560,0.010889599720637005
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,2560,3072,0.014932266871134438
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,2560,2560,0.026324266195297243
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,2560,2048,0.009193600217501322
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,2560,2048,0.0247925341129303
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,2560,2048,0.0124917338291804
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,2560,1536,0.0076330666740735366
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,2560,2560,0.013676800330479941
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,2560,1536,0.02259733279546102
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,2560,1024,0.006259199976921081
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,2560,1024,0.02038080096244812
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,2560,1024,0.009586133559544881
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,2560,768,0.005570133527119955
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,2560,768,0.018604799111684164
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,2560,1536,0.01128000020980835
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,2560,512,0.004680533210436503
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,2560,512,0.01843520005544027
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,2560,768,0.008966400225957235
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,2560,256,0.004204800228277842
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,2560,512,0.008398933211962382
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,2560,256,0.016546133160591125
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,2560,128,0.00346666673819224
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,2560,128,0.01662720044453939
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,2560,256,0.008004266520341237
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,2560,64,0.0034048000971476236
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,2560,128,0.00791786660750707
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,2560,32,0.0033930666744709016
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,2560,64,0.015245866775512696
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,2560,32,0.016661333044370015
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,2048,65536,0.1825503985087077
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,2048,65536,0.286570676167806
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,2048,16384,0.04589866797129313
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,2048,16384,0.06902720133463541
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,3072,65536,0.15766506195068358
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,2048,12288,0.03510080178578694
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,2048,12288,0.05150400002797445
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,2048,12288,0.036322132746378584
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,2048,10240,0.029942399263381957
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,2048,10240,0.04762666622797648
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,2048,16384,0.045349331696828206
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,2048,8192,0.02364586591720581
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,2048,8192,0.0392192006111145
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,2048,10240,0.032384000221888226
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,2048,7168,0.020643200476964316
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,2048,7168,0.039187200864156085
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,2048,8192,0.02694186568260193
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,2048,6144,0.017760000626246133
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,2048,6144,0.034984532992045084
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,2048,7168,0.025169066588083905
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,2048,5120,0.01572266618410746
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,2048,5120,0.03151786724726359
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,2048,6144,0.02276479999224345
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,2048,4096,0.012710400422414145
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,2048,4096,0.027845333019892376
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,2048,5120,0.020692267020543418
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,2048,3584,0.011723732948303223
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,2048,3584,0.027640533447265626
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,2048,4096,0.017460266749064125
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,2048,3072,0.01023466686407725
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,2048,3072,0.02677759925524394
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,2048,3584,0.016581333676973977
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,2048,2560,0.009041066964467366
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,2048,2560,0.02675519982973735
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,2048,3072,0.014988799889882406
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,2048,2048,0.007924266656239827
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,2048,2048,0.022658133506774904
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,2560,65536,0.1568138599395752
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,2048,1536,0.006504533191521962
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,2048,1536,0.02177493373552958
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,2048,2560,0.013680000106493631
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,2048,1024,0.005464533468087515
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,2048,1024,0.02044266661008199
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,2048,2048,0.012030933300654094
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,2048,768,0.0046741331617037455
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,2048,768,0.018888533115386963
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,2048,1536,0.010843732953071594
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,2048,512,0.00425493319829305
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,2048,512,0.01728640000025431
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,2048,1024,0.009964799880981446
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,2048,768,0.009186133742332458
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,2048,256,0.003483733286460241
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,2048,256,0.016370133558909098
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,2048,512,0.008361599842707316
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,2048,128,0.003356799980004629
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,2048,128,0.016170666615168253
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,2048,256,0.008303999900817871
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,2048,64,0.003052799900372823
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,2048,128,0.007949866851170858
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,2048,64,0.015738667050997416
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,2048,32,0.003033600002527237
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,2048,32,0.014662399888038635
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,1536,65536,0.14815573692321776
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,1536,65536,0.2576618671417236
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,1536,16384,0.03838613430658976
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,1536,16384,0.06354133288065592
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,1536,12288,0.028278400500615437
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,1536,12288,0.04833600123723348
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,1536,12288,0.035410133997599284
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,1536,10240,0.023560533920923867
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,1536,10240,0.04380160172780355
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,1536,16384,0.04607253472010295
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,1536,8192,0.019745065768559774
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,1536,8192,0.038194131851196286
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,1536,10240,0.03141013383865356
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,1536,7168,0.017504000663757326
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,1536,7168,0.03649493455886841
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,1536,8192,0.02691733241081238
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,1536,6144,0.015656532843907674
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,1536,6144,0.03346879879633586
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,1536,7168,0.02445546587308248
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,1536,5120,0.01325653294722239
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,1536,5120,0.030088533957799275
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,1536,6144,0.022711465756098427
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,1536,4096,0.01130986710389455
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,1536,4096,0.026909865935643512
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,1536,4096,0.017425066232681273
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,1536,3584,0.010375466942787171
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,1536,5120,0.020334933201471964
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,1536,3584,0.02685760060946147
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,1536,3072,0.009227733810742695
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,1536,3072,0.026290132602055864
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,1536,3072,0.015015467007954916
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,1536,2560,0.008090666433175405
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,1536,2560,0.024782933791478477
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,1536,2560,0.013401599725087485
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,1536,2048,0.00713919997215271
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,1536,2048,0.02291413346926371
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,1536,2048,0.012097066640853882
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,1536,1536,0.006344533463319142
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,1536,1536,0.020897066593170165
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,2048,65536,0.1565290609995524
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,1536,1024,0.005031466484069824
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,1536,1024,0.019798400004704793
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,1536,1536,0.011201066772143047
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,1536,768,0.004640000065167745
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,1536,768,0.019645865758260092
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,1536,1024,0.009212799866994222
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,1536,512,0.003857066730658213
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,1536,768,0.008761599659919739
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,1536,512,0.01655893325805664
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,1536,256,0.0034506666163603462
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,1536,256,0.016962132851282754
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,1536,512,0.00876586635907491
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,1536,128,0.0034080001215140024
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,1536,256,0.008030933141708375
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,1536,128,0.017336533466974894
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,1536,64,0.003033600002527237
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,1536,128,0.0074879998962084455
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,1536,32,0.002998399982849757
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,1536,32,0.015618133544921874
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,1024,65536,0.09745813210805257
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,1024,65536,0.2435904026031494
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,1536,65536,0.15509227116902669
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,1024,16384,0.031548800071080525
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,1024,16384,0.05882773399353027
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,1024,16384,0.04486720164616902
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,1024,12288,0.023818665742874147
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,1024,12288,0.04578133424123128
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,1024,12288,0.03588693141937256
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,1024,10240,0.020230400562286376
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,1024,10240,0.041523198286692306
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,1024,65536,0.154584534962972
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,1024,8192,0.016743467251459757
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,1024,8192,0.035906132062276205
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,1024,10240,0.031413332621256514
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,1024,7168,0.015321600437164306
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,1024,7168,0.03521600166956584
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,1024,8192,0.026129066944122314
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,1024,6144,0.013337600231170654
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,1024,6144,0.032595199346542356
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,1024,6144,0.02227413256963094
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,1024,5120,0.011684266726175944
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,1024,5120,0.028985599676767986
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,1024,7168,0.024155733982721965
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,1024,4096,0.010010666648546855
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,1024,4096,0.02547946572303772
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,1024,4096,0.017432532707850137
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,1024,3584,0.00906986693541209
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,1024,5120,0.019859200716018675
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,1024,3584,0.025785599152247114
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,1024,3072,0.00800960014263789
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,1024,3072,0.024842667579650878
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,1024,3584,0.01593066652615865
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,1024,2560,0.007313066720962524
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,1024,2560,0.024330667654673257
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,1024,3072,0.014147200187047324
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,1024,2048,0.006323199967543285
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,1024,2048,0.02134079933166504
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,1024,2560,0.013272533814112345
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,1024,1536,0.00550186683734258
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,1024,1536,0.02071040074030558
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,1024,2048,0.011980799833933513
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,1024,1024,0.004614399870236715
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,1024,1024,0.018667733669281004
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,1024,1536,0.010787199934323628
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,1024,768,0.004205866654713949
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,1024,768,0.018713599443435668
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,1024,1024,0.00955733358860016
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,1024,512,0.0034261333445707956
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,1024,768,0.008779733379681905
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,1024,512,0.016876800855000814
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,1024,256,0.003389866650104523
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,1024,256,0.015995732943216958
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,1024,512,0.008729599912961324
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,1024,128,0.0030080000559488933
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,1024,256,0.008280533552169799
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,1024,128,0.015821866194407144
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,1024,64,0.003005866706371307
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,1024,64,0.014674133062362671
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,1024,128,0.007880533238252004
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,1024,32,0.003045333425203959
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,1024,32,0.014713600277900696
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,768,65536,0.0832319974899292
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,768,65536,0.2281386693318685
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,768,16384,0.024991999069849648
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,768,16384,0.05362879832585653
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,768,16384,0.04408106803894043
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,768,12288,0.018552533785502114
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,768,12288,0.04312533140182495
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,768,12288,0.035035733381907144
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,768,10240,0.01627626617749532
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,768,10240,0.039420799414316816
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,768,10240,0.03129600087801616
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,768,8192,0.013436800241470337
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,768,8192,0.0341536005338033
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,768,8192,0.026421332359313966
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,768,7168,0.012314666310946147
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,768,7168,0.03300799926122029
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,768,7168,0.023974400758743287
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,768,6144,0.010939733187357584
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,768,6144,0.030800000826517744
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,768,6144,0.021910399198532104
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,768,5120,0.009675733248392741
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,768,5120,0.02699306607246399
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,768,5120,0.019833600521087645
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,768,4096,0.00795306662718455
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,768,4096,0.02469546596209208
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,768,65536,0.15400106112162273
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,768,4096,0.01695786714553833
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,768,3584,0.007202133536338806
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,768,3584,0.0247488001982371
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,768,3072,0.006739200154940288
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,768,3072,0.023987199862798056
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,768,3584,0.01612160007158915
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,768,2560,0.005891199906667074
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,768,3072,0.014467199643452963
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,768,2560,0.02294293244679769
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,768,2048,0.005534933507442474
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,768,2048,0.022073600689570108
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,768,2560,0.01322773297627767
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,768,1536,0.0047082667549451195
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,768,2048,0.011985066533088683
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,768,1536,0.020360533396402994
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,768,1024,0.003884800026814143
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,768,1024,0.01854613423347473
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,768,1536,0.01090666651725769
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,768,768,0.003756800045569738
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,768,1024,0.009128533800443013
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,768,768,0.01810773412386576
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,768,512,0.003385599950949351
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,768,512,0.017889066537221273
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,768,768,0.008758399883906047
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,768,256,0.002985599885384242
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,768,256,0.015364266435305276
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,768,512,0.008374399940172831
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,768,128,0.002995199958483378
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,768,128,0.015388799707094827
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,768,256,0.007897600034872691
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,768,64,0.0026261332134405774
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,768,128,0.00788373351097107
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,768,64,0.014477866888046264
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,768,32,0.0026805333793163298
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,768,32,0.014416000247001648
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,512,65536,0.05418453216552734
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,512,65536,0.22200533548990883
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,512,16384,0.018304000298182167
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,512,16384,0.054004267851511634
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,512,16384,0.044547200202941895
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,512,12288,0.017194666465123496
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,512,12288,0.0420576016108195
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,512,12288,0.03501546780268351
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,512,10240,0.015361066659291586
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,512,10240,0.03771093289057414
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,512,10240,0.03130666613578796
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,512,8192,0.012853333353996277
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,512,8192,0.03291626572608948
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,512,8192,0.02635519901911418
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,512,7168,0.01164906620979309
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,512,7168,0.03220799962679545
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,512,7168,0.023924267292022704
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,512,6144,0.010001066327095031
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,512,6144,0.028989867369333906
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,512,65536,0.15434346199035645
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,512,5120,0.008874666690826417
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,512,5120,0.027723733584086103
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,512,6144,0.02181653380393982
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,512,4096,0.007516799867153168
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,512,4096,0.02478613257408142
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,512,5120,0.019853866100311278
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,512,3584,0.006724266707897187
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,512,3584,0.02477226654688517
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,512,4096,0.0173632005850474
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,512,3072,0.006105599800745646
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,512,3072,0.024564266204833984
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,512,3584,0.015681067109107973
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,512,2560,0.005499733487764994
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,512,2560,0.024216532707214355
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,512,3072,0.014085333546002707
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,512,2048,0.005039999882380167
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,512,2048,0.02097813288370768
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,512,2560,0.013217066725095114
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,512,1536,0.004279466470082601
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,512,2048,0.01199679970741272
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,512,1536,0.02056106726328532
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,512,1024,0.0037717332442601522
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,512,1024,0.0185973326365153
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,512,1536,0.01039466659228007
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,512,768,0.0033941333492596946
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,512,1024,0.00960213343302409
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,512,768,0.017473065853118898
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,512,512,0.002979200085004171
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,512,512,0.017299199104309083
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,512,768,0.00876693328221639
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,512,256,0.0026378666361172995
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,512,512,0.008312533299128216
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,512,256,0.01562879979610443
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,512,128,0.0025397333006064097
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,512,128,0.014498133460680643
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,512,256,0.007990399996439617
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,512,64,0.002218666672706604
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,512,128,0.007890133559703827
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,512,64,0.014458666245142618
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,512,32,0.002313599983851115
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,512,32,0.01446293294429779
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,256,65536,0.03463253180185954
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,256,65536,0.20616426467895507
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,256,16384,0.012601600090662638
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,256,16384,0.04733973344167074
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,256,16384,0.043721600373586016
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,256,12288,0.01204266647497813
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,256,12288,0.04028693437576294
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,256,12288,0.03539733489354451
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,256,10240,0.011387733618418376
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,256,10240,0.037212800979614255
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,256,10240,0.030908799171447753
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,256,8192,0.011934933066368104
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,256,8192,0.0325162669022878
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,256,8192,0.026372265815734864
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,256,7168,0.010845866799354554
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,256,7168,0.031159466505050658
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,256,7168,0.023858133951822916
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,256,6144,0.009567999839782714
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,256,6144,0.028843732674916585
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,256,65536,0.1545141379038493
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,256,5120,0.008346666892369587
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,256,5120,0.028056534131368
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,256,6144,0.021899733940760294
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,256,4096,0.007073066631952922
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,256,4096,0.024986666440963746
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,256,5120,0.01985493302345276
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,256,3584,0.006743466854095459
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,256,3584,0.024529065688451132
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,256,4096,0.017313067118326822
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,256,3072,0.005857066810131073
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,256,3072,0.02286400000254313
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,256,3584,0.015737600127855935
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,256,2560,0.005491200089454651
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,256,2560,0.02291626731554667
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,256,3072,0.014124799768129984
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,256,2048,0.0046293333172798155
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,256,2048,0.020658133427302043
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,256,2560,0.013202133774757385
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,256,1536,0.004223999877770742
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,256,2048,0.011963733037312825
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,256,1536,0.021677867571512858
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,256,1024,0.0034186666210492453
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,256,1024,0.01881706714630127
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,256,1536,0.010340266426404317
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,256,768,0.003028266628583272
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,256,1024,0.00957546631495158
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,256,768,0.017615999778111777
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,256,512,0.002961066613594691
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,256,768,0.008777599533398945
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,256,512,0.017366399367650352
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,256,512,0.008380800485610962
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,256,256,0.0026357332865397137
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,256,256,0.015301332871119181
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,256,256,0.00800960014263789
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,256,128,0.002194133400917053
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,256,128,0.014989866813023885
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,256,128,0.007523199915885926
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,256,64,0.002219733347495397
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,256,64,0.014595199624697366
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,256,32,0.0022357332209746042
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,256,32,0.014826666315396628
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,128,65536,0.028061866760253906
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,128,65536,0.20483733812967936
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,128,16384,0.010238933563232421
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,128,16384,0.047251200675964354
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,128,16384,0.04395733277002971
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,128,12288,0.011609599987665812
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,128,12288,0.03909333149592082
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,128,12288,0.03500373363494873
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,128,10240,0.010434133807818095
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,128,10240,0.03698026736577352
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,128,10240,0.031111466884613036
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,128,8192,0.009111467003822326
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,128,8192,0.03304426670074463
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,128,65536,0.15434986750284832
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,128,7168,0.008730666836102803
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,128,7168,0.03132479985555013
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,128,8192,0.026332799593607587
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,128,6144,0.007965866724650066
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,128,6144,0.028834132353464763
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,128,7168,0.023913600047429404
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,128,5120,0.007563733557860057
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,128,5120,0.026866134007771807
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,128,6144,0.021869866053263347
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,128,4096,0.006728533407052357
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,128,4096,0.023991467555363973
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,128,5120,0.01979413429896037
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,128,3584,0.006259199976921081
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,128,3584,0.024663466215133666
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,128,4096,0.01738133430480957
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,128,3072,0.005566933254400889
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,128,3072,0.02443199952443441
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,128,3584,0.016139733791351318
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,128,2560,0.005085866649945577
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,128,2560,0.022714666525522866
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,128,3072,0.01404906709988912
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,128,2048,0.00462719996770223
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,128,2560,0.012849066654841104
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,128,2048,0.02071253259976705
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,128,1536,0.0042303999265035
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,128,1536,0.01955733299255371
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,128,2048,0.011988266309102377
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,128,1024,0.003435733417669932
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,128,1536,0.010411733388900756
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,128,1024,0.018642133474349974
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,128,768,0.003050666550795237
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,128,768,0.018771199385325114
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,128,1024,0.009117866555849712
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,128,512,0.003009066730737686
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,128,512,0.016516266266504924
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,128,768,0.008724266290664673
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,128,256,0.0026047999660174055
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,128,512,0.007904000083605449
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,128,256,0.015099733074506124
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,128,128,0.002205866575241089
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,128,256,0.00795413355032603
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,128,128,0.014491732915242514
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,128,64,0.002491733431816101
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,128,128,0.0074442664782206226
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,128,64,0.014476799964904785
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,128,32,0.002551466723283132
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,128,32,0.014471466342608133
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,64,65536,0.029226666688919066
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,64,16384,0.009403733412424724
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,64,16384,0.043186132113138834
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,64,12288,0.007993599772453308
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,64,12288,0.03715840180714925
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,64,10240,0.007494399944941203
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,64,10240,0.03628480037053426
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,64,8192,0.007154133419195812
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,64,8192,0.031652265787124635
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,64,7168,0.006671999891599019
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,64,7168,0.03097813328107198
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,64,6144,0.006322133541107178
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,64,6144,0.0287989338239034
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,64,5120,0.006683733562628429
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,64,5120,0.026728532711664837
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,64,4096,0.006690133114655812
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,64,4096,0.02443199952443441
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,64,3584,0.006235733131567637
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,64,3584,0.024069333076477052
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,64,3072,0.005464533468087515
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,64,3072,0.02340373396873474
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,64,2560,0.005020800232887268
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,64,2560,0.022578134139378866
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,64,2048,0.004401066899299621
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,64,2048,0.020593067010243736
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,64,1536,0.0038304001092910765
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,64,1536,0.01919893423716227
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,64,1024,0.0034944000343481696
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,64,1024,0.01908479928970337
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,64,768,0.003031466652949651
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,64,768,0.017254400253295898
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,64,512,0.002726399898529053
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,64,512,0.016446933150291443
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,64,256,0.002647466709216436
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,64,256,0.015558399756749473
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,64,128,0.002243199944496155
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,64,128,0.014881066481272378
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,64,64,0.0021759999295075734
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,64,64,0.014439466595649719
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,64,32,0.002221866697072983
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,64,32,0.014980266491572062
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,32,65536,0.02326186696688334
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,32,65536,0.19843840599060059
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,32,16384,0.011098666985829671
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,32,16384,0.043398400147755936
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,32,12288,0.00901759962240855
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,32,12288,0.03721173206965129
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,32,10240,0.008318933347860973
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,32,10240,0.03497173388799031
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,32,8192,0.007483733197053273
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,32,8192,0.031061333417892457
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,32,7168,0.0070709332823753355
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,32,7168,0.03096853295962016
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,32,6144,0.0063360000650088
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,32,6144,0.028862933317820232
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,32,5120,0.005896533528963724
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,32,5120,0.026862933238347368
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,32,4096,0.005904000004132589
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,32,4096,0.024729599555333458
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,32,3584,0.005942399799823761
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,32,3584,0.024683733781178795
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,32,3072,0.005465599894523621
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,32,3072,0.023549866676330567
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,32,2560,0.005027199784914652
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,32,2560,0.0230240007241567
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,32,2048,0.005266133447488149
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,32,2048,0.02067413330078125
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,32,1536,0.004673066735267639
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,32,1536,0.019374932845433554
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,32,1024,0.003738666574160258
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,32,1024,0.018952532609303793
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,32,768,0.003050666550795237
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,32,768,0.016645333170890807
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,32,512,0.0026410666604836782
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,32,512,0.016505600015322367
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,32,256,0.0025759999950726825
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,32,256,0.01628266672293345
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,32,128,0.0021802666286627453
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,32,128,0.014865066607793173
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,32,64,0.0021888000269730886
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,32,64,0.014289066195487976
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1024,32,32,0.0021461332837740582
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,32,32,0.015222400426864624
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,65536,16384,0.9640928268432617
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,65536,16384,0.6483402887980143
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1024,1536,3584,0.01577279965082804
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,65536,12288,0.7299999872843425
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,65536,12288,0.5447327931722005
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,1536,64,0.015258666872978211
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,65536,10240,0.6560810724894206
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,65536,10240,0.3997973442077637
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1024,64,65536,0.20345813433329263
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,65536,8192,0.5236831982930501
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,65536,8192,0.30662720998128257
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,65536,10240,0.23858027458190917
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,65536,7168,0.4498698552449544
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,65536,7168,0.30778773625691735
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,65536,8192,0.19573973019917806
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,65536,6144,0.40654827753702805
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,65536,6144,0.21466026306152344
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,65536,12288,0.2776255925496419
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,65536,5120,0.3250794728597005
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,65536,5120,0.17465707461039226
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,65536,16384,0.35651626586914065
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,65536,4096,0.2601792017618815
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,65536,4096,0.15624319712320964
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,65536,7168,0.17211626370747884
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,65536,3584,0.2250805377960205
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,65536,3584,0.13181653022766113
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,65536,5120,0.12698240280151368
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,65536,3072,0.1879466692606608
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,65536,3072,0.1247221310933431
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,65536,6144,0.15200746854146321
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,65536,2560,0.16926612854003906
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,65536,2560,0.11147200266520182
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,65536,4096,0.10745920340220134
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,65536,2048,0.13489279747009278
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,65536,2048,0.09482773145039877
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,65536,3584,0.0964298645655314
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,65536,1536,0.10508586565653484
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,65536,1536,0.07733866373697916
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,65536,3072,0.08409706751505533
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,65536,1024,0.07336853345235189
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,65536,1024,0.05825813213984171
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,65536,2560,0.07527573108673095
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,65536,768,0.06010026534398397
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,65536,768,0.05264533360799154
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,65536,1024,0.04220586617787679
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,65536,512,0.042606933911641436
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,65536,512,0.04718399842580159
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,65536,768,0.03733760118484497
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,65536,256,0.031310933828353885
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,65536,256,0.04355520009994507
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,65536,256,0.02688639958699544
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,65536,128,0.021998933951059976
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,65536,128,0.03224426706631978
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,65536,1536,0.053998935222625735
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,65536,64,0.02009493311246236
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,65536,64,0.03451093435287476
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,65536,32,0.019153066476186118
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,65536,32,0.03500373363494873
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,16384,65536,0.9774037043253581
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,16384,65536,0.7566346486409505
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,65536,128,0.024027733008066814
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,16384,16384,0.2652650674184164
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,16384,16384,0.17147307395935057
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,65536,512,0.032485334078470866
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,16384,12288,0.18984427452087402
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,16384,12288,0.12501440048217774
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,65536,2048,0.06447999874750773
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,16384,10240,0.17251413663228352
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,16384,10240,0.11929492950439453
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,16384,10240,0.07345813115437826
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,16384,8192,0.1377141316731771
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,16384,8192,0.08697706858317057
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,16384,12288,0.08216213385264079
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,16384,7168,0.11765120029449463
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,16384,7168,0.07987840175628662
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,16384,16384,0.10982293287913006
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,16384,6144,0.10370453198750813
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,16384,6144,0.06773973306020101
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,16384,8192,0.0603061318397522
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,16384,5120,0.08844160238901774
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,16384,5120,0.06427839994430543
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,16384,7168,0.05324586629867554
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,16384,4096,0.06553066571553548
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,16384,4096,0.05550293525060018
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,16384,6144,0.04846186637878418
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,16384,3584,0.05984640121459961
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,16384,3584,0.04939200083414714
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,16384,5120,0.04291520118713379
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,16384,3072,0.05111466646194458
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,16384,3072,0.04500373204549153
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,16384,4096,0.03573866685231526
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,16384,2560,0.041637333234151204
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,16384,2560,0.040909866491953537
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,16384,2560,0.02712106704711914
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,16384,2048,0.03387413422266643
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,16384,2048,0.037085866928100585
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,16384,3584,0.0329802672068278
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,16384,3072,0.030059732993443805
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,16384,1536,0.03298239906628926
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,16384,1536,0.020962133010228475
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,16384,1024,0.01956160068511963
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,16384,1024,0.02904213269551595
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,16384,2048,0.023666133483250938
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,16384,768,0.015128533045450846
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,16384,768,0.026703999439875288
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,16384,1024,0.01691306630770365
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,16384,512,0.012602667013804117
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,16384,512,0.025560534000396727
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,16384,768,0.014476799964904785
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,16384,256,0.01001706620057424
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,16384,256,0.02267199953397115
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,16384,512,0.012880000472068786
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,16384,128,0.009346133470535279
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,16384,128,0.021357866128285725
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,16384,256,0.011690666278203327
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,16384,64,0.008757332960764568
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,16384,64,0.022408533096313476
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,16384,32,0.008841600020726521
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,16384,32,0.02219946583112081
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,16384,128,0.011815466483434041
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,12288,65536,0.6779434839884441
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,12288,65536,0.5170111974080404
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,12288,16384,0.1753738721211751
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,12288,16384,0.12491306463877361
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,12288,16384,0.07885119915008545
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,12288,12288,0.1316362698872884
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,12288,12288,0.09714346726735433
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,12288,12288,0.06129706700642904
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,12288,10240,0.12424000104268391
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,12288,10240,0.08857386906941732
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,16384,1536,0.027240532636642455
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,12288,8192,0.10192000071207682
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,12288,8192,0.07429013252258301
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,12288,10240,0.05321173270543417
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,12288,7168,0.08852266470591227
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,12288,7168,0.06943573156992594
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,12288,8192,0.04437226851781209
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,12288,6144,0.07665493488311767
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,12288,6144,0.06022293170293173
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,12288,7168,0.04096320072809855
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,12288,5120,0.06400426626205444
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,12288,5120,0.05288639863332113
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,12288,6144,0.036690131823221846
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,12288,4096,0.05225813388824463
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,12288,4096,0.049075198173522946
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,12288,5120,0.032560000816981
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,12288,3584,0.045060265064239505
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,12288,3584,0.044948267936706546
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,12288,4096,0.027746133009592694
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,12288,3072,0.039153067270914714
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,12288,3072,0.040965334574381514
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,12288,3584,0.025674666961034136
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,12288,2560,0.03285226623217265
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,12288,2560,0.03720320065816243
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,12288,3072,0.02327573299407959
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,12288,2048,0.027153066794077557
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,12288,2048,0.03471999963124593
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,12288,2560,0.021441066265106203
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,12288,1536,0.020846933126449585
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,12288,1536,0.029450666904449464
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,12288,2048,0.01872746745745341
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,12288,1024,0.015612799922625223
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,12288,1024,0.0253930668036143
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,12288,1024,0.01362666686375936
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,12288,768,0.012782933314641318
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,12288,768,0.02300373315811157
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,12288,1536,0.016250666975975037
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,12288,512,0.009629866480827332
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,12288,512,0.021134932835896812
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,12288,512,0.01046399970849355
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,12288,256,0.006667733192443848
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,12288,768,0.012019200126330058
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,12288,256,0.018900267283121743
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,12288,128,0.006263466676076253
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,12288,128,0.01880853374799093
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,12288,128,0.0091839998960495
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,12288,64,0.005436799923578898
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,12288,64,0.017013333241144814
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,12288,32,0.00550186683734258
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,12288,32,0.017314134041468303
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,12288,256,0.009781333804130554
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,10240,65536,0.6042794545491537
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,10240,65536,0.4530933380126953
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,10240,16384,0.15607360204060872
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,10240,16384,0.11953492959340413
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,16384,65536,0.40929174423217773
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,10240,12288,0.1187872012456258
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,10240,12288,0.10053973197937012
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,10240,16384,0.07629653612772623
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,10240,10240,0.10150612990061443
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,10240,10240,0.0775221347808838
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,12288,65536,0.2891093254089355
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,10240,8192,0.08363626797993978
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,10240,8192,0.06814826329549153
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,10240,12288,0.06011840105056763
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,10240,7168,0.07471360365549723
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,10240,7168,0.05957653522491455
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,10240,10240,0.05174933274586996
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,10240,6144,0.06540266672770181
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,10240,6144,0.05303039948145548
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,10240,7168,0.03968853155771891
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,10240,5120,0.05293759902318319
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,10240,5120,0.04951893488566081
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,10240,8192,0.0427509347597758
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,10240,4096,0.04170986811319987
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,10240,4096,0.04358933369318645
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,10240,6144,0.035930665334065755
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,10240,3584,0.03821546634038289
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,10240,3584,0.04137920141220093
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,10240,3584,0.02481813430786133
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,10240,3072,0.03327253262201945
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,10240,3072,0.037298134962717694
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,10240,5120,0.03176746765772502
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,10240,4096,0.026753065983454387
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,10240,2560,0.027553067604700728
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,10240,2560,0.033473066488901776
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,10240,2048,0.021983999013900756
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,10240,2048,0.030820266405741377
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,10240,2048,0.018260266383488974
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,10240,1536,0.017669334014256795
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,10240,1536,0.027875200907389326
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,10240,2560,0.020708266894022623
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,10240,1024,0.013130666812260947
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,10240,1024,0.025126399596532185
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,10240,3072,0.022796799739201866
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,10240,768,0.010982400178909302
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,10240,768,0.024743467569351196
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,10240,1024,0.012894933422406515
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,10240,512,0.00885973374048869
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,10240,512,0.022750933965047203
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,10240,768,0.011590400338172912
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,10240,256,0.0071285332242647815
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,10240,256,0.02094506621360779
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,10240,1536,0.015727999806404113
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,10240,128,0.005868799984455109
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,10240,128,0.019401599963506065
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,10240,256,0.009637332955996196
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,10240,64,0.006502399841944377
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,10240,64,0.01867093245188395
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,10240,32,0.006663466493288676
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,10240,32,0.01882986625035604
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,10240,128,0.009147733449935913
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,8192,65536,0.4946751912434896
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,8192,65536,0.3927722613016764
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,10240,512,0.010413866241772969
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,8192,16384,0.12691840330759685
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,8192,16384,0.09980266888936361
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,8192,12288,0.0982421318689982
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,8192,12288,0.08077546755472818
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,8192,12288,0.059740801652272545
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,8192,10240,0.08682773113250733
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,8192,10240,0.07159466743469238
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,8192,16384,0.07606080373128256
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,8192,8192,0.06788480281829834
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,8192,8192,0.05905813376108805
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,8192,10240,0.05162986516952515
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,8192,7168,0.0608234683672587
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,8192,7168,0.055129599571228025
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,8192,8192,0.043171199162801106
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,8192,6144,0.05152213176091512
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,8192,6144,0.049457065264383954
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,8192,7168,0.03923413356145223
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,8192,5120,0.043986133734385174
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,8192,5120,0.04402666489283244
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,8192,6144,0.03538453181584676
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,8192,4096,0.034358398119608564
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,8192,4096,0.039628799756368
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,8192,5120,0.031421866019566855
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,8192,3584,0.03101866642634074
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,8192,3584,0.03601706822713216
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,8192,4096,0.026898133754730224
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,8192,3072,0.026445867617925008
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,8192,3072,0.03497813145319621
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,8192,3584,0.024961066246032716
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,8192,2560,0.023037866751352946
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,8192,2560,0.03178986708323161
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,8192,3072,0.022744532426198324
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,8192,2048,0.01882666746775309
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,8192,2048,0.029242666562398274
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,8192,2560,0.020530132452646892
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,8192,1536,0.01513706644376119
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,8192,1536,0.02688213388125102
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,8192,1536,0.015787733594576518
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,8192,1024,0.01145919958750407
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,8192,1024,0.022909865776697794
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,8192,2048,0.017978666226069133
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,8192,768,0.00928106705347697
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,8192,768,0.021998933951059976
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,8192,1024,0.012925866246223449
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,8192,512,0.007534933090209961
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,8192,512,0.020692267020543418
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,8192,768,0.011593600114186604
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,8192,256,0.006727466483910878
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,8192,256,0.019980800151824952
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,8192,256,0.009598933657010396
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,8192,128,0.005937066674232483
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,8192,128,0.019978666305541994
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,8192,512,0.010392533739407857
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,8192,64,0.00556160012880961
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,8192,64,0.0183786670366923
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,8192,32,0.005739733576774597
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,8192,32,0.01845759948094686
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,8192,128,0.009192533294359843
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,7168,65536,0.4290293375651042
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,7168,65536,0.35857601165771485
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,7168,16384,0.10764479637145996
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,7168,16384,0.08737813631693522
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,10240,65536,0.2804949442545573
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,7168,12288,0.08206079800923666
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,7168,12288,0.07101546923319499
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,7168,16384,0.07617920239766439
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,7168,10240,0.07287680308024089
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,7168,10240,0.06780906518300375
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,7168,12288,0.060661331812540686
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,7168,8192,0.059374932448069254
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,7168,8192,0.05622186660766602
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,7168,10240,0.05184853474299113
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,7168,7168,0.05129493474960327
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,7168,7168,0.049218134085337324
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,7168,8192,0.04364159901936849
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,7168,6144,0.04476266702016195
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,7168,6144,0.044130134582519534
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,7168,7168,0.039297068119049074
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,7168,5120,0.03686293363571167
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,7168,5120,0.03956053256988525
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,7168,6144,0.03519466718037923
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,7168,4096,0.030481066306432086
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,7168,4096,0.03702400128046672
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,7168,5120,0.031325866778691605
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,7168,3584,0.026705066363016765
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,7168,3584,0.0352789322535197
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,7168,4096,0.02690453330675761
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,7168,3072,0.023563732703526817
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,7168,3072,0.03285546700159709
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,7168,3584,0.02483200033505758
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,7168,2560,0.019860267639160156
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,7168,2560,0.030750934282938642
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,8192,65536,0.2754613240559896
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,7168,2048,0.015524267156918844
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,7168,2048,0.02683626612027486
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,7168,3072,0.022201599677403767
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,7168,1536,0.013426132996877036
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,7168,1536,0.02509760061899821
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,7168,2560,0.02029973268508911
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,7168,1024,0.009711999694506328
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,7168,1024,0.022715733448664347
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,7168,2048,0.017480534315109254
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,7168,768,0.007990399996439617
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,7168,768,0.02076373298962911
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,7168,1536,0.01525973379611969
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,7168,512,0.00631466656923294
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,7168,512,0.01888426740964254
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,7168,768,0.010998400052388509
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,7168,256,0.0054666668176651
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,7168,256,0.018518400192260743
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,7168,1024,0.012437333663304646
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,7168,128,0.004673066735267639
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,7168,128,0.016710400581359863
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,7168,256,0.009171199798583985
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,7168,64,0.00421973317861557
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,7168,128,0.00879146655400594
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,7168,32,0.004237866898377737
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,7168,64,0.01722453236579895
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,7168,512,0.009939199686050415
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,7168,32,0.01655893325805664
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,6144,65536,0.3656789461771647
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,6144,65536,0.3183306694030762
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,6144,16384,0.08973653316497802
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,6144,16384,0.08399466673533121
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,6144,12288,0.07035626570383707
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,6144,12288,0.06880533695220947
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,6144,12288,0.03684480190277099
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,6144,10240,0.06099626620610556
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,6144,10240,0.06019519964853922
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,6144,16384,0.045135998725891115
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,6144,8192,0.047977598508199056
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,6144,8192,0.05293759902318319
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,6144,10240,0.032891732454299924
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,6144,7168,0.04385066827138265
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,6144,7168,0.047363201777140304
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,6144,8192,0.027271467447280883
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,6144,6144,0.036524800459543864
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,6144,6144,0.04131306807200114
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,6144,7168,0.025770666201909383
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,6144,5120,0.0321237325668335
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,6144,5120,0.0370741327603658
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,6144,6144,0.023501867055892946
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,6144,4096,0.02531519929567973
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,6144,4096,0.03529173135757446
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,6144,5120,0.021195733547210695
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,6144,3584,0.023423999547958374
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,6144,3584,0.03403840065002441
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,6144,4096,0.01839253306388855
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,6144,3072,0.019797333081563315
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,6144,3072,0.031498666604359946
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,6144,3584,0.01737706661224365
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,6144,2560,0.017799466848373413
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,6144,2560,0.0288810670375824
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,7168,65536,0.27969172795613606
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,6144,2048,0.01490239997704824
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,6144,2048,0.02893226742744446
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,6144,3072,0.015731199582417806
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,6144,1536,0.011895466844240825
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,6144,1536,0.02476693391799927
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,6144,2560,0.014552533626556396
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,6144,1024,0.009492266178131103
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,6144,1024,0.021335466702779134
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,6144,2048,0.013229866822560629
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,6144,768,0.008146133522192638
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,6144,768,0.020906666914621987
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,6144,1536,0.011292800307273865
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,6144,512,0.005998933315277099
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,6144,512,0.018784000476201376
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,6144,1024,0.009581866860389709
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,6144,256,0.005036800106366476
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,6144,256,0.01652906636397044
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,6144,768,0.00912000040213267
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,6144,128,0.004293333490689596
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,6144,128,0.01688746611277262
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,6144,512,0.008745599786440532
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,6144,128,0.007949866851170858
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,6144,64,0.003881600002447764
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,6144,32,0.00386559988061587
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,6144,256,0.008343467116355896
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,6144,64,0.016521599888801575
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,6144,32,0.016537599762280784
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,5120,65536,0.3183253288269043
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,5120,16384,0.08121279875437418
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,5120,65536,0.3222506523132324
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,5120,16384,0.08970879713694255
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,5120,12288,0.06427733500798544
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,5120,12288,0.07220053672790527
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,5120,12288,0.03657493193944295
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,5120,10240,0.05358613332112631
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,5120,10240,0.05860586563746134
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,5120,16384,0.044395732879638675
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,5120,8192,0.04456959962844849
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,5120,8192,0.05150933265686035
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,5120,10240,0.03226133386294047
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,5120,7168,0.03874239921569824
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,5120,7168,0.04538559913635254
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,5120,8192,0.026950399080912273
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,5120,6144,0.03423999945322673
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,5120,6144,0.04186346530914307
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,5120,7168,0.02508266568183899
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,5120,5120,0.028742400805155437
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,5120,5120,0.03829653263092041
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,5120,6144,0.022743467489878336
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,5120,4096,0.023834667603174844
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,5120,4096,0.03525333404541016
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,5120,5120,0.02076266606648763
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,5120,3584,0.021210666497548422
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,5120,3584,0.03290026585261027
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,5120,4096,0.018295466899871826
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,5120,3072,0.01884053349494934
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,5120,3072,0.031142399708429976
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,5120,3584,0.016684800386428833
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,5120,2560,0.016516266266504924
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,5120,2560,0.028819199403127032
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,5120,3072,0.01506239970525106
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,5120,2048,0.014032000303268432
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,5120,2048,0.026929066578547163
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,5120,2560,0.013738666971524557
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,5120,1536,0.011592533191045125
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,5120,1536,0.02273599902788798
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,5120,1536,0.010811733206113179
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,5120,1024,0.009150933225949604
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,5120,1024,0.020708266894022623
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,5120,2048,0.012391466895739238
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,5120,768,0.00754559983809789
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,5120,768,0.021011199553807577
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,5120,1024,0.00958720048268636
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,5120,512,0.006354133288065593
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,5120,512,0.01884053349494934
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,5120,768,0.008782933155695598
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,5120,256,0.005110399921735128
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,5120,256,0.01657600005467733
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,5120,256,0.00788800021012624
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,5120,128,0.004273066421349844
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,5120,128,0.01651946703592936
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,5120,512,0.008705066641171773
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,5120,64,0.0042015999555587765
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,5120,64,0.017539199193318686
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,5120,128,0.007507200042406718
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,5120,32,0.00510506679614385
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,5120,32,0.018147200345993042
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,4096,65536,0.25951786041259767
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,4096,65536,0.28693332672119143
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,4096,16384,0.06759359836578369
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,4096,16384,0.0715231974919637
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,6144,65536,0.16217919985453289
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,4096,12288,0.05006080071131388
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,4096,12288,0.05645546515782675
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,4096,16384,0.04406506617863973
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,4096,10240,0.043338668346405027
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,4096,10240,0.051201065381368004
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,4096,12288,0.03524586757024129
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,4096,8192,0.03425600131352742
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,4096,8192,0.04375893274943034
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,4096,8192,0.026474666595458985
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,4096,7168,0.0304202675819397
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,4096,7168,0.041061333815256756
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,4096,10240,0.03128213286399841
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,4096,6144,0.02664320071538289
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,4096,6144,0.037138132254282634
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,4096,7168,0.02445333401362101
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,4096,5120,0.022525866826375328
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,4096,5120,0.03491520086924235
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,4096,6144,0.022434133291244506
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,4096,4096,0.01872746745745341
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,4096,4096,0.032409600416819256
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,4096,4096,0.017435733477274576
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,4096,3584,0.01705706715583801
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,4096,3584,0.03095573385556539
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,4096,5120,0.020360533396402994
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,4096,3072,0.014891733725865683
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,4096,3072,0.02885119915008545
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,4096,3072,0.014964266618092855
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,4096,2560,0.012935466567675271
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,4096,2560,0.028146133820215864
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,4096,3584,0.01613866686820984
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,4096,2048,0.011283199985822041
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,4096,2048,0.024684800704320272
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,4096,2560,0.01381226678689321
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,4096,1536,0.009195733070373534
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,4096,2048,0.012012799580891926
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,4096,1536,0.022709333896636964
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,4096,1024,0.007121066749095917
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,4096,1024,0.020680532852808634
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,4096,1024,0.009239466985066731
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,4096,768,0.006266666452089946
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,4096,768,0.018879999717076622
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,4096,1536,0.010469333330790202
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,4096,512,0.005145599941412607
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,4096,512,0.0185205340385437
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,4096,768,0.008736000458399455
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,4096,256,0.004699733356634776
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,4096,256,0.016618667046229045
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,4096,512,0.0082997332016627
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,4096,128,0.00421013335386912
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,4096,128,0.016511999567349753
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,4096,128,0.0075914666056633
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,4096,256,0.00795413355032603
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,4096,64,0.0038624001046021783
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,4096,64,0.01754666765530904
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,4096,32,0.0039018665750821433
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,4096,32,0.016566399733225504
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,3584,65536,0.22437440554300941
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,3584,65536,0.2683338801066081
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,3584,16384,0.056898132960001624
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,3584,16384,0.0671018679936727
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,5120,65536,0.15240853627522785
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,3584,12288,0.042045867443084715
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,3584,12288,0.053193600972493496
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,3584,12288,0.0346399982770284
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,3584,10240,0.035563735167185466
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,3584,10240,0.0463861346244812
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,3584,16384,0.04393173456192016
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,3584,8192,0.029665066798528032
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,3584,8192,0.041155199209849044
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,3584,10240,0.030587732791900635
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,3584,7168,0.025568000475565594
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,3584,7168,0.038378667831420896
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,3584,8192,0.026475733518600462
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,3584,6144,0.022818134228388468
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,3584,6144,0.03649813334147135
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,3584,7168,0.02398186723391215
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,3584,5120,0.01911146640777588
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,3584,5120,0.03249066670735677
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,3584,6144,0.02214399973551432
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,3584,4096,0.016182399789492288
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,3584,4096,0.030878933270772298
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,3584,5120,0.019849600394566853
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,3584,3584,0.01420266628265381
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,3584,3584,0.028194133440653486
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,3584,4096,0.01744640072186788
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,3584,3072,0.01295360028743744
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,3584,3072,0.027046400308609008
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,3584,3584,0.016198399662971496
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,3584,2560,0.011053867141405741
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,3584,2560,0.024759467442830405
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,4096,65536,0.14839040438334147
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,3584,2048,0.009669333696365356
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,3584,2048,0.023056000471115112
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,3584,3072,0.01458453337351481
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,3584,1536,0.008019199967384339
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,3584,1536,0.021681066354115805
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,3584,2560,0.013313066959381104
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,3584,1024,0.006299733122189839
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,3584,1024,0.018626133600870766
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,3584,1536,0.010821333527565003
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,3584,768,0.005457066496213277
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,3584,2048,0.01164479951063792
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,3584,768,0.019832533597946168
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,3584,512,0.004645333190759023
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,3584,512,0.01842666665712992
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,3584,1024,0.009447466333707173
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,3584,256,0.003788800040880839
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,3584,256,0.01723946730295817
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,3584,512,0.008371200164159138
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,3584,128,0.003504000107447306
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,3584,128,0.017417599757512413
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,3584,768,0.008716799815495809
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,3584,64,0.0033642667035261786
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,3584,64,0.01446399986743927
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,3584,32,0.0033781332274278007
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,3584,128,0.007550933460394542
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,3584,32,0.01444906691710154
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,3072,65536,0.19221760431925455
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,3072,65536,0.24727039337158202
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,3072,16384,0.04933546781539917
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,3072,16384,0.06212906837463379
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,3072,16384,0.04323306481043498
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,3072,12288,0.03843733469645182
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,3072,12288,0.05125333468119303
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,3072,12288,0.03429866631825765
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,3072,10240,0.032654933134714764
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,3072,10240,0.04538453420003255
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,3584,65536,0.14654293060302734
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,3072,8192,0.025650133689244587
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,3072,8192,0.03844906489054362
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,3072,10240,0.02997973362604777
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,3072,7168,0.023252266645431518
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,3072,7168,0.03704640070597331
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,3072,7168,0.023468800385793052
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,3072,6144,0.020248534282048543
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,3072,6144,0.03359573284784953
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,3072,8192,0.026051199436187743
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,3072,5120,0.01755626598993937
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,3072,5120,0.031496532758076984
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,3072,5120,0.019475199778874717
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,3072,4096,0.014518400033315023
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,3072,4096,0.030053333441416426
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,3072,6144,0.021690666675567627
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,3072,3584,0.013100799918174744
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,3072,3584,0.027124265829722088
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,3072,3584,0.01574186682701111
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,3072,3072,0.011534933249155681
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,3072,3072,0.02553706765174866
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,3072,4096,0.01691733400026957
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,3072,2560,0.010218666990598042
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,3072,2560,0.024780799945195518
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,3072,2560,0.01332373321056366
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,3072,3072,0.014565333724021912
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,3072,2048,0.008666666348775227
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,3072,2048,0.022660267353057862
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,3072,1536,0.007638399799664815
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,3072,1536,0.020843732357025146
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,3072,2048,0.011958400408426922
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,3072,1024,0.005524266759554545
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,3072,1024,0.019370667139689126
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,3072,1536,0.010455466310183207
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,3072,768,0.0051360001166661584
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,3072,768,0.01914773384730021
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,3072,768,0.008785067001978557
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,3072,1024,0.009165866176287334
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,3072,512,0.004286933441956838
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,3072,512,0.01662613352139791
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,3072,256,0.003419733295838038
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,3072,256,0.0160778671503067
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,3072,256,0.007914666831493378
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,3072,128,0.0030591999491055804
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,3072,128,0.014516266187032065
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,3072,512,0.008379733562469483
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,3072,64,0.0030026666820049284
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,3072,64,0.016938666502634682
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,3072,32,0.0030741333961486817
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,3072,128,0.00754559983809789
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,3072,32,0.015757866700490317
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,2560,65536,0.1727786699930827
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,2560,65536,0.23375040690104165
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,2560,16384,0.04536853233973186
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,2560,16384,0.05960959990819296
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,2560,16384,0.0426528016726176
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,2560,12288,0.03452479839324951
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,2560,12288,0.04768106540044149
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,3072,65536,0.14673706690470378
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,2560,10240,0.028253867228825884
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,2560,10240,0.04312320152918498
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,2560,10240,0.030613332986831665
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,2560,8192,0.02289066712061564
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,2560,8192,0.03829013506571452
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,2560,12288,0.03423146804173787
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,2560,7168,0.021141332387924195
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,2560,7168,0.03542506694793701
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,2560,8192,0.02605973283449809
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,2560,6144,0.01805866758028666
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,2560,7168,0.02317439913749695
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,2560,6144,0.03328640063603719
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,2560,5120,0.015717333555221556
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,2560,5120,0.030771199862162275
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,2560,5120,0.019475199778874717
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,2560,4096,0.01301653285821279
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,2560,4096,0.028940800825754804
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,2560,6144,0.02140586574872335
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,2560,3584,0.011921067039171855
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,2560,3584,0.027509333690007527
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,2560,4096,0.016951467593510947
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,2560,3072,0.010785067081451416
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,2560,3584,0.015313067038853965
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,2560,3072,0.027027199665705364
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,2560,2560,0.00942186713218689
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,2560,2560,0.024296534061431885
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,2560,2560,0.013316266735394797
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,2560,2048,0.00795306662718455
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,2560,2048,0.023439999421437582
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,2560,3072,0.014150399963061014
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,2560,1536,0.006822399795055389
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,2560,1536,0.020730666319529214
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,2560,1536,0.010821333527565003
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,2560,1024,0.0058378666639328
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,2560,1024,0.0186901330947876
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,2560,2048,0.011640533804893494
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,2560,768,0.0051018665234247845
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,2560,768,0.018627200524012247
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,2560,768,0.008778666456540424
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,2560,512,0.004266666869322458
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,2560,1024,0.00920746624469757
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,2560,512,0.017026132345199584
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,2560,256,0.0037429332733154297
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,2560,256,0.016536532839139303
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,2560,256,0.007978666822115581
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,2560,128,0.0033834666013717652
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,2560,512,0.008322133123874665
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,2560,128,0.016270933548609416
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,2560,64,0.0030805334448814393
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,2560,64,0.017097600301106772
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,2560,32,0.0033600000043710076
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,2560,128,0.007499733567237854
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,2560,32,0.017101866006851197
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,2048,65536,0.15179519653320311
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,2048,65536,0.21399359703063964
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,2048,16384,0.03883626858393351
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,2048,16384,0.05576426585515341
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,2048,16384,0.042700799306233723
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,2048,12288,0.02953280011812846
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,2048,12288,0.04522879918416341
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,2048,12288,0.034230399131774905
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,2048,10240,0.02491733431816101
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,2048,10240,0.04156479835510254
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,2560,65536,0.14562880198160807
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,2048,8192,0.020623999834060668
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,2048,8192,0.03731946547826131
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,2048,10240,0.029981867472330732
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,2048,7168,0.018241065740585326
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,2048,7168,0.034509865442911784
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,2048,8192,0.025260800123214723
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,2048,6144,0.01588266690572103
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,2048,6144,0.030881067117055256
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,2048,7168,0.023119999965031942
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,2048,5120,0.013877333203951518
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,2048,5120,0.029127466678619384
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,2048,6144,0.021125332514444987
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,2048,4096,0.011714133620262145
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,2048,4096,0.026866134007771807
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,2048,5120,0.019298134247461955
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,2048,3584,0.010498133301734925
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,2048,3584,0.02634773254394531
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,2048,4096,0.01657386620839437
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,2048,3072,0.009609599908192951
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,2048,3072,0.024881066878636678
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,2048,3584,0.015332266688346863
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,2048,2560,0.008356266220410665
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,2048,2560,0.023069866498311362
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,2048,3072,0.014170666535695394
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,2048,2048,0.007157333195209503
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,2048,2048,0.021949867407480873
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,2048,2560,0.012889599800109864
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,2048,1536,0.006307200094064077
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,2048,1536,0.020709333817164104
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,2048,2048,0.011637333035469054
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,2048,1024,0.005448533097902933
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,2048,1024,0.018742400407791137
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,2048,1536,0.010798933108647664
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,2048,768,0.004682666560014089
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,2048,768,0.01837546626726786
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,2048,1024,0.00893440047899882
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,2048,512,0.004218666752179464
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,2048,512,0.01755733291308085
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,2048,768,0.008806399504343669
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,2048,256,0.0038431999584039056
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,2048,256,0.01664746701717377
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,2048,512,0.007979733248551685
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,2048,128,0.0033759998778502146
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,2048,128,0.015973333517710367
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,2048,256,0.007993599772453308
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,2048,64,0.0030559999247392017
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,2048,64,0.014797866344451904
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,2048,128,0.0075573335091273
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,2048,32,0.0034143999218940735
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,2048,32,0.014502400159835815
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,1536,65536,0.1253376007080078
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,1536,65536,0.20108480453491212
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,1536,16384,0.03334826628367106
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,1536,16384,0.05133333206176758
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,1536,16384,0.04206186532974243
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,1536,12288,0.02481493353843689
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,1536,12288,0.04283413489659627
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,2048,65536,0.1456501324971517
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,1536,10240,0.021211733420689903
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,1536,10240,0.0391103982925415
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,1536,12288,0.03305066625277202
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,1536,8192,0.01759999990463257
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,1536,8192,0.0350165327390035
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,1536,10240,0.02932586669921875
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,1536,7168,0.015710933009783427
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,1536,7168,0.03298133412996928
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,1536,7168,0.02247999906539917
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,1536,6144,0.013980799913406372
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,1536,6144,0.030808534224828082
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,1536,8192,0.02476373314857483
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,1536,5120,0.012014933427174886
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,1536,5120,0.029131732384363812
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,1536,5120,0.018643200397491455
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,1536,4096,0.010098133484522502
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,1536,4096,0.02595306634902954
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,1536,6144,0.020772266387939452
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,1536,3584,0.009317333499590557
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,1536,3584,0.026451200246810913
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,1536,3584,0.01499626636505127
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,1536,3072,0.008348799745241801
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,1536,3072,0.024599466721216837
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,1536,4096,0.016178133090337117
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,1536,2560,0.007166933516661327
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,1536,2560,0.023893332481384276
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,3584,256,0.007949866851170858
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,1536,2048,0.006325333317120869
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,1536,2560,0.012467199563980102
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,1536,2048,0.021460266908009847
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,1536,3072,0.013704533378283182
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,1536,1536,0.005489066739877065
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,1536,1536,0.02036906679471334
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,1536,1024,0.004636799792448679
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,1536,1024,0.01853760083516439
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,1536,1024,0.009226666887601216
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,1536,768,0.0042015999555587765
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,1536,1536,0.010353066523869832
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,1536,768,0.016710400581359863
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,1536,512,0.0034965333839257562
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,1536,512,0.018032000462214152
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,1536,2048,0.011640533804893494
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,1536,256,0.0029919999341169994
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,1536,256,0.015796266992886863
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,1536,768,0.00872320036093394
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,1536,512,0.008291199803352356
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,1536,128,0.0026346666117509207
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,1536,128,0.015057067076365152
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,1536,256,0.007582933207352956
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,1536,64,0.0026208000878492994
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,1536,128,0.0075541332364082335
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,1536,32,0.002632533262173335
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,1536,32,0.014855466286341348
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,1024,65536,0.08457706769307455
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,1024,65536,0.1804693380991618
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,1024,16384,0.02526400089263916
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,1024,16384,0.05003413359324137
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,1024,16384,0.04158506790796916
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,1024,12288,0.020551466941833497
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,1024,12288,0.04062186479568482
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,1024,12288,0.03310400048891703
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,1024,10240,0.017293866475423178
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,1024,10240,0.03696426550547282
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,1536,65536,0.14617279370625813
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,1024,8192,0.015520000457763672
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,1024,8192,0.03493760029474895
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,1024,10240,0.02929813265800476
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,1024,7168,0.013459199666976928
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,1024,7168,0.030306132634480794
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,1024,8192,0.024357332785924276
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,1024,6144,0.011860266327857971
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,1024,6144,0.027462400992711383
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,1024,7168,0.022705066204071044
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,1024,5120,0.010384000341097514
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,1024,6144,0.02069013317426046
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,1024,5120,0.025730133056640625
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,1024,4096,0.008254933357238769
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,1024,4096,0.025808000564575197
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,1024,5120,0.01867093245188395
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,1024,3584,0.007179733117421467
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,1024,4096,0.01607146660486857
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,1024,3584,0.025265065828959148
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,1024,3072,0.006733866532643636
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,1024,3072,0.023088000218073525
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,1024,3072,0.014137599865595499
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,1024,2560,0.005946666498978933
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,1024,2560,0.022501333554585775
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,1024,2560,0.012538666526476542
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,1024,2048,0.0052490666508674625
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,1024,2048,0.020869332551956176
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,1024,2048,0.011642666657765706
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,1024,1536,0.004665599763393402
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,1024,1536,0.019242666165033975
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,1024,1536,0.010398933291435241
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,1024,1024,0.0038143999874591826
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,1024,1024,0.018605866034825645
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,1024,1024,0.008795733253161114
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,1024,768,0.003492266684770584
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,1024,768,0.016597333550453185
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,1024,768,0.008762666583061218
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,1024,512,0.0030847998956839246
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,1024,512,0.016436266899108886
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,1024,512,0.007973333199818928
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,1024,256,0.002995199958483378
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,1024,256,0.014882133404413859
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,1024,256,0.007956266899903615
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,1024,128,0.0029343999922275543
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,1024,128,0.016357333461443583
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,1024,128,0.007763200004895528
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,1024,64,0.0026186667382717133
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,1024,64,0.01454080045223236
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,1024,32,0.002648533384005229
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,1024,32,0.015837867061297098
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,768,65536,0.0614741325378418
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,768,65536,0.17134826978047688
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,1024,65536,0.1453877290089925
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,768,16384,0.018937599658966065
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,768,16384,0.04789653221766154
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,768,16384,0.041653335094451904
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,768,12288,0.017849600315093993
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,768,12288,0.041206399599711105
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,768,12288,0.03302293419837952
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,768,10240,0.015014400084813436
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,768,10240,0.037062398592631024
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,768,10240,0.028869332869847615
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,768,8192,0.012939733266830445
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,768,8192,0.032076799869537355
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,768,8192,0.024753065903981526
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,768,7168,0.011667199929555257
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,768,7168,0.03057493368784587
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,768,7168,0.02234986623128255
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,768,6144,0.010445866982142131
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,768,6144,0.02827306588490804
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,768,6144,0.020618667205174766
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,768,5120,0.008845866719881693
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,768,5120,0.02614826758702596
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,768,5120,0.018633600076039633
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,768,4096,0.007546666761239369
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,768,4096,0.02477546731630961
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,768,65536,0.14600106875101726
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,768,3584,0.006785066425800323
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,768,3584,0.023868799209594727
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,768,4096,0.016107733050982155
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,768,3072,0.006287999947865804
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,768,3584,0.01490239997704824
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,768,3072,0.02268799940745036
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,768,3072,0.014127999544143677
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,768,2560,0.0055189331372578945
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,768,2560,0.021741867065429688
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,768,2048,0.005144533514976501
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,768,2048,0.02078826626141866
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,768,2560,0.01284160017967224
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,768,1536,0.004285866518815359
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,768,1536,0.018793600797653198
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,768,2048,0.011623467008272808
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,768,1024,0.0037418665985266366
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,768,1024,0.018595200777053834
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,768,1536,0.010381866494814556
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,768,768,0.003402666747570038
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,768,768,0.017942400773366292
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,768,1024,0.009166933099428813
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,768,512,0.0029887999097506206
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,768,512,0.016178133090337117
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,768,768,0.00870293378829956
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,768,256,0.0026506667335828146
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,768,512,0.008296533425649007
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,768,256,0.016404267152150473
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,768,128,0.0025418666501839954
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,768,128,0.014688000082969666
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,768,256,0.00790826678276062
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,768,64,0.002253866692384084
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,768,128,0.007730133334795634
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,768,64,0.014484266440073649
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,768,32,0.0025429333249727885
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,768,32,0.014791466792424521
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,512,65536,0.045287466049194335
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,512,65536,0.16517972946166992
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,512,16384,0.016679465770721436
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,512,16384,0.04568320115407308
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,512,16384,0.041530664761861166
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,512,12288,0.01437333325544993
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,512,12288,0.03930453459421794
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,512,12288,0.03269760012626648
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,512,10240,0.014509866635004679
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,512,10240,0.03503679831822713
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,512,10240,0.028874667485555013
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,512,8192,0.012942933042844138
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,512,8192,0.03279146750768026
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,512,8192,0.024638932943344117
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,512,7168,0.011250133315722149
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,512,7168,0.031522132953008014
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,512,7168,0.02236586610476176
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,512,6144,0.010029866298039754
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,512,6144,0.028732800483703615
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,512,6144,0.020652800798416138
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,512,5120,0.008380800485610962
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,512,5120,0.02616426746050517
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,512,5120,0.01857173244158427
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,512,4096,0.007189333438873291
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,512,4096,0.025088000297546386
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,512,65536,0.14638080596923828
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,512,3584,0.007121066749095917
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,512,3584,0.023550933599472045
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,512,4096,0.016139733791351318
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,512,3072,0.006332799792289734
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,512,3584,0.014883200327555338
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,512,3072,0.022796799739201866
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,512,2560,0.005453866720199585
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,512,2560,0.02071466644605001
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,512,3072,0.01405333379904429
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,512,2560,0.01251520017782847
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,512,2048,0.00506986677646637
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,512,2048,0.020881066719690956
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,512,1536,0.004283733169237773
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,512,1536,0.018662399053573607
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,512,2048,0.011238400141398113
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,512,1024,0.003788800040880839
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,512,1536,0.01040000021457672
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,512,1024,0.018553600708643595
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,512,768,0.003421866645415624
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,512,768,0.016659200191497803
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,512,1024,0.009131733576456707
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,512,512,0.003049599876006444
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,512,768,0.008372267087300617
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,512,512,0.0173909326394399
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,512,256,0.0027093333502610523
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,512,256,0.014618666966756186
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,512,512,0.008338133494059246
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,512,128,0.002537599951028824
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,512,256,0.007550933460394542
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,512,128,0.015931733449300132
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,512,64,0.002197333425283432
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,512,128,0.007516799867153168
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,512,64,0.014491732915242514
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,512,32,0.0025962665677070618
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,512,32,0.014947199821472168
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,256,65536,0.031864533821741745
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,256,65536,0.1586901346842448
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,256,16384,0.012088533242543538
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,256,16384,0.040190935134887695
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,256,16384,0.04112533330917358
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,256,12288,0.010403199990590414
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,256,12288,0.0351093331972758
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,256,12288,0.0329205334186554
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,256,10240,0.00940053363641103
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,256,10240,0.03446186780929565
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,256,10240,0.02927253246307373
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,256,8192,0.008356266220410665
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,256,8192,0.030913066864013673
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,256,65536,0.1451050599416097
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,256,7168,0.007974400122960409
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,256,7168,0.029509333769480388
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,256,8192,0.02477653423945109
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,256,6144,0.007548800110816956
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,256,6144,0.0269813338915507
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,256,7168,0.0226474662621816
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,256,5120,0.00856213370958964
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,256,5120,0.02566186587015788
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,256,6144,0.02095253268877665
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,256,4096,0.007067733506361644
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,256,4096,0.023798400163650514
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,256,5120,0.01856000026067098
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,256,3584,0.006303999821345012
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,256,3584,0.023835732539494833
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,256,4096,0.016122666994730632
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,256,3072,0.005902933577696482
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,256,3072,0.022771199544270836
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,256,3584,0.014988799889882406
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,256,2560,0.005115733544031779
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,256,2560,0.021385600169499717
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,256,3072,0.01406719982624054
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,256,2048,0.004625066618124644
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,256,2048,0.0206496000289917
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,256,2560,0.012822399536768595
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,256,1536,0.004194133480389913
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,256,1536,0.019128533204396565
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,256,2048,0.011566932996114094
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,256,1024,0.003479466587305069
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,256,1024,0.01873706579208374
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,256,1536,0.010058666268984478
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,256,768,0.003403733422358831
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,256,1024,0.009180800120035807
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,256,768,0.016824533541997276
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,256,512,0.0030239999294281008
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,256,512,0.016506666938463845
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,256,768,0.008345599969228108
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,256,256,0.002621866762638092
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,256,256,0.014797866344451904
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,256,512,0.00829013337691625
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,256,128,0.0022709332406520844
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,256,128,0.01628373364607493
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,256,128,0.007478400071461995
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,256,64,0.0022783999641736346
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,256,64,0.016526933511098227
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,256,32,0.002414933343728383
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,256,32,0.014573867122332254
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,128,65536,0.024112000068028768
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,128,65536,0.15326719284057616
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,128,65536,0.14512640635172527
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,128,16384,0.011239467064539592
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,128,16384,0.03922239939371745
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,128,16384,0.041155199209849044
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,128,12288,0.00960533320903778
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,128,12288,0.036033066113789876
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,128,12288,0.0329205334186554
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,128,10240,0.008729599912961324
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,128,10240,0.03380053440729777
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,128,10240,0.028446932633717854
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,128,8192,0.00798933357000351
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,128,8192,0.030869332949320476
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,128,8192,0.024333866437276204
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,128,7168,0.007565866907437642
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,128,7168,0.028860799471537274
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,128,7168,0.02235520084698995
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,128,6144,0.007129600147406261
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,128,6144,0.027713066339492796
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,128,6144,0.020207999149958293
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,128,5120,0.007141333321730297
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,128,5120,0.025114667415618897
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,128,5120,0.018225065867106118
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,128,4096,0.0066890666882197065
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,128,4096,0.024691200256347655
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,128,4096,0.015702399611473083
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,128,3584,0.006297599772612255
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,128,3584,0.02302079995473226
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,128,3584,0.014884266257286071
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,128,3072,0.0055167997876803074
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,128,3072,0.022617600361506143
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,128,3072,0.013613866766293845
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,128,2560,0.0051146666208903
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,128,2560,0.020910932620366415
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,128,2560,0.012456533312797547
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,128,2048,0.004620799918969473
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,128,2048,0.020753065745035805
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,128,2048,0.011218133568763732
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,128,1536,0.003857066730658213
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,128,1536,0.019163733720779418
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,128,1536,0.01002239982287089
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,128,1024,0.003368533402681351
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,128,1024,0.018294399976730345
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,128,1024,0.008734933535257975
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,128,768,0.0030752000709374744
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,128,768,0.01767680048942566
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,128,768,0.008351999521255492
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,128,512,0.0028021333118279776
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,128,512,0.016940800348917644
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,128,512,0.00792746643225352
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,128,256,0.002656000107526779
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,128,256,0.01563093364238739
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,128,256,0.007526400188604991
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,128,128,0.0022656001150608064
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,128,128,0.014533332983652749
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,128,128,0.0074890668193499255
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,128,64,0.0022837333381175995
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,128,64,0.014589866995811463
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,128,32,0.002282666663328807
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,128,32,0.014929067095120749
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,64,65536,0.02214399973551432
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,64,65536,0.1525045394897461
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,64,16384,0.008332799871762593
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,64,16384,0.03708159923553467
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,64,12288,0.007156266768773396
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,64,12288,0.033763198057810466
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,64,10240,0.006680533289909363
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,64,10240,0.03302293419837952
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,64,8192,0.0062720000743865965
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,64,8192,0.030938667058944703
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,64,7168,0.005919999877611796
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,64,7168,0.02877013285954793
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,64,6144,0.006298666695753734
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,64,6144,0.027437865734100342
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,64,5120,0.006681600213050842
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,64,5120,0.02505279978116353
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,64,4096,0.0062613333264986675
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,64,4096,0.024361600478490196
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,64,3584,0.005846400062243144
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,64,3584,0.023014400402704874
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,64,3072,0.005530666808287303
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,64,3072,0.02272426684697469
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,64,2560,0.005436799923578898
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,64,2560,0.021364265680313112
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,64,2048,0.0046165332198143
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,64,2048,0.020470400651295982
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,64,1536,0.0038634667793909705
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,64,1536,0.018780799706776936
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,64,1024,0.0034101332227389016
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,64,1024,0.017633066574732462
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,64,768,0.003014400104681651
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,64,768,0.016565333803494772
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,64,512,0.0026367999613285064
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,64,512,0.015847466389338174
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,64,256,0.0025301332275072734
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,64,256,0.014642133315404256
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,64,128,0.0022613334159056345
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,64,128,0.01609386702378591
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,64,64,0.0021162666380405426
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,64,64,0.014896000425020853
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,64,32,0.002203733225663503
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,64,32,0.01458026667435964
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,32,65536,0.01646080017089844
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,32,65536,0.1547221342722575
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,32,16384,0.009913600484530131
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,32,16384,0.03719573418299357
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,32,12288,0.008278400202592214
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,32,12288,0.034031999111175534
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,32,10240,0.007547733187675476
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,32,10240,0.03256319959958394
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,32,8192,0.006658133367697399
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,32,8192,0.030479999383290608
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,32,7168,0.006293333570162455
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,32,7168,0.028794666131337483
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,32,6144,0.006307200094064077
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,32,6144,0.026885332663853963
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,32,5120,0.005868799984455109
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,32,5120,0.025498666365941364
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,32,4096,0.006263466676076253
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,32,4096,0.024476800362269083
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,32,3584,0.005904000004132589
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,32,3584,0.023099732398986817
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,32,3072,0.005683200061321258
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,32,3072,0.023117866118748984
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,32,2560,0.005449600021044413
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,32,2560,0.02134400010108948
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,32,2048,0.004353066782156626
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,32,2048,0.021025067567825316
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,32,1536,0.004289066791534424
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,32,1536,0.01917866667111715
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,32,1024,0.0034645333886146545
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,32,1024,0.017256534099578856
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,32,768,0.003033600002527237
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,32,768,0.017241599162419637
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,32,512,0.002600533266862233
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,32,512,0.01618026693662008
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,32,256,0.0025631998976071674
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,32,256,0.014589866995811463
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,32,128,0.002221866697072983
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,32,128,0.014640000462532044
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,32,64,0.002195200075705846
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,32,64,0.014498133460680643
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,768,32,32,0.0021898667017618816
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,32,32,0.01492800017197927
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,65536,16384,0.6563135782877605
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,65536,16384,0.4779061317443848
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,768,1536,64,0.014433067043622336
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,65536,12288,0.5022794723510742
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,65536,12288,0.28016640345255533
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,1024,3584,0.015284267067909241
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,65536,10240,0.42354027430216473
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,65536,10240,0.2340031941731771
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,768,256,256,0.007942399879296621
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,65536,8192,0.3408245404561361
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,65536,8192,0.1867029349009196
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,65536,10240,0.15894932746887208
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,65536,7168,0.2926080067952474
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,65536,7168,0.16738559405008951
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,65536,8192,0.12815679709116617
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,65536,6144,0.2520714600880941
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,65536,6144,0.14496639569600422
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,65536,12288,0.18261440594991046
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,65536,5120,0.2114720026652018
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,65536,5120,0.12030719916025798
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,65536,16384,0.24043520291646323
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,65536,4096,0.17112000783284503
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,65536,4096,0.10403947035471599
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,65536,6144,0.10048960049947102
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,65536,3584,0.14858026504516603
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,65536,3584,0.09067626794179282
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,65536,7168,0.11565759976704915
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,65536,3072,0.13228267033894855
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,65536,3072,0.08281599680582682
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,65536,5120,0.08518826961517334
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,65536,2560,0.1076543966929118
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,65536,2560,0.06893760363260905
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,65536,4096,0.07224000295003255
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,65536,2048,0.09122986793518066
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,65536,2048,0.060974931716918944
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,65536,3072,0.058565334479014075
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,65536,1536,0.06890347003936767
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,65536,1536,0.05118720134099325
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,65536,2560,0.050861867268880215
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,65536,1024,0.04666453202565511
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,65536,1024,0.04277333418528239
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,65536,3584,0.06523093382517496
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,65536,768,0.037996800740559895
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,65536,768,0.037146667639414474
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,65536,2048,0.04462186495463054
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,65536,512,0.028249599536259967
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,65536,512,0.03310186664263408
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,65536,768,0.02687893311182658
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,65536,256,0.018589866161346436
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,65536,256,0.029594665765762328
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,65536,512,0.023718400796254476
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,65536,128,0.01585706671079
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,65536,128,0.02704106569290161
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,65536,256,0.019745065768559774
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,65536,64,0.013726933797200521
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,65536,64,0.02687893311182658
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,65536,32,0.014086400469144186
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,65536,32,0.026869332790374754
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,65536,1024,0.030538666248321533
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,16384,65536,0.6542698542277019
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,16384,65536,0.42934614817301436
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,65536,128,0.017986132701237997
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,16384,16384,0.16655893325805665
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,16384,16384,0.11207786401112874
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,16384,12288,0.12667307058970134
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,16384,12288,0.09399680296579996
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,65536,1536,0.038159998257954915
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,16384,10240,0.11131306489308675
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,16384,10240,0.08239999612172445
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,16384,10240,0.05208959976832071
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,16384,8192,0.08979199727376302
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,16384,8192,0.06474026838938395
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,16384,12288,0.05913706620534261
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,16384,7168,0.0790773312250773
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,16384,7168,0.061186134815216064
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,16384,16384,0.07424213091532389
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,16384,6144,0.0654698650042216
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,16384,6144,0.05782613356908163
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,16384,7168,0.03938560088475545
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,16384,5120,0.057885865370432533
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,16384,5120,0.04989866813023885
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,16384,8192,0.04339413245519002
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,16384,4096,0.04645866552988688
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,16384,4096,0.04362560113271077
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,16384,4096,0.027385600407918292
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,16384,3584,0.041196799278259276
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,16384,3584,0.041042133172353105
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,16384,6144,0.03490133285522461
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,16384,3072,0.03494826555252075
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,16384,3072,0.037342933813730876
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,16384,5120,0.03115306695302327
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,16384,2560,0.03068266709645589
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,16384,2560,0.03507733345031738
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,16384,2560,0.020487467447916664
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,16384,2048,0.025229867299397784
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,16384,2048,0.03130879998207092
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,16384,3584,0.024975999196370443
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,16384,1536,0.02013546625773112
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,16384,1536,0.027963733673095702
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,16384,3072,0.02230506738026937
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,16384,1024,0.014468266566594442
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,16384,1024,0.02524799903233846
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,16384,1024,0.01250879963239034
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,16384,768,0.011589333415031433
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,16384,768,0.02365013360977173
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,16384,1536,0.01622719963391622
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,16384,512,0.009569066762924194
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,16384,512,0.022589866320292154
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,16384,512,0.010494933525721232
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,16384,256,0.007430399954319
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,16384,2048,0.018263467152913413
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,16384,768,0.011181867122650147
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,16384,128,0.005936000247796377
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,16384,128,0.018593066930770875
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,16384,256,0.009586133559544881
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,16384,64,0.006538666784763336
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,16384,64,0.018898133436838785
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,16384,32,0.006682666639486949
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,16384,32,0.018894932667414346
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,12288,65536,0.5220949490865071
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,12288,65536,0.3391925175984701
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,16384,256,0.020669867595036827
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,16384,128,0.008763733506202697
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,12288,16384,0.09300159613291423
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,12288,16384,0.1327829360961914
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,12288,12288,0.09532159964243571
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,12288,12288,0.07555306752522786
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,12288,12288,0.05963519811630249
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,12288,10240,0.08758506774902344
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,12288,10240,0.06577066580454508
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,12288,16384,0.07399253050486246
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,12288,8192,0.06790613333384196
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,12288,8192,0.05226453145345052
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,12288,10240,0.05170559883117676
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,12288,7168,0.06211093266805014
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,12288,7168,0.051203199227650965
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,12288,8192,0.042260265350341795
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,12288,6144,0.05148693323135376
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,12288,6144,0.048951466878255204
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,12288,7168,0.03879679838816325
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,12288,5120,0.045126398404439286
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,12288,5120,0.04233920176823934
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,16384,65536,0.2803573290506999
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,12288,4096,0.03511679967244466
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,12288,4096,0.038525867462158206
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,12288,6144,0.03466666539510091
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,12288,3584,0.030573866764704388
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,12288,3584,0.035588268438975015
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,12288,5120,0.03112000028292338
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,12288,3072,0.02775999903678894
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,12288,3072,0.03391466538111369
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,12288,4096,0.026745599508285523
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,12288,2560,0.023642667134602866
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,12288,2560,0.031206399202346802
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,12288,3584,0.024413865804672242
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,12288,2048,0.018883200486501057
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,12288,2048,0.028526933987935384
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,12288,3072,0.02232853372891744
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,12288,1536,0.01555519998073578
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,12288,1536,0.026632533470789595
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,12288,2048,0.017760000626246133
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,12288,1024,0.011390933394432068
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,12288,1024,0.022443733612696328
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,12288,2560,0.02021226684252421
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,12288,768,0.009783466657002766
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,12288,768,0.020705066124598184
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,12288,1536,0.01529813309510549
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,12288,512,0.0075914666056633
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,12288,512,0.02050559918085734
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,12288,768,0.011290666460990906
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,12288,256,0.006670933465162914
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,12288,256,0.01984213391939799
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,12288,1024,0.012498133381207784
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,12288,128,0.0058890665570894875
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,12288,128,0.01660053332646688
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,12288,256,0.009202133615811665
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,12288,64,0.00552106648683548
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,12288,128,0.009198932846387228
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,12288,64,0.018451199928919474
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,12288,32,0.005508266886075338
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,12288,32,0.018548266092936198
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,12288,512,0.010067199667294819
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,10240,65536,0.4351680119832356
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,10240,65536,0.3049994786580404
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,10240,16384,0.11073173681894939
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,10240,16384,0.08901653289794922
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,10240,12288,0.08581866423288981
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,10240,12288,0.07398293018341065
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,12288,65536,0.2632383982340495
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,10240,10240,0.07233813603719076
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,10240,10240,0.06090773344039917
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,10240,10240,0.05078826745351156
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,10240,8192,0.05822399854660034
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,10240,8192,0.04966613451639811
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,10240,12288,0.05932053327560425
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,10240,7168,0.05162560145060221
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,10240,7168,0.04772160053253174
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,10240,16384,0.07383573055267334
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,10240,6144,0.04357226689656575
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,10240,6144,0.04477546612421672
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,10240,7168,0.038541865348815915
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,10240,5120,0.03725866476694743
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,10240,5120,0.04101119836171468
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,10240,8192,0.04210346539815267
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,10240,4096,0.030504532655080158
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,10240,4096,0.03531946738560994
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,10240,6144,0.03423253297805786
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,10240,3584,0.026019199689229326
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,10240,3584,0.03288960059483846
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,10240,4096,0.02649173339207967
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,10240,3072,0.023425066471099855
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,10240,3072,0.030921600262324017
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,10240,5120,0.030602665742238362
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,10240,2560,0.02025173306465149
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,10240,2560,0.028810666004816694
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,10240,3584,0.024203733603159586
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,10240,2048,0.016189866264661155
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,10240,2048,0.027076266209284466
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,10240,2560,0.019891200462977092
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,10240,1536,0.01341759959856669
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,10240,1536,0.02486506700515747
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,10240,3072,0.02190933426221212
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,10240,1024,0.010025599598884582
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,10240,1024,0.02188053329785665
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,10240,1024,0.01209173301855723
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,10240,768,0.008387200037638346
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,10240,768,0.020581332842508952
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,10240,1536,0.014792533715566
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,10240,512,0.006679466863473256
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,10240,512,0.020403200387954713
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,10240,768,0.01123413344224294
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,10240,256,0.005493333439032236
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,10240,256,0.017509333292643228
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,10240,512,0.010016000270843506
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,10240,128,0.005009066561857859
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,10240,128,0.01664426624774933
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,10240,256,0.009179733196894328
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,10240,64,0.004678399860858917
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,10240,64,0.01684160033861796
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,10240,32,0.004735999802748362
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,10240,32,0.016720000902811685
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,10240,2048,0.016908800601959227
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,10240,128,0.008762666583061218
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,8192,65536,0.25968000094095867
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,8192,65536,0.3415594736735026
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,8192,16384,0.08542400201161703
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,8192,16384,0.07404266993204753
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,8192,12288,0.06605013211568198
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,8192,12288,0.06205013195673624
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,8192,12288,0.035512534777323405
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,8192,10240,0.056569600105285646
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,8192,10240,0.0536629319190979
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,8192,16384,0.04342720111211141
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,8192,8192,0.04451733430226644
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,8192,8192,0.04545493523279826
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,8192,10240,0.03163413405418396
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,8192,7168,0.04057173331578572
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,8192,7168,0.043350398540496826
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,8192,8192,0.0264682670434316
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,8192,6144,0.03467626571655273
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,8192,6144,0.04017813205718994
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,8192,7168,0.024385066827138265
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,8192,5120,0.030230400959650676
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,8192,5120,0.03679573138554891
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,8192,6144,0.02227413256963094
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,8192,4096,0.024222934246063234
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,8192,4096,0.0332202672958374
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,8192,5120,0.02039466698964437
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,8192,3584,0.02223680019378662
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,8192,3584,0.03225066661834717
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,8192,4096,0.017947733402252197
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,8192,3072,0.019181867440541588
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,8192,3072,0.029329067468643187
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,8192,3584,0.016596266627311708
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,8192,2560,0.017139200369517008
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,8192,2560,0.028118399779001872
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,8192,3072,0.014598400394121806
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,8192,2048,0.014393599828084311
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,8192,2048,0.025198932488759356
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,8192,2560,0.013361066579818726
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,8192,1536,0.012184533476829528
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,8192,1536,0.023565866549809775
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,8192,1536,0.010461866855621338
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,8192,1024,0.009348266323407491
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,8192,1024,0.021228800217310585
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,8192,2048,0.012063999970753986
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,8192,768,0.007608533402283986
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,8192,768,0.01873813271522522
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,8192,1024,0.009572266538937887
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,8192,512,0.006043733159701029
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,8192,512,0.018705066045125326
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,8192,768,0.008835200468699138
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,8192,256,0.004763733347256978
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,8192,256,0.01761066714922587
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,8192,256,0.007628799974918365
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,8192,128,0.004215466479460398
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,8192,128,0.015723733107248943
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,8192,512,0.008345599969228108
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,8192,64,0.004228266576925913
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,8192,64,0.01781546672185262
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,8192,32,0.005099733173847198
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,8192,32,0.0171615997950236
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,8192,128,0.007500799993673961
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,7168,65536,0.3095594724019369
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,7168,65536,0.25828906695048015
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,7168,16384,0.07839146455128988
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,7168,16384,0.07069546381632487
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,7168,16384,0.04279893239339193
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,7168,12288,0.05910826524098715
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,7168,12288,0.059412264823913576
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,10240,65536,0.26363520622253417
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,7168,10240,0.050987732410430905
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,7168,10240,0.05145493348439535
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,7168,12288,0.034519465764363606
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,7168,8192,0.040746665000915526
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,7168,8192,0.04328213135401408
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,7168,10240,0.030989867448806763
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,7168,7168,0.03699520031611125
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,7168,7168,0.04062079985936483
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,7168,8192,0.026013867060343428
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,7168,6144,0.03131306568781535
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,7168,6144,0.037783467769622804
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,8192,65536,0.1498143990834554
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,7168,5120,0.02649066646893819
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,7168,5120,0.03448853492736816
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,7168,7168,0.02402879993120829
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,7168,4096,0.0221781333287557
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,7168,4096,0.032365866502126056
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,7168,6144,0.02207146684328715
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,7168,3584,0.019287467002868652
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,7168,3584,0.029709867636362713
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,7168,5120,0.019900800784428914
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,7168,3072,0.017473065853118898
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,7168,3072,0.02967039942741394
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,7168,4096,0.017387733856836955
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,7168,2560,0.01497706671555837
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,7168,2560,0.026807467142740887
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,7168,3072,0.014643200238545737
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,7168,2048,0.012708266576131185
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,7168,2048,0.026075732707977296
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,7168,3584,0.01599360009034475
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,7168,1536,0.01018453339735667
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,7168,1536,0.022504534324010214
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,7168,2560,0.013427199920018515
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,7168,1024,0.007975466549396515
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,7168,1024,0.019682133197784425
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,7168,1024,0.009559466441472372
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,7168,768,0.006742399930953979
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,7168,768,0.019027199347813925
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,7168,1536,0.010774399836858113
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,7168,512,0.00547626664241155
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,7168,512,0.01707520087560018
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,7168,768,0.00878613293170929
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,7168,256,0.004295466840267182
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,7168,512,0.008316799998283386
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,7168,256,0.015485866864522298
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,7168,128,0.0038431999584039056
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,7168,2048,0.012035199999809265
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,7168,128,0.015752533078193666
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,7168,64,0.003409066547950109
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,7168,64,0.01585813363393148
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,7168,32,0.0034634667138258614
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,7168,32,0.016521599888801575
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,7168,128,0.007501866420110066
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,6144,65536,0.26557653745015464
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,7168,256,0.007520000139872233
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,6144,65536,0.23705387115478516
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,6144,16384,0.06638293266296387
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,6144,16384,0.06186453501383463
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,6144,12288,0.05087039868036906
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,6144,12288,0.05241599877675375
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,6144,12288,0.03429866631825765
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,6144,10240,0.0439306656519572
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,6144,10240,0.045399467150370285
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,6144,16384,0.04242240190505982
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,6144,8192,0.03431679805119832
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,6144,8192,0.03977813323338826
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,6144,10240,0.030282666285832722
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,6144,7168,0.031491200129191085
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,6144,7168,0.03889919916788737
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,6144,8192,0.025614933172861738
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,6144,6144,0.02672213315963745
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,6144,6144,0.03555946747461955
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,6144,7168,0.024010666211446128
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,6144,5120,0.023614933093388878
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,6144,5120,0.03360106547673543
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,6144,6144,0.021375999848047892
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,6144,4096,0.018687999248504637
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,6144,4096,0.0298634668191274
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,7168,65536,0.144377597173055
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,6144,3584,0.017072000106175742
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,6144,3584,0.029021867116292316
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,6144,5120,0.019323732455571493
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,6144,3072,0.015416533748308817
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,6144,3072,0.029340799649556475
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,6144,3584,0.015397333105405173
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,6144,2560,0.013303466637929282
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,6144,2560,0.026124799251556398
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,6144,4096,0.01742186745007833
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,6144,2048,0.011240532994270325
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,6144,2048,0.022966400782267252
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,6144,3072,0.014505599935849508
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,6144,1536,0.00965119997660319
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,6144,1536,0.02304746707280477
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,6144,2048,0.011657599608103435
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,6144,1024,0.007467733323574066
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,6144,2560,0.012859732906023661
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,6144,1024,0.018742400407791137
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,6144,768,0.0062826668222745255
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,6144,768,0.01864853302637736
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,6144,1536,0.010424533486366272
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,6144,512,0.005498666564623515
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,6144,512,0.018066134055455527
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,6144,768,0.008724266290664673
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,6144,256,0.004614399870236715
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,6144,256,0.016536532839139303
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,6144,512,0.0083146666487058
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,6144,128,0.004215466479460398
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,6144,128,0.016566399733225504
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,6144,1024,0.009178666273752849
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,6144,64,0.003902933249870936
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,6144,128,0.007536000013351441
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,6144,32,0.0038773333032925926
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,6144,256,0.007575466732184092
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,6144,32,0.016523733735084534
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,5120,65536,0.2171018600463867
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,5120,65536,0.21825812657674154
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,5120,16384,0.05692053238550822
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,5120,16384,0.06031573216120402
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,5120,16384,0.041902931531270345
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,5120,12288,0.044359465440114335
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,5120,12288,0.05119786659876505
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,5120,12288,0.03342506488164266
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,5120,10240,0.03764266570409139
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,5120,10240,0.04328639904658
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,5120,10240,0.02977493405342102
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,5120,8192,0.030324266354242964
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,5120,8192,0.037859201431274414
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,5120,8192,0.025443200270334882
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,5120,7168,0.027135999997456868
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,5120,7168,0.03666666746139526
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,6144,65536,0.1413973331451416
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,5120,6144,0.022587732474009196
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,5120,6144,0.03323946595191955
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,5120,6144,0.021630932887395225
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,5120,5120,0.01978773276011149
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,5120,5120,0.031777065992355344
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,5120,7168,0.02363199989000956
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,5120,4096,0.016570666432380678
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,5120,4096,0.02900480031967163
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,5120,5120,0.019053866465886436
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,5120,3584,0.01493119994799296
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,5120,3584,0.02759360074996948
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,5120,4096,0.016555733482042947
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,5120,3072,0.013315199812253316
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,5120,3072,0.026840533812840777
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,5120,3072,0.013773866494496665
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,5120,2560,0.011796266833941142
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,5120,2560,0.024971733490626015
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,5120,3584,0.01573013365268707
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,5120,2048,0.00993066628774007
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,5120,2048,0.022648533185323082
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,5120,2048,0.011269332965215047
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,5120,1536,0.00837546686331431
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,5120,1536,0.0206112007300059
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,5120,2560,0.012910933295885722
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,5120,1024,0.006757333377997081
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,5120,1024,0.01981653372446696
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,5120,1024,0.008844799796740214
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,5120,768,0.005977599819501241
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,5120,768,0.018576000134150186
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,5120,1536,0.010010666648546855
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,5120,512,0.0046847999095916745
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,5120,512,0.018769067525863648
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,5120,512,0.008327466746171314
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,5120,768,0.008738133311271667
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,5120,256,0.0038463999827702843
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,5120,256,0.01753600041071574
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,5120,128,0.003516799956560135
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,5120,128,0.015361066659291586
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,5120,128,0.007513600091139476
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,5120,64,0.0034122665723164878
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,5120,256,0.007922133306662242
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,5120,32,0.0033749334514141084
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,5120,64,0.014672000209490457
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,5120,32,0.016132266322771708
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,4096,65536,0.19090453783671063
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,4096,65536,0.18531626065572102
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,4096,16384,0.04855146805445353
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,4096,16384,0.05554666519165039
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,5120,65536,0.1378335952758789
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,4096,12288,0.037163734436035156
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,4096,12288,0.044869331518809
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,4096,16384,0.041016534964243574
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,4096,10240,0.03237333297729492
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,4096,10240,0.03884799877802531
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,4096,10240,0.029106134176254274
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,4096,8192,0.026282666126887004
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,4096,8192,0.034550400575002034
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,4096,12288,0.03289706707000732
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,4096,7168,0.02254293362299601
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,4096,7168,0.03299199938774109
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,4096,8192,0.02504533330599467
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,4096,7168,0.02268799940745036
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,4096,6144,0.01858773430188497
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,4096,6144,0.03127786715825399
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,4096,5120,0.016775466998418174
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,4096,5120,0.029624533653259278
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,4096,5120,0.018959999084472656
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,4096,4096,0.01367573340733846
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,4096,4096,0.02694186568260193
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,6144,64,0.016518400112787882
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,4096,3584,0.012362666924794515
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,4096,3584,0.027324799696604413
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,4096,6144,0.020957867304484047
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,4096,3072,0.010787199934323628
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,4096,3072,0.02342826724052429
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,4096,3072,0.013707733154296875
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,4096,2560,0.009608532985051472
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,4096,2560,0.02323946754137675
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,4096,3584,0.015358933806419372
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,4096,2048,0.008327466746171314
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,4096,2048,0.02103253404299418
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,4096,4096,0.016207999984423318
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,4096,1536,0.006861866513888042
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,4096,1536,0.019751467307408652
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,4096,2048,0.011311999956766764
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,4096,1536,0.010379733641942342
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,4096,1024,0.0051125332713127134
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,4096,1024,0.018458666404088338
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,4096,768,0.004628266890843709
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,4096,2560,0.012483200430870056
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,4096,768,0.01699840029080709
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,4096,512,0.004203733305136363
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,4096,512,0.0166101336479187
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,4096,512,0.008338133494059246
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,4096,256,0.0034474665919939675
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,4096,256,0.015378133455912272
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,4096,768,0.00830506682395935
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,4096,128,0.003386666625738144
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,4096,128,0.014514133334159851
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,4096,1024,0.009099733829498292
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,4096,256,0.007582933207352956
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,4096,128,0.007514666517575581
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,4096,64,0.003045333425203959
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,4096,32,0.003054933249950409
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,4096,64,0.015498666961987814
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,4096,32,0.01567893326282501
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,3584,65536,0.1637781302134196
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,3584,16384,0.04314133326212565
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,3584,65536,0.18298986752827961
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,3584,16384,0.052476799488067626
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,3584,12288,0.03292693297068278
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,3584,12288,0.04234879811604818
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,3584,12288,0.032587732871373495
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,3584,10240,0.02826026678085327
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,3584,10240,0.0371946652730306
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,3584,16384,0.041331199804941814
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,3584,8192,0.0237119992574056
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,3584,8192,0.03326933383941651
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,3584,10240,0.028926932811737062
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,3584,7168,0.019985065857569376
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,3584,7168,0.031897600491841635
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,3584,8192,0.02438933253288269
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,3584,6144,0.017553067207336424
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,3584,6144,0.030738133192062377
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,3584,7168,0.022357332706451415
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,3584,5120,0.014750933647155762
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,3584,5120,0.028599466880162554
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,3584,6144,0.020708266894022623
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,3584,4096,0.01251520017782847
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,3584,4096,0.02490026752154032
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,4096,65536,0.137500794728597
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,3584,3584,0.011245866616566975
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,3584,3584,0.024782933791478477
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,3584,5120,0.018224000930786133
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,3584,3072,0.010014933347702027
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,3584,3072,0.022973867257436116
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,3584,4096,0.01620693306128184
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,3584,2560,0.008797867099444072
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,3584,2560,0.021606399615605672
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,3584,3584,0.014999467134475707
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,3584,2048,0.007567999760309856
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,3584,2048,0.021898667017618813
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,3584,3072,0.013306666413942972
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,3584,1536,0.006018133461475372
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,3584,1536,0.01945599913597107
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,3584,2560,0.012498133381207784
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,3584,1024,0.00469760000705719
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,3584,1024,0.017223467429478966
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,3584,1536,0.00995306670665741
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,3584,768,0.004301866888999939
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,3584,768,0.01676266590754191
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,3584,2048,0.01123199959595998
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,3584,512,0.0038783999780813852
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,3584,512,0.018129066626230875
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,3584,768,0.008321066697438557
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,3584,256,0.003386666625738144
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,3584,256,0.014503467082977294
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,3584,1024,0.009198932846387228
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,3584,128,0.0030154667794704436
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,3584,128,0.015375999609629312
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,3584,512,0.008298666775226593
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,3584,256,0.00754559983809789
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,3584,64,0.0029994666576385496
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,3584,64,0.014814933141072592
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,3584,128,0.007513600091139476
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,3584,32,0.0029898665845394133
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,3584,32,0.014885333180427552
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,3072,65536,0.15374719301859538
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,3072,65536,0.16164266268412272
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,3072,16384,0.04143786827723185
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,3072,16384,0.049404799938201904
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,3072,12288,0.032381866375605264
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,3072,12288,0.04164906740188599
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,3072,12288,0.032790400584538776
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,3072,10240,0.02717439929644267
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,3072,10240,0.037514666716257736
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,3072,16384,0.040175998210906984
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,3072,8192,0.021554134289423623
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,3072,8192,0.03284693360328674
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,3072,10240,0.028880000114440918
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,3072,7168,0.019232000907262167
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,3072,7168,0.03272320032119751
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,3072,8192,0.024410667022069295
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,3072,6144,0.016794667641321818
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,3072,6144,0.031011199951171874
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,3072,7168,0.022171733776728313
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,3072,5120,0.014615466197331747
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,3072,5120,0.028970666726430255
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,3072,6144,0.020356265703837077
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,3072,4096,0.012145066261291504
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,3072,4096,0.02521173357963562
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,3072,5120,0.018260266383488974
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,3072,3584,0.01112000048160553
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,3072,3584,0.024965333938598632
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,3072,4096,0.01592639982700348
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,3072,3072,0.009806933005650838
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,3072,3072,0.022685867547988892
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,3072,3584,0.014569600423177084
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,3072,2560,0.008858666817347208
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,3072,2560,0.022819199164708457
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,3072,3072,0.013315199812253316
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,3072,2048,0.007594666878382365
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,3072,2048,0.020623999834060668
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,3072,2560,0.01214400033156077
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,3072,1536,0.006313600142796834
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,3072,1536,0.021062399943669638
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,3072,2048,0.011241599917411804
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,3072,1024,0.005048533280690512
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,3072,1024,0.0182261327902476
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,3072,1024,0.00876800020535787
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,3072,768,0.005041066805521647
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,3072,1536,0.010032000144322713
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,3072,768,0.01773866613705953
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,3072,512,0.004249600072701773
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,3072,512,0.016620799899101257
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,3072,512,0.008026666442553202
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,3072,256,0.0034080001215140024
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,3072,768,0.008358400066693623
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,3072,256,0.01453013320763906
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,3072,128,0.003419733295838038
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,3072,128,0.014497066537539164
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,3072,256,0.007931733131408691
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,3072,128,0.007496533294518788
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,3072,64,0.003094399968783061
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,3072,32,0.0030410667260487873
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,3072,64,0.0159061332543691
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,3072,32,0.016221867005030314
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,2560,65536,0.12623146375020344
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,2560,65536,0.1490517298380534
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,2560,16384,0.03318293293317159
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,2560,16384,0.047348264853159586
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,3584,65536,0.1382794698079427
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,2560,12288,0.02547520001729329
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,2560,12288,0.041102933883666995
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,2560,12288,0.03254719972610474
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,2560,10240,0.02115839918454488
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,2560,10240,0.03553280035654704
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,2560,16384,0.04001386562983195
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,2560,8192,0.01767573356628418
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,2560,8192,0.0315391997496287
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,2560,10240,0.028142933050791425
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,2560,7168,0.01565439999103546
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,2560,7168,0.030868266026179
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,2560,8192,0.023933867613474526
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,2560,6144,0.01423466702302297
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,2560,6144,0.02941333254178365
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,2560,6144,0.019937066237131755
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,2560,7168,0.021898667017618813
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,2560,5120,0.012058666348457337
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,2560,5120,0.027913600206375122
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,2560,4096,0.01060693363348643
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,2560,4096,0.02547093431154887
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,2560,4096,0.016173866391181946
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,2560,3584,0.009611733754475911
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,2560,3584,0.024409600098927817
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,2560,5120,0.01825493375460307
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,2560,3072,0.008680533369382222
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,2560,3072,0.023011199633280435
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,2560,3584,0.014652799566586813
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,2560,2560,0.007573333382606506
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,2560,2560,0.023517866929372154
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,2560,3072,0.01330880026022593
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,2560,2048,0.006970666845639547
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,2560,2048,0.021150932709376017
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,2560,2560,0.012170666456222534
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,2560,1536,0.005862399935722351
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,2560,1536,0.01935360034306844
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,3072,65536,0.1369376023610433
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,2560,1024,0.004664533336957296
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,2560,1024,0.016800000270207723
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,2560,1536,0.00995199978351593
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,2560,768,0.004260266820589701
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,2560,768,0.017258665959040322
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,2560,2048,0.01120960017045339
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,2560,512,0.003853866706291834
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,2560,512,0.016611199577649435
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,2560,1024,0.008769067128499348
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,2560,256,0.0033834666013717652
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,2560,256,0.014531200130780539
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,2560,512,0.00804799993832906
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,2560,768,0.008386133114496867
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,2560,128,0.003017599880695343
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,2560,256,0.007533866663773854
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,2560,64,0.0030016000072161358
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,2560,128,0.015003732840220132
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,2560,32,0.003092266619205475
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,2560,64,0.014547200004259745
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,2560,32,0.015343999862670899
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,2560,128,0.007158400118350982
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,2048,65536,0.09702293078104654
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,2048,16384,0.03340799808502197
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,2048,65536,0.13374719619750977
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,2048,16384,0.04414933522542318
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,2048,12288,0.025035732984542848
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,2048,12288,0.039435732364654544
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,2048,12288,0.03196159998575847
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,2048,10240,0.021246933937072755
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,2048,10240,0.034798932075500486
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,2048,16384,0.04020693302154541
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,2048,8192,0.01796906590461731
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,2048,8192,0.03030400077501933
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,2048,10240,0.028062933683395387
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,2048,7168,0.015624533096949259
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,2048,7168,0.028807467222213744
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,2048,8192,0.02399359941482544
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,2048,6144,0.014241066575050355
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,2048,6144,0.0287882665793101
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,2048,7168,0.021895466248194377
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,2048,5120,0.012116266290346782
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,2048,5120,0.0253440002600352
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,2048,6144,0.019911466042200725
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,2048,4096,0.009671466549237569
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,2048,4096,0.023928532997767128
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,2048,5120,0.017874133586883546
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,2048,3584,0.008841600020726521
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,2048,3584,0.02291413346926371
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,2048,4096,0.01541759967803955
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,2048,3072,0.007844266792138416
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,2048,3072,0.022529067595799764
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,2048,3584,0.014137599865595499
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,2048,2560,0.006878933310508728
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,2048,2560,0.02069013317426046
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,2048,3072,0.01337386667728424
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,2048,2048,0.006152533491452535
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,2048,2048,0.02049493392308553
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,2048,2560,0.012054399649302164
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,2048,1536,0.005111466844876607
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,2048,1536,0.018718934059143065
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,2560,65536,0.13708693186442059
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,2048,1024,0.004364799956480662
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,2048,2048,0.011170132954915365
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,2048,1024,0.018125865856806436
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,2048,768,0.0038293334345022834
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,2048,768,0.0167413334051768
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,2048,1536,0.009914666414260864
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,2048,512,0.003492266684770584
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,2048,512,0.015726932883262636
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,2048,768,0.008329600095748901
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,2048,256,0.0029877332349618276
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,2048,512,0.00787306676308314
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,2048,1024,0.009164800246556599
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,2048,256,0.014651733636856078
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,2048,128,0.0029898665845394133
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,2048,128,0.01474240024884542
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,2048,64,0.0029696000119050344
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,2048,64,0.014738133549690247
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,2048,256,0.0074976002176602675
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,2048,32,0.0029834667841593427
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,2048,32,0.015389866630236306
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,1536,65536,0.08356800079345703
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,1536,65536,0.12724800109863282
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,1536,16384,0.02598186731338501
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,1536,16384,0.04313493172327677
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,1536,16384,0.039971200625101726
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,1536,12288,0.01880533297856649
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,1536,12288,0.03569813172022502
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,1536,12288,0.03135786652565002
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,1536,10240,0.016177067160606386
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,1536,10240,0.03262506723403931
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,1536,10240,0.02804906765619914
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,1536,8192,0.013529599706331889
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,1536,8192,0.028908799091974895
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,2048,65536,0.13723626136779785
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,1536,7168,0.012798933188120523
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,1536,7168,0.028597333033879596
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,1536,8192,0.023679999510447185
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,1536,6144,0.011084799965222675
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,1536,6144,0.0268010675907135
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,1536,7168,0.02191466689109802
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,1536,5120,0.009925333658854167
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,1536,5120,0.026062933603922527
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,1536,5120,0.017870932817459106
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,1536,4096,0.008739200234413148
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,1536,4096,0.02304960091908773
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,1536,6144,0.019853866100311278
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,1536,3584,0.008035199840863545
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,1536,3584,0.022768000761667885
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,1536,3584,0.014535466829935709
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,1536,3072,0.007089066505432129
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,1536,3072,0.021398399273554483
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,1536,4096,0.015388799707094827
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,1536,2560,0.006619733572006225
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,1536,2560,0.021150932709376017
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,1536,3072,0.013264000415802002
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,1536,2048,0.005469866593678792
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,1536,2560,0.012108799815177918
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,1536,2048,0.01865066687266032
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,1536,1536,0.005025066435337067
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,1536,1536,0.018652800718943277
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,1536,1536,0.01002239982287089
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,1536,1024,0.004203733305136363
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,1536,1024,0.016713599363962807
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,1536,2048,0.010898133118947346
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,1536,768,0.0037600000699361167
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,1536,768,0.017409066359202065
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,1536,1024,0.008826667070388794
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,1536,512,0.0030026666820049284
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,1536,768,0.008709333340326945
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,1536,512,0.01750719944636027
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,1536,256,0.0029653333127498626
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,1536,256,0.014536533753077188
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,1536,256,0.007550933460394542
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,1536,512,0.007926400005817413
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,1536,128,0.002614400039116542
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,1536,128,0.014705066879590353
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,1536,64,0.002630399912595749
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,1536,64,0.014541866381963095
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,1536,128,0.007494399944941203
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,1536,32,0.0026880001028378804
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,1536,32,0.014453333616256715
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,1024,65536,0.054067198435465494
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,1024,65536,0.11688106854756672
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,1024,16384,0.018716800212860107
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,1024,16384,0.040456533432006836
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,1536,65536,0.1364031950632731
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,1024,12288,0.017621332406997682
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,1024,12288,0.03498986562093099
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,1024,16384,0.03951253493626912
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,1024,10240,0.015267200271288552
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,1024,10240,0.030938667058944703
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,1024,12288,0.031380265951156616
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,1024,8192,0.013100799918174744
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,1024,8192,0.027723733584086103
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,1024,10240,0.02772480050722758
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,1024,7168,0.011704533298810323
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,1024,7168,0.02683626612027486
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,1024,8192,0.02362346649169922
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,1024,6144,0.010477866729100544
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,1024,6144,0.026126933097839356
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,1024,7168,0.021527467171351115
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,1024,5120,0.009011200070381165
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,1024,5120,0.024201599756876628
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,1024,6144,0.01947093407313029
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,1024,4096,0.007869866490364075
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,1024,4096,0.022526933749516805
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,1024,5120,0.017805866400400796
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,1024,3584,0.006748799979686737
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,1024,3584,0.022025599082310995
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,1024,4096,0.015689599514007568
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,1024,3072,0.006268799801667531
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,1024,3072,0.02074026664098104
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,1024,3584,0.014498133460680643
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,1024,2560,0.005453866720199585
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,1024,2560,0.020568533738454183
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,1024,3072,0.012950399518013
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,1024,2048,0.005054933329423269
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,1024,2560,0.01202880044778188
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,1024,2048,0.018603734175364175
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,1024,1536,0.004278400043646494
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,1024,1536,0.018337066968282065
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,1024,1536,0.010044800241788228
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,1024,1024,0.0034474665919939675
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,1024,2048,0.010778666536013285
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,1024,1024,0.017139200369517008
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,1024,768,0.003373866776625315
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,1024,768,0.016445866227149962
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,1024,1024,0.00881706674893697
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,1024,768,0.008377599716186523
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,1024,512,0.0029898665845394133
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,1024,512,0.014993066589037577
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,1024,256,0.0026346666117509207
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,1024,512,0.007925333579381307
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,1024,256,0.015335466464360556
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,1024,256,0.007919999957084655
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,1024,128,0.002571733295917511
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,1024,128,0.014433067043622336
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,1024,128,0.007209600011507671
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,1024,64,0.002571733295917511
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,1024,64,0.014847999811172486
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,1024,32,0.0026015999416510267
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,1024,32,0.014482133587201438
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,768,65536,0.04793493350346883
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,768,65536,0.11233279705047608
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,768,16384,0.017155200242996216
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,768,16384,0.039191468556722
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,768,16384,0.0395583987236023
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,768,12288,0.014437333742777506
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,768,12288,0.03389973243077596
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,1024,65536,0.13684479395548504
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,768,10240,0.013545599579811097
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,768,10240,0.031023999055226643
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,768,12288,0.03136746684710185
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,768,8192,0.01320319970448812
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,768,8192,0.0269322673479716
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,768,10240,0.02769706646601359
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,768,7168,0.011252267162005107
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,768,7168,0.02722773353258769
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,768,8192,0.023510400454203287
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,768,6144,0.009976533055305482
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,768,6144,0.024899200598398844
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,768,7168,0.021553067366282146
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,768,5120,0.008785067001978557
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,768,6144,0.019513599077860513
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,768,5120,0.024889600276947022
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,768,4096,0.007494399944941203
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,768,4096,0.02276159922281901
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,768,4096,0.015685333808263143
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,768,5120,0.01781439979871114
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,768,3584,0.006740266581376393
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,768,3584,0.021633066733678184
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,768,3072,0.006320000191529592
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,768,3072,0.020601600408554077
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,768,3584,0.014841600259145101
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,768,2560,0.0059008002281188965
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,768,3072,0.013115732868512472
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,768,2560,0.02087999979654948
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,768,2048,0.00510506679614385
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,768,2048,0.020453333854675293
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,768,2560,0.012467199563980102
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,768,1536,0.004304000238577525
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,768,2048,0.011161599556605022
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,768,1536,0.018681599696477254
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,768,1024,0.0037930667400360107
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,768,1024,0.01657600005467733
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,768,1536,0.010010666648546855
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,768,768,0.003517866631348928
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,768,1024,0.008736000458399455
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,768,768,0.018321067094802856
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,768,512,0.0030080000559488933
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,768,512,0.016012799739837647
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,768,768,0.008719999591509502
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,768,256,0.0029898665845394133
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,768,512,0.008263466755549113
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,768,256,0.015179733435312906
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,768,128,0.0025418666501839954
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,768,128,0.015330132842063905
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,768,256,0.007516799867153168
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,768,64,0.0025429333249727885
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,768,128,0.007464533547560374
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,768,64,0.01530026694138845
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,768,32,0.002586666742960612
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,512,65536,0.03214400013287862
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,768,32,0.014617600043614707
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,512,65536,0.10895360310872396
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,512,16384,0.013027200102806091
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,512,16384,0.03691093524297078
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,768,65536,0.13613333702087402
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,512,12288,0.011956266562143962
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,512,12288,0.033854933579762776
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,512,16384,0.039206401507059736
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,512,10240,0.01050986647605896
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,512,10240,0.029585067431132
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,512,12288,0.03139946659406026
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,512,10240,0.027614933252334595
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,512,8192,0.011733333269755047
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,512,8192,0.02680319945017497
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,512,7168,0.010858666896820069
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,512,7168,0.026754132906595868
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,512,7168,0.02151893377304077
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,512,6144,0.009483733773231506
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,512,6144,0.024711465835571288
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,512,8192,0.023486934105555215
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,512,5120,0.008091733356316884
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,512,5120,0.023137066761652628
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,512,5120,0.018077866236368815
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,512,4096,0.007121066749095917
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,512,6144,0.019822933276494346
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,512,4096,0.022223999102910362
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,512,3584,0.0063498665889104204
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,512,3584,0.02130026618639628
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,512,3584,0.014123732844988504
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,512,4096,0.015347199638684592
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,512,3072,0.005859200159708659
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,512,3072,0.020707199970881142
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,512,3072,0.013276799519856771
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,512,2560,0.005098666747411092
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,512,2560,0.01931519905726115
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,512,2560,0.012458667159080505
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,512,2048,0.0046165332198143
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,512,2048,0.01858133276303609
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,512,1536,0.00429013321797053
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,512,1536,0.01755839983622233
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,512,2048,0.011252267162005107
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,512,1024,0.0034304000437259674
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,512,1024,0.016711467504501344
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,512,1536,0.010009599725405376
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,512,768,0.0030645333230495454
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,512,768,0.016396799683570863
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,512,1024,0.009181867043177288
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,512,512,0.002994133283694585
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,512,512,0.014677332838376364
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,512,768,0.00832533339659373
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,512,256,0.0025962665677070618
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,512,256,0.014621866742769876
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,512,512,0.007970133423805236
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,512,128,0.0022453332940737406
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,512,256,0.007905066510041555
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,512,128,0.013715199629465737
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,512,64,0.0021536000072956084
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,512,64,0.015273599823315939
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,512,128,0.007109333574771881
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,512,32,0.002164266755183538
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,512,32,0.013854933778444925
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,256,65536,0.02348053256670634
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,256,65536,0.10082879861195881
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,256,16384,0.010005333026250203
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,256,16384,0.03497173388799031
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,256,16384,0.03912426630655925
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,256,12288,0.011687466502189636
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,256,12288,0.03100053270657857
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,512,65536,0.13716799418131512
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,256,10240,0.010407466689745586
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,256,10240,0.029604266087214153
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,256,12288,0.03130346735318502
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,256,8192,0.009476266304651896
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,256,8192,0.026718932390213012
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,256,10240,0.027639466524124145
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,256,7168,0.008764800429344178
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,256,7168,0.02646506627400716
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,256,8192,0.02350613276163737
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,256,6144,0.007979733248551685
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,256,7168,0.02147093415260315
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,256,6144,0.02486506700515747
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,256,5120,0.007289599875609081
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,256,5120,0.02328746716181437
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,256,6144,0.019463467597961425
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,256,5120,0.01773866613705953
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,256,4096,0.006675200164318084
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,256,4096,0.021657600005467733
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,256,3584,0.0062613333264986675
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,256,3584,0.020999467372894286
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,256,4096,0.015734400351842245
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,256,3072,0.005865600208441416
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,256,3584,0.014478933811187745
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,256,3072,0.02070080041885376
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,256,2560,0.005085866649945577
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,256,2560,0.020988800128300986
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,256,3072,0.01321386694908142
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,256,2048,0.0046741331617037455
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,256,2048,0.01877973278363546
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,256,2560,0.012242133418718975
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,256,1536,0.003861333429813385
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,256,1536,0.018628267447153728
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,256,2048,0.011226666967074077
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,256,1024,0.0033770665526390077
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,256,1024,0.017056000232696534
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,256,1536,0.010027733445167542
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,256,768,0.003045333425203959
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,256,768,0.015994667013486227
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,256,1024,0.00876800020535787
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,256,65536,0.13664533297220866
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,256,512,0.0025792000194390613
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,256,768,0.008370133241017659
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,256,512,0.014781866470972696
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,256,256,0.002583466718594233
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,256,512,0.00790719985961914
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,256,256,0.014990933736165366
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,256,128,0.0022122666239738466
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,256,256,0.00783679982026418
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,256,128,0.013818666338920593
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,256,64,0.0021322667598724365
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,256,32,0.002229333420594533
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,256,128,0.00709440012772878
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,256,64,0.013961600263913474
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,128,65536,0.01562879979610443
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,256,32,0.014939733346303306
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,128,16384,0.009187199672063192
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,128,65536,0.10139093399047852
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,128,16384,0.033056000868479415
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,128,12288,0.007941333452860515
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,128,12288,0.030829866727193195
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,128,16384,0.03911360104878743
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,128,12288,0.031300266583760575
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,128,10240,0.007247999807198842
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,128,10240,0.028905600309371948
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,128,8192,0.007355733215808869
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,128,8192,0.027800534168879194
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,128,10240,0.027194666862487792
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,128,7168,0.006684799989064534
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,128,7168,0.02571306626001994
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,128,8192,0.02353386680285136
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,128,6144,0.006342400113741558
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,128,6144,0.025150932868321735
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,128,7168,0.02144320011138916
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,128,5120,0.006664533416430156
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,128,5120,0.022872533400853476
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,128,65536,0.13541332880655926
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,128,4096,0.006763733426729838
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,128,6144,0.019411200284957887
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,128,4096,0.02226026654243469
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,128,3584,0.005905066430568695
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,128,3584,0.021989333629608154
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,128,5120,0.01771093408266703
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,128,3072,0.005470933516820272
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,128,4096,0.015668267011642457
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,128,3072,0.02063039938608805
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,128,2560,0.005036800106366476
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,128,3584,0.014133333166440328
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,128,2560,0.01908586621284485
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,128,2048,0.004474666714668274
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,128,3072,0.01295146644115448
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,128,2048,0.018782933553059898
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,128,1536,0.003868799904982249
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,128,2560,0.011641599734624227
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,128,1536,0.01771413286526998
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,128,1024,0.0033930666744709016
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,128,2048,0.010863999525705974
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,128,1024,0.017002665996551515
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,128,768,0.0029919999341169994
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,128,1536,0.009959466258684794
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,128,1024,0.008770133058230083
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,128,768,0.015196800231933594
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,128,512,0.0026208000878492994
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,128,768,0.008318933347860973
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,128,512,0.0147189329067866
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,128,256,0.002586666742960612
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,128,256,0.015255467096964518
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,128,512,0.007969066500663757
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,128,128,0.002163200080394745
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,128,256,0.007549866537253062
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,128,128,0.014549332857131957
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,128,64,0.0021546666820844015
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,128,128,0.007067733506361644
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,128,32,0.002197333425283432
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,128,64,0.014495999614397685
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,64,65536,0.015356799960136414
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,128,32,0.013379200299580892
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,64,16384,0.006326400240262349
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,64,12288,0.006687999765078227
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,64,16384,0.030905600388844805
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,64,10240,0.006244266529877981
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,64,12288,0.028838400046030683
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,64,8192,0.005849599838256836
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,64,10240,0.028782933950424194
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,64,7168,0.005459199845790863
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,64,8192,0.0258240004380544
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,64,6144,0.005895466605822245
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,64,7168,0.026770132780075073
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,64,5120,0.006311466793219249
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,64,6144,0.024690133333206177
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,64,4096,0.0059349333246548975
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,64,5120,0.02461973428726196
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,64,3584,0.0058442667126655575
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,64,4096,0.022660267353057862
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,64,3072,0.005435733497142792
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,64,3584,0.021210666497548422
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,64,2560,0.00510506679614385
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,64,3072,0.020909865697224937
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,64,2048,0.005063466727733612
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,64,2560,0.01902186671892802
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,64,1536,0.0038058665891488397
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,64,2048,0.018872533241907755
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,64,1024,0.0030464000999927522
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,64,1536,0.016785067319869996
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,64,768,0.002962133288383484
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,64,1024,0.016825600465138753
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,64,512,0.002570666621128718
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,64,768,0.016425599654515587
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,64,256,0.002218666672706604
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,64,512,0.015731199582417806
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,64,128,0.0021674667795499166
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,64,256,0.014494933684666953
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,64,64,0.0021461332837740582
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,64,128,0.015146666765213012
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,64,32,0.002169599880774816
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,64,64,0.014856533209482829
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,64,32,0.014476799964904785
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,32,65536,0.027287467320760088
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,32,16384,0.0075103998184204105
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,32,65536,0.0936896006266276
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,32,16384,0.030859732627868654
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,32,12288,0.0070592001080513
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,32,12288,0.029018666346867877
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,32,10240,0.006663466493288676
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,32,10240,0.028091732660929365
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,32,8192,0.00625493327776591
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,32,8192,0.025229867299397784
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,32,7168,0.005849599838256836
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,32,7168,0.025785599152247114
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,32,6144,0.005890133480230967
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,32,6144,0.024710400899251302
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,32,5120,0.005513600011666616
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,32,5120,0.024588799476623534
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,32,4096,0.005870933334032694
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,32,4096,0.02276479999224345
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,32,3584,0.0058677335580190025
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,32,3584,0.020781866709391274
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,32,3072,0.005458133419354757
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,32,3072,0.020807466904322305
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,32,2560,0.00505920002857844
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,32,2560,0.018557866414388023
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,32,2048,0.004682666560014089
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,32,2048,0.0185589333375295
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,32,1536,0.0037920000652472176
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,32,1536,0.017658666769663493
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,32,1024,0.003033600002527237
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,32,1024,0.016546133160591125
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,32,768,0.00297173336148262
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,32,768,0.015341867009798685
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,32,512,0.002570666621128718
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,32,512,0.014726400375366211
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,32,256,0.002681600054105123
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,32,128,0.0022986667851607008
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,32,256,0.01553600033124288
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,32,64,0.002196266750494639
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,32,128,0.01460693379243215
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,32,64,0.014501333236694336
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,512,32,32,0.0021578667064507803
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,32,32,0.01442026694615682
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,65536,16384,0.2942805290222168
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,65536,16384,0.5248341242472331
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,65536,12288,0.40703573226928713
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,65536,12288,0.2524895985921224
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,512,2048,128,0.007129600147406261
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,65536,10240,0.3375946680704752
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,65536,10240,0.18749119440714518
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,512,64,65536,0.09208213488260905
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,65536,8192,0.25740267435709635
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,65536,8192,0.16194987297058105
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,65536,10240,0.13927680651346844
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,65536,7168,0.22952639261881508
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,65536,7168,0.13852373758951825
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,65536,12288,0.16575679779052735
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,65536,6144,0.19678293863932292
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,65536,6144,0.11920853455861409
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,65536,8192,0.11227626800537109
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,65536,5120,0.16869333585103352
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,65536,5120,0.10090453624725342
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,65536,16384,0.2119914690653483
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,65536,4096,0.1297760009765625
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,65536,4096,0.08449386755625407
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,65536,6144,0.08827520211537679
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,65536,3584,0.12422826290130615
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,65536,3584,0.07775893211364746
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,65536,7168,0.10042986869812012
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,65536,3072,0.10736853281656902
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,65536,3072,0.06779733498891195
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,65536,5120,0.07574506600697836
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,65536,2560,0.08889919916788737
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,65536,2560,0.06104106505711874
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,65536,3584,0.05739839871724447
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,65536,2048,0.07276906967163085
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,65536,2048,0.05253119866053263
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,65536,3072,0.050894931952158606
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,65536,1536,0.056281598409016934
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,65536,1536,0.045517865816752115
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,65536,2560,0.04530986547470093
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,65536,1024,0.038489600022633866
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,65536,1024,0.037843199570973714
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,65536,4096,0.0623306671778361
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,65536,768,0.030048000812530517
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,65536,768,0.033115732669830325
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,65536,768,0.023125332593917847
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,65536,512,0.022592000166575112
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,65536,512,0.029766400655110676
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,65536,1024,0.02651840051015218
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,65536,256,0.015371732910474143
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,65536,256,0.030418133735656737
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,65536,256,0.017035732666651406
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,65536,128,0.013638400038083396
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,65536,128,0.027113600571950273
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,65536,512,0.021115734179814657
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,65536,64,0.013673599561055502
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,65536,64,0.026741333802541095
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,65536,32,0.01406826674938202
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,65536,128,0.015713066856066386
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,65536,32,0.02684906721115112
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,16384,65536,0.5242591857910156
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,16384,65536,0.33125759760538737
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,16384,16384,0.1294495979944865
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,16384,16384,0.09555839697519938
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,65536,1536,0.03308159907658895
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,16384,12288,0.10480213165283203
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,16384,12288,0.0762997309366862
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,65536,2048,0.039580798149108885
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,16384,10240,0.1045354684193929
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,16384,10240,0.07536853154500325
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,16384,10240,0.051252265771230064
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,16384,8192,0.0861407995223999
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,16384,8192,0.0656544009844462
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,16384,12288,0.05870933135350546
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,16384,7168,0.07564480304718017
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,16384,7168,0.0588266650835673
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,16384,16384,0.07408426602681478
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,16384,6144,0.06444586515426635
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,16384,6144,0.05370133320490519
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,16384,7168,0.038395734628041585
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,16384,5120,0.05648106733957926
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,16384,5120,0.047882668177286786
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,16384,8192,0.04282666842142741
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,16384,4096,0.045626668135325114
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,16384,4096,0.04192533493041992
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,16384,6144,0.034456535180409746
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,16384,3584,0.03899946610132853
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,16384,3584,0.037463466326395675
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,16384,4096,0.02645866672197978
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,16384,3072,0.033989334106445314
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,16384,3072,0.03529599905014038
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,16384,5120,0.03017173409461975
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,16384,2560,0.02877440055211385
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,16384,2560,0.031514666477839154
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,16384,2560,0.020253866910934448
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,16384,2048,0.023756800095240276
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,16384,2048,0.02914133270581563
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,16384,3072,0.02230293353398641
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,16384,1536,0.018142932653427125
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,16384,1536,0.025544534126917522
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,16384,3584,0.02399253249168396
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,16384,1024,0.013041067123413085
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,16384,1024,0.022886399428049722
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,16384,1024,0.01248426636060079
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,16384,768,0.01092693308989207
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,16384,768,0.02064639925956726
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,16384,1536,0.015780267119407655
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,16384,512,0.008586666981379191
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,16384,512,0.01962666710217794
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,16384,768,0.011272533734639486
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,16384,256,0.006694399813810985
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,16384,256,0.018677333990732826
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,16384,512,0.010423466563224792
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,16384,128,0.005890133480230967
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,16384,128,0.016537599762280784
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,16384,256,0.009196799993515015
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,16384,64,0.005514666438102722
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,16384,64,0.016849066813786825
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,16384,32,0.00586346685886383
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,16384,32,0.017450666427612303
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,12288,65536,0.38142080307006837
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,16384,128,0.009145599603652955
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,12288,65536,0.29603306452433265
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,12288,16384,0.1058303991953532
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,12288,16384,0.09236799875895182
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,16384,2048,0.01814613342285156
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,12288,12288,0.07937173048655191
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,12288,12288,0.06383786598841348
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,12288,12288,0.03601280053456624
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,12288,10240,0.07284053166707358
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,12288,10240,0.06016000111897787
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,12288,16384,0.044793601830800375
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,12288,8192,0.05846079985300699
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,12288,8192,0.055257598559061684
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,12288,10240,0.03198933402697245
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,12288,7168,0.05217066605885824
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,12288,7168,0.05095680157343546
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,12288,8192,0.026873600482940675
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,12288,6144,0.045118931929270426
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,12288,6144,0.04521706501642863
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,12288,7168,0.025115732351938886
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,12288,5120,0.03874666690826416
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,12288,5120,0.039789867401123044
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,12288,6144,0.022706133127212525
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,12288,4096,0.03145280083020528
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,12288,4096,0.034526932239532473
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,12288,5120,0.020307199160257975
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,12288,3584,0.028216532866160077
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,12288,3584,0.033149866263071696
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,12288,4096,0.017566933234532674
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,12288,3072,0.024052266279856363
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,12288,3072,0.03004266619682312
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,12288,3584,0.016659200191497803
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,12288,2560,0.020802134275436403
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,12288,2560,0.029332266251246138
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,12288,2560,0.01418346663316091
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,12288,2048,0.01705706715583801
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,12288,2048,0.02669973373413086
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,12288,3072,0.01565226713816325
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,12288,1536,0.013247999548912048
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,12288,1536,0.023829332987467446
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,12288,1536,0.010891733566919963
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,12288,1024,0.0100682665904363
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,12288,1024,0.021896533171335855
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,12288,2048,0.012548266847928365
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,12288,768,0.008739200234413148
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,12288,768,0.019236266613006592
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,12288,768,0.00918826659520467
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,12288,512,0.006681600213050842
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,12288,512,0.019234132766723634
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,12288,1024,0.009239466985066731
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,12288,256,0.0050570666790008545
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,12288,256,0.017760000626246133
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,12288,256,0.007891199986139933
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,12288,128,0.004222933451334635
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,12288,128,0.01554026703039805
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,12288,512,0.00871573289235433
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,12288,64,0.003924266745646795
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,12288,64,0.01648319959640503
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,12288,32,0.00420906643072764
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,12288,32,0.016300800442695617
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,12288,128,0.007542400062084198
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,10240,65536,0.35164267222086587
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,10240,65536,0.2547797362009684
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,10240,16384,0.08609066804250082
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,10240,16384,0.07847999731699626
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,10240,16384,0.04263999859491984
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,10240,12288,0.0652010679244995
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,10240,12288,0.06278613408406576
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,10240,12288,0.03451626698176066
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,10240,10240,0.05507200161616007
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,10240,10240,0.058872532844543454
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,16384,65536,0.27199678421020507
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,10240,8192,0.046216531594594316
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,10240,8192,0.0487125317255656
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,10240,10240,0.030534400542577105
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,10240,7168,0.039717332522074385
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,10240,7168,0.04323413372039795
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,10240,8192,0.025546665986378985
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,10240,6144,0.0357258677482605
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,10240,6144,0.04146666526794433
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,12288,65536,0.1572223981221517
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,10240,5120,0.030191999673843384
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,10240,5120,0.03572053511937459
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,10240,7168,0.023605332771937052
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,10240,4096,0.024401066700617473
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,10240,4096,0.032979200283686325
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,10240,6144,0.021901865800221763
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,10240,3584,0.02241706649462382
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,10240,3584,0.03123840093612671
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,10240,5120,0.019799466927846274
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,10240,3072,0.01990933418273926
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,10240,3072,0.028895999987920123
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,10240,4096,0.01742186745007833
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,10240,2560,0.016710400581359863
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,10240,2560,0.026877866188685103
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,10240,3584,0.016157866517702738
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,10240,2048,0.014678399761517844
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,10240,2048,0.025697066386540728
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,10240,2560,0.01328000028928121
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,10240,1536,0.012071466445922852
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,10240,1536,0.022771199544270836
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,10240,3072,0.014945066968599954
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,10240,1024,0.009583999713261921
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,10240,1024,0.020525866746902467
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,10240,1024,0.009574400385220845
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,10240,768,0.007982933521270752
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,10240,768,0.019961599508921304
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,10240,1536,0.010460799932479859
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,10240,512,0.00664213349421819
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,10240,512,0.018229333559672038
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,10240,2048,0.011738666892051696
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,10240,256,0.005096533397833506
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,10240,256,0.018490666151046754
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,10240,256,0.007563733557860057
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,10240,128,0.004657066861788432
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,10240,128,0.01727573275566101
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,10240,768,0.008739200234413148
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,10240,64,0.004252799848715464
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,10240,64,0.016845866044362386
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,10240,32,0.005453866720199585
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,10240,32,0.016549332936604818
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,10240,512,0.008318933347860973
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,10240,128,0.00749120016892751
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,8192,65536,0.31083733240763345
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,8192,16384,0.0812725305557251
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,8192,65536,0.21765333811442056
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,8192,16384,0.07132799625396728
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,8192,12288,0.06004266738891602
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,8192,12288,0.05624959866205851
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,8192,12288,0.033831465244293216
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,8192,10240,0.052279468377431235
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,8192,10240,0.0510101318359375
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,8192,16384,0.04140479962031047
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,8192,8192,0.04207786719004313
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,8192,8192,0.045057066281636554
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,8192,10240,0.029951999584833782
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,8192,7168,0.03720320065816243
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,8192,7168,0.041162665685017905
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,8192,8192,0.025102933247884113
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,8192,6144,0.032630399862925215
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,8192,6144,0.03700480063756307
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,8192,7168,0.023107200860977173
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,8192,5120,0.028220800558725993
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,10240,65536,0.14548266728719075
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,8192,5120,0.033521068096160886
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,8192,4096,0.023121066888173423
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,8192,4096,0.030846933523813885
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,8192,6144,0.021180800596872964
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,8192,3584,0.02000853419303894
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,8192,3584,0.02908693353335063
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,8192,4096,0.01695893406867981
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,8192,3072,0.017658666769663493
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,8192,3072,0.026977066198984784
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,8192,5120,0.01914986570676168
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,8192,2560,0.015195733308792115
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,8192,2560,0.02688746651013692
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,8192,3584,0.01575146714846293
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,8192,2048,0.01266986628373464
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,8192,2048,0.023222400744756063
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,8192,2560,0.01330560048421224
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,8192,1536,0.010628267129262289
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,8192,1536,0.02257386644681295
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,8192,3072,0.0145087997118632
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,8192,2048,0.011587199568748475
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,8192,1024,0.00825386643409729
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,8192,1024,0.01883626580238342
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,8192,768,0.0067210664351781205
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,8192,768,0.017944532632827758
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,8192,768,0.00879146655400594
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,8192,512,0.005500799914201101
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,8192,512,0.019080533583958944
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,8192,1536,0.010066133737564088
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,8192,1024,0.00913706620534261
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,8192,256,0.005038933455944061
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,8192,256,0.019237333536148073
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,8192,128,0.004214400053024292
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,8192,128,0.01630506714185079
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,8192,128,0.007514666517575581
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,8192,64,0.003832533210515976
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,8192,64,0.01578986644744873
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,8192,32,0.003899733225504557
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,8192,512,0.00795306662718455
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,8192,256,0.007944533228874206
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,8192,32,0.016356266538302102
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,7168,65536,0.25159573554992676
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,7168,65536,0.20414506594340004
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,7168,16384,0.06581653356552124
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,7168,16384,0.06346346537272135
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,7168,12288,0.05059946775436401
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,7168,12288,0.05196160078048706
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,7168,12288,0.03313493331273397
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,7168,10240,0.044061867396036784
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,7168,10240,0.046173866589864096
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,7168,16384,0.04079573154449463
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,7168,8192,0.03490560054779053
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,7168,8192,0.04080959955851237
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,7168,10240,0.0294922669728597
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,7168,7168,0.03140160044034322
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,7168,7168,0.03689599831899007
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,7168,8192,0.024703999360402427
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,7168,6144,0.026929066578547163
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,7168,6144,0.03400319814682007
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,7168,7168,0.022977066040039063
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,7168,5120,0.023413334290186563
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,7168,5120,0.031005867322285968
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,7168,6144,0.02104960083961487
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,7168,4096,0.018693333864212035
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,7168,4096,0.028952533006668092
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,7168,5120,0.01904319922129313
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,7168,3584,0.01699733336766561
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,7168,3584,0.027588266134262084
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,7168,4096,0.01623040040334066
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,7168,3072,0.015131733814875283
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,7168,3072,0.02536746660868327
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,7168,3584,0.015222400426864624
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,7168,2560,0.01288106640179952
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,7168,2560,0.024791467189788818
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,7168,3072,0.014106667041778565
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,7168,2048,0.010681600371996561
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,7168,2048,0.022715733448664347
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,7168,2560,0.012476799885431926
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,7168,1536,0.009166933099428813
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,7168,1536,0.021203200022379555
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,7168,1536,0.010097066561381023
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,7168,1024,0.007132799923419952
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,7168,1024,0.01989333430926005
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,7168,2048,0.011212799946467083
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,7168,768,0.005861333509286245
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,7168,768,0.018428800503412883
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,7168,1024,0.008840533097585042
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,7168,512,0.005063466727733612
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,7168,512,0.017768534024556477
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,7168,768,0.00886079967021942
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,7168,256,0.004214400053024292
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,7168,256,0.015989333391189575
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,7168,256,0.00758186678091685
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,7168,128,0.0037920000652472176
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,7168,512,0.0079434668024381
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,7168,128,0.014648532867431641
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,7168,64,0.0033941333492596946
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,7168,64,0.014532267053922018
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,7168,32,0.0034506666163603462
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,7168,32,0.015026133259137472
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,7168,128,0.007462400197982788
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,6144,65536,0.23334506352742515
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,6144,65536,0.17811733881632488
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,6144,16384,0.06094933350880941
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,6144,16384,0.06160000165303549
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,8192,65536,0.1378378709157308
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,6144,12288,0.04728000164031983
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,6144,12288,0.04956053495407105
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,6144,16384,0.040831998984018965
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,6144,10240,0.03951253493626912
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,6144,10240,0.04320640166600545
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,6144,12288,0.033148799339930216
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,6144,8192,0.032706133524576825
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,6144,8192,0.03905920187632243
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,6144,8192,0.024936532974243163
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,6144,7168,0.029007999102274577
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,6144,7168,0.036183468500773114
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,6144,10240,0.028827732801437377
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,6144,6144,0.024810665845870973
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,6144,6144,0.03267520070075989
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,6144,6144,0.020777599016825358
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,6144,5120,0.02106879949569702
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,6144,5120,0.029175466299057005
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,6144,7168,0.023146667083104453
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,6144,4096,0.017078399658203125
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,6144,4096,0.028509867191314698
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,6144,5120,0.018758400281270345
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,6144,3584,0.015408000349998474
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,6144,3584,0.02646399935086568
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,6144,4096,0.016280532876650492
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,6144,3072,0.013772799571355184
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,6144,3072,0.025008000930150348
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,6144,3072,0.0139957328637441
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,6144,2560,0.011975466211636861
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,6144,2560,0.022858667373657226
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,6144,3584,0.014919466773668923
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,6144,2048,0.010001066327095031
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,6144,2048,0.020758400360743202
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,6144,2560,0.012998400131861368
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,6144,1536,0.008391466736793519
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,6144,2048,0.011190399527549744
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,6144,1536,0.019294933478037516
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,6144,1024,0.006362666686375936
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,6144,1024,0.01851946711540222
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,7168,65536,0.13525546391805013
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,6144,768,0.0052490666508674625
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,6144,768,0.01746666630109151
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,6144,1024,0.009201066692670186
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,6144,512,0.0042805333932240805
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,6144,512,0.016568533579508462
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,6144,768,0.008422399560610454
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,6144,256,0.0034506666163603462
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,6144,1536,0.010388267040252686
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,6144,256,0.015799466768900552
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,6144,128,0.003436800092458725
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,6144,128,0.014542933305104574
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,6144,512,0.008390399813652038
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,6144,64,0.003070933371782303
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,6144,64,0.014662399888038635
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,6144,128,0.007561600208282471
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,6144,32,0.0030613332986831666
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,6144,32,0.014475733041763306
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,6144,256,0.007547733187675476
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,5120,65536,0.2018943945566813
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,5120,16384,0.05342719952265421
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,5120,65536,0.16633280118306476
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,5120,16384,0.05785173177719116
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,5120,12288,0.041247999668121337
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,5120,12288,0.04586666822433472
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,5120,12288,0.0321343998114268
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,5120,10240,0.03504960139592488
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,5120,10240,0.04044266541798909
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,5120,16384,0.040350933869679764
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,5120,8192,0.028670932849248248
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,5120,8192,0.03575679858525594
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,5120,10240,0.028498133023579914
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,5120,7168,0.02525866627693176
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,5120,7168,0.03460053205490112
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,5120,8192,0.02441920042037964
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,5120,6144,0.021861332654953002
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,5120,6144,0.03134933312733968
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,5120,7168,0.02254826625188192
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,5120,5120,0.018372267484664917
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,5120,5120,0.03010666569073995
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,5120,6144,0.02041706641515096
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,5120,4096,0.01515733301639557
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,5120,4096,0.027085866530736285
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,5120,5120,0.018694400787353516
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,5120,3584,0.01378986636797587
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,5120,3584,0.02621866663297017
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,5120,4096,0.01620693306128184
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,5120,3072,0.01222933332125346
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,5120,3072,0.024768000841140746
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,5120,3584,0.014750933647155762
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,5120,2560,0.0108842670917511
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,5120,2560,0.022759467363357544
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,5120,3072,0.013336533308029174
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,5120,2048,0.009443199634552002
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,5120,2048,0.020642133553822835
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,5120,2560,0.01251413325468699
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,6144,65536,0.13406826655069987
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,5120,1536,0.00790826678276062
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,5120,1536,0.020546134312947592
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,5120,1024,0.00606826643149058
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,5120,1024,0.019035732746124266
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,5120,2048,0.011192533373832702
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,5120,768,0.0055061335364977515
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,5120,768,0.016577066977818807
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,5120,1024,0.00912000040213267
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,5120,512,0.004626133541266123
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,5120,512,0.016526933511098227
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,5120,1536,0.009965866804122925
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,5120,256,0.003828266759713491
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,5120,256,0.01605226695537567
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,5120,256,0.007534933090209961
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,5120,128,0.0034080001215140024
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,5120,128,0.01683733264605204
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,5120,768,0.008711466193199157
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,5120,512,0.008371200164159138
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,5120,64,0.003005866706371307
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,5120,32,0.0030591999491055804
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,5120,64,0.01578986644744873
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,5120,32,0.014457600315411887
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,5120,128,0.007142400244871776
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,4096,65536,0.15733013153076172
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,4096,16384,0.04227840105692546
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,4096,65536,0.15143574078877767
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,4096,16384,0.05431040128072103
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,4096,12288,0.03251840074857076
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,4096,12288,0.0425866683324178
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,4096,12288,0.03215679923693339
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,4096,10240,0.02768426736195882
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,4096,10240,0.038889598846435544
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,4096,16384,0.04002666473388672
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,4096,8192,0.022740266720453896
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,4096,8192,0.03291306694348653
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,4096,10240,0.028498133023579914
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,4096,7168,0.019801600774129232
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,4096,7168,0.0316703995068868
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,4096,8192,0.023592533667882283
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,4096,6144,0.017357865969340004
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,4096,6144,0.029204267263412475
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,4096,7168,0.022403200467427574
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,4096,5120,0.014867200454076131
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,4096,5120,0.02702186703681946
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,4096,6144,0.01992320020993551
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,4096,4096,0.012571733196576437
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,4096,4096,0.024977066119511924
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,4096,5120,0.018234666188557944
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,4096,3584,0.011397332946459452
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,4096,3584,0.023638399442036946
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,4096,4096,0.015366400281588236
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,4096,3072,0.01030293305714925
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,4096,3072,0.023260800043741862
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,4096,3584,0.014473600188891092
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,4096,2560,0.009197866916656494
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,4096,2560,0.02089386582374573
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,4096,3072,0.013266133268674216
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,4096,2048,0.007830399771531422
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,4096,2048,0.020860799153645835
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,4096,2560,0.012089600165685017
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,4096,1536,0.0063296000162760425
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,4096,1536,0.020504534244537354
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,4096,2048,0.010856533050537109
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,4096,1024,0.005452799797058106
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,4096,1024,0.017358932892481485
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,4096,1536,0.009939199686050415
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,4096,768,0.004795733094215393
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,4096,768,0.016657066345214844
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,4096,768,0.008378666639328004
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,4096,512,0.004262400170167288
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,4096,512,0.016661333044370015
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,4096,1024,0.009157333771387737
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,4096,256,0.0034602666894594826
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,4096,256,0.015398400028546652
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,4096,256,0.007561600208282471
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,4096,128,0.003386666625738144
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,4096,128,0.014443733294804893
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,4096,512,0.007950933277606964
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,4096,64,0.0030517332255840302
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,4096,64,0.014739200472831726
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,4096,128,0.007080533107121785
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,4096,32,0.0031818665564060213
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,4096,32,0.0144405335187912
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,3584,65536,0.13423147201538085
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,3584,65536,0.13802560170491535
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,3584,16384,0.03438613414764404
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,3584,16384,0.050612266858418783
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,5120,65536,0.13320639928181965
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,3584,12288,0.026282666126887004
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,3584,12288,0.03984640041987102
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,3584,16384,0.039154132207234696
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,3584,10240,0.02291839917500814
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,3584,10240,0.0370688001314799
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,3584,12288,0.03135040005048116
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,3584,8192,0.018266665935516357
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,3584,8192,0.03095146616299947
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,3584,8192,0.02395306626955668
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,3584,7168,0.01604373355706533
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,3584,7168,0.028868265946706134
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,3584,10240,0.02775786717732747
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,3584,6144,0.014652799566586813
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,3584,6144,0.02738986611366272
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,3584,6144,0.019796266158421835
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,3584,5120,0.012339199582735699
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,3584,5120,0.025832533836364746
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,3584,7168,0.021869866053263347
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,3584,4096,0.010710400342941285
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,3584,4096,0.02450986703236898
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,3584,4096,0.015434666474660238
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,3584,3584,0.009820800026257832
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,3584,3584,0.023372799158096313
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,3584,5120,0.017667200167973837
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,3584,3072,0.008589866757392883
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,3584,3072,0.021782400210698445
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,4096,65536,0.13267199993133544
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,3584,2560,0.007955199976762136
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,3584,2560,0.021422932545344033
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,3584,3584,0.014526933431625366
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,3584,2048,0.006704000135262807
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,3584,2560,0.012057600418726604
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,3584,2048,0.01892906626065572
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,3584,1536,0.005532800157864889
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,3584,1536,0.01787839929262797
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,3584,3072,0.012834133704503379
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,3584,1024,0.00432533323764801
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,3584,1024,0.016714666287104288
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,3584,1024,0.008813866972923278
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,3584,768,0.0038719999293486277
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,3584,768,0.01839359998703003
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,3584,1536,0.009621333082516987
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,3584,512,0.0034602666894594826
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,3584,512,0.01608746647834778
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,3584,768,0.008362666765848795
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,3584,256,0.003044266750415166
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,3584,256,0.015939199924468996
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,3584,512,0.007961600025494894
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,3584,2048,0.011222400267918905
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,3584,128,0.002974933385848999
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,3584,128,0.01565439999103546
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,3584,64,0.0029909332593282064
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,3584,64,0.014636799693107605
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,3584,32,0.003058133274316788
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,3584,128,0.0074879998962084455
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,3584,32,0.014631467064221701
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,3584,256,0.007483733197053273
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,3072,65536,0.1345098654429118
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,3072,16384,0.035898665587107345
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,3072,65536,0.12733866373697916
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,3072,16384,0.044905598958333334
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,3072,12288,0.02770133415857951
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,3072,12288,0.03684586683909098
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,3072,12288,0.03158613244692485
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,3072,10240,0.0234933336575826
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,3072,10240,0.033556266625722246
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,3072,16384,0.03877653280893962
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,3072,8192,0.019403733809789023
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,3072,8192,0.03198506633440654
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,3072,10240,0.02795413335164388
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,3072,7168,0.017110399405161538
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,3072,7168,0.02969920039176941
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,3072,8192,0.023566933472951253
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,3072,6144,0.015222400426864624
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,3072,6144,0.02890239953994751
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,3072,7168,0.02123946746190389
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,3072,5120,0.012915199995040894
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,3072,5120,0.026346667607625322
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,3072,6144,0.019530665874481202
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,3072,4096,0.011290666460990906
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,3072,4096,0.024821333090464272
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,3072,5120,0.01781546672185262
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,3072,3584,0.010021332899729412
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,3072,3584,0.022709333896636964
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,3072,4096,0.015706666310628257
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,3072,3072,0.009154133001963298
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,3072,3072,0.02134400010108948
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,3072,3584,0.014198399583498635
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,3072,2560,0.007896533111731212
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,3072,2560,0.020089600483576456
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,3072,3072,0.012941867113113403
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,3072,2048,0.0070933332045873005
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,3072,2048,0.020500266551971437
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,3072,2560,0.012117333213488261
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,3072,1536,0.005864533285299936
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,3072,1536,0.018447999159495035
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,3584,65536,0.13313813209533693
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,3072,1024,0.004695466657479604
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,3072,1024,0.016846932967503867
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,3072,2048,0.010826667149861652
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,3072,768,0.004231466849644979
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,3072,768,0.016793600718180337
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,3072,1536,0.01000426709651947
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,3072,512,0.0034453332424163817
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,3072,512,0.01508799990018209
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,3072,512,0.00788800021012624
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,3072,256,0.003031466652949651
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,3072,768,0.008357333143552144
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,3072,256,0.014442666371663412
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,3072,128,0.002611200014750163
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,3072,128,0.014470400412877402
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,3072,1024,0.009126399954160053
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,3072,256,0.007485866546630859
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,3072,128,0.0071839998165766404
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,3072,64,0.002566399921973546
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,3072,32,0.002631466587384542
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,3072,64,0.014646400014559427
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,3072,32,0.01447466711203257
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,2560,65536,0.11104106903076172
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,2560,65536,0.12109973430633544
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,2560,16384,0.03561600049336751
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,2560,16384,0.047566934426625566
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,2560,12288,0.026733867327372235
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,2560,12288,0.03602666854858398
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,2560,12288,0.03138986627260844
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,2560,10240,0.022362667322158813
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,2560,10240,0.03312746683756511
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,2560,16384,0.03896426757176717
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,2560,8192,0.018557866414388023
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,2560,8192,0.030216532945632934
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,2560,10240,0.027383466561635334
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,2560,7168,0.01627306640148163
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,2560,7168,0.027893332640329997
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,2560,8192,0.023550933599472045
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,2560,6144,0.014625066518783569
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,2560,6144,0.027394133806228637
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,2560,7168,0.02109439969062805
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,2560,5120,0.01279146671295166
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,2560,5120,0.024708267052968344
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,2560,6144,0.019090133905410766
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,2560,4096,0.009851732850074768
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,2560,4096,0.023589332898457847
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,2560,5120,0.017061332861582436
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,2560,3584,0.00897706647713979
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,2560,3584,0.022283732891082764
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,2560,4096,0.015305599570274353
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,2560,3072,0.008235733211040496
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,2560,3072,0.021112533410390218
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,2560,3584,0.014138666788736978
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,2560,3072,0.012899200121561686
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,2560,2560,0.007084799806276958
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,2560,2560,0.0210805336634318
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,2560,2048,0.0067669332027435304
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,2560,2048,0.019654399156570433
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,2560,2048,0.01090239981810252
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,2560,1536,0.005551999807357788
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,2560,2560,0.012029866377512615
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,2560,1536,0.018557866414388023
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,2560,1024,0.004654933512210846
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,2560,1024,0.017476266622543334
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,3072,65536,0.13215573628743488
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,2560,768,0.004231466849644979
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,2560,1024,0.008683733145395915
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,2560,768,0.016705065965652466
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,2560,512,0.003469866762558619
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,2560,512,0.01759679913520813
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,2560,1536,0.010002133250236512
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,2560,256,0.003066666672627131
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,2560,256,0.015614933768908181
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,2560,512,0.007944533228874206
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,2560,128,0.0029781334102153777
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,2560,256,0.007900799810886382
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,2560,768,0.008339200417200725
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,2560,128,0.014432000120480857
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,2560,64,0.0029824001093705496
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,2560,32,0.0029887999097506206
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,2560,64,0.014459733168284097
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,2560,32,0.015094400445620219
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,2560,128,0.007100800176461537
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,2048,65536,0.09837653636932372
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,2048,16384,0.0252618670463562
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,2048,65536,0.10941440264383953
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,2048,16384,0.04464106559753418
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,2048,12288,0.019222400585810342
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,2048,12288,0.03297386765480041
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,2048,12288,0.030561065673828124
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,2048,10240,0.016378666957219443
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,2048,10240,0.031111466884613036
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,2048,16384,0.03880639870961507
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,2048,8192,0.014179199934005737
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,2048,8192,0.029688533147176104
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,2048,10240,0.027217066287994383
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,2048,7168,0.012541866302490235
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,2048,7168,0.02796906630198161
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,2048,8192,0.022738132874170938
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,2048,6144,0.011406933267911274
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,2048,6144,0.026471465826034546
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,2048,7168,0.021144533157348634
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,2048,5120,0.010073600212732951
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,2048,5120,0.02446720004081726
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,2048,6144,0.019091200828552247
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,2048,4096,0.008782933155695598
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,2048,4096,0.023018666108449302
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,2048,5120,0.017413334051767985
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,2048,3584,0.007960533102353413
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,2048,3584,0.02072746753692627
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,2048,4096,0.015004799763361613
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,2048,3072,0.007146666447321574
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,2048,3072,0.020984532435735066
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,2048,3584,0.014104533195495605
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,2048,2560,0.0063391998410224915
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,2048,3072,0.012938666343688964
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,2048,2560,0.020822399854660036
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,2048,2048,0.005786666770776113
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,2048,2048,0.018662399053573607
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,2048,2560,0.011724799871444702
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,2048,1536,0.005064533154169718
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,2048,1536,0.019010132551193236
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,2048,2048,0.010934399565060933
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,2048,1024,0.004222933451334635
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,2048,1024,0.016747732957204185
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,2048,1536,0.01000853379567464
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,2048,768,0.0038719999293486277
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,2048,768,0.017773866653442383
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,2048,1024,0.008749866485595703
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,2048,512,0.003453866640726725
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,2048,512,0.015927466750144958
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,2048,768,0.008359466989835102
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,2048,256,0.0029685333371162414
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,2048,256,0.014920533696810404
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,2048,512,0.008371200164159138
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,2048,128,0.0030048000315825146
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,2048,128,0.01448853313922882
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,2048,256,0.0075530668099721265
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,2048,64,0.002661333233118057
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,2560,65536,0.13261653582255045
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,2048,64,0.014504533012708029
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,2048,32,0.003053866575161616
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,2048,128,0.007482666770617168
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,2048,32,0.014570666352907815
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,1536,65536,0.0710197369257609
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,1536,65536,0.09848852952321371
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,1536,16384,0.022457599639892578
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,1536,16384,0.039538133144378665
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,1536,12288,0.018683733542760213
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,1536,12288,0.032637866338094075
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,1536,12288,0.030535467465718585
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,1536,10240,0.015260799725850423
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,1536,10240,0.0302346666653951
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,1536,16384,0.0387231985727946
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,1536,8192,0.012917332847913108
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,1536,8192,0.028910932938257854
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,1536,10240,0.02695786754290263
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,1536,7168,0.011693867047627766
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,1536,7168,0.026892799139022826
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,1536,8192,0.022809600830078124
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,1536,6144,0.010838400324185688
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,1536,6144,0.024892799059549966
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,1536,7168,0.02068586746851603
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,1536,5120,0.009287466605504353
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,1536,5120,0.02281279961268107
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,1536,6144,0.01913386583328247
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,1536,4096,0.007619200150171916
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,1536,4096,0.022692267100016275
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,1536,5120,0.01738026738166809
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,1536,3584,0.006878933310508728
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,1536,3584,0.020687999327977498
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,1536,4096,0.015377066532770791
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,1536,3072,0.0063296000162760425
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,1536,3072,0.02129279971122742
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,1536,3584,0.014110933740933737
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,1536,2560,0.00553599993387858
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,1536,2560,0.02028053402900696
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,1536,3072,0.0129120002190272
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,1536,2048,0.0050357331832249965
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,1536,2048,0.01862506667772929
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,1536,2560,0.011940266688664753
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,1536,1536,0.004276266694068909
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,1536,1536,0.017118932803471883
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,1536,2048,0.010825600226720173
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,1536,1024,0.0038453333079814913
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,2048,65536,0.1318389336268107
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,1536,1024,0.016619732975959776
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,1536,768,0.0033621333539485933
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,1536,768,0.018065067132314046
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,1536,1536,0.0102101335922877
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,1536,512,0.0030250666042168934
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,1536,512,0.01488746702671051
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,1536,768,0.008329600095748901
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,1536,256,0.0026719999810059865
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,1536,1024,0.008762666583061218
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,1536,256,0.01623146633307139
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,1536,128,0.002573866645495097
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,1536,512,0.007916800181070964
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,1536,128,0.014544000228246054
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,1536,64,0.0025557334224383037
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,1536,256,0.007530666887760162
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,1536,32,0.00258240004380544
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,1536,64,0.014450132846832275
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,1536,128,0.007338666419188182
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,1536,32,0.014533332983652749
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,1024,65536,0.04992853403091431
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,1024,16384,0.01735466718673706
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,1024,65536,0.09461440245310465
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,1024,16384,0.037732267379760744
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,1024,12288,0.014694399634997048
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,1024,12288,0.033243733644485476
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,1024,12288,0.030580266316731768
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,1024,10240,0.01405333379904429
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,1024,10240,0.02909119923909505
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,1024,16384,0.03805866638819377
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,1024,8192,0.013416533668835958
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,1024,8192,0.02714346647262573
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,1024,10240,0.026840533812840777
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,1024,7168,0.012057600418726604
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,1024,7168,0.02667733430862427
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,1024,7168,0.020757333437601725
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,1024,6144,0.010445866982142131
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,1024,6144,0.02481386661529541
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,1024,8192,0.022707200050354003
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,1024,5120,0.009091200431187947
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,1024,5120,0.02275200088818868
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,1024,5120,0.017107200622558594
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,1024,4096,0.00784746656815211
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,1024,6144,0.019032533963521323
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,1024,4096,0.021754666169484457
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,1024,3584,0.006804266571998596
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,1024,3584,0.020680532852808634
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,1024,3584,0.014114133516947427
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,1024,3072,0.0063274666666984555
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,1024,3072,0.01947306593259176
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,1024,4096,0.014923733472824097
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,1024,2560,0.00555626650651296
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,1024,2560,0.019797333081563315
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,1024,2560,0.01209920048713684
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,1024,2048,0.005062399804592133
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,1024,2048,0.018665599822998046
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,1024,2048,0.01086186667283376
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,1024,1536,0.004644266764322917
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,1024,1536,0.01924053430557251
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,1536,65536,0.13161813418070475
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,1024,1024,0.003913599997758865
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,1024,1024,0.017195733388264973
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,1024,1536,0.009566932916641235
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,1024,1024,0.009134933352470398
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,1024,768,0.0034773332377274835
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,1024,768,0.01590506633122762
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,1024,512,0.003067733347415924
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,1024,512,0.016408532857894897
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,1024,768,0.008382933338483174
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,1024,256,0.00297173336148262
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,1024,512,0.007991466422875721
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,1024,256,0.01623040040334066
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,1024,128,0.002679466704527537
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,1024,128,0.015639467040697734
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,1024,256,0.007538133362929027
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,1024,64,0.0026186667382717133
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,1024,128,0.007503999769687653
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,1024,64,0.014632532993952433
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,1024,32,0.0025973332424958544
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,1024,32,0.01456106702486674
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,768,65536,0.08970452944437662
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,1024,65536,0.1327296018600464
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,768,16384,0.013622400164604188
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,768,16384,0.03509120146433513
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,768,16384,0.0380298654238383
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,768,12288,0.011228799819946289
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,768,12288,0.030987733602523805
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,768,12288,0.030547199646631877
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,768,10240,0.011866666873296102
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,768,10240,0.029113600651423138
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,768,65536,0.13248213132222492
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,768,10240,0.02681279977162679
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,768,8192,0.012268799543380737
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,768,8192,0.026997333765029906
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,768,7168,0.010845866799354554
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,768,7168,0.025146667162577314
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,768,7168,0.020707199970881142
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,768,6144,0.009943466385205586
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,768,6144,0.024990934133529662
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,768,8192,0.022690133253733317
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,768,5120,0.008725333213806152
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,768,5120,0.023127466440200806
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,768,6144,0.01902186671892802
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,768,5120,0.01695786714553833
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,768,4096,0.00713919997215271
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,768,4096,0.021934932470321654
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,768,3584,0.006713599960009258
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,768,3584,0.020957867304484047
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,768,3584,0.014040533701578775
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,768,3072,0.00588266650835673
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,768,4096,0.015003732840220132
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,768,3072,0.02086720069249471
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,768,2560,0.005496533215045929
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,768,2560,0.01871573328971863
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,768,2560,0.012056533495585125
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,768,2048,0.004692266881465912
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,768,3072,0.0128330667813619
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,768,2048,0.018620800971984864
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,768,1536,0.004244266450405121
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,768,1536,0.01732906699180603
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,768,1536,0.01002239982287089
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,768,2048,0.010779733459154766
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,768,1024,0.0034602666894594826
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,768,1024,0.016548267006874083
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,768,768,0.003091199944416682
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,768,768,0.01676693360010783
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,768,1024,0.008756267031033833
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,768,512,0.0030303999781608583
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,768,768,0.008326399823029835
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,768,512,0.015579733252525329
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,768,512,0.008335999647776286
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,768,256,0.0025685332715511323
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,768,256,0.014522666732470194
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,768,256,0.0075914666056633
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,768,128,0.0022240000466505687
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,768,128,0.014451199769973755
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,768,128,0.007530666887760162
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,768,64,0.0021589333812395734
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,768,64,0.01400106648604075
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,768,32,0.0022272000710169475
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,768,32,0.013857066631317139
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,512,65536,0.031062400341033934
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,512,65536,0.0817525307337443
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,512,16384,0.012709333499272665
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,512,16384,0.03319786588350932
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,512,16384,0.03798506657282512
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,512,12288,0.010481066505114238
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,512,12288,0.028937600056330365
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,512,12288,0.030137600501378377
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,512,10240,0.009610666831334432
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,512,10240,0.028998400767644244
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,512,10240,0.02648426691691081
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,512,8192,0.008668800195058186
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,512,8192,0.02681279977162679
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,512,8192,0.022728532552719116
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,512,7168,0.0081535999973615
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,512,7168,0.024996266762415568
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,512,7168,0.020687999327977498
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,512,6144,0.007469866673151653
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,512,6144,0.024446932474772136
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,512,6144,0.018977065881093345
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,512,5120,0.008571733037630718
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,512,5120,0.02283946673075358
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,512,65536,0.13175040086110432
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,512,4096,0.007212799787521362
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,512,4096,0.021027199427286782
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,512,5120,0.017373865842819212
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,512,3584,0.00636053333679835
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,512,3584,0.020617600282033285
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,512,3584,0.01402666668097178
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,512,4096,0.014916266997655234
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,512,3072,0.005922133227189382
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,512,3072,0.019275732835133872
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,512,2560,0.005117866893609365
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,512,2560,0.018977065881093345
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,512,3072,0.012826666235923767
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,512,2048,0.0046741331617037455
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,512,2560,0.01202239990234375
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,512,2048,0.018600533405939736
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,512,1536,0.004182399809360504
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,512,1536,0.017118932803471883
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,512,2048,0.010769066214561463
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,512,1536,0.009947733084360758
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,512,1024,0.0034261333445707956
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,512,1024,0.016692266861597697
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,512,768,0.0030879999200503034
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,512,768,0.016833066940307617
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,512,1024,0.008714666962623597
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,512,512,0.002977066735426585
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,512,512,0.014524799585342408
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,512,768,0.00810346653064092
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,512,256,0.0026047999660174055
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,512,256,0.014524799585342408
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,512,512,0.007874133189519246
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,512,128,0.002254933367172877
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,512,256,0.0075552001595497135
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,512,128,0.01437440017859141
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,512,64,0.002221866697072983
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,512,128,0.007145600020885467
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,512,64,0.014172800381978354
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,512,32,0.0025301332275072734
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,256,65536,0.02143146594365438
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,512,32,0.014509866635004679
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,256,65536,0.07636586825052896
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,256,16384,0.011632000406583149
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,256,16384,0.03173973361651103
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,256,16384,0.037939198811848956
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,256,12288,0.009591466188430786
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,256,12288,0.028727465867996217
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,256,12288,0.030477867523829145
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,256,10240,0.008764800429344178
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,256,10240,0.028350933392842607
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,256,10240,0.026552534103393553
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,256,8192,0.007904000083605449
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,256,8192,0.02690133253733317
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,256,8192,0.022742400566736855
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,256,7168,0.007500799993673961
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,256,7168,0.02487679918607076
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,256,7168,0.020610133806864418
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,256,6144,0.0070592001080513
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,256,6144,0.024114133914311726
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,256,65536,0.13100693225860596
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,256,5120,0.007177599767843883
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,256,5120,0.022362667322158813
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,256,6144,0.018998400370279948
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,256,4096,0.006708266834417979
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,256,4096,0.021828265984853108
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,256,5120,0.01692053278287252
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,256,3584,0.0062943999965985615
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,256,3584,0.020590933163960774
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,256,4096,0.014922666549682616
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,256,3584,0.013719466328620911
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,256,3072,0.005543466905752818
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,256,3072,0.02016106645266215
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,256,3072,0.012829867005348206
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,256,2560,0.005096533397833506
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,256,2560,0.01876266598701477
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,256,2560,0.012053333719571431
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,256,2048,0.004734933376312256
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,256,2048,0.018578133980433145
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,256,1536,0.0038965334494908653
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,256,1536,0.016806399822235106
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,256,2048,0.01083733340104421
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,256,1536,0.009596799810727436
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,256,1024,0.0034634667138258614
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,256,1024,0.0166485329469045
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,256,768,0.0030730667213598887
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,256,768,0.017052799463272095
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,256,1024,0.008750933408737182
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,256,512,0.0026133333643277483
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,256,768,0.00830080012480418
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,256,512,0.014548266927401224
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,256,256,0.002583466718594233
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,256,512,0.00785813331604004
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,256,256,0.014573867122332254
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,256,128,0.0021717332303524016
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,256,256,0.00746666689713796
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,256,128,0.015289599696795145
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,256,64,0.0021546666820844015
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,256,128,0.0071722666422526045
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,256,64,0.014607999722162882
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,256,32,0.00222080002228419
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,128,65536,0.01976319948832194
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,256,32,0.0144896000623703
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,128,65536,0.07191999753316244
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,128,16384,0.008371200164159138
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,128,16384,0.030535467465718585
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,128,16384,0.03786666790644328
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,128,12288,0.007167999943097432
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,128,12288,0.027273599306742353
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,128,12288,0.030495999256769817
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,128,10240,0.006742399930953979
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,128,10240,0.026947200298309326
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,128,10240,0.02674986720085144
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,128,8192,0.006278400123119354
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,128,8192,0.026902399460474652
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,128,8192,0.022644267479578654
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,128,7168,0.006695466736952464
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,128,65536,0.12885653177897136
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,128,7168,0.025405865907669068
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,128,6144,0.0063285330931345625
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,128,6144,0.024740266799926757
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,128,7168,0.020552533864974975
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,128,5120,0.006663466493288676
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,128,5120,0.023818665742874147
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,128,6144,0.018578133980433145
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,128,4096,0.0059456000725428265
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,128,4096,0.02207039992014567
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,128,5120,0.016921599706014
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,128,3584,0.005625600119431814
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,128,3584,0.020946133136749267
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,128,4096,0.014478933811187745
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,128,3072,0.00547626664241155
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,128,3072,0.02020906607309977
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,128,3584,0.01362986663977305
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,128,2560,0.005027199784914652
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,128,3072,0.012432000041007996
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,128,2560,0.018598399559656777
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,128,2048,0.004534400006135305
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,128,2048,0.017574399709701538
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,128,2560,0.011646933356920878
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,128,1536,0.0039018665750821433
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,128,2048,0.010483200351397198
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,128,1536,0.0166293332974116
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,128,1024,0.0034122665723164878
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,128,1024,0.01703146696090698
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,128,1536,0.00959999958674113
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,128,768,0.00297173336148262
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,128,1024,0.008410666386286418
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,128,768,0.01597866714000702
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,128,512,0.002600533266862233
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,128,512,0.016040533781051636
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,128,768,0.008374399940172831
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,128,256,0.0025418666501839954
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,128,256,0.01476800044377645
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,128,512,0.007495466868082683
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,128,128,0.002178133279085159
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,128,128,0.014105600118637086
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,128,128,0.0070783997575442
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,128,64,0.0021418665846188864
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,128,64,0.0139082670211792
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,128,32,0.0021888000269730886
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,128,32,0.014782933394114175
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,64,65536,0.013672533631324767
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,64,65536,0.06334720055262247
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,64,16384,0.006292266647020976
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,64,16384,0.02906026641527812
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,64,12288,0.0057781333724657696
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,64,12288,0.02681066592534383
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,64,10240,0.005897599955399831
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,64,10240,0.02574933369954427
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,64,8192,0.005901866654555003
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,64,8192,0.024964267015457155
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,64,7168,0.00559146652619044
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,64,7168,0.025544534126917522
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,64,6144,0.006325333317120869
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,64,6144,0.02297919988632202
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,64,5120,0.00592853327592214
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,64,5120,0.02292693257331848
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,64,4096,0.005514666438102722
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,64,4096,0.02262186606725057
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,64,3584,0.005500799914201101
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,64,3584,0.02063573400179545
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,64,3072,0.005075199902057648
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,64,3072,0.018942934274673463
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,64,2560,0.005448533097902933
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,64,2560,0.0187008003393809
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,64,2048,0.005039999882380167
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,64,2048,0.016927999258041383
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,64,1536,0.003765333443880081
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,64,1536,0.017288533846537273
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,64,1024,0.003143466760714849
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,64,1024,0.01535040040810903
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,64,768,0.0030378667016824085
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,64,768,0.015012266238530478
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,64,512,0.0027242665489514667
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,64,512,0.014520532886187234
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,64,256,0.002316800008217494
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,64,256,0.015001599987347921
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,64,128,0.0021503999829292296
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,64,128,0.014356266458829245
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,64,64,0.0021536000072956084
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,64,64,0.01336853305498759
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,64,32,0.0021877333521842955
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,64,32,0.013609600067138673
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,32,65536,0.021452800432840983
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,32,65536,0.06248213450113932
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,32,16384,0.007084799806276958
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,32,16384,0.02908373276392619
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,32,12288,0.006250666578610738
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,32,12288,0.027306665976842243
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,32,10240,0.005862399935722351
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,32,10240,0.02601813276608785
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,32,8192,0.005842133363087972
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,32,8192,0.02474986712137858
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,32,7168,0.0058559998869895935
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,32,7168,0.024702932437260947
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,32,6144,0.005858133236567179
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,32,6144,0.022788266340891518
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,32,5120,0.005770666897296906
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,32,5120,0.02271359960238139
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,32,4096,0.005420800050099691
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,32,4096,0.021562665700912476
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,32,3584,0.005031466484069824
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,32,3584,0.02105813423792521
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,32,3072,0.005029333134492239
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,32,3072,0.01946880022684733
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,32,2560,0.005125333368778229
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,32,2560,0.019019732872645058
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,32,2048,0.004711466530958811
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,32,2048,0.019172267119089762
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,32,1536,0.003789866715669632
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,32,1536,0.016739199558893837
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,32,1024,0.0029824001093705496
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,32,1024,0.016705065965652466
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,32,768,0.002979200085004171
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,32,768,0.014713600277900696
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,32,512,0.0025578667720158894
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,32,512,0.01458453337351481
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,32,256,0.002229333420594533
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,32,256,0.016364799936612447
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,32,128,0.002162133405605952
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,32,128,0.014435199896494546
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,32,64,0.002161066730817159
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,32,64,0.014103466272354126
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,32,32,0.0021685334543387097
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,384,32,32,0.013656533757845559
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,65536,16384,0.4765365282694499
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,65536,16384,0.2737557411193848
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,1024,3072,0.012836266557375589
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,65536,12288,0.3577482541402181
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,65536,12288,0.2123583952585856
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,384,768,65536,0.040827735265096025
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,65536,10240,0.26132267316182456
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,65536,10240,0.15718293190002441
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,384,128,256,0.007246933380762736
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,65536,8192,0.21133012771606446
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,65536,8192,0.12676266829172772
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,65536,10240,0.09904320240020752
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,65536,7168,0.18351893424987792
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,65536,7168,0.1132853349049886
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,65536,8192,0.08099093437194824
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,65536,6144,0.16090453465779622
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,65536,6144,0.09730239709218344
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,65536,12288,0.1144330660502116
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,65536,5120,0.13213653564453126
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,65536,5120,0.08602453072865804
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,65536,16384,0.15207146008809408
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,65536,4096,0.10778559843699138
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,65536,4096,0.07495360374450684
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,65536,7168,0.07210773626963297
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,65536,3584,0.09414079984029135
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,65536,3584,0.06624426841735839
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,65536,6144,0.06336853504180909
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,65536,3072,0.08201173146565756
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,65536,3072,0.05665706793467203
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,65536,5120,0.05443840026855469
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,65536,2560,0.06774720350901285
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,65536,2560,0.048845867315928146
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,65536,4096,0.04689600070317586
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,65536,2048,0.05680213371912638
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,65536,2048,0.043816534678141277
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,65536,3072,0.03810346523920695
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,65536,1536,0.04084693193435669
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,65536,1536,0.0375765323638916
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,65536,3584,0.04230506817499797
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,65536,1024,0.028940800825754804
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,65536,1024,0.03264106710751851
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,65536,2560,0.03383573293685913
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,65536,768,0.02257919907569885
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,65536,768,0.02917226751645406
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,65536,2048,0.02990506688753764
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,65536,512,0.0173418660958608
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,65536,512,0.02666880091031392
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,65536,1024,0.020348799228668214
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,65536,256,0.012602667013804117
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,65536,256,0.026859732468922932
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,65536,768,0.018236800034840902
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,65536,128,0.010099200407663982
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,65536,128,0.022843732436498006
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,65536,1536,0.025073067347208662
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,65536,64,0.010056533416112264
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,65536,256,0.013491200407346091
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,65536,64,0.022844799359639487
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,65536,32,0.010452266534169514
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,65536,32,0.023857067028681435
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,16384,65536,0.40540266036987305
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,16384,65536,0.2589194615681966
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,16384,16384,0.10915626684824627
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,16384,16384,0.07773760159810385
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,65536,128,0.012050132950146992
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,16384,12288,0.07918293476104736
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,16384,12288,0.06642666657765707
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,65536,512,0.016199466586112977
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,16384,10240,0.06824959913889567
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,16384,10240,0.0602837324142456
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,16384,10240,0.03195306658744812
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,16384,8192,0.056385068098704014
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,16384,8192,0.05039573510487875
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,16384,12288,0.035955198605855304
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,16384,7168,0.04755413134892782
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,16384,7168,0.046212267875671384
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,16384,16384,0.045509334405263266
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,16384,6144,0.042583465576171875
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,16384,6144,0.04282346566518148
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,16384,8192,0.026980266968409224
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,16384,5120,0.036456533273061115
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,16384,5120,0.03860373497009277
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,16384,7168,0.024436267217000325
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,16384,4096,0.029849600791931153
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,16384,4096,0.033838931719462076
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,16384,5120,0.020101332664489747
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,16384,3584,0.026809600989023845
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,16384,3584,0.0310805340607961
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,16384,6144,0.02241493264834086
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,16384,3072,0.023729066054026283
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,16384,3072,0.03013226588567098
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,16384,4096,0.017401599884033205
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,16384,2560,0.02015893260637919
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,16384,2560,0.026820266246795656
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,16384,2560,0.013306666413942972
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,16384,2048,0.016522666811943053
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,16384,2048,0.024700800577799477
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,16384,3072,0.014686933159828186
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,16384,1536,0.013294933239618936
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,16384,1536,0.02471253275871277
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,16384,3584,0.015925332903862
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,16384,1024,0.010360532999038696
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,16384,1024,0.020861866076787312
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,16384,1024,0.00960213343302409
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,16384,768,0.008754133184750875
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,16384,768,0.018611200650533042
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,16384,1536,0.01088746686776479
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,16384,512,0.0072405333320299785
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,16384,512,0.019564799467722573
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,16384,768,0.00886079967021942
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,16384,256,0.004766933123270671
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,16384,256,0.016572800278663636
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,16384,512,0.007985066870848339
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,16384,128,0.004341333111127218
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,16384,128,0.01657600005467733
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,16384,128,0.007577600081761678
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,16384,64,0.004198400179545084
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,16384,64,0.016847999890645345
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,16384,32,0.005446400245030721
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,16384,32,0.016951467593510947
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,16384,256,0.0075647999842961625
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,12288,65536,0.2942410786946615
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,12288,65536,0.1886624018351237
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,16384,2048,0.012452266613642375
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,12288,16384,0.07582186857859294
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,12288,16384,0.061178668340047204
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,12288,12288,0.056119465827941896
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,12288,12288,0.054230399926503504
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,12288,12288,0.03344000180562337
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,12288,10240,0.04836373329162598
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,12288,10240,0.048503466447194415
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,12288,16384,0.04214293162027995
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,12288,8192,0.040134398142496745
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,12288,8192,0.04188693364461263
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,12288,10240,0.02935146689414978
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,12288,7168,0.03460799853006999
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,12288,7168,0.039297068119049074
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,12288,8192,0.025357866287231447
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,12288,6144,0.031267199913660684
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,12288,6144,0.03535679976145427
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,12288,7168,0.023086933294932048
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,12288,5120,0.02635626594225566
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,12288,5120,0.031165866057078044
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,12288,6144,0.021010132630666097
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,12288,4096,0.021653334299723305
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,12288,4096,0.02932800054550171
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,12288,5120,0.018649599949518838
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,12288,3584,0.018973867098490395
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,12288,3584,0.02751893401145935
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,12288,4096,0.016721065839131674
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,12288,3072,0.01694186727205912
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,12288,3072,0.027035733064015705
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,12288,3584,0.015265066425005594
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,12288,2560,0.014186666409174601
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,12288,2560,0.024864000082015992
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,12288,3072,0.014321066935857139
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,12288,2048,0.01252906620502472
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,12288,2048,0.023465599616368613
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,12288,2560,0.012862933675448099
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,12288,1536,0.010428800185521444
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,12288,1536,0.020772266387939452
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,12288,1536,0.010484266281127929
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,12288,1024,0.0083146666487058
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,12288,1024,0.019499733050664266
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,12288,2048,0.011799466609954835
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,12288,768,0.006838400165239971
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,12288,768,0.017758933703104655
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,12288,1024,0.009171199798583985
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,12288,512,0.005482666691144307
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,12288,512,0.017473065853118898
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,12288,768,0.00877333382765452
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,12288,256,0.0046304002404212955
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,12288,256,0.01644266645113627
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,12288,256,0.007547733187675476
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,12288,128,0.004206933577855428
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,12288,128,0.01634773313999176
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,12288,512,0.00830506682395935
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,12288,64,0.003804799914360046
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,12288,64,0.017257599035898845
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,12288,32,0.004211199780305227
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,12288,32,0.016551466782887776
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,12288,128,0.007528533538182576
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,10240,65536,0.2399967988332113
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,10240,65536,0.2065312067667643
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,10240,16384,0.06575359900792441
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,10240,16384,0.05688213507334391
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,10240,16384,0.040906667709350586
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,10240,12288,0.04963839848836263
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,10240,12288,0.05115093390146891
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,10240,12288,0.033078400293986
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,10240,10240,0.04268906513849895
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,10240,10240,0.045643734931945804
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,16384,65536,0.15404799779256184
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,10240,8192,0.03515306711196899
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,10240,8192,0.03897813161214193
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,10240,10240,0.029028266668319702
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,10240,7168,0.030453334252039593
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,10240,7168,0.03522239923477173
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,10240,8192,0.024765866994857787
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,10240,6144,0.027110399802525838
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,10240,6144,0.03300799926122029
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,12288,65536,0.1403765360514323
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,10240,5120,0.02305493354797363
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,10240,5120,0.030126933256785077
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,10240,7168,0.022915200392405192
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,10240,4096,0.018730666240056357
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,10240,4096,0.027105067173639936
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,10240,5120,0.018722132841746012
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,10240,3584,0.016619732975959776
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,10240,3584,0.027114667495091754
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,10240,6144,0.02103360096613566
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,10240,3072,0.014678399761517844
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,10240,3072,0.025618133942286174
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,10240,4096,0.016261333227157594
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,10240,2560,0.012823466459910074
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,10240,2560,0.024099200963974
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,10240,3584,0.015315199891726175
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,10240,2048,0.011160533626874287
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,10240,2048,0.02286826570828756
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,10240,3072,0.01381226678689321
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,10240,1536,0.009161600470542907
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,10240,1536,0.01907520095507304
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,10240,2560,0.0129120002190272
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,10240,1024,0.007112533350785573
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,10240,1024,0.018228266636530557
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,10240,2048,0.011136000355084736
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,10240,768,0.005901866654555003
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,10240,768,0.018049067258834837
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,10240,1536,0.010035199920336406
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,10240,512,0.004693333307902018
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,10240,512,0.01669013301531474
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,10240,1024,0.00921066701412201
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,10240,256,0.004218666752179464
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,10240,256,0.014570666352907815
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,10240,512,0.008098133405049642
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,10240,128,0.00383786658445994
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,10240,128,0.01644373337427775
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,10240,768,0.008796800176302593
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,10240,64,0.003432533393303553
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,10240,64,0.015161599715550741
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,10240,256,0.007530666887760162
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,10240,32,0.0034293333689371743
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,10240,32,0.01646933356920878
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,10240,128,0.007131733496983846
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,8192,65536,0.21168959935506187
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,8192,65536,0.17986987431844076
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,8192,16384,0.05760960181554159
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,8192,16384,0.0530837337176005
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,8192,12288,0.0426474650700887
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,8192,12288,0.045975466569264725
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,8192,12288,0.031959466139475506
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,8192,10240,0.03640640179316203
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,8192,10240,0.04097386598587036
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,8192,16384,0.03988266785939534
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,8192,8192,0.03025280038515727
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,8192,8192,0.034560000896453856
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,8192,10240,0.028544000784556073
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,8192,7168,0.026469333966573076
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,8192,7168,0.03269226749738057
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,8192,8192,0.024130133787790935
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,8192,6144,0.022177066405614218
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,8192,6144,0.030472532908121748
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,8192,7168,0.022107734282811483
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,8192,5120,0.01875306765238444
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,8192,5120,0.027793065706888838
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,8192,6144,0.020568533738454183
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,8192,4096,0.01542080044746399
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,8192,4096,0.0262442668279012
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,10240,65536,0.1331701358159383
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,8192,3584,0.013380266229311624
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,8192,3584,0.02426346739133199
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,8192,5120,0.0184608002503713
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,8192,3072,0.012386133273442585
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,8192,3072,0.022898133595784506
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,8192,4096,0.01588053305943807
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,8192,2560,0.010912000139554342
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,8192,2560,0.02232853372891744
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,8192,3584,0.014524799585342408
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,8192,2048,0.009218133489290873
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,8192,2048,0.019639466206232706
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,8192,3072,0.013533866405487061
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,8192,1536,0.007965866724650066
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,8192,1536,0.018627200524012247
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,8192,2560,0.01216426690419515
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,8192,1024,0.005962666869163513
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,8192,1024,0.01727466583251953
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,8192,1536,0.010034132997194927
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,8192,768,0.005099733173847198
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,8192,768,0.016708266735076905
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,8192,2048,0.010904533664385478
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,8192,512,0.004247466723124186
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,8192,512,0.017233065764109292
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,8192,768,0.008330666522185007
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,8192,256,0.003402666747570038
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,8192,256,0.015478400389353433
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,8192,1024,0.009217066566149394
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,8192,128,0.0034058667719364167
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,8192,128,0.014573867122332254
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,8192,256,0.007597866654396057
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,8192,64,0.002985599885384242
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,8192,64,0.016516266266504924
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,8192,32,0.003403733422358831
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,8192,128,0.0072053333123524976
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,8192,32,0.016378666957219443
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,7168,65536,0.1877845287322998
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,7168,65536,0.16169919967651367
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,7168,16384,0.05025920073191324
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,8192,512,0.007955199976762136
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,7168,16384,0.049407998720804855
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,7168,12288,0.038262399037679036
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,7168,12288,0.04424426555633545
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,7168,12288,0.03175999919573466
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,7168,10240,0.032874667644500734
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,7168,10240,0.038713598251342775
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,7168,16384,0.038696531454722086
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,7168,8192,0.026999467611312868
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,7168,8192,0.03245546619097392
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,7168,10240,0.02773759961128235
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,7168,7168,0.023883734146753946
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,7168,7168,0.030808534224828082
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,7168,8192,0.023655466238657632
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,7168,6144,0.020826667547225952
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,7168,6144,0.028837333122889202
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,7168,7168,0.021614933013916017
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,7168,5120,0.017208532492319743
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,7168,5120,0.0271232008934021
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,7168,6144,0.019887999693552653
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,7168,4096,0.013827199737230936
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,7168,4096,0.024729599555333458
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,7168,5120,0.01789120038350423
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,7168,3584,0.012577066818873087
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,7168,3584,0.023457066218058268
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,7168,4096,0.01495573321978251
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,7168,3072,0.011250133315722149
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,7168,3072,0.02283733288447062
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,7168,3584,0.014513066411018372
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,7168,2560,0.009760000308354696
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,7168,2560,0.02137493292490641
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,7168,3072,0.012882133324941
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,7168,2048,0.008663466572761536
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,7168,2048,0.02007466753323873
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,7168,2560,0.01211199959119161
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,7168,1536,0.007028266787528992
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,7168,1536,0.018014933665593466
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,7168,2048,0.010840533177057902
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,7168,1024,0.005197866757710775
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,7168,1024,0.01809599995613098
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,7168,1536,0.009660800298055012
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,7168,768,0.004524800181388855
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,7168,768,0.016692266861597697
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,7168,1024,0.008775466680526733
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,7168,512,0.00384853333234787
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,7168,512,0.015160533785820007
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,7168,512,0.008005333443482716
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,7168,256,0.0034005333979924522
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,7168,256,0.014457600315411887
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,7168,768,0.00839573343594869
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,7168,128,0.002976000060637792
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,7168,128,0.015352533260981242
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,7168,128,0.007118933399518331
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,7168,64,0.0030261332790056865
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,7168,64,0.015863466262817382
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,7168,32,0.003385599950949351
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,7168,256,0.007523199915885926
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,7168,32,0.016177067160606386
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,6144,65536,0.16615254084269207
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,6144,65536,0.13598400751749676
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,6144,16384,0.045840001106262206
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,6144,16384,0.04758293231328328
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,8192,65536,0.131988263130188
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,6144,12288,0.03515733480453491
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,6144,12288,0.041339735190073654
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,6144,12288,0.03089173237482707
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,6144,10240,0.03023253281911214
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,6144,10240,0.03696533441543579
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,6144,16384,0.03887146711349487
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,6144,8192,0.024910932779312132
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,6144,8192,0.03256426652272542
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,6144,8192,0.023516800006230673
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,6144,7168,0.02183786630630493
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,6144,7168,0.03035093347231547
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,6144,10240,0.027258666356404622
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,6144,6144,0.01882986625035604
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,6144,6144,0.028854399919509888
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,6144,7168,0.021476266781489055
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,6144,5120,0.016102400422096253
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,6144,5120,0.026764800151189167
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,6144,6144,0.019847466548283895
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,6144,4096,0.013421866297721862
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,6144,4096,0.025149865945180254
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,6144,4096,0.015016532937685647
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,6144,3584,0.012199466427167256
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,6144,3584,0.024425599972407022
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,6144,5120,0.017851734161376955
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,6144,3072,0.010897066195805867
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,6144,3072,0.022443733612696328
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,6144,3072,0.012836266557375589
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,6144,2560,0.009636267026265462
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,6144,2560,0.020119466384251914
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,7168,65536,0.1305525302886963
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,6144,2048,0.00839466651280721
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,6144,2048,0.021022933721542358
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,6144,2560,0.011693867047627766
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,6144,1536,0.006759466727574666
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,6144,1536,0.018784000476201376
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,6144,1536,0.00958079993724823
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,6144,1024,0.005514666438102722
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,6144,1024,0.016847999890645345
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,6144,2048,0.01086186667283376
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,6144,768,0.005156266689300537
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,6144,768,0.017091200749079386
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,6144,1024,0.008748799562454224
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,6144,512,0.004232533276081085
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,6144,512,0.01723519961039225
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,6144,768,0.008355200290679932
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,6144,256,0.0038101332883040107
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,6144,256,0.014978133638699851
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,6144,512,0.007905066510041555
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,6144,128,0.003390933324893316
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,6144,128,0.015305599570274353
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,6144,256,0.007569066683451335
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,6144,64,0.0030016000072161358
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,6144,64,0.014626133441925048
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,6144,32,0.003385599950949351
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,6144,128,0.007083733379840851
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,6144,32,0.015362133582433065
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,5120,65536,0.1415434678395589
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,5120,65536,0.13410132726033527
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,5120,16384,0.039109333356221514
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,5120,16384,0.04446719884872437
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,5120,16384,0.03804159959157308
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,5120,12288,0.029491200049718218
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,5120,12288,0.03841813405354817
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,5120,12288,0.03057173291842143
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,5120,10240,0.02504533330599467
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,5120,10240,0.03496640125910441
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,5120,10240,0.027267199754714967
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,5120,8192,0.02065920035044352
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,5120,8192,0.029586132367451983
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,6144,65536,0.1298858642578125
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,5120,7168,0.0179967999458313
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,5120,7168,0.029107199112574263
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,5120,7168,0.0214464008808136
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,5120,6144,0.015804800391197204
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,5120,6144,0.027741867303848266
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,5120,8192,0.023431466023127238
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,5120,5120,0.013648000359535218
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,5120,5120,0.026294400294621784
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,5120,5120,0.017273600896199545
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,5120,4096,0.011524266997973124
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,5120,4096,0.023938133319218954
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,5120,6144,0.019452800353368126
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,5120,3584,0.010377599795659383
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,5120,3584,0.022856533527374268
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,5120,3584,0.01370133360226949
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,5120,3072,0.009206400314966837
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,5120,3072,0.020666666825612388
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,5120,4096,0.014941866199175516
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,5120,2560,0.008425600330034892
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,5120,2560,0.020691200097401937
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,5120,2560,0.011709866921106975
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,5120,2048,0.007358933488527934
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,5120,2048,0.01882986625035604
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,5120,3072,0.01243946651617686
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,5120,1536,0.006157866617043813
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,5120,1536,0.018266665935516357
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,5120,1536,0.009581866860389709
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,5120,1024,0.00468800018231074
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,5120,1024,0.018659200270970663
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,5120,2048,0.010430933038393656
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,5120,768,0.004301866888999939
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,5120,768,0.01658986707528432
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,5120,1024,0.008801066875457763
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,5120,512,0.0037994667887687682
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,5120,512,0.015991466244061787
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,5120,768,0.008398933211962382
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,5120,256,0.003454933315515518
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,5120,256,0.014590932925542196
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,5120,256,0.007660800218582153
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,5120,128,0.0030250666042168934
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,5120,512,0.007979733248551685
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,5120,128,0.014654933412869772
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,5120,64,0.0029493334392706556
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,5120,64,0.01461120049158732
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,5120,32,0.003031466652949651
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,5120,32,0.014598400394121806
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,5120,128,0.00710399995247523
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,4096,65536,0.11244373321533203
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,4096,65536,0.10650560061136882
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,4096,16384,0.037487999598185225
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,4096,16384,0.04113173484802246
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,4096,16384,0.0376853346824646
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,4096,12288,0.028936533133188884
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,4096,12288,0.03633280197779338
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,5120,65536,0.12997972965240479
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,4096,10240,0.02481493353843689
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,4096,10240,0.03304426670074463
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,4096,12288,0.030268800258636475
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,4096,8192,0.02036693294843038
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,4096,8192,0.028830933570861816
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,4096,10240,0.02687679926554362
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,4096,7168,0.017595734198888144
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,4096,7168,0.0275221327940623
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,4096,7168,0.021100799242655434
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,4096,6144,0.015533866484959922
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,4096,6144,0.026202666759490966
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,4096,8192,0.02277759909629822
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,4096,5120,0.01330560048421224
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,6144,3584,0.014147200187047324
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,4096,5120,0.024946133295694985
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,4096,4096,0.009912533561388652
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,4096,4096,0.023212800423304238
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,4096,4096,0.014507733782132468
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,4096,3584,0.009192533294359843
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,4096,3584,0.021771732966105142
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,4096,5120,0.017016534010569254
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,4096,3072,0.008386133114496867
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,4096,3072,0.021955200036366782
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,4096,6144,0.01901866594950358
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,4096,2560,0.007559466858704884
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,4096,2560,0.01893226703008016
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,4096,2560,0.01176106631755829
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,4096,2048,0.006421333551406861
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,4096,2048,0.018678400913874307
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,4096,3584,0.013657599687576294
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,4096,3072,0.01299626628557841
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,4096,1536,0.005520000060399374
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,4096,1536,0.01816426714261373
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,4096,1024,0.004473599791526795
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,4096,1024,0.016873600085576375
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,4096,1536,0.009659733374913533
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,4096,768,0.003930666546026865
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,4096,768,0.017145599921544394
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,4096,2048,0.010801066954930622
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,4096,512,0.003479466587305069
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,4096,512,0.014591999848683677
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,4096,1024,0.00879146655400594
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,4096,256,0.003054933249950409
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,4096,256,0.015284267067909241
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,4096,768,0.008280533552169799
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,4096,128,0.002998399982849757
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,4096,128,0.015547733505566916
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,4096,256,0.007542400062084198
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,4096,64,0.0026378666361172995
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,4096,128,0.00714026689529419
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,4096,32,0.0030762667457262674
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,4096,64,0.015024000406265258
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,4096,512,0.00792746643225352
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,4096,32,0.01511253317197164
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,3584,65536,0.10265706380208332
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,3584,65536,0.10400533676147461
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,3584,16384,0.028883200883865357
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,3584,16384,0.03754133383433024
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,3584,12288,0.022056533892949422
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,3584,12288,0.033297065893809
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,3584,12288,0.03025919993718465
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,3584,10240,0.01874026656150818
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,3584,10240,0.030884265899658203
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,3584,16384,0.03761920134226481
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,3584,8192,0.01562346617380778
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,3584,8192,0.027853866418202717
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,3584,10240,0.026837333043416338
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,3584,7168,0.01368106702963511
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,3584,7168,0.026760532458623247
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,3584,8192,0.02302293380101522
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,3584,6144,0.01239359974861145
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,3584,6144,0.025550933678944905
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,3584,7168,0.02072319984436035
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,3584,5120,0.010725333293279012
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,3584,5120,0.023641600211461385
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,3584,6144,0.018716800212860107
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,3584,4096,0.00943893293539683
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,3584,4096,0.022294400135676067
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,4096,65536,0.12892800172170002
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,3584,3584,0.00860693355401357
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,3584,3584,0.02076906760533651
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,3584,5120,0.01655893325805664
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,3584,3072,0.0076000000039736434
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,3584,3072,0.020616533358891805
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,3584,3584,0.013734400272369385
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,3584,2560,0.006878933310508728
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,3584,2560,0.019690666596094766
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,3584,4096,0.01452906628449758
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,3584,2048,0.005850666761398315
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,3584,2048,0.017505067586898803
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,3584,3072,0.012574932972590127
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,3584,1536,0.005094400048255921
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,3584,1536,0.017834667364756265
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,3584,2560,0.011628799637158712
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,3584,1024,0.0038624001046021783
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,3584,2048,0.01076800028483073
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,3584,1024,0.01667733391125997
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,3584,768,0.0035413332283496858
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,3584,768,0.01605226695537567
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,3584,1536,0.009617066383361817
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,3584,512,0.0034058667719364167
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,3584,512,0.016075733304023742
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,3584,768,0.008302933474381765
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,3584,256,0.003009066730737686
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,3584,256,0.014598400394121806
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,3584,1024,0.008986666798591614
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,3584,128,0.0026026666164398193
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,3584,512,0.007894399762153625
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,3584,128,0.015377066532770791
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,3584,64,0.00257493332028389
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,3584,64,0.014591999848683677
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,3584,32,0.002525866776704788
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,3584,256,0.007560533285140991
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,3584,32,0.014946132898330688
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,3584,128,0.007101866602897644
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,3072,65536,0.09343893527984619
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,3072,16384,0.030054400364557903
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,3072,65536,0.09989759922027588
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,3072,16384,0.039390933513641355
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,3072,12288,0.020747733116149903
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,3072,12288,0.03295680085817973
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,3072,12288,0.030144000053405763
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,3072,10240,0.018191999197006224
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,3072,10240,0.030828799804051715
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,3072,16384,0.037913600603739425
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,3072,8192,0.014949333667755128
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,3072,8192,0.02718720038731893
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,3072,10240,0.02643199960390727
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,3072,7168,0.013081600268681845
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,3072,7168,0.027062400182088213
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,3072,8192,0.022359466552734374
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,3072,6144,0.01188266674677531
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,3072,6144,0.024795732895533242
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,3072,7168,0.020660267273585
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,3072,5120,0.01051093339920044
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,3072,5120,0.023985066016515098
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,3072,6144,0.018645334243774413
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,3072,4096,0.008983467022577922
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,3072,4096,0.02156160076459249
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,3072,5120,0.01702079971631368
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,3072,3584,0.008414933085441589
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,3072,3584,0.02074346741040548
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,3072,4096,0.014514133334159851
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,3072,3072,0.007542400062084198
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,3072,3072,0.020065067211786906
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,3072,3584,0.013492266337076822
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,3072,2560,0.006770133475462596
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,3072,2560,0.01928640007972717
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,3072,3072,0.012504532933235168
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,3072,2048,0.005853866537412008
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,3072,2048,0.018374399344126383
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,3072,2560,0.011704533298810323
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,3072,1536,0.005096533397833506
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,3072,1536,0.017489065726598106
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,3072,1536,0.009610666831334432
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,3072,1024,0.004198400179545084
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,3072,1024,0.0166485329469045
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,3072,1024,0.008646399776140849
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,3072,768,0.003774933268626531
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,3072,768,0.015901866555213928
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,3072,768,0.008382933338483174
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,3072,512,0.00340693344672521
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,3072,512,0.016320000092188515
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,3072,512,0.007965866724650066
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,3072,256,0.0030741333961486817
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,3072,256,0.015409066279729208
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,3072,256,0.007551999886830647
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,3072,128,0.0030069333811601003
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,3072,128,0.01609280010064443
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,3072,128,0.007214933137098948
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,3072,64,0.0026591998835404714
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,3584,65536,0.1297258694966634
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,3072,64,0.014600533246994018
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,3072,32,0.0026261332134405774
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,3072,32,0.014622933665911355
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,2560,65536,0.07455253601074219
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,2560,65536,0.08883306980133057
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,2560,16384,0.025226666529973345
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,2560,16384,0.03842133283615112
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,2560,16384,0.037444265683492024
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,2560,12288,0.02148159941037496
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,2560,12288,0.03296213348706563
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,2560,12288,0.030013867219289142
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,2560,10240,0.0245685338973999
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,2560,10240,0.03020159999529521
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,3072,65536,0.13000106811523438
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,2560,8192,0.014114133516947427
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,2560,8192,0.02684906721115112
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,2560,10240,0.02598293423652649
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,2560,7168,0.01258666714032491
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,2560,7168,0.026062933603922527
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,2560,7168,0.020282665888468422
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,2560,6144,0.011424000064531963
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,2560,6144,0.02481600046157837
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,2560,8192,0.022235733270645142
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,2560,5120,0.010026666522026061
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,2560,5120,0.022738132874170938
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,2560,5120,0.016614400347073875
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,2560,4096,0.00844693382581075
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,2560,4096,0.02068693240483602
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,2560,6144,0.01869973341623942
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,2560,3584,0.007981866598129272
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,2560,3584,0.020718934138615926
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,2560,3584,0.01383039951324463
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,2560,3072,0.0071381335457166035
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,2560,3072,0.020521599054336547
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,2560,4096,0.014597333470980325
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,2560,2560,0.006302933394908905
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,2560,2560,0.019337600469589232
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,2560,2560,0.01165120005607605
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,2560,2048,0.0055402666330337524
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,2560,2048,0.01787733236948649
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,2560,3072,0.012471466263135274
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,2560,1536,0.004871466755867004
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,2560,1536,0.017463467518488564
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,2560,1536,0.009964799880981446
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,2560,1024,0.0038730666041374207
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,2560,1024,0.017129600048065186
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,2560,2048,0.010878933469454448
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,2560,768,0.0038101332883040107
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,2560,768,0.014612266421318054
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,2560,1024,0.008849066495895386
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,2560,512,0.0034314667185147605
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,2560,512,0.014840533336003622
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,2560,768,0.00837546686331431
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,2560,256,0.0029877332349618276
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,2560,256,0.014486400286356607
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,2560,256,0.007527466615041096
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,2560,512,0.00798826664686203
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,2560,128,0.0025802666942278544
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,2560,128,0.014455466469128927
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,2560,128,0.007129600147406261
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,2560,64,0.0025770666698614756
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,2560,64,0.014510933558146158
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,2560,32,0.0025802666942278544
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,2560,32,0.01476693352063497
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,2048,65536,0.061043198903401694
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,2048,65536,0.08119466304779052
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,2048,16384,0.021236266692479452
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,2048,16384,0.03544640143712362
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,2048,16384,0.03749333222707112
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,2560,65536,0.1299168030420939
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,2048,12288,0.019206400712331137
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,2048,12288,0.032333866755167646
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,2048,10240,0.015734400351842245
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,2048,10240,0.029304534196853638
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,2048,10240,0.02606719930966695
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,2048,8192,0.013321600357691445
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,2048,8192,0.027317333221435546
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,2048,12288,0.02963520089785258
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,2048,7168,0.01209920048713684
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,2048,7168,0.024935466051101685
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,2048,8192,0.022375466426213582
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,2048,6144,0.010892800490061442
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,2048,6144,0.02405653397242228
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,2048,7168,0.02031893332799276
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,2048,5120,0.009590400258700053
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,2048,5120,0.023060266176859537
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,2048,5120,0.01694613297780355
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,2048,4096,0.00803306649128596
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,2048,4096,0.0206986665725708
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,2048,6144,0.018702934185663857
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,2048,3584,0.0076223999261856076
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,2048,3584,0.020770132541656494
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,2048,3584,0.013659733533859252
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,2048,3072,0.006750933329264323
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,2048,3072,0.0187285323937734
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,2048,4096,0.014578133821487427
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,2048,2560,0.0063274666666984555
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,2048,2560,0.019001599152882895
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,2048,2560,0.011654399832089742
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,2048,2048,0.00513919989267985
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,2048,3072,0.012466133634249369
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,2048,2048,0.018601600329081217
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,2048,1536,0.004277333120505015
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,2048,1536,0.016701867183049522
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,2048,1536,0.009609599908192951
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,2048,1024,0.00378560001651446
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,2048,1024,0.016731733083724977
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,2048,2048,0.010757333040237427
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,2048,768,0.0034346667428811393
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,2048,768,0.01694719990094503
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,2048,1024,0.008368000388145447
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,2048,512,0.0029887999097506206
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,2048,512,0.01579093337059021
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,2048,768,0.007955199976762136
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,2048,256,0.0026208000878492994
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,2048,256,0.014852266510327658
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,2048,512,0.007877333462238312
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,2048,256,0.007537066439787547
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,2048,128,0.0026464000344276427
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,2048,128,0.014545067151387533
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,2048,64,0.002549333373705546
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,2048,64,0.014518400033315023
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,2048,128,0.007124266525109608
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,2048,32,0.014644267161687216
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,1536,65536,0.04821226596832275
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,1536,65536,0.07597440083821615
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,2048,65536,0.1296789328257243
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,1536,16384,0.018659200270970663
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,1536,16384,0.03331093390782674
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,1536,16384,0.037164799372355145
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,1536,12288,0.015610667069753012
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,1536,12288,0.03224213322003682
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,1536,12288,0.02932479977607727
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,1536,10240,0.013800533612569174
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,1536,10240,0.028536534309387206
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,1536,10240,0.025570134321848553
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,1536,8192,0.013217066725095114
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,1536,8192,0.025131734212239583
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,1536,65536,0.12889599800109863
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,1536,7168,0.011636267105738323
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,1536,7168,0.02587839961051941
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,1536,8192,0.021921066443125407
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,1536,6144,0.010546132922172546
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,1536,6144,0.022937599817911783
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,1536,7168,0.020179200172424316
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,1536,5120,0.009218133489290873
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,1536,5120,0.023085866371790567
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,1536,6144,0.018196266889572144
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,1536,4096,0.007354666789372762
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,1536,4096,0.020895999670028687
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,1536,5120,0.016569599509239197
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,1536,3584,0.007176533341407776
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,1536,3584,0.02061333258946737
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,1536,3584,0.013733333349227906
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,1536,3072,0.006738133231798808
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,1536,3072,0.020463999112447104
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,1536,4096,0.014481066664059957
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,1536,3072,0.0124917338291804
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,1536,2560,0.005910400052865346
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,1536,2560,0.018920532862345376
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,1536,2048,0.005103999873002371
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,1536,2048,0.017698132991790773
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,1536,2560,0.011319466431935628
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,1536,1536,0.004327466587225596
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,1536,2048,0.010771200060844421
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,1536,1536,0.018476800123850504
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,1536,1024,0.0038090666135152185
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,1536,1024,0.016777600844701132
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,1536,1024,0.008755200107892354
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,1536,768,0.0034080001215140024
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,1536,1536,0.009573333462079366
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,1536,768,0.01700586676597595
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,1536,512,0.003035733352104823
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,1536,512,0.015194666385650635
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,1536,768,0.008323200047016144
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,1536,512,0.008012799918651581
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,1536,256,0.002587733417749405
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,1536,256,0.014725333452224732
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,1536,128,0.0025909334421157838
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,1536,128,0.014528000354766845
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,1536,256,0.007480533421039581
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,1536,64,0.0022304000953833262
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,1536,64,0.014247467120488485
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,1536,128,0.007136000196139018
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,1536,32,0.002552533398071925
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,1024,65536,0.035905067125956217
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,1536,32,0.014433067043622336
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,1024,65536,0.06655466556549072
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,1024,16384,0.013751467068990072
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,1024,16384,0.031128533681233722
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,1024,16384,0.03668800195058187
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,1024,12288,0.011987200379371643
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,1024,12288,0.02990613381067912
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,1024,12288,0.029050666093826293
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,1024,10240,0.010795733332633973
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,1024,10240,0.027002666393915815
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,1024,10240,0.025597866376241046
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,1024,8192,0.012079999844233195
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,1024,8192,0.02513386607170105
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,1024,8192,0.021893332401911415
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,1024,7168,0.011267200112342834
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,1024,7168,0.024171733856201173
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,1024,7168,0.020054399967193604
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,1024,6144,0.009990400075912476
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,1024,6144,0.02264960010846456
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,1024,6144,0.018219733238220216
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,1024,5120,0.008746666709582011
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,1024,5120,0.021651200453440347
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,1024,5120,0.016451199849446617
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,1024,4096,0.007143466671307881
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,1024,4096,0.020428800582885744
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,1024,4096,0.014133333166440328
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,1024,3584,0.006376533210277558
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,1024,3584,0.02025493383407593
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,1024,65536,0.12846506436665853
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,1024,3072,0.005931733548641205
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,1024,3072,0.019769599040349327
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,1024,3584,0.013678933183352152
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,1024,2560,0.005095466474692027
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,1024,2560,0.019444266955057778
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,1024,3072,0.012479999661445617
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,1024,2048,0.004660266637802124
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,1024,2048,0.016952532529830932
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,1024,2560,0.011223466197649638
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,1024,1536,0.004198400179545084
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,1024,1536,0.0180565337340037
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,1024,2048,0.010733866691589355
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,1024,1024,0.0034656000634034475
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,1024,1024,0.016109866897265117
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,1024,1536,0.009532800316810608
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,1024,768,0.002995199958483378
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,1024,768,0.014586666226387024
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,1024,1024,0.009112532933553059
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,1024,512,0.0030271999537944795
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,1024,768,0.008296533425649007
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,1024,512,0.015209600329399109
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,1024,256,0.002624000112215678
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,1024,256,0.0144896000623703
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,1024,512,0.007942399879296621
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,1024,128,0.002221866697072983
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,1024,256,0.007523199915885926
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,1024,128,0.01388053297996521
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,1024,64,0.0021727999051411946
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,1024,128,0.007118933399518331
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,1024,64,0.01474240024884542
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,1024,32,0.0021920000513394673
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,1024,32,0.014538666605949402
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,768,65536,0.029501867294311524
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,768,65536,0.0657034675280253
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,768,16384,0.012845866878827415
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,768,16384,0.02902719974517822
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,768,16384,0.03660906553268432
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,768,12288,0.010890666643778484
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,768,12288,0.028709334135055543
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,768,12288,0.02889813383420308
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,768,10240,0.010019200046857198
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,768,10240,0.026959999402364092
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,768,10240,0.025240532557169598
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,768,8192,0.00879039963086446
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,768,8192,0.02501759926478068
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,3072,2048,0.01076479951540629
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,768,7168,0.008364799618721008
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,768,7168,0.023316266139348348
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,768,8192,0.02184106707572937
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,768,6144,0.007901866734027863
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,768,6144,0.02223466634750366
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,768,7168,0.020188800493876138
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,768,5120,0.008709333340326945
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,768,5120,0.021447465817133585
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,768,6144,0.018556799491246542
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,768,4096,0.00674773355325063
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,768,4096,0.02031146685282389
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,768,5120,0.016611199577649435
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,768,3584,0.006347733239332835
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,768,4096,0.014453333616256715
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,768,3584,0.01990613341331482
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,768,3072,0.005866666634877523
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,768,3072,0.018606932957967122
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,768,3584,0.013297067085901896
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,768,2560,0.005130666494369507
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,768,3072,0.01239359974861145
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,768,2560,0.018281600872675576
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,768,2048,0.004621866842110952
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,768,2048,0.01827413241068522
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,768,2560,0.011682132879892986
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,768,2048,0.01039466659228007
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,768,1536,0.004198400179545084
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,768,1536,0.01653333306312561
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,768,1024,0.0034101332227389016
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,768,1024,0.014737066626548768
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,768,65536,0.1271999994913737
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,768,768,0.0034602666894594826
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,768,768,0.014761599898338317
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,768,1536,0.009197866916656494
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,768,512,0.002976000060637792
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,768,1024,0.00872320036093394
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,768,512,0.014681599537531533
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,768,256,0.002567466596762339
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,768,768,0.008348799745241801
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,768,256,0.014444800217946372
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,768,128,0.0025813333690166474
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,768,512,0.007592533528804779
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,768,128,0.014075733224550881
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,768,256,0.007145600020885467
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,768,64,0.002195200075705846
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,768,32,0.002233600119749705
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,768,128,0.007052800059318543
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,768,64,0.014587733149528503
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,768,32,0.014492799838383993
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,512,65536,0.02402026653289795
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,512,16384,0.010439466436703999
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,512,65536,0.060179201761881504
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,512,16384,0.029179733991622925
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,512,12288,0.012075733145078022
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,512,12288,0.027956267197926838
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,512,12288,0.02892586588859558
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,512,10240,0.010776533683141073
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,512,10240,0.026709334055582686
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,512,16384,0.03667733271916707
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,512,8192,0.00944640040397644
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,512,8192,0.02492799957593282
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,512,10240,0.02560639977455139
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,512,7168,0.008539733290672303
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,512,7168,0.022765866915384927
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,512,8192,0.021901865800221763
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,512,6144,0.00800960014263789
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,512,6144,0.02273919979731242
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,512,7168,0.019843200842539467
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,512,5120,0.007180800040562947
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,512,5120,0.02071040074030558
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,512,6144,0.018262400229771932
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,512,4096,0.006730666756629944
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,512,4096,0.0209824005762736
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,512,5120,0.016225066781044007
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,512,3584,0.00631466656923294
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,512,3584,0.01904639999071757
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,512,4096,0.014512000481287637
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,512,3072,0.0054837331175804135
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,512,3072,0.019672532876332603
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,512,3584,0.013699199755986533
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,512,2560,0.00506986677646637
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,512,2560,0.01741866668065389
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,512,3072,0.01243839959303538
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,512,2048,0.004610133171081543
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,512,2048,0.017156267166137697
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,512,2560,0.011662933230400085
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,512,1536,0.003864533454179764
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,512,1536,0.01768746574719747
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,512,2048,0.010801066954930622
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,512,1024,0.003387733300526937
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,512,1024,0.016037333011627197
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,512,1536,0.009630933403968811
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,512,768,0.0030346666773160298
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,512,1024,0.008750933408737182
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,512,768,0.016396799683570863
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,512,512,0.0026421333352724713
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,512,512,0.014563199877738953
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,512,768,0.007971199850241344
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,512,512,0.007567999760309856
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,512,256,0.0026378666361172995
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,512,256,0.01452906628449758
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,512,256,0.007490133245786031
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,512,128,0.0023306667804718017
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,512,128,0.014380799730618796
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,512,128,0.007144533097743988
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,512,64,0.002219733347495397
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,512,64,0.014365866780281067
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,512,32,0.002274133265018463
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,512,32,0.014427733421325684
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,256,65536,0.015612799922625223
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,256,65536,0.05748053391774496
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,256,16384,0.009760000308354696
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,256,16384,0.02690453330675761
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,256,16384,0.036702934900919595
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,256,12288,0.00837546686331431
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,256,12288,0.026495999097824095
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,512,65536,0.12694506645202636
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,256,12288,0.029215999444325763
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,256,10240,0.007594666878382365
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,256,10240,0.025228800376256307
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,256,8192,0.007167999943097432
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,256,8192,0.025226666529973345
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,256,10240,0.025605332851409913
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,256,8192,0.021887999773025513
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,256,7168,0.006675200164318084
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,256,7168,0.022856533527374268
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,256,6144,0.006306133170922597
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,256,6144,0.022523732980092366
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,256,7168,0.019853866100311278
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,256,5120,0.006715733309586842
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,256,6144,0.018145066499710084
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,256,5120,0.021572266022364298
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,256,4096,0.006676266590754191
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,256,4096,0.01990506649017334
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,256,5120,0.01653439998626709
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,256,3584,0.005914666752020518
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,256,4096,0.014135467012723288
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,256,3584,0.018582399686177573
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,256,3072,0.005460266768932342
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,256,3072,0.018935465812683107
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,256,65536,0.12503360112508138
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,256,3584,0.01328213314215342
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,256,2560,0.005077333251635233
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,256,2560,0.01834133267402649
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,256,3072,0.012423466642697651
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,256,2048,0.0045973335703214015
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,256,2048,0.017573332786560057
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,256,1536,0.003836799909671148
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,256,1536,0.017194666465123496
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,256,2560,0.011582932869593303
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,256,1024,0.0034314667185147605
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,256,2048,0.01039466659228007
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,256,1024,0.016498133540153503
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,256,768,0.0029919999341169994
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,256,1536,0.0091839998960495
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,256,768,0.01644159952799479
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,256,1024,0.008348799745241801
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,256,512,0.0026122666895389557
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,256,512,0.014467199643452963
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,256,768,0.007964799801508587
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,256,256,0.0025578667720158894
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,256,512,0.007923200229803721
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,256,256,0.014535466829935709
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,256,128,0.002205866575241089
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,256,256,0.007080533107121785
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,256,128,0.014628266294797262
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,256,64,0.002162133405605952
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,256,128,0.007122133175532024
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,256,32,0.0021727999051411946
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,256,64,0.014129066467285156
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,128,65536,0.015356799960136414
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,256,32,0.01453013320763906
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,128,16384,0.006353066861629486
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,128,65536,0.0478432019551595
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,128,16384,0.025517867008845015
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,128,12288,0.006717866659164429
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,128,12288,0.02654186685880025
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,128,16384,0.036584532260894774
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,128,12288,0.02903573314348857
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,128,10240,0.006267733375231425
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,128,10240,0.02571733395258586
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,128,8192,0.005926399926344554
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,128,8192,0.024600533644358318
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,128,10240,0.025523199637730913
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,128,7168,0.005436799923578898
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,128,8192,0.021512534221013388
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,128,7168,0.022797866662343343
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,128,6144,0.005850666761398315
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,128,6144,0.022295467058817544
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,128,7168,0.020247467358907065
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,128,5120,0.006673066814740498
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,128,6144,0.018219733238220216
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,128,5120,0.0218122661113739
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,128,4096,0.005871999760468801
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,128,4096,0.020870399475097657
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,128,5120,0.016074666380882265
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,128,3584,0.0057087997595469155
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,128,4096,0.014425599575042724
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,128,3584,0.019030400117238364
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,128,3584,0.012969600160916648
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,128,3072,0.00555733342965444
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,128,3072,0.019027199347813925
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,128,65536,0.124236798286438
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,128,3072,0.01241386632124583
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,128,2560,0.005123200019200643
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,128,2560,0.018986666202545167
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,128,2048,0.004701866706212362
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,128,2048,0.01765759984652201
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,128,2560,0.01130986710389455
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,128,1536,0.003833599885304769
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,128,1536,0.01775253415107727
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,128,2048,0.010412800312042236
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,128,1024,0.003366400053103765
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,128,1024,0.01600106656551361
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,128,1536,0.009197866916656494
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,128,768,0.0029696000119050344
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,128,1024,0.008335999647776286
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,128,768,0.014786133170127868
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,128,512,0.002609066665172577
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,128,512,0.014681599537531533
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,128,512,0.0076117331782976794
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,128,256,0.002583466718594233
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,128,256,0.014335999886194864
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,128,128,0.0021418665846188864
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,128,256,0.007245866457621257
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,128,128,0.01537493367989858
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,128,64,0.0021685334543387097
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,128,64,0.014537599682807923
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,128,128,0.006773333251476288
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,128,32,0.0021738665799299877
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,64,65536,0.010943999886512757
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,128,32,0.014600533246994018
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,64,65536,0.04479680061340332
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,64,16384,0.006295466423034668
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,64,12288,0.0055861334005991616
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,64,16384,0.02523626685142517
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,64,10240,0.00586346685886383
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,64,12288,0.024860799312591553
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,64,8192,0.005874133110046387
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,64,10240,0.02477653423945109
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,64,7168,0.005513600011666616
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,64,8192,0.022793600956598915
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,64,7168,0.022657066583633423
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,64,6144,0.005468800167242686
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,64,5120,0.0051136001944541935
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,64,6144,0.022295467058817544
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,64,4096,0.005460266768932342
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,64,5120,0.021322667598724365
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,64,3584,0.005338666836420695
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,64,4096,0.021233065923055013
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,64,3072,0.005028266708056132
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,64,3584,0.019834667444229126
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,64,2560,0.005107200145721436
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,64,3072,0.018580265839894614
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,64,2048,0.005047466854254405
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,64,2560,0.01773973306020101
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,64,1536,0.0037461332976818085
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,64,2048,0.01811093290646871
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,64,1024,0.0029824001093705496
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,64,1536,0.016780799627304076
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,64,768,0.002977066735426585
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,64,1024,0.01618559956550598
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,64,512,0.002603733291228612
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,64,768,0.014574933052062988
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,64,256,0.0022069332500298817
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,64,512,0.014468266566594442
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,64,128,0.0021578667064507803
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,64,256,0.01458133359750112
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,64,64,0.002186666677395503
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,64,128,0.015072000026702882
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,64,32,0.0021407999098300933
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,64,64,0.01474240024884542
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,64,32,0.014597333470980325
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,32,65536,0.016106667121251424
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,32,16384,0.006674133241176605
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,32,65536,0.044930132230122884
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,32,12288,0.005449600021044413
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,32,12288,0.024756266673405965
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,32,10240,0.005857066810131073
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,32,10240,0.02436586618423462
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,32,8192,0.005857066810131073
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,32,8192,0.024317866563796996
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,32,7168,0.005497600138187409
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,32,7168,0.022658133506774904
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,32,6144,0.005435733497142792
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,32,6144,0.02239146629969279
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,32,5120,0.005041066805521647
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,32,5120,0.0226474662621816
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,32,4096,0.005437866846720377
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,32,4096,0.022205867369969687
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,32,3584,0.00506986677646637
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,32,3584,0.018756266434987387
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,32,3072,0.004741333425045013
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,32,3072,0.01983893314997355
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,32,2560,0.005074133475621542
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,32,2560,0.017884800831476845
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,32,2048,0.004637866715590159
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,32,2048,0.017875200510025023
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,32,1536,0.003881600002447764
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,32,1536,0.016668800512949625
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,32,1024,0.003125333289305369
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,32,1024,0.01609599987665812
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,32,768,0.0030410667260487873
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,32,768,0.014920533696810404
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,32,512,0.002656000107526779
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,32,512,0.014492799838383993
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,32,256,0.002253866692384084
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,32,256,0.014520532886187234
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,32,128,0.0022314667701721193
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,32,128,0.01504853367805481
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,32,64,0.0022367998957633973
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,32,64,0.014688000082969666
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,32,32,0.002163200080394745
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,32,32,0.0145087997118632
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,65536,16384,0.40131839116414386
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,65536,16384,0.230348793665568
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,256,2048,32,0.0026880001028378804
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,65536,12288,0.29541759490966796
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,65536,12288,0.17663466135660807
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,256,128,768,0.00798826664686203
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,65536,10240,0.24699840545654297
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,65536,10240,0.14531413714090985
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,256,32,16384,0.02485226591428121
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,65536,8192,0.19736852645874023
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,65536,8192,0.12000853220621746
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,65536,8192,0.07981546719868979
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,65536,7168,0.17049813270568848
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,65536,7168,0.10747520128885905
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,65536,12288,0.11666239897410076
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,65536,6144,0.14997119903564454
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,65536,6144,0.09466559886932373
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,65536,10240,0.09983466466267904
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,65536,5120,0.12592853705088297
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,65536,5120,0.08286293347676596
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,65536,16384,0.15377813975016277
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,65536,4096,0.10173439979553223
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,65536,4096,0.07069653669993083
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,65536,5120,0.05488959948221842
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,65536,3584,0.089464537302653
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,65536,3584,0.06298133134841918
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,65536,6144,0.0634335994720459
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,65536,3072,0.07753067016601563
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,65536,3072,0.05685333410898844
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,65536,4096,0.04594560066858928
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,65536,2560,0.06543039878209432
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,65536,2560,0.048997334639231366
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,65536,7168,0.07055253187815348
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,65536,2048,0.05345919926961264
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,65536,2048,0.043281066417694095
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,65536,3584,0.041960533459981283
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,65536,1536,0.041492267449696855
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,65536,1536,0.038571735223134354
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,65536,3072,0.03700799942016601
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,65536,1024,0.02937813401222229
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,65536,1024,0.03123626708984375
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,65536,1536,0.02466986576716105
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,65536,768,0.023256532351175942
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,65536,768,0.027211733659108478
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,65536,2048,0.0287445326646169
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,65536,512,0.016454399625460307
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,65536,512,0.02325119972229004
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,65536,768,0.0181877334912618
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,65536,256,0.01060693363348643
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,65536,256,0.022707200050354003
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,65536,1024,0.019973333676656088
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,65536,128,0.00858133335908254
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,65536,128,0.020703999201456706
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,65536,2560,0.033217066526412965
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,65536,64,0.0071487997968991595
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,65536,64,0.020758400360743202
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,65536,32,0.007524266839027405
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,65536,32,0.01878933310508728
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,16384,65536,0.3771519978841146
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,16384,65536,0.23309653600056968
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,65536,128,0.01170133352279663
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,16384,16384,0.10162346363067627
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,16384,16384,0.0777728001276652
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,65536,512,0.015732266505559287
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,16384,12288,0.07497920195261637
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,65536,256,0.012938666343688964
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,16384,12288,0.061844265460968016
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,16384,10240,0.064628267288208
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,16384,10240,0.05433493455251058
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,16384,10240,0.031326933701833086
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,16384,8192,0.05338346560796102
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,16384,8192,0.04742613236109416
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,16384,12288,0.03548906644185384
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,16384,7168,0.04579306840896606
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,16384,7168,0.04338560104370117
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,16384,16384,0.044601599375406906
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,16384,6144,0.03984640041987102
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,16384,6144,0.04029759963353475
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,16384,8192,0.02610879937807719
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,16384,5120,0.034474666913350424
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,16384,5120,0.035011200110117595
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,16384,7168,0.023706666628519692
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,16384,4096,0.028039467334747315
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,16384,4096,0.03221759994824727
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,16384,5120,0.019433599710464478
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,16384,3584,0.025054933627446492
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,16384,3584,0.02969920039176941
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,16384,4096,0.01687893271446228
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,16384,3072,0.02212693293889364
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,16384,3072,0.027302400271097822
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,16384,6144,0.022004266579945884
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,16384,2560,0.019016534090042114
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,16384,2560,0.026308266321818034
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,16384,3584,0.01552959978580475
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,16384,2048,0.015398400028546652
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,16384,2048,0.024413865804672242
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,16384,2560,0.013582932949066161
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,16384,1536,0.01246399978796641
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,16384,1536,0.022813866535822548
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,16384,3072,0.014533332983652749
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,16384,1024,0.009628799557685853
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,16384,1024,0.01986666719118754
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,16384,1024,0.00914026697476705
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,16384,768,0.008056533336639405
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,16384,768,0.01802560091018677
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,16384,1536,0.010380799571673077
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,16384,512,0.006345599889755249
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,16384,512,0.01702186663945516
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,16384,768,0.008827733000119527
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,16384,256,0.004293333490689596
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,16384,256,0.018339200814565023
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,16384,512,0.008411733309427898
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,16384,128,0.003833599885304769
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,16384,128,0.016446933150291443
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,16384,128,0.007166933516661327
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,16384,64,0.004238933324813843
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,16384,64,0.016545066237449647
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,16384,32,0.004297600189844767
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,16384,2048,0.011991467078526814
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,16384,32,0.016421332955360413
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,16384,256,0.0075434664885203045
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,12288,65536,0.2719029426574707
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,12288,65536,0.20717120170593262
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,12288,16384,0.07476692994435628
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,12288,16384,0.06776426633199056
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,12288,12288,0.05480106671651205
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,12288,12288,0.055446398258209226
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,12288,12288,0.03329066634178161
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,12288,10240,0.047858134905497236
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,12288,10240,0.04923839966456096
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,12288,16384,0.041050668557484946
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,12288,8192,0.039204267660776775
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,12288,8192,0.04139093160629272
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,12288,10240,0.029428267478942872
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,12288,7168,0.03375680049260457
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,12288,7168,0.0371946652730306
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,12288,8192,0.024513065814971924
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,12288,6144,0.02987733284632365
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,12288,6144,0.033411200841267905
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,12288,7168,0.02255893349647522
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,12288,5120,0.026001065969467163
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,12288,5120,0.030932267506917317
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,12288,6144,0.02034986615180969
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,12288,4096,0.02102186679840088
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,12288,4096,0.02684906721115112
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,12288,5120,0.01912533243497213
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,12288,3584,0.018708266814549766
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,12288,3584,0.02704106569290161
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,12288,4096,0.01618773341178894
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,12288,3072,0.01614293356736501
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,12288,3072,0.0245088001092275
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,12288,3584,0.01530346671740214
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,12288,2560,0.014390400052070618
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,12288,2560,0.023974400758743287
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,12288,3072,0.014129066467285156
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,12288,2048,0.012000000476837159
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,12288,2048,0.02256960074106852
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,12288,2560,0.01284160017967224
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,12288,1536,0.00999679962793986
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,12288,1536,0.02068693240483602
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,12288,1536,0.010366933544476827
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,12288,1024,0.00798826664686203
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,12288,1024,0.020041600863138834
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,12288,2048,0.01086293359597524
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,12288,768,0.00633493314186732
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,12288,768,0.016532267133394875
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,12288,768,0.008386133114496867
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,12288,512,0.005044266581535339
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,12288,512,0.016309332847595216
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,12288,1024,0.00885653297106425
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,12288,256,0.00413973331451416
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,12288,256,0.016691199938456216
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,12288,256,0.007931733131408691
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,12288,128,0.0034101332227389016
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,12288,128,0.014551466703414917
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,12288,128,0.007539199789365132
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,12288,64,0.003369600077470144
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,12288,64,0.015642666816711427
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,12288,32,0.003390933324893316
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,12288,32,0.014587733149528503
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,12288,512,0.007906133433183034
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,10240,65536,0.30088427861531575
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,10240,65536,0.1788991928100586
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,10240,16384,0.08219733238220214
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,10240,16384,0.06522026856740316
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,10240,16384,0.0403658668200175
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,10240,12288,0.06293866634368897
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,10240,12288,0.05305600166320801
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,10240,12288,0.03241066733996074
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,10240,10240,0.05123200019200643
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,10240,10240,0.0461952010790507
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,16384,65536,0.15227200190226237
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,10240,8192,0.03905920187632243
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,10240,8192,0.03906453450520833
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,10240,10240,0.028112000226974486
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,10240,7168,0.03742293516794841
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,10240,7168,0.03730666637420654
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,12288,65536,0.13708052635192872
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,10240,6144,0.032681600252787275
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,10240,6144,0.0340992013613383
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,10240,8192,0.02438933253288269
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,10240,5120,0.026437334219614667
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,10240,5120,0.03041599988937378
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,10240,7168,0.02207039992014567
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,10240,4096,0.02241493264834086
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,10240,4096,0.028593067328135175
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,10240,6144,0.02031573255856832
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,10240,3584,0.021116799116134642
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,10240,3584,0.027284266551335652
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,10240,5120,0.01853760083516439
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,10240,3072,0.017128533124923705
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,10240,3072,0.026339199145634967
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,10240,4096,0.016152532895406087
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,10240,2560,0.01576746702194214
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,10240,2560,0.022706133127212525
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,10240,3072,0.013700266679128012
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,10240,2048,0.012758400042851767
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,10240,2048,0.02173653244972229
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,10240,3584,0.014914133151372275
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,10240,1536,0.011086933811505635
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,10240,1536,0.020440532763799032
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,10240,2560,0.012095999717712403
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,10240,1024,0.007411199808120728
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,10240,1024,0.01845759948094686
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,10240,1024,0.008835200468699138
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,10240,768,0.006307200094064077
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,10240,1536,0.009989333152770997
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,10240,768,0.01655466655890147
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,10240,512,0.005030400057633718
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,10240,512,0.017744000752766928
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,10240,512,0.00795306662718455
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,10240,256,0.004220800101757049
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,10240,256,0.017672532796859743
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,10240,768,0.00839573343594869
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,10240,128,0.0038602667550245917
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,10240,2048,0.011219200491905213
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,10240,128,0.015221333503723145
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,10240,64,0.003390933324893316
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,10240,64,0.014707199732462563
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,10240,32,0.003399466723203659
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,10240,256,0.007548800110816956
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,10240,32,0.01514240006605784
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,8192,65536,0.18676053682963054
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,10240,128,0.00716480016708374
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,8192,65536,0.15677013397216796
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,8192,16384,0.0521237333615621
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,8192,16384,0.05403733253479004
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,8192,12288,0.038729600111643475
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,8192,12288,0.051394132773081455
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,8192,12288,0.03152746756871541
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,8192,10240,0.0339466651280721
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,8192,10240,0.04527999957402547
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,8192,16384,0.039628799756368
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,8192,8192,0.027731200059254963
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,8192,8192,0.03928639888763428
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,8192,10240,0.027668267488479614
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,8192,7168,0.024258132775624594
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,8192,7168,0.031181865930557252
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,8192,8192,0.023990400632222495
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,8192,6144,0.021643733978271483
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,8192,6144,0.028518400589625043
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,8192,7168,0.0216757337252299
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,8192,5120,0.018398932615915933
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,8192,5120,0.026837333043416338
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,10240,65536,0.13310720125834147
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,8192,4096,0.01490239997704824
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,8192,4096,0.024948267141977946
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,8192,6144,0.020004266500473024
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,8192,3584,0.013530666629473368
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,8192,3584,0.02299413283665975
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,8192,5120,0.017861332496007284
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,8192,3072,0.012098133563995361
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,8192,3072,0.022964266935984294
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,8192,4096,0.015742933750152587
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,8192,2560,0.010822400450706482
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,8192,2560,0.02101759910583496
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,8192,3584,0.015050666530927024
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,8192,2048,0.009269332885742188
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,8192,2048,0.018548266092936198
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,8192,3072,0.013401599725087485
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,8192,1536,0.007572266459465027
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,8192,1536,0.018849066893259683
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,8192,2560,0.011708799997965496
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,8192,2048,0.010869333148002624
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,8192,1024,0.005881600081920624
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,8192,1024,0.01753173271814982
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,8192,768,0.0045738667249679565
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,8192,768,0.017259732882181803
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,8192,1536,0.009986133376757304
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,8192,512,0.003975466638803482
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,8192,512,0.016532267133394875
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,8192,768,0.008724266290664673
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,8192,256,0.003402666747570038
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,8192,256,0.016269866625467935
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,8192,1024,0.008743466933568318
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,8192,128,0.0034133332471052804
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,8192,128,0.0151637335618337
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,8192,256,0.007589333256085714
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,8192,64,0.0034229333202044168
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,8192,64,0.014728533228238425
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,8192,32,0.003446399917205175
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,8192,128,0.0070826664566993715
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,8192,512,0.007863466441631318
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,7168,65536,0.1750592072804769
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,7168,65536,0.13382186889648437
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,7168,16384,0.04907413323720296
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,7168,16384,0.055490132172902426
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,7168,16384,0.03881066640218099
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,7168,12288,0.03673066695531209
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,7168,12288,0.04537493387858073
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,7168,12288,0.03097386757532756
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,7168,10240,0.03174720009167989
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,7168,10240,0.044523731867472334
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,7168,10240,0.0273087998231252
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,7168,8192,0.02643946607907613
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,7168,8192,0.038278400897979736
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,8192,65536,0.13081066608428954
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,7168,7168,0.022694400946299233
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,7168,8192,0.023528534173965453
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,7168,7168,0.028937600056330365
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,7168,6144,0.02009920080502828
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,7168,6144,0.027040000756581622
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,7168,6144,0.019859200716018675
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,7168,5120,0.017030400037765504
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,7168,5120,0.02526613275210063
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,7168,7168,0.0216757337252299
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,7168,4096,0.014467199643452963
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,7168,4096,0.025219200054804485
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,7168,5120,0.017864533265431724
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,7168,3584,0.012888532876968384
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,7168,3584,0.0229312002658844
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,7168,4096,0.01580693324406942
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,7168,3072,0.011641599734624227
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,7168,3072,0.021356799205144248
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,7168,3072,0.0126720001300176
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,7168,2560,0.010433066884676616
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,7168,2560,0.02070186734199524
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,7168,3584,0.013940266768137612
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,7168,2048,0.008916266759236654
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,7168,2048,0.019843200842539467
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,7168,2048,0.010899200042088827
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,7168,1536,0.007210666437943776
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,7168,1536,0.01911679903666178
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,7168,2560,0.011667199929555257
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,7168,1024,0.006330666442712148
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,7168,1024,0.01834026575088501
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,7168,1024,0.008771199981371562
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,7168,768,0.005551999807357788
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,7168,768,0.016999467213948568
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,7168,1536,0.009656533598899841
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,7168,512,0.0046634669105211895
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,7168,512,0.014588800072669984
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,7168,512,0.007589333256085714
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,7168,256,0.0034485332667827605
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,7168,256,0.014520532886187234
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,7168,768,0.008017066617806752
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,7168,128,0.0030623999734719592
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,7168,128,0.015358933806419372
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,7168,128,0.007195733487606049
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,7168,64,0.002976000060637792
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,7168,64,0.014693333705266317
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,7168,256,0.0075103998184204105
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,7168,32,0.014479999740918478
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,6144,65536,0.1748821258544922
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,6144,65536,0.11989759604136149
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,8192,32,0.016461867094039916
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,6144,16384,0.047670400142669676
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,6144,16384,0.05011200110117594
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,7168,65536,0.13048213322957355
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,6144,12288,0.03664106527964274
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,6144,12288,0.041212801138559976
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,6144,16384,0.03848426739374797
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,6144,10240,0.0310698668162028
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,6144,10240,0.0357045332590739
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,6144,12288,0.030482133229573567
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,6144,8192,0.025612799326578777
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,6144,8192,0.0313429335753123
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,6144,10240,0.02725119988123576
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,6144,7168,0.022743467489878336
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,6144,7168,0.029923200607299805
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,6144,8192,0.023162666956583658
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,6144,6144,0.019800533850987755
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,6144,6144,0.027539199590682982
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,6144,7168,0.021523199478785195
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,6144,5120,0.016759467124938966
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,6144,5120,0.026500266790390015
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,6144,5120,0.017433599630991618
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,6144,4096,0.013729066650072733
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,6144,4096,0.02438186605771383
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,6144,6144,0.019517866770426433
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,6144,3584,0.012473600109418233
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,6144,3584,0.02309760053952535
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,6144,3584,0.014100266496340432
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,6144,3072,0.011266133189201355
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,6144,3072,0.021177599827448525
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,6144,4096,0.015006933609644571
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,6144,2560,0.009807999928792317
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,6144,2560,0.02061226765314738
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,6144,2560,0.011665067076683045
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,6144,2048,0.008400000135103862
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,6144,2048,0.018632533152898152
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,6144,3072,0.01290986637274424
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,6144,1536,0.007064533233642578
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,6144,1536,0.016682666540145875
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,6144,1536,0.00999679962793986
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,6144,1024,0.0051136001944541935
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,6144,1024,0.016109866897265117
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,6144,2048,0.010804266730944315
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,6144,768,0.004249600072701773
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,6144,768,0.016603733102480568
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,6144,768,0.008427733182907104
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,6144,512,0.003828266759713491
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,6144,512,0.015733333428700765
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,6144,1024,0.00876586635907491
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,6144,256,0.002994133283694585
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,6144,256,0.014471466342608133
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,6144,256,0.007494399944941203
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,6144,128,0.0029696000119050344
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,6144,128,0.014512000481287637
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,6144,512,0.007957333326339721
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,6144,64,0.002611200014750163
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,6144,64,0.014509866635004679
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,6144,32,0.0026591998835404714
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,6144,32,0.01453439990679423
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,6144,128,0.007134933272997539
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,5120,65536,0.15793066024780272
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,5120,65536,0.11910400390625
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,5120,16384,0.0452458659807841
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,5120,16384,0.051949866612752284
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,5120,16384,0.03831786712010701
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,5120,12288,0.034595199426015216
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,5120,12288,0.03898239930470784
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,5120,12288,0.030946133534113567
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,5120,10240,0.029656533400217695
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,5120,10240,0.03405226469039917
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,6144,65536,0.12933119932810466
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,5120,8192,0.024282666047414143
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,5120,8192,0.03018346627553304
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,7168,32,0.003018666555484136
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,5120,7168,0.021370667219161987
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,5120,7168,0.029450666904449464
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,5120,10240,0.027276800076166792
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,5120,6144,0.018786134322484334
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,5120,6144,0.026798933744430542
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,5120,8192,0.02285439968109131
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,5120,5120,0.01590293347835541
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,5120,5120,0.025065600872039795
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,5120,7168,0.021357866128285725
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,5120,4096,0.01328213314215342
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,5120,4096,0.02371413310368856
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,5120,5120,0.016636799772580466
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,5120,3584,0.012190933028856914
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,5120,3584,0.022155733903249104
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,5120,6144,0.019661867618560792
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,5120,3072,0.010843732953071594
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,5120,3072,0.020985599358876547
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,5120,4096,0.014528000354766845
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,5120,2560,0.009399466713269551
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,5120,2560,0.019899733861287437
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,5120,3584,0.013686399658521017
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,5120,2048,0.0082997332016627
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,5120,2048,0.018309332927068076
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,5120,3072,0.012499200304349263
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,5120,1536,0.006769066552321117
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,5120,1536,0.01688533425331116
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,5120,2560,0.011711999773979187
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,5120,1024,0.004683733483155568
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,5120,1024,0.017505067586898803
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,5120,1024,0.008799999952316284
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,5120,768,0.004266666869322458
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,5120,768,0.01730026602745056
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,5120,2048,0.010486400127410889
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,5120,512,0.003822933385769526
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,5120,512,0.014657066265741984
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,5120,768,0.008414933085441589
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,5120,256,0.0033770665526390077
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,5120,512,0.007565866907437642
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,5120,256,0.014599466323852539
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,5120,128,0.0029866665601730345
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,5120,128,0.014556800325711569
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,5120,128,0.007057066758473713
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,5120,64,0.002682666728893916
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,5120,256,0.007513600091139476
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,5120,64,0.015191466609636942
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,5120,32,0.0029919999341169994
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,5120,32,0.0146506667137146
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,4096,65536,0.1087615966796875
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,4096,65536,0.09943573474884033
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,4096,16384,0.03749119838078817
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,4096,16384,0.04212373495101929
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,4096,16384,0.03749226729075114
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,4096,12288,0.028459733724594115
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,4096,12288,0.035045333703358966
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,4096,12288,0.030151466528574627
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,4096,10240,0.02444159984588623
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,4096,10240,0.030829866727193195
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,5120,65536,0.13013652960459393
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,4096,8192,0.01967573364575704
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,4096,8192,0.02892586588859558
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,4096,10240,0.026845866441726686
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,4096,7168,0.017454934120178223
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,4096,7168,0.026717867453893023
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,4096,8192,0.022699733575185142
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,4096,6144,0.015401599804560342
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,4096,6144,0.024762666225433348
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,4096,7168,0.020658133427302043
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,4096,5120,0.013302399714787801
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,4096,5120,0.022744532426198324
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,4096,6144,0.018662399053573607
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,4096,4096,0.010860799749692281
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,4096,4096,0.02141759991645813
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,4096,5120,0.016593066851298015
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,4096,3584,0.008474666873613994
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,4096,3584,0.020884267489115396
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,4096,4096,0.014593066771825156
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,4096,3072,0.007957333326339721
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,4096,3072,0.020258132616678873
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,4096,3584,0.013719466328620911
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,4096,2560,0.006738133231798808
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,4096,2560,0.018212266763051353
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,4096,3072,0.012516267100969949
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,4096,2048,0.006330666442712148
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,4096,2048,0.017591466506322227
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,4096,2048,0.010456533233324686
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,4096,1536,0.005089066425959269
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,4096,1536,0.016900267203648886
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,4096,2560,0.01160426636536916
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,4096,1024,0.0038474666575590765
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,4096,1024,0.01580693324406942
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,4096,1024,0.008730666836102803
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,4096,768,0.0034783999125162757
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,4096,768,0.015090133746465048
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,4096,1536,0.009928533434867859
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,4096,512,0.0030762667457262674
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,4096,512,0.015403733650843302
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,4096,768,0.008362666765848795
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,4096,256,0.0026581334571043652
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,4096,256,0.01446613371372223
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,4096,512,0.007919999957084655
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,4096,128,0.002976000060637792
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,4096,128,0.014567466576894126
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,4096,256,0.007485866546630859
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,4096,128,0.007130666573842366
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,4096,64,0.002570666621128718
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,4096,32,0.003014400104681651
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,4096,64,0.014575999975204468
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,4096,32,0.01444586714108785
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,3584,65536,0.10257066885630291
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,3584,65536,0.09247146447499593
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,3584,16384,0.030947200457255047
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,3584,16384,0.04392533302307129
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,3584,16384,0.03712426821390788
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,3584,12288,0.029561599095662434
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,3584,12288,0.034246400992075605
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,4096,65536,0.1287391980489095
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,3584,10240,0.02432960073153178
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,3584,10240,0.03155200084050496
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,3584,10240,0.026782933870951337
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,3584,8192,0.020485333601633706
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,3584,8192,0.029099732637405396
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,3584,12288,0.030107732613881426
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,3584,7168,0.0180021325747172
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,3584,7168,0.027116799354553224
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,3584,7168,0.020412800709406535
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,3584,6144,0.015638400117556253
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,3584,6144,0.022950400908788048
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,3584,8192,0.02232746680577596
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,3584,5120,0.013528533776601157
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,3584,5120,0.02217386762301127
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,3584,6144,0.01818986733754476
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,3584,4096,0.011171199878056844
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,3584,4096,0.02025173306465149
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,3584,5120,0.016617600123087564
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,3584,3584,0.010211199522018433
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,3584,3584,0.019230933984120686
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,3584,3584,0.013413332899411521
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,3584,3072,0.007550933460394542
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,3584,3072,0.020415999492009482
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,3584,4096,0.01455466647942861
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,3584,2560,0.006780800223350525
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,3584,2560,0.017476266622543334
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,3584,2560,0.011614933609962463
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,3584,3072,0.012376532951990763
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,3584,2048,0.005670399963855743
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,3584,2048,0.01690773367881775
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,3584,1536,0.005286400020122528
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,3584,1536,0.01691840092341105
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,3584,2048,0.010392533739407857
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,3584,1024,0.003857066730658213
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,3584,1024,0.01515733301639557
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,3584,1536,0.009607467055320739
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,3584,768,0.003765333443880081
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,3584,768,0.01602133313814799
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,3584,1024,0.008354133367538452
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,3584,512,0.0033813332517941795
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,3584,512,0.01628159979979197
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,3584,768,0.008323200047016144
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,3584,256,0.0029887999097506206
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,3584,256,0.014506666858990987
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,3584,512,0.007915733257929485
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,3584,128,0.0025994665920734406
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,3584,128,0.015624533096949259
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,3584,256,0.007482666770617168
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,3584,64,0.002603733291228612
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,3584,128,0.0070816000302632645
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,3584,64,0.014500266313552857
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,3584,32,0.002609066665172577
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,3584,32,0.01461120049158732
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,3072,65536,0.0791647990544637
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,3072,65536,0.08722986380259196
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,3072,16384,0.028668800989786785
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,3072,16384,0.035028266906738284
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,3072,16384,0.037522133191426596
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,3072,12288,0.019831466674804687
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,3072,12288,0.03106879989306132
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,5120,1536,0.00962453285853068
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,3072,10240,0.016986666123072307
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,3072,10240,0.028962133328119914
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,3584,65536,0.13017919858296711
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,3072,8192,0.01390506625175476
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,3072,8192,0.026959999402364092
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,3072,12288,0.029782400528589888
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,3072,7168,0.012643200159072877
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,3072,7168,0.026570665836334228
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,3072,10240,0.02646613319714864
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,3072,6144,0.011260799566904704
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,3072,6144,0.02349546750386556
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,3072,7168,0.0202293336391449
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,3072,5120,0.009988266229629516
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,3072,5120,0.022760534286499025
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,3072,8192,0.021993599335352578
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,3072,4096,0.008514133095741273
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,3072,4096,0.020962133010228475
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,3072,5120,0.016204800208409628
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,3072,3584,0.007956266899903615
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,3072,3584,0.020106667280197145
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,3072,4096,0.014620799819628397
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,3072,3072,0.007162666817506154
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,3072,3072,0.0188810666402181
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,3072,6144,0.018194133043289186
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,3072,2560,0.006284800171852112
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,3072,2560,0.017805866400400796
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,3072,2560,0.011662933230400085
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,3072,2048,0.005454933146635691
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,3072,2048,0.017683200041453042
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,3072,3072,0.01251200040181478
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,3072,1536,0.004314666489760081
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,3072,3584,0.013289599617322286
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,3072,1536,0.017097600301106772
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,3072,1024,0.0038293334345022834
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,3072,1024,0.015713066856066386
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,3072,1024,0.008865066369374593
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,3072,768,0.003401600072781245
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,3072,768,0.015770666797955833
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,3072,1536,0.009642666578292847
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,3072,512,0.0030432000756263735
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,3072,512,0.014569600423177084
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,3072,2048,0.010443733135859171
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,3072,256,0.0025813333690166474
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,3072,256,0.014487466216087342
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,3072,768,0.008334933718045553
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,3072,128,0.0025685332715511323
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,3072,512,0.007941333452860515
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,3072,256,0.007170133292675018
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,3072,128,0.0151146670182546
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,3072,64,0.002197333425283432
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,3072,32,0.0022367998957633973
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,3072,64,0.014447999993960061
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,3072,32,0.014458666245142618
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,2560,65536,0.0815562645594279
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,3072,128,0.007189333438873291
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,2560,16384,0.026666667064030963
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,2560,65536,0.08077546755472818
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,2560,16384,0.03828159968058269
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,2560,12288,0.02211306691169739
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,2560,12288,0.03086506724357605
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,2560,12288,0.029704533020655316
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,2560,10240,0.01636693378289541
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,2560,10240,0.029014400641123456
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,2560,16384,0.03744853337605794
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,2560,8192,0.013860266407330832
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,2560,8192,0.026547199487686156
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,2560,10240,0.0260917325814565
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,2560,7168,0.012719999750455221
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,2560,7168,0.027350399891535444
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,2560,8192,0.02210986614227295
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,2560,6144,0.01135040024916331
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,2560,6144,0.024225066105524697
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,2560,7168,0.02035520076751709
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,2560,5120,0.009913600484530131
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,2560,5120,0.021221333742141725
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,2560,6144,0.018636800845464072
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,2560,4096,0.008781866232554118
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,2560,4096,0.02058560053507487
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,2560,5120,0.01660480002562205
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,2560,3584,0.008003200093905132
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,2560,4096,0.014115200440088908
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,2560,3584,0.02095680038134257
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,2560,3072,0.007122133175532024
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,2560,3072,0.018887466192245482
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,3072,65536,0.12917333443959553
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,2560,2560,0.006298666695753734
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,2560,2560,0.018439465761184694
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,2560,3072,0.012574932972590127
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,2560,2048,0.005500799914201101
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,2560,2560,0.011632000406583149
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,2560,2048,0.01726400057474772
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,2560,3584,0.013687466581662497
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,2560,1536,0.005027199784914652
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,2560,1536,0.016545066237449647
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,2560,1024,0.0042581334710121155
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,2560,1024,0.01722559928894043
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,2560,1024,0.00840106705824534
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,2560,768,0.0038133333126703895
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,2560,768,0.014954666296641031
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,2560,1536,0.009938133756319682
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,2560,512,0.0034229333202044168
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,2560,512,0.01503679951032003
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,2560,2048,0.010431999961535137
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,2560,768,0.008328533172607422
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,2560,256,0.002994133283694585
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,2560,512,0.007851733267307282
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,2560,128,0.0026047999660174055
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,2560,256,0.014453333616256715
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,2560,128,0.014505599935849508
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,2560,64,0.002603733291228612
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,2560,64,0.014532267053922018
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,2560,32,0.0025792000194390613
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,2560,128,0.007068799932797749
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,2560,256,0.007517866790294647
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,2560,32,0.014535466829935709
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,2048,65536,0.06517226696014404
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,2048,65536,0.07350827058156331
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,2048,16384,0.021281067530314127
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,2048,16384,0.03486933310826619
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,2048,12288,0.017097600301106772
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,2048,12288,0.031227733691533404
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,2048,12288,0.02974826693534851
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,2048,10240,0.015362133582433065
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,2048,10240,0.027314132452011107
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,2048,16384,0.037265066305796304
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,2048,8192,0.013738666971524557
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,2048,8192,0.024843732515970864
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,2048,10240,0.0255786657333374
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,2048,7168,0.012300800283749897
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,2048,7168,0.024370133876800537
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,2048,8192,0.021945599714914957
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,2048,6144,0.011061333616574605
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,2048,6144,0.02374293406804403
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,2048,7168,0.02010986606280009
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,2048,5120,0.009938133756319682
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,2048,5120,0.0217141330242157
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,2048,6144,0.01871466636657715
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,2048,4096,0.008551466464996337
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,2048,4096,0.019973333676656088
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,2048,5120,0.016545066237449647
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,2048,3584,0.007513600091139476
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,2048,3584,0.019742933909098308
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,2048,4096,0.014593066771825156
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,2048,3072,0.006691200037797292
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,2048,3072,0.018106667200724284
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,2048,3584,0.01325440009435018
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,2048,2560,0.0058005332946777345
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,2048,2560,0.017518933614095053
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,2048,3072,0.012446932991345723
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,2048,2048,0.004715733230113983
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,2048,2048,0.017214934031168617
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,2048,2560,0.01125440001487732
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,2048,1536,0.004231466849644979
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,2048,1536,0.018773333231608073
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,2048,2048,0.010830932855606079
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,2048,1024,0.0034741332133611047
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,2048,1024,0.016140799721082053
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,2048,1536,0.009618133306503296
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,2048,768,0.003409066547950109
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,2048,768,0.016315733393033348
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,2048,1024,0.008785067001978557
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,2048,512,0.003031466652949651
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,2048,768,0.007999999821186066
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,2048,512,0.01460906664530436
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,2048,256,0.0026517334083716077
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,2048,256,0.014493866761525472
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,2048,256,0.007563733557860057
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,2048,512,0.007930666704972585
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,2048,128,0.002569599946339925
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,2048,128,0.014751999576886495
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,2048,128,0.007157333195209503
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,2048,64,0.002254933367172877
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,2048,64,0.014446933070818582
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,2048,32,0.002569599946339925
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,2048,32,0.014484266440073649
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,1536,65536,0.04536853233973186
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,2560,65536,0.12915093104044598
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,1536,65536,0.06794772942860922
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,1536,16384,0.015533866484959922
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,1536,16384,0.03220053315162659
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,1536,12288,0.013925333817799887
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,1536,12288,0.028930133581161498
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,1536,12288,0.02937920093536377
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,1536,10240,0.012487467130025227
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,1536,10240,0.02677546739578247
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,1536,16384,0.036710401376088456
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,1536,8192,0.012462932864824932
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,1536,8192,0.024689066410064697
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,1536,10240,0.025642667214075727
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,1536,7168,0.011300266782442728
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,1536,7168,0.02300373315811157
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,1536,8192,0.021597866217295328
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,1536,6144,0.010098133484522502
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,1536,6144,0.022627200682957968
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,1536,7168,0.020283732811609903
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,1536,5120,0.008760533730189006
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,1536,5120,0.02068159977595011
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,1536,5120,0.016566399733225504
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,1536,4096,0.0075530668099721265
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,1536,4096,0.020188800493876138
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,1536,4096,0.014598400394121806
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,2048,65536,0.1278112014134725
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,1536,3584,0.006781866649786632
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,1536,3584,0.018797866503397622
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,1536,3072,0.006309333443641663
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,1536,3072,0.018809600671132406
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,1536,3584,0.013377066453297934
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,1536,3072,0.012469333410263062
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,1536,2560,0.0054666668176651
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,1536,2560,0.01800533334414164
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,1536,2048,0.004730666677157084
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,1536,2048,0.01772800087928772
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,1536,2560,0.011273599664370219
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,1536,2048,0.01034986674785614
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,1536,1536,0.004238933324813843
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,1536,1536,0.016572800278663636
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,1536,1024,0.003475199888149897
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,1536,1024,0.01695893406867981
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,1536,1024,0.008716799815495809
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,1536,768,0.0030613332986831666
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,1536,1536,0.009621333082516987
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,1536,768,0.015148799618085226
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,1536,512,0.0029696000119050344
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,1536,512,0.016596266627311708
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,1536,768,0.007948799928029378
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,1536,512,0.007541333138942718
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,1536,256,0.00257493332028389
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,1536,256,0.01444906691710154
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,1536,128,0.002553600072860718
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,1536,128,0.014544000228246054
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,1536,256,0.007506133119265239
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,1536,64,0.0021685334543387097
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,1536,128,0.00713919997215271
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,1536,64,0.012985600034395852
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,1536,32,0.0022698665658632917
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,1536,32,0.014059733351071677
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,1024,65536,0.036347734928131106
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,1024,65536,0.060850131511688235
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,1024,16384,0.013274666666984559
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,1024,16384,0.030379732449849445
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,1024,16384,0.03673706849416097
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,1024,12288,0.011356799801190694
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,1024,12288,0.028832000494003297
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,1024,12288,0.029341866572697956
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,1024,10240,0.010406399766604107
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,1024,10240,0.024946133295694985
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,1536,65536,0.1291690667470296
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,1024,8192,0.00956053336461385
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,1024,8192,0.024250666300455727
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,1024,10240,0.025569067398707075
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,1024,7168,0.008780800302823384
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,1024,7168,0.02295359969139099
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,1024,7168,0.01989013353983561
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,1024,6144,0.010089600086212158
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,1024,6144,0.02118933399518331
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,1024,8192,0.021942400932312013
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,1024,5120,0.008588799834251403
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,1024,5120,0.020940800507863365
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,1024,6144,0.01862186590830485
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,1024,4096,0.007490133245786031
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,1024,4096,0.018641066551208497
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,1024,4096,0.014523733655611673
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,1024,3584,0.006676266590754191
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,1024,3584,0.018807466824849448
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,1024,3584,0.01346560021241506
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,1024,3072,0.005924266576766968
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,1024,3072,0.018488534291585288
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,1024,3072,0.012134400010108948
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,1024,2560,0.005448533097902933
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,1024,2560,0.017837866147359212
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,1024,2560,0.01167466640472412
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,1024,2048,0.004614399870236715
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,1024,2048,0.016698666413625083
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,1024,65536,0.1273034652074178
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,1024,1536,0.00420053352912267
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,1024,1536,0.016709333658218382
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,1024,2048,0.010407466689745586
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,1024,1024,0.003402666747570038
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,1024,1024,0.015430399775505066
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,1024,1536,0.00918506681919098
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,1024,768,0.003068800022204717
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,1024,768,0.015412267049153647
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,1024,1024,0.008732799688975017
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,1024,512,0.0029781334102153777
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,1024,512,0.014670933286348978
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,1024,768,0.0083146666487058
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,1024,256,0.00258240004380544
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,1024,256,0.014595199624697366
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,1024,512,0.007941333452860515
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,1024,128,0.0022474666436513263
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,1024,128,0.014452266693115234
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,1024,256,0.007502933343251546
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,1024,64,0.002233600119749705
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,1024,128,0.007101866602897644
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,1024,64,0.014622933665911355
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,1024,32,0.0022250667214393617
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,1024,32,0.014379733800888061
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,768,65536,0.0279968003431956
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,768,65536,0.05614293416341146
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,768,16384,0.015041066209475198
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,768,16384,0.030927999814351397
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,768,16384,0.03641386826833089
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,768,12288,0.012641066312789917
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,768,12288,0.02746986746788025
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,768,12288,0.028905600309371948
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,768,10240,0.011390933394432068
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,768,10240,0.026225066184997557
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,768,10240,0.02566080093383789
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,768,8192,0.009992532928784688
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,768,8192,0.02476373314857483
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,768,8192,0.021528534094492593
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,768,7168,0.009484799702962239
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,768,7168,0.022734934091567995
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,768,7168,0.01945599913597107
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,768,6144,0.008344533046086629
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,768,6144,0.020968532562255858
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,768,6144,0.01819733381271362
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,768,5120,0.007536000013351441
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,768,5120,0.020810665686925252
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,768,5120,0.016578132907549538
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,768,4096,0.007066666583220164
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,768,4096,0.021163733800252278
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,768,4096,0.0144896000623703
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,768,3584,0.006664533416430156
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,768,3584,0.01876693367958069
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,768,65536,0.1279221296310425
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,768,3072,0.005548800031344095
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,768,3072,0.018463999032974243
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,768,3584,0.013223466277122498
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,768,2560,0.005130666494369507
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,768,2560,0.01817493240038554
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,768,3072,0.012389333049456278
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,768,2048,0.004632533093293508
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,768,2048,0.016577066977818807
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,768,2560,0.011222400267918905
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,768,1536,0.0038912000755469004
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,768,1536,0.016590933005015053
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,768,2048,0.010357333223025005
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,768,1024,0.0033728001018365227
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,768,1024,0.015897599856058757
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,768,1536,0.009244799613952637
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,768,768,0.0031146667897701263
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,768,1024,0.008738133311271667
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,768,768,0.014615466197331747
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,768,512,0.003337600082159042
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,768,512,0.014846932888031007
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,768,768,0.007964799801508587
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,768,256,0.002571733295917511
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,768,512,0.007874133189519246
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,768,256,0.014843733112017313
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,768,128,0.0022069332500298817
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,768,128,0.013358933726946512
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,768,256,0.007542400062084198
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,768,64,0.0022005334496498106
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,768,64,0.014549332857131957
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,768,128,0.007474133372306823
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,768,32,0.00222080002228419
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,768,32,0.014425599575042724
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,512,65536,0.02268480062484741
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,512,65536,0.05355519851048788
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,512,16384,0.011924266815185547
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,512,16384,0.02805226643880208
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,512,16384,0.03628586530685425
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,512,12288,0.009973333279291789
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,512,12288,0.025245867172876996
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,512,12288,0.028889600435892743
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,512,10240,0.009156266848246258
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,512,10240,0.024949334065119424
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,512,10240,0.02560960054397583
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,512,8192,0.00793280005455017
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,512,8192,0.023893332481384276
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,512,8192,0.02150826652844747
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,512,7168,0.007502933343251546
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,512,7168,0.02260800004005432
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,512,7168,0.01989013353983561
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,512,6144,0.007075199981530507
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,512,6144,0.020662399133046468
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,512,65536,0.12479039827982585
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,512,5120,0.007246933380762736
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,512,5120,0.02086720069249471
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,512,6144,0.01779200037320455
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,512,4096,0.0067456002036730455
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,512,4096,0.019745065768559774
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,512,5120,0.016084266702334087
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,512,3584,0.006306133170922597
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,512,3584,0.01920426686604818
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,512,4096,0.014439466595649719
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,512,3072,0.005523199836413065
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,512,3072,0.0169813334941864
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,512,3584,0.013267200191815695
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,512,2560,0.005085866649945577
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,512,2560,0.017191465695699057
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,512,3072,0.012057600418726604
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,512,2048,0.004640000065167745
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,512,2048,0.01651306649049123
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,512,2560,0.011201066772143047
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,512,1536,0.003869866579771042
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,512,1536,0.016666666666666666
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,512,2048,0.01040000021457672
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,512,1024,0.003419733295838038
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,512,1024,0.0144896000623703
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,512,1536,0.009576533238093059
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,512,768,0.0030346666773160298
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,512,768,0.016620799899101257
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,512,1024,0.008311466872692108
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,512,512,0.0026176000634829206
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,512,512,0.014450132846832275
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,512,768,0.008356266220410665
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,512,256,0.0025759999950726825
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,512,512,0.007495466868082683
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,512,256,0.007502933343251546
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,512,128,0.0022079999248186748
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,512,128,0.013416533668835958
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,512,128,0.006932266553243001
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,512,64,0.002209066599607468
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,512,64,0.013288533687591553
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,512,32,0.002195200075705846
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,512,32,0.014450132846832275
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,256,65536,0.019806933403015137
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,256,65536,0.0453439990679423
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,256,65536,0.1245685338973999
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,256,16384,0.008801066875457763
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,256,16384,0.02621866663297017
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,256,16384,0.03622293472290039
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,256,12288,0.007153066496054332
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,256,12288,0.024894932905832924
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,256,12288,0.028776532411575316
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,256,10240,0.006697600086530049
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,256,10240,0.02471253275871277
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,256,10240,0.02512960036595662
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,256,8192,0.006247466802597046
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,256,8192,0.022667733828226726
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,256,8192,0.021474132935206093
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,256,7168,0.005868799984455109
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,256,7168,0.022525866826375328
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,256,7168,0.019831466674804687
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,256,6144,0.006296533346176148
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,256,6144,0.021412267287572225
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,256,6144,0.01812373399734497
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,256,5120,0.006671999891599019
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,256,5120,0.020660267273585
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,256,5120,0.01607253352801005
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,256,4096,0.006234666705131531
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,256,4096,0.018786134322484334
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,256,4096,0.01411946713924408
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,256,3584,0.0058559998869895935
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,256,3584,0.0186901330947876
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,256,3584,0.013204266627629599
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,256,3072,0.0055167997876803074
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,256,3072,0.01845973332722982
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,256,3072,0.012005333105723064
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,256,2560,0.0050570666790008545
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,256,2560,0.01689599951108297
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,256,2560,0.011246933539708456
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,256,2048,0.0042805333932240805
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,256,2048,0.01698453426361084
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,256,2048,0.01034879982471466
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,256,1536,0.003832533210515976
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,256,1536,0.015432533621788026
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,256,1536,0.009237333138783773
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,256,1024,0.0033887999753157297
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,256,1024,0.0164490669965744
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,256,1024,0.008319999774297078
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,256,768,0.0029674666623274487
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,256,768,0.014569600423177084
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,256,768,0.008345599969228108
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,256,512,0.002548266698916753
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,256,512,0.014964266618092855
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,256,512,0.007621333499749501
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,256,256,0.002625066787004471
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,256,256,0.013844266533851624
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,256,256,0.007512533167997997
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,256,128,0.002205866575241089
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,256,128,0.014055466651916504
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,256,128,0.0071381335457166035
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,256,64,0.002178133279085159
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,256,64,0.013801599542299906
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,256,32,0.0022133332987626392
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,256,32,0.013548800349235534
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,128,65536,0.013898666699727377
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,128,65536,0.04026453495025635
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,128,65536,0.12403733730316162
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,128,16384,0.006321066617965698
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,128,16384,0.025305600961049397
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,128,16384,0.03631466627120972
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,128,12288,0.0056533331672350565
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,128,12288,0.025384533405303954
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,128,12288,0.0287882665793101
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,128,10240,0.005901866654555003
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,128,10240,0.02304640014966329
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,128,10240,0.025509333610534667
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,128,8192,0.005871999760468801
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,128,8192,0.022884267568588256
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,128,8192,0.021463465690612794
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,128,7168,0.005542399982611338
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,128,7168,0.022932267189025878
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,128,7168,0.019826134045918785
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,128,6144,0.006258133550484974
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,128,6144,0.021014400323232017
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,128,6144,0.01807039976119995
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,128,5120,0.0058442667126655575
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,128,5120,0.020988800128300986
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,128,5120,0.016170666615168253
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,128,4096,0.005464533468087515
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,128,4096,0.02031146685282389
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,128,4096,0.01439466675122579
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,128,3584,0.005448533097902933
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,128,3584,0.018551466862360637
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,128,3584,0.013225600123405457
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,128,3072,0.005087999999523163
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,128,3072,0.017668267091115318
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,128,3072,0.012363732854525248
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,128,2560,0.005418666700522105
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,128,2560,0.018236800034840902
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,128,2560,0.010845866799354554
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,128,2048,0.005037866532802582
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,128,2048,0.0185205340385437
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,128,2048,0.010388267040252686
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,128,1536,0.00382080003619194
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,128,1536,0.01662613352139791
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,128,1536,0.009544533491134644
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,128,1024,0.003068800022204717
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,128,1024,0.014542933305104574
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,128,1024,0.00840106705824534
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,128,768,0.002958933264017105
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,128,768,0.014663466811180114
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,128,768,0.00786240001519521
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,128,512,0.00262719988822937
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,128,512,0.014477866888046264
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,128,512,0.007533866663773854
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,128,256,0.0021674667795499166
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,128,256,0.014885333180427552
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,128,256,0.007076266904671986
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,128,128,0.002148266633351644
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,128,128,0.014406399925549826
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,128,128,0.006666666766007741
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,128,64,0.002162133405605952
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,128,64,0.014046933253606161
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,128,32,0.002136533210674922
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,128,32,0.014453333616256715
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,64,65536,0.01244160036245982
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,64,65536,0.03890986839930217
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,64,16384,0.006712533533573151
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,64,16384,0.02563626567522685
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,64,12288,0.005492266515890757
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,64,12288,0.023119999965031942
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,64,10240,0.005860266586144766
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,64,10240,0.022712532679239908
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,64,8192,0.005843199789524078
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,64,8192,0.02224640051523844
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,64,7168,0.005475200215975444
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,64,7168,0.021651200453440347
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,64,6144,0.00547733356555303
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,64,6144,0.020677334070205687
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,64,5120,0.005095466474692027
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,64,5120,0.02063466707865397
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,64,4096,0.004698666433493296
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,64,4096,0.020297600825627645
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,64,3584,0.005389866729577383
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,64,3584,0.018617600202560425
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,64,3072,0.0050346667567888895
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,64,3072,0.01781546672185262
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,64,2560,0.005188266436258951
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,64,2560,0.01739306648572286
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,64,2048,0.005021866659323374
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,64,2048,0.016615466276804606
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,64,1536,0.0037930667400360107
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,64,1536,0.016062933206558227
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,64,1024,0.0032042667269706728
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,64,1024,0.015517866611480713
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,64,768,0.0030293333033720653
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,64,768,0.014656000336011252
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,64,512,0.0026975999275843303
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,64,512,0.01486186683177948
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,64,256,0.0022912000616391497
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,64,256,0.015356799960136414
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,64,128,0.0022304000953833262
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,64,128,0.013871999581654868
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,64,64,0.0022122666239738466
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,64,64,0.013809067010879517
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,64,32,0.0022250667214393617
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,64,32,0.014498133460680643
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,32,65536,0.01301653285821279
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,32,65536,0.04113599856694539
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,32,16384,0.0067104001839955645
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,32,16384,0.02493120034535726
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,32,12288,0.005490133166313171
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,32,12288,0.023781333367029825
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,32,10240,0.005860266586144766
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,32,10240,0.022669865687688192
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,32,8192,0.005464533468087515
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,32,8192,0.02249280015627543
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,32,7168,0.005073066552480062
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,32,7168,0.02239146629969279
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,32,6144,0.005425066749254862
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,32,6144,0.02157333294550578
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,32,5120,0.005074133475621542
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,32,5120,0.020853332678476968
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,32,4096,0.004701866706212362
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,32,4096,0.020691200097401937
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,32,3584,0.005269333223501841
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,32,3584,0.01855573256810506
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,32,3072,0.0050335998336474095
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,32,3072,0.017698132991790773
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,32,2560,0.005097599824269613
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,32,2560,0.016874667008717856
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,32,2048,0.004680533210436503
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,32,2048,0.01657600005467733
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,32,1536,0.003807999938726425
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,32,1536,0.01564479966958364
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,32,1024,0.002979200085004171
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,32,1024,0.014516266187032065
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,32,768,0.003050666550795237
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,32,768,0.014490666985511779
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,32,512,0.002549333373705546
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,32,512,0.015441067020098367
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,32,256,0.0021589333812395734
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,32,256,0.01395840048789978
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,32,128,0.0021514666577180227
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,32,128,0.013662933309872945
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,32,64,0.002161066730817159
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,32,64,0.013990400234858194
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,192,32,32,0.0021664001047611236
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,32,32,0.014470400412877402
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,65536,16384,0.35987841288248695
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,65536,16384,0.20906666119893394
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,1536,6144,0.018143999576568603
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,65536,12288,0.2623765309651693
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,65536,12288,0.16056639353434246
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,192,1024,5120,0.016457600394884746
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,65536,10240,0.2373802661895752
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,65536,10240,0.14124800364176432
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,192,512,256,0.014473600188891092
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,65536,8192,0.19253120422363282
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,65536,8192,0.11589120229085285
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,65536,10240,0.10170026620229085
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,65536,7168,0.1651807943979899
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,65536,7168,0.1029098669687907
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,65536,8192,0.08852480252583822
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,65536,6144,0.14444586435953777
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,65536,6144,0.09098666508992513
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,65536,12288,0.11668159961700439
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,65536,5120,0.1207808017730713
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,65536,5120,0.07899626890818277
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,65536,16384,0.1739722728729248
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,65536,4096,0.0977450688680013
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,65536,4096,0.06677866776784261
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,65536,6144,0.06732479731241861
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,65536,3584,0.08593173027038574
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,65536,3584,0.06193493207295736
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,65536,7168,0.0776149352391561
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,65536,3072,0.07401386896769205
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,65536,3072,0.05412266651789347
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,65536,5120,0.05706773201624552
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,65536,2560,0.06292159954706827
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,65536,2560,0.04770666758219401
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,65536,4096,0.04754240115483602
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,65536,2048,0.05197439988454183
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,65536,2048,0.0424074649810791
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,65536,2560,0.03360426823298136
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,65536,1536,0.0394976019859314
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,65536,1536,0.03531519969304402
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,65536,3072,0.03797866503397624
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,65536,1024,0.02765866716702779
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,65536,1024,0.029073067506154376
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,65536,3584,0.04261866807937622
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,65536,768,0.021769599119822184
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,65536,768,0.026667733987172444
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,65536,768,0.017869865894317626
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,65536,512,0.015357866883277893
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,65536,512,0.02445440093676249
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,65536,1024,0.019656533002853395
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,65536,256,0.009667199850082398
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,65536,256,0.020678400993347168
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,65536,256,0.012910933295885722
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,65536,128,0.007550933460394542
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,65536,128,0.0187285323937734
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,65536,128,0.011707733074824016
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,65536,64,0.006742399930953979
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,65536,64,0.02107306718826294
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,65536,32,0.007115733126799266
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,65536,32,0.023018666108449302
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,65536,512,0.015528532862663268
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,16384,65536,0.371290651957194
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,16384,65536,0.23589866956075034
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,65536,1536,0.024218666553497314
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,16384,16384,0.10294613043467205
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,16384,16384,0.07225066820780436
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,16384,12288,0.07395946979522705
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,16384,12288,0.05833813349405924
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,65536,2048,0.02901653250058492
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,16384,10240,0.0632213314374288
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,16384,10240,0.05197120110193888
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,16384,10240,0.032739200194676715
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,16384,8192,0.0519594669342041
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,16384,8192,0.0452394684155782
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,16384,12288,0.037240533034006755
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,16384,7168,0.044804267088572186
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,16384,7168,0.04202666680018107
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,16384,16384,0.048434134324391684
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,16384,6144,0.04081386725107829
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,16384,6144,0.038737066586812335
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,16384,7168,0.02407146692276001
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,16384,5120,0.03366719881693522
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,16384,5120,0.03463360071182251
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,16384,8192,0.02683200041453044
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,16384,4096,0.027294933795928955
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,16384,4096,0.03058133323987325
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,16384,6144,0.02181653380393982
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,16384,3584,0.024855466683705647
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,16384,3584,0.028732800483703615
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,16384,4096,0.016953599452972413
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,16384,3072,0.021619200706481934
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,16384,3072,0.026526933908462523
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,16384,5120,0.019463467597961425
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,16384,2560,0.01837973395983378
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,16384,2560,0.024754132827123007
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,16384,2560,0.013684266805648803
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,16384,2048,0.015281066298484802
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,16384,2048,0.024075732628504435
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,16384,3584,0.015802666544914246
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,16384,1536,0.01229759951432546
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,16384,1536,0.021053866545359293
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,16384,1536,0.010919466614723206
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,16384,1024,0.009513599673906963
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,16384,1024,0.0195413331190745
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,16384,2048,0.012239999572436015
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,16384,768,0.00795306662718455
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,16384,768,0.01657386620839437
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,16384,1024,0.009688533345858256
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,16384,512,0.005985066791375478
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,16384,512,0.016676266988118492
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,16384,768,0.008809600273768108
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,16384,256,0.004242133100827535
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,16384,256,0.014539733529090881
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,16384,512,0.008361599842707316
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,16384,128,0.0037791999677817024
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,16384,128,0.01532586713631948
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,16384,128,0.0071839998165766404
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,16384,64,0.003432533393303553
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,16384,64,0.014582399527231851
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,16384,32,0.0034485332667827605
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,16384,32,0.01456000010172526
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,16384,256,0.007901866734027863
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,12288,65536,0.26808853149414064
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,12288,65536,0.18794560432434082
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,12288,16384,0.07517759799957276
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,12288,16384,0.06204479932785034
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,12288,16384,0.041155199209849044
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,12288,12288,0.05506346623102824
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,12288,12288,0.04957760175069173
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,12288,12288,0.0330570658047994
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,12288,10240,0.04739733139673869
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,12288,10240,0.0444757342338562
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,16384,3072,0.01434879998366038
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,12288,8192,0.03851626714070638
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,12288,8192,0.039154132207234696
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,12288,10240,0.028598399957021077
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,12288,7168,0.03333866596221924
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,12288,7168,0.035010135173797606
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,12288,8192,0.024652800957361855
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,12288,6144,0.029632000128428142
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,12288,6144,0.03272213339805603
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,12288,7168,0.022753065824508666
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,12288,5120,0.025550933678944905
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,12288,5120,0.029687466224034627
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,12288,6144,0.020900267362594604
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,12288,4096,0.02086720069249471
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,12288,4096,0.02682346701622009
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,12288,5120,0.018617600202560425
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,12288,3584,0.0185098667939504
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,12288,3584,0.0252895991007487
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,12288,4096,0.0165802667538325
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,12288,3072,0.016314666469891867
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,12288,3072,0.024625066916147867
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,12288,3584,0.01539306640625
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,12288,2560,0.014084266622861228
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,12288,2560,0.022805333137512207
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,12288,2560,0.012493866682052612
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,12288,2048,0.011990400155385335
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,12288,2048,0.02140373388926188
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,12288,3072,0.01416853368282318
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,12288,1536,0.009956266482671101
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,12288,1536,0.019715199867884316
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,12288,1536,0.010505599776903789
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,12288,1024,0.00793280005455017
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,12288,1024,0.018630399306615194
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,12288,2048,0.010938666264216105
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,12288,768,0.006345599889755249
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,12288,768,0.01664746701717377
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,12288,1024,0.008825600147247314
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,12288,512,0.005166933437188466
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,12288,512,0.016548267006874083
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,12288,768,0.008412800232569377
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,12288,256,0.0038634667793909705
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,12288,256,0.014521599809328715
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,12288,256,0.007612800101439159
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,12288,128,0.0034282666941483817
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,12288,128,0.015060266852378846
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,12288,512,0.008063999811808269
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,12288,64,0.003398400048414866
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,12288,64,0.014786133170127868
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,12288,32,0.003390933324893316
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,12288,128,0.007133866846561432
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,12288,32,0.01499626636505127
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,10240,65536,0.2397600015004476
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,10240,65536,0.17759146690368652
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,10240,16384,0.06649706761042276
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,10240,16384,0.06456746657689413
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,16384,65536,0.16122026443481446
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,10240,12288,0.0495904008547465
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,10240,12288,0.05119040012359619
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,10240,16384,0.04021120071411133
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,10240,10240,0.04269546667734782
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,10240,10240,0.044854398568471274
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,12288,65536,0.13933439254760743
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,10240,8192,0.03551146586736043
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,10240,8192,0.04170346657435099
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,10240,12288,0.032263465722401935
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,10240,7168,0.030611199140548707
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,10240,7168,0.03409066597620646
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,10240,10240,0.028626133998235066
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,10240,6144,0.02717120051383972
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,10240,6144,0.031125332911809283
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,10240,8192,0.023721599578857423
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,10240,5120,0.023165865739186605
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,10240,5120,0.02881706754366557
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,10240,7168,0.0225055992603302
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,10240,4096,0.01904319922129313
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,10240,4096,0.027048534154891966
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,10240,6144,0.02027413249015808
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,10240,3584,0.01699840029080709
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,10240,3584,0.025363200902938844
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,10240,4096,0.016179200013478598
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,10240,3072,0.01492800017197927
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,10240,3072,0.02386346658070882
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,10240,5120,0.018542933464050292
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,10240,2560,0.013173333803812661
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,10240,2560,0.0230240007241567
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,10240,3584,0.015043200055758158
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,10240,2048,0.011638399958610535
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,10240,2048,0.022102399667104086
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,10240,2048,0.011282133062680562
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,10240,1536,0.009528533617655436
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,10240,1536,0.01885120073954264
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,10240,2560,0.012077866991360982
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,10240,1024,0.00758186678091685
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,10240,1024,0.01715946594874064
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,10240,3072,0.013688533504803976
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,10240,768,0.005952000121275584
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,10240,768,0.016561067104339598
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,10240,1024,0.008803199728329976
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,10240,512,0.004897066454092661
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,10240,512,0.016574933131535848
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,10240,768,0.008406399687131246
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,10240,256,0.0039018665750821433
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,10240,256,0.016673066218694053
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,10240,1536,0.010041600465774536
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,10240,128,0.003475199888149897
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,10240,128,0.014456533392270408
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,10240,256,0.007575466732184092
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,10240,64,0.003005866706371307
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,10240,64,0.014684800306955972
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,10240,32,0.003373866776625315
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,10240,128,0.007115733126799266
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,10240,32,0.01456000010172526
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,10240,512,0.007973333199818928
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,8192,65536,0.17639253934224447
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,8192,65536,0.14573225975036622
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,8192,16384,0.0494485338528951
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,8192,16384,0.04928640127182007
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,8192,12288,0.03721173206965129
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,8192,12288,0.04022186597188314
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,8192,12288,0.03127146760622661
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,8192,10240,0.03216533263524373
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,8192,10240,0.03552853266398112
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,8192,16384,0.039955198764801025
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,8192,8192,0.026405332485834758
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,8192,8192,0.031228800614674885
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,8192,10240,0.027340799570083618
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,8192,7168,0.023155200481414794
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,8192,7168,0.028990934292475384
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,8192,8192,0.023715200026830037
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,8192,6144,0.020372267564137778
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,8192,6144,0.028707200288772584
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,8192,7168,0.021947733561197915
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,8192,5120,0.01734613378842672
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,8192,5120,0.026308266321818034
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,8192,6144,0.020204800367355346
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,8192,4096,0.014170666535695394
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,8192,4096,0.0250271995862325
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,8192,5120,0.018237866957982383
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,8192,3584,0.013004799683888754
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,8192,3584,0.022771199544270836
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,8192,4096,0.015846400459607442
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,8192,3072,0.01129919985930125
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,8192,3072,0.02304746707280477
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,8192,3584,0.014544000228246054
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,8192,2560,0.010194133718808491
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,8192,2560,0.020917334159215293
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,8192,3072,0.01297599971294403
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,8192,2048,0.008789333701133727
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,8192,2048,0.019318399826685588
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,8192,2560,0.01202880044778188
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,8192,1536,0.007562666634718577
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,8192,1536,0.01874986688296
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,8192,1536,0.00973653296629588
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,8192,1024,0.005957333246866862
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,8192,1024,0.01658986707528432
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,8192,2048,0.010967466235160827
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,8192,768,0.00470719983180364
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,8192,768,0.01563093364238739
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,8192,768,0.008333866794904072
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,8192,512,0.003774933268626531
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,8192,512,0.016739199558893837
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,8192,1024,0.009111467003822326
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,8192,256,0.0030378667016824085
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,8192,256,0.014617600043614707
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,8192,256,0.007528533538182576
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,8192,128,0.002997333308060964
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,8192,128,0.01586560010910034
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,8192,512,0.008345599969228108
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,8192,64,0.002648533384005229
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,8192,64,0.01477013329664866
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,8192,32,0.002776533365249634
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,8192,32,0.014648532867431641
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,8192,128,0.007152000069618225
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,7168,65536,0.16513919830322266
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,7168,65536,0.12745280265808107
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,7168,16384,0.04629546801249186
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,7168,16384,0.05548906723658244
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,10240,65536,0.13209280172983806
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,7168,12288,0.03419093290964763
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,7168,12288,0.03840106725692749
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,7168,16384,0.03885973294576009
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,7168,10240,0.029950932661692305
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,7168,10240,0.03476159969965617
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,7168,12288,0.031035733222961426
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,7168,8192,0.024770132700602212
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,7168,8192,0.03025599916776021
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,7168,10240,0.02725226680437724
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,7168,7168,0.021594667434692384
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,7168,7168,0.02894933422406514
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,7168,8192,0.02318933407465617
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,7168,6144,0.01932586630185445
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,7168,6144,0.02699306607246399
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,7168,7168,0.021527467171351115
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,7168,5120,0.016422399878501893
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,7168,5120,0.02508693337440491
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,7168,6144,0.019636267423629762
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,7168,4096,0.013507200280825296
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,7168,4096,0.0228383998076121
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,7168,5120,0.017863466342290243
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,7168,3584,0.01220906674861908
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,7168,3584,0.022850133975346885
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,7168,4096,0.015356799960136414
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,7168,3072,0.010853333274523418
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,7168,3072,0.021677867571512858
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,7168,3584,0.01420799990495046
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,7168,2560,0.009750399986902874
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,7168,2560,0.02049386699994405
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,8192,65536,0.13079360326131184
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,7168,2048,0.00853653351465861
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,7168,2048,0.0182751993338267
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,7168,3072,0.012921599547068277
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,7168,1536,0.0070826664566993715
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,7168,1536,0.01810773412386576
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,7168,2560,0.011662933230400085
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,7168,1024,0.005211733281612396
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,7168,1024,0.017694934209187826
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,7168,1024,0.008900266885757447
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,7168,768,0.005482666691144307
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,7168,768,0.017066667477289833
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,7168,2048,0.010470400253931682
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,7168,512,0.0042581334710121155
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,7168,512,0.01604480048020681
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,7168,1536,0.009549867113431294
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,7168,256,0.003402666747570038
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,7168,256,0.01452906628449758
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,7168,256,0.0075765331586201985
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,7168,128,0.003018666555484136
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,7168,128,0.014506666858990987
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,7168,768,0.008371200164159138
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,7168,64,0.002661333233118057
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,7168,64,0.014550399780273438
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,7168,32,0.00296426663796107
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,7168,32,0.014569600423177084
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,7168,512,0.007960533102353413
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,7168,128,0.0071733335653940845
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,6144,65536,0.11478400230407715
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,6144,65536,0.14858880043029785
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,6144,16384,0.04839466810226441
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,6144,16384,0.04760746558507283
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,6144,12288,0.03623893260955811
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,6144,12288,0.03922346830368042
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,6144,12288,0.030315732955932616
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,6144,10240,0.03150933384895325
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,6144,10240,0.035104000568389894
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,6144,16384,0.038337067763010664
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,6144,8192,0.025854933261871337
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,6144,8192,0.03108479976654053
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,6144,10240,0.027164800961812334
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,6144,7168,0.02241920034090678
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,6144,7168,0.028832000494003297
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,6144,8192,0.022985599438349404
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,6144,6144,0.020028799772262573
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,6144,6144,0.027939200401306152
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,6144,7168,0.0215445339679718
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,6144,5120,0.0169322669506073
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,6144,5120,0.025999999046325682
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,6144,6144,0.01950613260269165
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,6144,4096,0.013776000340779623
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,6144,4096,0.023093332846959434
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,6144,5120,0.017612799008687337
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,6144,3584,0.012435199817021687
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,6144,3584,0.022695465882619222
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,6144,4096,0.014937600493431092
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,6144,3072,0.011203199625015259
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,6144,3072,0.021282132466634116
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,6144,3584,0.01418880025545756
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,6144,2560,0.010016000270843506
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,6144,2560,0.02065920035044352
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,6144,3072,0.012479999661445617
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,6144,2048,0.008371200164159138
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,6144,2048,0.0186463991800944
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,6144,2560,0.011698133746782939
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,7168,65536,0.13046720027923583
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,6144,1536,0.007064533233642578
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,6144,1536,0.018012799819310508
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,6144,1024,0.005083733300367991
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,6144,1024,0.01795413295427958
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,6144,1024,0.008806399504343669
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,6144,768,0.004594133297602335
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,6144,768,0.01704533298810323
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,6144,1536,0.009685333569844563
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,6144,512,0.003869866579771042
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,6144,512,0.015434666474660238
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,6144,768,0.008319999774297078
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,6144,256,0.0030720000465710956
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,6144,256,0.015191466609636942
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,6144,2048,0.010780800382296245
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,6144,512,0.007945600152015685
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,6144,128,0.0025962665677070618
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,6144,128,0.01474666694800059
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,6144,64,0.0030293333033720653
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,6144,64,0.014607999722162882
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,6144,32,0.002921599894762039
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,6144,32,0.014602667093276978
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,6144,128,0.0071050668756167095
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,6144,256,0.007501866420110066
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,5120,65536,0.15808320045471191
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,5120,65536,0.10903039773305256
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,5120,16384,0.04753066698710124
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,5120,16384,0.0522325317064921
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,5120,12288,0.03447573184967041
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,5120,12288,0.037385598818461103
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,5120,12288,0.030089600880940752
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,5120,10240,0.02948586742083232
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,5120,10240,0.03395413160324097
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,5120,16384,0.03784960110982259
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,5120,8192,0.02448106606801351
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,5120,8192,0.030595199267069502
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,5120,10240,0.026850134134292603
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,5120,7168,0.021091200908025107
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,5120,7168,0.028155734141667683
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,5120,8192,0.022796799739201866
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,5120,6144,0.018776534001032512
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,5120,6144,0.02722559968630473
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,5120,7168,0.021336533625920615
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,5120,5120,0.015675733486811318
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,5120,5120,0.024810665845870973
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,5120,6144,0.0190720001856486
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,5120,4096,0.013362133502960205
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,5120,4096,0.023468800385793052
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,5120,5120,0.017428267002105712
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,5120,3584,0.011934933066368104
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,5120,3584,0.021860265731811525
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,5120,4096,0.014599466323852539
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,5120,3072,0.010995200276374817
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,5120,3072,0.020660267273585
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,5120,3584,0.013722667098045349
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,5120,2560,0.009820800026257832
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,5120,2560,0.018773333231608073
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,5120,3072,0.012453333536783854
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,5120,2048,0.00820906658967336
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,5120,2048,0.018574933211008705
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,5120,2560,0.011666133006413778
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,5120,1536,0.006635733445485433
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,5120,1536,0.01767359972000122
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,5120,1536,0.009697066744168599
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,5120,1024,0.0047978664437929785
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,5120,1024,0.01695893406867981
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,5120,2048,0.01050986647605896
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,5120,768,0.004351999859015147
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,5120,768,0.01758613387743632
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,5120,1024,0.008467200398445129
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,5120,512,0.00384853333234787
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,5120,512,0.016940800348917644
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,5120,768,0.007965866724650066
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,5120,256,0.003398400048414866
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,5120,256,0.014613333344459533
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,5120,256,0.007502933343251546
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,5120,128,0.0029557332396507262
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,5120,128,0.015914666652679443
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,5120,512,0.007876266539096833
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,5120,64,0.002573866645495097
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,5120,64,0.014470400412877402
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,5120,32,0.0025557334224383037
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,5120,32,0.014473600188891092
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,5120,128,0.007209600011507671
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,4096,65536,0.0990773359934489
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,4096,65536,0.09648959636688233
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,4096,16384,0.03183466593424479
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,4096,16384,0.04090346495310466
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,6144,65536,0.12921173572540284
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,4096,12288,0.02595626711845398
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,4096,12288,0.030188800891240437
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,4096,12288,0.03025066653887431
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,4096,10240,0.019402666886647543
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,4096,10240,0.02861119906107585
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,4096,16384,0.03755946556727092
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,4096,8192,0.015962666273117064
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,4096,8192,0.02672959963480632
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,4096,8192,0.022733867168426514
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,4096,7168,0.014176000157992044
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,4096,7168,0.024903466304143272
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,4096,10240,0.025993599494298296
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,4096,6144,0.012549333771069846
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,4096,6144,0.02435093323389689
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,4096,7168,0.020717867215474448
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,4096,5120,0.011169067025184632
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,4096,5120,0.022565333048502605
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,4096,6144,0.018626133600870766
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,4096,4096,0.009405866265296936
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,4096,4096,0.020012799898783365
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,4096,5120,0.017063466707865398
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,4096,3584,0.008377599716186523
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,4096,3584,0.021243733167648316
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,4096,4096,0.014481066664059957
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,4096,3072,0.007842133442560833
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,4096,3072,0.018463999032974243
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,4096,3072,0.012477866808573405
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,4096,2560,0.0067456002036730455
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,4096,2560,0.01816640098889669
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,4096,3584,0.013324800133705138
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,4096,2048,0.005949866771697998
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,4096,2048,0.017426133155822754
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,4096,2048,0.010376532872517902
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,4096,1536,0.005184000233809153
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,4096,2560,0.011643733580907185
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,4096,1536,0.016681599617004394
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,4096,1024,0.0038592000802357995
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,4096,1024,0.017935999234517417
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,4096,1024,0.008841600020726521
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,4096,768,0.003521066655715307
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,4096,768,0.0155157337586085
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,4096,1536,0.00957973301410675
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,4096,512,0.003065599997838338
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,4096,512,0.015478400389353433
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,4096,768,0.008403199911117553
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,4096,256,0.0029696000119050344
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,4096,256,0.014691199858983359
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,4096,512,0.007551999886830647
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,4096,128,0.002603733291228612
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,4096,128,0.014523733655611673
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,4096,256,0.00761706680059433
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,4096,128,0.007091199855009715
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,4096,64,0.0023573334018389385
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,5120,65536,0.13011840184529622
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,4096,32,0.00264533335963885
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,4096,64,0.014481066664059957
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,4096,32,0.014579199751218162
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,3584,65536,0.09089919726053873
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,3584,65536,0.08619306882222494
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,3584,16384,0.028797866900761922
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,3584,16384,0.03812586863835653
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,3584,12288,0.023331199089686075
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,3584,12288,0.02937493324279785
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,3584,12288,0.030023467540740967
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,3584,10240,0.023252266645431518
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,3584,10240,0.027670399347941084
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,3584,16384,0.037934935092926024
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,3584,8192,0.016114133596420287
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,3584,8192,0.02757866581281026
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,3584,10240,0.026054400205612182
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,3584,7168,0.014140799641609192
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,3584,7168,0.02442880074183146
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,3584,8192,0.021934932470321654
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,3584,6144,0.012548266847928365
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,3584,6144,0.02477653423945109
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,3584,7168,0.020472532510757445
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,3584,5120,0.01095466713110606
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,3584,5120,0.021477333704630532
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,3584,6144,0.018258132537206016
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,3584,4096,0.010880000392595927
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,3584,4096,0.019789866606394448
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,3584,5120,0.01662826637427012
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,3584,3584,0.00804373323917389
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,3584,3584,0.01861013372739156
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,3584,4096,0.014651733636856078
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,3584,3072,0.007562666634718577
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,3584,3072,0.01867093245188395
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,4096,65536,0.12845760186513264
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,3584,2560,0.006739200154940288
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,3584,2560,0.01880000034968058
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,3584,3584,0.013276799519856771
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,3584,2048,0.0054848000407218935
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,3584,2048,0.017240534226099648
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,3584,3072,0.012820266683896384
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,3584,1536,0.0046079998215039575
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,3584,1536,0.01728640000025431
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,3584,2560,0.011686399579048157
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,3584,1024,0.003668266783157984
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,3584,1024,0.015412267049153647
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,3584,1536,0.009614933530489604
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,3584,768,0.0034602666894594826
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,3584,768,0.015568000078201295
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,3584,2048,0.010371200243631999
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,3584,512,0.003044266750415166
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,3584,512,0.014593066771825156
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,3584,512,0.007561600208282471
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,3584,256,0.0030016000072161358
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,3584,256,0.01448319951693217
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,3584,768,0.008385066191355388
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,3584,128,0.002625066787004471
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,3584,1024,0.008763733506202697
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,3584,128,0.014492799838383993
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,3584,64,0.0025429333249727885
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,3584,64,0.014638933539390563
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,3584,32,0.002569599946339925
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,3584,128,0.0070783997575442
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,3584,32,0.014782933394114175
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,3584,256,0.0074986666440963745
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,3072,65536,0.07308373451232911
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,3072,16384,0.024971733490626015
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,3072,65536,0.08087466557820638
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,3072,16384,0.03537813425064087
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,3072,12288,0.019363200664520262
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,3072,12288,0.03102506597836812
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,3072,12288,0.029719465970993043
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,3072,10240,0.020784000555674233
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,3072,10240,0.028823467095692952
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,3072,16384,0.03760639826456706
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,3072,8192,0.016942934195200602
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,3072,8192,0.02691733241081238
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,3072,10240,0.026339199145634967
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,3072,7168,0.015152000387509666
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,3072,7168,0.026337067286173504
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,3072,8192,0.021949867407480873
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,3072,6144,0.013538133104642233
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,3072,6144,0.024770132700602212
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,3072,7168,0.019986132780710854
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,3072,5120,0.011461333433787028
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,3072,5120,0.022871466477711995
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,3072,6144,0.01858453353246053
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,3072,4096,0.009608532985051472
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,3072,4096,0.02087999979654948
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,3072,5120,0.016583466529846193
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,3072,3584,0.008472533027331034
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,3072,3584,0.018902399142583213
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,3072,4096,0.014545067151387533
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,3072,3072,0.007718400160471599
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,3072,3072,0.01958613395690918
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,3072,3584,0.013350400328636169
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,3072,2560,0.006836266815662384
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,3072,2560,0.0187882661819458
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,3072,3072,0.01250879963239034
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,3072,2560,0.011681066950162251
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,3072,2048,0.005845333139101664
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,3072,2048,0.016621866822242738
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,3584,65536,0.12806613445281984
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,3072,1536,0.005072000126043955
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,3072,1536,0.017195733388264973
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,3072,1024,0.003966933240493139
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,3072,1024,0.016531200210253397
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,3072,1024,0.008829866846402485
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,3072,768,0.003489066660404205
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,3072,768,0.015838932991027833
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,3072,1536,0.00962453285853068
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,3072,512,0.003102933367093404
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,3072,2048,0.010410666465759277
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,3072,512,0.015333333611488342
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,3072,256,0.0025898667673269907
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,3072,256,0.014551466703414917
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,3072,256,0.0072053333123524976
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,3072,128,0.0025898667673269907
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,3072,512,0.007893333335717519
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,3072,768,0.008307200173536937
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,3072,64,0.002268799891074498
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,3072,128,0.0070826664566993715
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,3072,32,0.0025642665723959604
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,3072,32,0.014523733655611673
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,2560,65536,0.06532053152720133
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,2560,65536,0.07596266269683838
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,3072,65536,0.12903892993927002
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,2560,16384,0.022373332579930624
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,2560,16384,0.033907198905944826
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,2560,16384,0.03744213183720906
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,2560,12288,0.01753066579500834
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,2560,12288,0.03088639974594116
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,2560,12288,0.029781333605448407
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,2560,10240,0.01615466674168905
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,2560,10240,0.026829866568247478
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,2560,65536,0.12829333146413166
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,2560,8192,0.015554133057594299
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,2560,8192,0.02477653423945109
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,2560,10240,0.025832533836364746
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,2560,7168,0.01416000028451284
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,2560,7168,0.02463360031445821
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,2560,8192,0.021949867407480873
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,2560,6144,0.012074666221936543
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,2560,6144,0.023729066054026283
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,2560,7168,0.01989333430926005
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,2560,5120,0.011011200149854024
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,2560,5120,0.02217280069986979
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,2560,6144,0.018210132916768394
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,2560,4096,0.009358933568000794
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,2560,4096,0.019485867023468016
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,2560,5120,0.01663253307342529
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,2560,3584,0.008913066983222962
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,2560,3584,0.019947733481725058
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,2560,4096,0.0145087997118632
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,2560,3072,0.007573333382606506
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,2560,3072,0.019237333536148073
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,2560,3584,0.01302826702594757
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,2560,2560,0.006673066814740498
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,2560,2560,0.01953279972076416
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,2560,2560,0.011564800143241882
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,2560,2048,0.005714133381843567
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,2560,2048,0.016821332772572837
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,2560,3072,0.012458667159080505
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,2560,1536,0.005382399757703145
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,2560,1536,0.017410133282343546
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,2560,1536,0.009559466441472372
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,2560,1024,0.004228266576925913
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,2560,1024,0.015063466628392539
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,2560,2048,0.01037013332049052
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,2560,768,0.004137599964936575
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,2560,768,0.01592639982700348
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,2560,768,0.007897600034872691
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,2560,512,0.0034005333979924522
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,2560,1024,0.008724266290664673
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,2560,512,0.015169066190719605
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,2560,256,0.0029834667841593427
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,2560,256,0.014819199840227762
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,2560,256,0.007523199915885926
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,2560,128,0.00257493332028389
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,2560,512,0.007546666761239369
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,2560,128,0.01453013320763906
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,2560,64,0.00223786657055219
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,2560,64,0.014455466469128927
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,2560,32,0.00257493332028389
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,2560,128,0.007106133302052816
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,2560,32,0.014410666624704995
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,2048,65536,0.05426880121231079
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,2048,65536,0.06860693295796713
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,2048,16384,0.01743146578470866
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,2048,16384,0.03145280083020528
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,2048,16384,0.03712746699651082
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,2048,12288,0.014531200130780539
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,2048,12288,0.028897066911061604
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,2048,12288,0.029365332921346028
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,2048,10240,0.013140267133712769
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,2048,10240,0.027134933074315387
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,2048,10240,0.025701334079106648
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,2048,8192,0.012470400333404541
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,2048,8192,0.02487466732660929
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,2048,8192,0.02198186715443929
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,2048,7168,0.011636267105738323
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,2048,7168,0.02341653307278951
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,2048,7168,0.01998186707496643
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,2048,6144,0.010831999778747558
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,2048,6144,0.021951999266942343
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,2048,6144,0.018237866957982383
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,2048,5120,0.01037013332049052
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,2048,5120,0.020478934049606323
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,2048,5120,0.01658986707528432
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,2048,4096,0.008666666348775227
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,2048,4096,0.01960853338241577
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,2048,4096,0.014132266243298849
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,2048,3584,0.008027733365694682
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,2048,3584,0.019033600886662803
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,2048,3584,0.013369599978129068
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,2048,3072,0.00633493314186732
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,2048,3072,0.017914666732152303
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,2048,65536,0.12803946336110433
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,2048,2560,0.005706666906674703
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,2048,2560,0.01770240068435669
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,2048,2560,0.011192533373832702
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,2048,2048,0.005021866659323374
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,2048,2048,0.016618667046229045
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,2048,3072,0.01204800009727478
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,2048,1536,0.004276266694068909
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,2048,1536,0.016714666287104288
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,2048,1536,0.009544533491134644
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,2048,1024,0.003409066547950109
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,2048,1024,0.015565866231918335
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,2048,2048,0.01035520037015279
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,2048,768,0.003386666625738144
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,2048,768,0.01453013320763906
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,2048,768,0.008360532919565837
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,2048,1024,0.008707200487454731
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,2048,512,0.0029696000119050344
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,2048,512,0.014740266402562461
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,2048,256,0.0025621332228183747
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,2048,256,0.0144896000623703
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,2048,256,0.007478400071461995
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,2048,128,0.002419200042883555
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,2048,128,0.014536533753077188
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,2048,128,0.006716800232728322
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,2048,64,0.0021877333521842955
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,2048,64,0.014461867014567056
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,2048,32,0.0022463999688625337
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,2048,32,0.014426666498184203
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,1536,65536,0.04184639851252238
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,1536,65536,0.06138773361841837
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,1536,65536,0.12756160100301106
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,1536,16384,0.014533332983652749
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,1536,16384,0.03254613280296326
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,3072,128,0.014494933684666953
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,1536,12288,0.013522133231163025
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,1536,12288,0.028759467601776122
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,1536,12288,0.02906773289044698
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,1536,10240,0.01237226625283559
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,1536,10240,0.027003733317057292
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,1536,16384,0.03629333178202311
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,1536,8192,0.012446932991345723
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,1536,8192,0.024265599250793458
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,1536,10240,0.025587199131647746
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,1536,7168,0.011266133189201355
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,1536,7168,0.023511467377344768
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,1536,8192,0.021502933899561563
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,1536,6144,0.010246400038401287
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,1536,6144,0.021627734104792275
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,1536,7168,0.01985493302345276
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,1536,5120,0.008763733506202697
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,1536,5120,0.02087786595026652
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,1536,6144,0.018196266889572144
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,1536,4096,0.007233066856861115
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,1536,4096,0.018632533152898152
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,1536,5120,0.016525866587956746
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,1536,3584,0.007417599856853485
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,1536,3584,0.019593600432078043
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,1536,4096,0.014052266875902811
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,1536,3072,0.0058890665570894875
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,1536,3072,0.01832746664683024
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,1536,3584,0.013300266861915589
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,1536,2560,0.005520000060399374
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,1536,2560,0.01890986760457357
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,1536,3072,0.012032000223795573
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,1536,2048,0.004640000065167745
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,1536,2048,0.016521599888801575
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,1536,2560,0.011591466267903645
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,1536,1536,0.004253866771856943
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,1536,2048,0.010351999600728353
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,1536,1536,0.01730133295059204
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,1536,1024,0.003443199892838796
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,1536,1024,0.0159850666920344
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,1536,1024,0.008745599786440532
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,1536,768,0.003398400048414866
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,1536,1536,0.009204266468683879
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,1536,768,0.014871467153231302
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,1536,512,0.002963199963172277
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,1536,512,0.014575999975204468
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,1536,512,0.007869866490364075
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,1536,768,0.008297599852085114
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,1536,256,0.002632533262173335
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,1536,256,0.014362667004267374
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,1536,256,0.007515733440717061
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,1536,128,0.0022367998957633973
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,1536,128,0.013767466942469279
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,1536,128,0.007090133428573608
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,1536,64,0.002178133279085159
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,1536,64,0.014670933286348978
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,1536,32,0.002179199953873952
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,1536,32,0.013527466853459676
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,1024,65536,0.03022613326708476
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,1024,65536,0.05420053402582804
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,1024,16384,0.01518933375676473
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,1024,16384,0.0302293340365092
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,1024,16384,0.036266668637593584
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,1024,12288,0.012384000420570373
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,1024,12288,0.02685119907061259
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,1024,12288,0.028920533259709676
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,1024,10240,0.011367467045783997
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,1024,10240,0.025050665934880572
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,1024,10240,0.025621332724889118
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,1024,8192,0.010018133123715718
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,1024,8192,0.027781333525975543
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,1024,8192,0.021502933899561563
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,1024,7168,0.00925546685854594
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,1024,7168,0.026778666178385417
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,1024,7168,0.019897600015004478
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,1024,6144,0.008809600273768108
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,1024,6144,0.024963200092315674
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,1024,6144,0.01816426714261373
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,1024,5120,0.008970666925112407
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,1024,5120,0.023532799879709878
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,1024,5120,0.01658560037612915
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,1024,4096,0.011646933356920878
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,1024,4096,0.021869866053263347
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,1024,4096,0.014173866311709086
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,1024,3584,0.011034666498502096
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,1024,3584,0.019355734189351402
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,1024,65536,0.12559146881103517
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,1024,3072,0.009750399986902874
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,1024,3072,0.017141334215799966
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,1024,3584,0.013271466890970866
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,1024,2560,0.0086218665043513
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,1024,3072,0.012437333663304646
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,1024,2560,0.01732800006866455
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,1024,2048,0.007422933479150136
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,1024,2048,0.016702934106191
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,1024,2560,0.011248000462849935
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,1024,1536,0.004211199780305227
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,1024,1536,0.01692906618118286
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,1024,2048,0.010434133807818095
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,1024,1024,0.0034944000343481696
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,1024,1024,0.016708266735076905
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,1024,1024,0.008701866865158081
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,1024,1536,0.009538132945696514
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,1024,768,0.0034101332227389016
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,1024,768,0.014600533246994018
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,1024,512,0.0029578665892283124
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,1024,512,0.01455573340257009
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,1024,768,0.008154666423797608
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,1024,256,0.0026378666361172995
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,1024,512,0.0075434664885203045
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,1024,256,0.014451199769973755
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,1024,128,0.002566399921973546
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,1024,128,0.014590932925542196
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,1024,256,0.007495466868082683
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,1024,64,0.002164266755183538
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,1024,128,0.007136000196139018
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,1024,64,0.014437333742777506
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,1024,32,0.002526933451493581
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,768,65536,0.026401066780090333
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,768,65536,0.05512533187866211
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,768,65536,0.12572480042775472
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,768,16384,0.012180266777674358
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,768,16384,0.02961919903755188
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,768,16384,0.03620586792627971
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,768,12288,0.012808533509572348
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,768,12288,0.026787199576695758
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,768,12288,0.0288480003674825
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,768,10240,0.011237333218256634
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,768,10240,0.025518933931986492
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,768,10240,0.025161600112915038
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,768,8192,0.00995199978351593
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,768,8192,0.02447893420855204
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,768,8192,0.02104640007019043
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,768,7168,0.009169066945711773
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,768,7168,0.022593067089716593
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,768,7168,0.019825067122777304
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,768,6144,0.008755200107892354
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,768,6144,0.021418666839599608
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,768,6144,0.01784106691678365
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,768,5120,0.007524266839027405
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,768,5120,0.020795732736587524
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,768,5120,0.016130133469899496
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,768,4096,0.006714666883150737
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,768,4096,0.01932586630185445
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,768,4096,0.014134400089581809
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,768,3584,0.006301866471767425
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,768,3584,0.018564265966415406
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,768,3584,0.012905599673589072
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,768,3072,0.00582826683918635
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,768,3072,0.018594133853912353
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,768,3072,0.012070399522781373
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,768,2560,0.005102933446566264
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,768,2560,0.016786134243011473
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,768,2560,0.011206400394439698
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,768,2048,0.004633600016434988
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,768,2048,0.01665066679318746
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,768,2048,0.009970133503278095
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,768,1536,0.004192000130812327
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,768,1536,0.016570666432380678
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,768,1536,0.009525332848230999
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,768,1024,0.0033887999753157297
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,768,1024,0.016645333170890807
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,768,1024,0.008796800176302593
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,768,768,0.0031061333914597826
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,768,768,0.01476800044377645
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,768,768,0.007924266656239827
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,768,512,0.0029909332593282064
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,768,512,0.014612266421318054
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,768,512,0.007939200103282928
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,768,256,0.0025813333690166474
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,768,256,0.015007999539375306
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,768,256,0.007493333518505096
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,768,128,0.002199466774861018
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,768,128,0.014516266187032065
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,768,128,0.007114666700363159
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,768,64,0.0021727999051411946
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,768,64,0.014386133352915446
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,768,32,0.002203733225663503
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,768,32,0.014406399925549826
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,512,65536,0.019169066349665323
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,512,65536,0.04842880169550578
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,512,65536,0.12442879676818848
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,512,16384,0.012079999844233195
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,512,16384,0.028870399792989093
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,512,16384,0.03631360133488973
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,512,12288,0.00997440020243327
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,512,12288,0.02532373269399007
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,512,12288,0.02885013421376546
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,512,10240,0.009193600217501322
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,512,10240,0.024201599756876628
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,512,10240,0.025196800629297893
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,512,8192,0.008301866551240284
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,512,8192,0.02323626677195231
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,512,8192,0.021547732750574748
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,512,7168,0.0075328002373377485
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,512,7168,0.02248319983482361
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,512,7168,0.019811199108759562
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,512,6144,0.007117866476376851
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,512,6144,0.020692267020543418
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,512,6144,0.018142932653427125
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,512,5120,0.006667733192443848
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,512,5120,0.020716800292332967
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,512,5120,0.016130133469899496
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,512,4096,0.006662400066852569
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,512,4096,0.018845866123835243
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,512,4096,0.01411626636981964
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,512,3584,0.006248533229033152
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,512,3584,0.018133334318796792
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,512,3584,0.013226667046546936
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,512,3072,0.006316799918810527
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,512,3072,0.018157867590586345
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,512,3072,0.012010666728019714
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,512,2560,0.0056202664971351625
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,512,2560,0.016874667008717856
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,512,2560,0.010820266604423524
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,512,2048,0.004603733122348785
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,512,2048,0.01651306649049123
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,512,2048,0.010358400146166484
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,512,1536,0.003851733356714249
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,512,1536,0.016365866859753928
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,512,1536,0.009159466624259949
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,512,1024,0.0033642667035261786
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,512,1024,0.017114667097727458
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,512,1024,0.008318933347860973
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,512,768,0.003048533449570338
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,512,768,0.016272000471750894
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,512,768,0.007940266529719036
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,512,512,0.00264533335963885
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,512,512,0.015165866414705912
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,512,512,0.007939200103282928
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,512,256,0.0026133333643277483
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,512,256,0.015160533785820007
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,512,256,0.0070826664566993715
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,512,128,0.002209066599607468
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,512,128,0.014501333236694336
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,512,128,0.007054933408896129
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,512,64,0.0021162666380405426
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,512,64,0.01432319978872935
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,512,32,0.002201599876085917
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,512,32,0.01458560029665629
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,256,65536,0.017806933323542277
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,256,65536,0.043188265959421795
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,256,65536,0.12417706648508708
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,256,16384,0.0084906667470932
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,256,16384,0.02980159918467204
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,256,16384,0.03624853293100993
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,256,12288,0.0071285332242647815
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,256,12288,0.024408533175786336
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,256,12288,0.028880000114440918
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,256,10240,0.006885333359241486
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,256,10240,0.02481386661529541
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,256,10240,0.0252074658870697
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,256,8192,0.009139200051625569
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,256,8192,0.023605332771937052
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,256,8192,0.02146880030632019
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,256,7168,0.008277333279450735
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,256,7168,0.022641066710154215
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,256,7168,0.01983893314997355
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,256,6144,0.006333866715431213
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,256,6144,0.02087786595026652
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,256,6144,0.018179200092951455
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,256,5120,0.006681600213050842
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,256,5120,0.020399999618530274
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,256,5120,0.016515200336774193
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,256,4096,0.006258133550484974
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,256,4096,0.018900267283121743
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,256,4096,0.014043733477592468
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,256,3584,0.005850666761398315
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,256,3584,0.019646932681401573
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,256,3584,0.01290986637274424
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,256,3072,0.00550186683734258
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,256,3072,0.017268266280492148
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,256,3072,0.012038399775822956
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,256,2560,0.0050016000866889955
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,256,2560,0.01694613297780355
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,256,2560,0.011210667093594869
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,256,2048,0.004621866842110952
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,256,2048,0.016966400543848674
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,256,2048,0.010361599922180175
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,256,1536,0.003852800031503042
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,256,1536,0.016012799739837647
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,256,1536,0.009159466624259949
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,256,1024,0.0033845332761605583
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,256,1024,0.01672853430112203
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,256,1024,0.008686932921409606
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,256,768,0.0029813334345817565
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,256,768,0.01544533371925354
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,256,768,0.007948799928029378
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,256,512,0.002587733417749405
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,256,512,0.01597760021686554
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,256,512,0.007496533294518788
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,256,256,0.0025610665480295816
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,256,256,0.01585493286450704
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,256,256,0.007074133555094402
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,256,128,0.00222080002228419
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,256,128,0.014432000120480857
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,256,128,0.006647466619809468
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,256,64,0.002203733225663503
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,256,64,0.01453013320763906
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,256,32,0.0021759999295075734
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,256,32,0.01477226714293162
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,128,65536,0.0151146670182546
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,128,65536,0.03902933200200399
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,128,65536,0.12387200196584065
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,128,16384,0.008932266632715862
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,128,16384,0.025285333395004272
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,128,16384,0.03617066542307536
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,128,12288,0.007910400132338206
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,128,12288,0.02477546731630961
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,128,12288,0.028806400299072266
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,128,10240,0.006266666452089946
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,128,10240,0.023001599311828613
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,128,10240,0.025203200181325276
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,128,8192,0.005895466605822245
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,128,8192,0.022663466135660806
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,128,8192,0.021383466323216756
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,128,7168,0.00547733356555303
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,128,7168,0.021661865711212158
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,128,7168,0.019402666886647543
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,128,6144,0.00631039987007777
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,128,6144,0.021547732750574748
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,128,6144,0.017787732680638633
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,128,5120,0.005858133236567179
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,128,5120,0.021248000860214233
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,128,5120,0.016127999623616537
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,128,4096,0.00553706685702006
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,128,4096,0.020536533991495767
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,128,4096,0.014071466525395713
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,128,3584,0.005390933156013489
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,128,3584,0.018539732694625853
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,128,3584,0.012874666849772134
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,128,3072,0.005087999999523163
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,128,3072,0.018382932742436728
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,128,3072,0.012055466572443645
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,128,2560,0.005107200145721436
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,128,2560,0.01800959904988607
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,128,2560,0.01081706682840983
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,128,2048,0.004985600213209788
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,128,2048,0.016546133160591125
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,128,2048,0.009975467125574749
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,128,1536,0.0037621334195137024
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,128,1536,0.016862932840983072
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,128,1536,0.009236266215642292
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,128,1024,0.0030218665798505146
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,128,1024,0.014614400267601014
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,128,1024,0.008697600166002909
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,128,768,0.0030410667260487873
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,128,768,0.01467519998550415
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,128,768,0.007957333326339721
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,128,512,0.002865066627661387
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,128,512,0.014708266655604044
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,128,512,0.007584000130494435
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,128,256,0.0024010665714740754
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,128,256,0.014640000462532044
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,128,256,0.007211733361085255
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,128,128,0.0022624000906944276
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,128,128,0.015205333630243937
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,128,128,0.006764799853165944
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,128,64,0.002179199953873952
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,128,64,0.014469333489735923
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,128,32,0.002164266755183538
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,128,32,0.014563199877738953
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,64,65536,0.013795199990272521
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,64,65536,0.03717973232269287
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,64,16384,0.007091199855009715
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,64,16384,0.024855466683705647
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,64,12288,0.005883733431498209
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,64,12288,0.023009065786997476
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,64,10240,0.005926399926344554
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,64,10240,0.022899200518925987
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,64,8192,0.005881600081920624
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,64,8192,0.021466666460037233
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,64,7168,0.005444266895453135
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,64,7168,0.021233065923055013
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,64,6144,0.005452799797058106
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,64,6144,0.02130240003267924
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,64,5120,0.005076266825199127
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,64,5120,0.020658133427302043
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,64,4096,0.004730666677157084
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,64,4096,0.019358932971954346
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,64,3584,0.0054293334484100345
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,64,3584,0.01854613423347473
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,64,3072,0.005109333495299021
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,64,3072,0.017513600985209148
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,64,2560,0.0054293334484100345
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,64,2560,0.017946666479110716
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,64,2048,0.005025066435337067
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,64,2048,0.016587733229001363
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,64,1536,0.0037823999921480812
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,64,1536,0.016611199577649435
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,64,1024,0.0031317333380381264
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,64,1024,0.01646719972292582
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,64,768,0.0029653333127498626
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,64,768,0.01609280010064443
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,64,512,0.0026378666361172995
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,64,512,0.014536533753077188
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,64,256,0.002221866697072983
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,64,256,0.014393599828084311
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,64,128,0.002201599876085917
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,64,128,0.014454399545987448
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,64,64,0.0021162666380405426
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,64,64,0.014450132846832275
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,64,32,0.0021301334102948504
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,64,32,0.014465066790580749
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,32,65536,0.010456533233324686
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,32,65536,0.036925868193308516
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,32,16384,0.0067114666104316715
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,32,16384,0.024804266293843587
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,32,12288,0.005542399982611338
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,32,12288,0.023226666450500488
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,32,10240,0.0058442667126655575
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,32,10240,0.022685867547988892
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,32,8192,0.005448533097902933
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,32,8192,0.020965333779652914
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,32,7168,0.0053962667783101406
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,32,7168,0.021205333868662517
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,32,6144,0.005128533144791921
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,32,6144,0.02140799959500631
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,32,5120,0.005014400184154511
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,32,5120,0.02073600093523661
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,32,4096,0.0046623999873797095
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,32,4096,0.020245333512624107
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,32,3584,0.005086933573087057
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,32,3584,0.019008000691731773
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,32,3072,0.00498879998922348
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,32,3072,0.017384533087412515
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,32,2560,0.0050911997755368555
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,32,2560,0.01705706715583801
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,32,2048,0.004644266764322917
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,32,2048,0.016667733589808144
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,32,1536,0.0038346665600935614
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,32,1536,0.015969066818555196
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,32,1024,0.003035733352104823
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,32,1024,0.01560426652431488
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,32,768,0.0029674666623274487
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,32,768,0.018526933590571084
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,32,512,0.0026047999660174055
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,32,512,0.014913066228230795
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,32,256,0.002414933343728383
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,32,256,0.01448853313922882
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,32,128,0.00222080002228419
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,32,128,0.014437333742777506
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,32,64,0.002102400114138921
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,32,64,0.01527679959932963
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,160,32,32,0.002221866697072983
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,32,32,0.014487466216087342
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,65536,16384,0.3740394592285156
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,65536,16384,0.2130570729573568
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,160,2048,512,0.007531733314196269
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,65536,12288,0.276147206624349
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,65536,12288,0.16686293284098308
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,3072,64,0.014471466342608133
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,65536,10240,0.2380842685699463
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,65536,10240,0.14194879531860352
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,160,1024,32,0.014477866888046264
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,65536,8192,0.19381866455078126
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,65536,8192,0.1182751973470052
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,65536,10240,0.07410666942596436
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,65536,7168,0.16553173065185547
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,65536,7168,0.10282560189565022
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,65536,8192,0.06253439982732137
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,65536,6144,0.14665172894795736
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,65536,6144,0.09007253646850585
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,65536,12288,0.08591360251108805
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,65536,5120,0.12032106717427571
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,65536,5120,0.07929279804229736
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,65536,5120,0.0423583984375
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,65536,4096,0.09878186384836832
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,65536,4096,0.06708906491597494
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,65536,7168,0.0547104001045227
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,65536,3584,0.08614719708760579
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,65536,3584,0.06052373250325521
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,65536,6144,0.04819200038909912
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,65536,3072,0.07452267011006673
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,65536,3072,0.0534112016359965
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,65536,16384,0.11249919732411702
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,65536,2560,0.06368106603622437
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,65536,2560,0.04779200156529744
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,65536,3072,0.028621866305669146
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,65536,2048,0.05321493148803711
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,65536,2048,0.043772800763448076
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,65536,2560,0.025548799832661943
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,65536,1536,0.04023146629333496
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,65536,1536,0.03530559937159221
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,65536,3584,0.0316703995068868
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,65536,1024,0.027889066934585573
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,65536,1024,0.02941653331120809
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,65536,1024,0.014502400159835815
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,65536,768,0.021769599119822184
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,65536,768,0.025913600126902265
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,65536,4096,0.03481386502583821
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,65536,512,0.015337600310643514
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,65536,512,0.022950400908788048
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,65536,1536,0.01801066597302755
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,65536,256,0.010012800494829815
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,65536,256,0.021244800090789794
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,65536,256,0.009622400005658466
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,65536,128,0.007162666817506154
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,65536,128,0.01875093380610148
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,65536,768,0.013250133395195008
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,65536,64,0.0066890666882197065
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,65536,64,0.02060479919115702
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,65536,32,0.008746666709582011
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,65536,32,0.021240532398223877
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,16384,65536,0.3197237332661947
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,16384,65536,0.2405973275502523
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,65536,128,0.00881173312664032
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,16384,16384,0.08786880175272624
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,16384,16384,0.07287466526031494
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,65536,512,0.010877866546312969
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,16384,12288,0.06248213450113932
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,16384,12288,0.058005332946777344
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,65536,2048,0.02147946755091349
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,16384,10240,0.053489065170288085
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,16384,10240,0.051750401655832924
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,16384,10240,0.029764266808827718
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,16384,8192,0.043694933255513504
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,16384,8192,0.04457279841105143
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,16384,12288,0.03402560154596965
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,16384,7168,0.03870400190353394
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,16384,7168,0.04032426675160726
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,16384,16384,0.04235519965489705
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,16384,6144,0.03419306675593058
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,16384,6144,0.034780800342559814
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,16384,8192,0.025164800882339477
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,16384,5120,0.029339732726414998
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,16384,5120,0.031547733147939044
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,16384,7168,0.02277440031369527
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,16384,4096,0.02407360076904297
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,16384,4096,0.028042666117350262
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,16384,6144,0.02028586665789286
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,16384,3584,0.021574399868647256
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,16384,3584,0.02696320017178853
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,16384,4096,0.016226133704185484
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,16384,3072,0.018837332725524902
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,16384,3072,0.024756266673405965
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,16384,5120,0.018272000551223754
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,16384,2560,0.01612160007158915
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,16384,2560,0.022887466351191203
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,16384,2560,0.012811733285586038
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,16384,2048,0.013246933619181315
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,16384,2048,0.022836265961329143
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,16384,3584,0.014939733346303306
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,16384,1536,0.010838400324185688
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,16384,1536,0.02054506738980611
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,16384,3072,0.01372160017490387
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,16384,1024,0.00823359986146291
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,16384,1024,0.016717867056528727
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,16384,1536,0.010073600212732951
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,16384,1024,0.0086517333984375
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,16384,768,0.0067562664548556015
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,16384,768,0.01650773286819458
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,16384,512,0.005267199873924255
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,16384,512,0.015808000167210897
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,16384,2048,0.01122773289680481
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,16384,256,0.003753600021203359
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,16384,256,0.017044266064961754
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,16384,768,0.008340266346931458
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,16384,128,0.003352533280849457
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,16384,512,0.007997866471608479
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,16384,128,0.015159466862678527
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,16384,64,0.003028266628583272
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,16384,64,0.014494933684666953
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,16384,32,0.0031456001102924346
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,16384,32,0.015147733688354491
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,12288,65536,0.23939733505249022
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,16384,128,0.007134933272997539
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,12288,65536,0.18816852569580078
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,12288,16384,0.06743893623352051
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,12288,16384,0.058898135026295984
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,16384,256,0.007238399982452392
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,12288,12288,0.048434134324391684
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,12288,12288,0.0480565349260966
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,12288,12288,0.031982932488123575
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,12288,10240,0.04174826542536418
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,12288,10240,0.04315733512242635
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,12288,16384,0.03991359869639079
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,12288,8192,0.033852799733479814
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,12288,8192,0.03725226720174153
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,12288,10240,0.02804479996363322
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,12288,7168,0.029815467198689778
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,12288,7168,0.03252159953117371
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,12288,8192,0.023475199937820435
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,12288,6144,0.02608426610628764
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,12288,6144,0.029794132709503172
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,12288,7168,0.021287467082341513
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,12288,5120,0.022882133722305298
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,12288,5120,0.027040000756581622
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,12288,6144,0.019802665710449217
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,12288,4096,0.01836693286895752
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,12288,4096,0.02490880091985067
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,12288,5120,0.017805866400400796
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,12288,3584,0.01665493349234263
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,12288,3584,0.02435520092646281
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,12288,4096,0.015758933623631795
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,12288,3072,0.014403200149536133
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,12288,3072,0.02441920042037964
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,12288,3584,0.014541866381963095
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,12288,2560,0.012482133507728577
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,12288,2560,0.021860265731811525
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,12288,3072,0.013278933366139731
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,12288,2048,0.01051200032234192
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,12288,2048,0.02127893368403117
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,12288,2560,0.012095999717712403
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,12288,1536,0.008844799796740214
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,12288,1536,0.018722132841746012
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,12288,1536,0.009622400005658466
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,12288,1024,0.006726400057474773
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,12288,1024,0.0164874662955602
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,12288,2048,0.010889599720637005
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,12288,768,0.005434666574001312
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,12288,768,0.017439999183019004
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,12288,1024,0.008598400155703227
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,12288,512,0.004275199770927429
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,12288,512,0.016108799974123636
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,12288,768,0.008391466736793519
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,12288,256,0.0038399999340375268
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,12288,256,0.015845333536465965
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,12288,512,0.007577600081761678
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,12288,128,0.0034048000971476236
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,12288,128,0.014838400483131408
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,12288,256,0.007490133245786031
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,12288,64,0.003033600002527237
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,12288,64,0.015523200233777365
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,12288,32,0.0033728001018365227
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,12288,32,0.014587733149528503
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,12288,128,0.007099733253320058
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,10240,65536,0.20071040789286293
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,10240,65536,0.1505013306935628
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,10240,16384,0.05635519822438558
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,10240,16384,0.052781867980957034
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,10240,16384,0.039136000474294025
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,10240,12288,0.04039040009180705
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,10240,12288,0.04328426520029704
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,10240,12288,0.030954666932423908
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,10240,10240,0.03516159852345784
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,10240,10240,0.0392522652943929
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,16384,65536,0.14477866490681965
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,10240,8192,0.028577067454655963
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,10240,8192,0.03386559883753459
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,10240,10240,0.02731093366940816
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,10240,7168,0.025464532772699992
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,10240,7168,0.03009066581726074
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,10240,8192,0.023190399010976158
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,10240,6144,0.02231360077857971
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,10240,6144,0.028204800685246785
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,12288,65536,0.1332192023595174
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,10240,5120,0.019474132855733236
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,10240,5120,0.026461867491404216
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,10240,7168,0.021121066808700562
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,10240,4096,0.015363199512163797
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,10240,4096,0.02369813323020935
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,10240,6144,0.019351466496785482
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,10240,3584,0.013942399621009826
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,10240,3584,0.02315839926401774
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,10240,5120,0.017406932512919106
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,10240,3072,0.012452266613642375
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,10240,3072,0.022637865940729775
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,10240,4096,0.015383467078208923
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,10240,2560,0.010805333654085796
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,10240,2560,0.021427200237909953
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,10240,3584,0.014479999740918478
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,10240,2048,0.009230933586756388
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,10240,2048,0.018897066513697304
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,10240,3072,0.0129013329744339
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,10240,1536,0.007835733393828075
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,10240,1536,0.017240534226099648
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,10240,2560,0.011672533551851908
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,10240,1024,0.005633066594600678
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,10240,1024,0.016540799538294473
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,10240,1024,0.008782933155695598
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,10240,768,0.004391466577847799
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,10240,768,0.01584106683731079
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,10240,1536,0.00965119997660319
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,10240,512,0.0039264000952243805
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,10240,512,0.016217600305875143
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,10240,2048,0.010591999689737955
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,10240,256,0.0034805332620938623
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,10240,256,0.014595199624697366
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,10240,512,0.007948799928029378
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,10240,128,0.0030389333764712016
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,10240,256,0.0074869334697723385
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,10240,768,0.008331733445326488
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,10240,128,0.015012266238530478
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,10240,64,0.003019733230272929
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,10240,32,0.0030965333183606463
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,10240,64,0.014491732915242514
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,10240,32,0.014478933811187745
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,8192,65536,0.1838304042816162
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,8192,65536,0.12717440128326415
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,10240,128,0.007459199925263722
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,8192,16384,0.04923093318939209
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,8192,16384,0.04681599934895833
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,8192,12288,0.0362175981203715
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,8192,12288,0.03937173287073771
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,8192,12288,0.030497066179911297
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,8192,10240,0.03177066644032796
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,8192,10240,0.035225598017374675
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,8192,16384,0.03847786585489909
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,8192,8192,0.026425600051879883
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,8192,8192,0.030359466870625813
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,8192,10240,0.026768000920613606
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,8192,7168,0.02272319992383321
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,8192,7168,0.028679466247558592
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,8192,8192,0.022668800751368203
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,8192,6144,0.020413867632548013
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,8192,6144,0.026446932554244997
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,8192,7168,0.02104533314704895
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,8192,5120,0.01735573410987854
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,8192,5120,0.02487679918607076
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,8192,6144,0.019041067361831664
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,8192,4096,0.014324266711870828
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,8192,4096,0.022756266593933105
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,10240,65536,0.1300437370936076
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,8192,3584,0.011789866288503011
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,8192,3584,0.022052266200383506
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,8192,5120,0.017464532454808553
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,8192,3072,0.010619733730951946
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,8192,3072,0.02098133365313212
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,8192,3584,0.013793067137400309
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,8192,2560,0.009539199868837993
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,8192,2560,0.020489599307378134
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,8192,4096,0.014980266491572062
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,8192,2048,0.00832533339659373
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,8192,2048,0.019248000780741372
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,8192,3072,0.012518399953842163
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,8192,1536,0.0070720002055168155
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,8192,1536,0.017960532506306966
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,8192,2560,0.01167680025100708
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,8192,1024,0.0047882666190465295
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,8192,1024,0.01673813263575236
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,8192,2048,0.010453333457310993
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,8192,768,0.0038986665507157645
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,8192,768,0.016548267006874083
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,8192,1024,0.008776533603668212
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,8192,512,0.00345920001467069
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,8192,512,0.014519466956456503
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,8192,1536,0.009689600268999735
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,8192,256,0.0032373333970705668
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,8192,256,0.014691199858983359
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,8192,512,0.007613866527875264
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,8192,128,0.002985599885384242
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,8192,768,0.008339200417200725
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,8192,128,0.014454399545987448
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,8192,64,0.002997333308060964
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,8192,64,0.014758400122324624
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,8192,32,0.002997333308060964
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,8192,256,0.0074890668193499255
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,8192,32,0.014538666605949402
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,8192,128,0.0071050668756167095
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,7168,65536,0.16183147430419922
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,7168,65536,0.12803200085957844
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,7168,16384,0.04049386580785115
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,7168,16384,0.041255466143290204
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,7168,12288,0.030051199595133464
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,7168,12288,0.03511999845504761
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,7168,12288,0.030052266518274945
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,7168,10240,0.025983999172846477
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,7168,10240,0.030929066737492877
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,7168,16384,0.03775466680526733
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,7168,8192,0.021116799116134642
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,7168,8192,0.028032000859578448
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,7168,10240,0.02645866672197978
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,7168,7168,0.018645334243774413
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,7168,7168,0.02662186622619629
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,7168,8192,0.022452267011006673
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,7168,6144,0.016436266899108886
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,7168,6144,0.024791467189788818
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,7168,7168,0.021081600586573282
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,7168,5120,0.014133333166440328
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,7168,5120,0.023346134026845298
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,7168,6144,0.018990933895111084
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,7168,4096,0.011621333161989848
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,7168,4096,0.023177599906921385
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,7168,5120,0.01697173317273458
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,7168,3584,0.010546132922172546
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,7168,3584,0.020719999074935914
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,7168,4096,0.014549332857131957
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,7168,3072,0.009594666957855224
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,7168,3072,0.01999680002530416
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,7168,3584,0.01411626636981964
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,7168,2560,0.008362666765848795
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,7168,2560,0.0186901330947876
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,7168,3072,0.01250879963239034
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,7168,2048,0.007507200042406718
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,7168,2048,0.017299199104309083
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,7168,2560,0.011264000336329143
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,7168,1536,0.005993600189685822
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,7168,1536,0.016505600015322367
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,7168,2048,0.010428800185521444
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,7168,1024,0.004279466470082601
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,7168,1024,0.014971733093261719
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,7168,1536,0.009562666217486065
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,7168,768,0.003700266778469086
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,7168,768,0.017813332875569663
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,7168,1024,0.008393599589665731
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,7168,512,0.0033941333492596946
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,7168,512,0.014492799838383993
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,7168,512,0.007976533472537994
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,7168,256,0.0029845332105954488
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,7168,768,0.008346666892369587
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,7168,256,0.014460800091425577
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,7168,128,0.0025920001169045764
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,7168,128,0.014507733782132468
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,7168,128,0.00709440012772878
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,7168,64,0.0025557334224383037
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,7168,64,0.0144896000623703
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,7168,32,0.0029525332152843474
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,7168,256,0.0071733335653940845
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,7168,32,0.014458666245142618
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,6144,65536,0.10485119819641113
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,6144,65536,0.14332906405131024
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,6144,16384,0.0365013321240743
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,6144,16384,0.03924479881922404
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,8192,65536,0.12842986583709717
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,6144,12288,0.02733653386433919
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,6144,12288,0.033297065893809
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,6144,12288,0.02932800054550171
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,6144,10240,0.024104533592859904
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,6144,10240,0.03089066743850708
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,6144,16384,0.03732586701711019
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,6144,8192,0.01941439906756083
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,6144,8192,0.02834879954655965
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,6144,8192,0.022465066115061442
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,6144,10240,0.02602880001068115
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,6144,7168,0.01688533425331116
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,6144,7168,0.02534186641375224
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,6144,6144,0.01513813336690267
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,6144,6144,0.02481493353843689
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,6144,6144,0.019056000312169395
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,6144,5120,0.013318399588267008
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,6144,5120,0.0231605331103007
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,6144,7168,0.02099626660346985
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,6144,4096,0.010777599612871806
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,6144,4096,0.02165013353029887
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,6144,5120,0.016845866044362386
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,6144,3584,0.010034132997194927
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,6144,3584,0.02118933399518331
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,6144,4096,0.01448319951693217
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,6144,3072,0.009005866448084513
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,6144,3072,0.019590399662653604
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,6144,3072,0.012523733576138816
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,6144,2560,0.007991466422875721
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,6144,3584,0.01288746694723765
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,6144,2560,0.01856000026067098
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,6144,2048,0.00695253312587738
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,6144,2048,0.016660267114639284
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,6144,2048,0.010871466994285584
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,6144,1536,0.005505066613356272
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,6144,2560,0.011595732967058818
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,6144,1536,0.016541866461435954
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,7168,65536,0.12838079929351806
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,6144,1024,0.004262400170167288
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,6144,1024,0.017203199863433837
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,6144,768,0.003786666691303253
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,6144,768,0.014745600024859109
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,6144,1024,0.008813866972923278
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,6144,512,0.003435733417669932
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,6144,512,0.014794666568438211
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,6144,768,0.008370133241017659
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,6144,256,0.002977066735426585
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,6144,256,0.014903466900189719
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,6144,1536,0.00920960009098053
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,6144,128,0.0030229332546393077
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,6144,128,0.014468266566594442
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,6144,256,0.0075093333919843035
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,6144,64,0.0025685332715511323
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,6144,64,0.014439466595649719
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,6144,32,0.002979200085004171
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,6144,128,0.007157333195209503
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,6144,512,0.007899733384450276
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,6144,32,0.014476799964904785
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,5120,65536,0.11441067059834797
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,5120,65536,0.09419199625651041
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,5120,16384,0.03630079825719197
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,5120,16384,0.04284480015436808
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,5120,12288,0.030076799790064494
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,5120,12288,0.031176533301671343
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,5120,12288,0.029603199164072676
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,5120,10240,0.02182719906171163
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,5120,10240,0.029292800029118854
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,5120,16384,0.03663146495819092
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,5120,8192,0.01788053313891093
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,5120,8192,0.026757333676020307
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,5120,10240,0.02605866591135661
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,5120,7168,0.015451733271280924
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,5120,7168,0.025564799706141155
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,5120,8192,0.022283732891082764
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,5120,6144,0.014058666427930197
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,5120,6144,0.024998400608698526
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,5120,7168,0.020722132921218873
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,5120,5120,0.012077866991360982
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,5120,5120,0.022725333770116173
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,5120,6144,0.018651733795801796
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,5120,4096,0.010078932841618855
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,5120,4096,0.021489065885543824
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,5120,5120,0.016613333423932394
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,5120,3584,0.009194667140642803
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,5120,3584,0.01989439924558004
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,5120,4096,0.014197333653767904
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,5120,3072,0.008337066570917765
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,5120,3072,0.018649599949518838
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,5120,3584,0.013271466890970866
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,5120,2560,0.007520000139872233
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,5120,2560,0.017785600821177163
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,5120,3072,0.012146133184432983
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,5120,2048,0.006381866832574208
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,5120,2048,0.016692266861597697
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,5120,2560,0.011611732840538024
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,5120,1536,0.005102933446566264
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,5120,1536,0.017836799224217735
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,5120,2048,0.010621866583824158
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,5120,1024,0.004025600105524063
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,5120,1024,0.016437333822250367
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,5120,1536,0.009644800424575805
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,5120,768,0.0038463999827702843
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,5120,768,0.015018666783968607
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,5120,1024,0.008841600020726521
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,5120,512,0.0035114665826161706
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,5120,512,0.017043199141820273
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,5120,512,0.007952000200748443
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,5120,256,0.0029898665845394133
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,5120,256,0.014935466647148132
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,5120,256,0.007565866907437642
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,5120,128,0.002639999985694885
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,5120,128,0.014614400267601014
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,5120,128,0.007073066631952922
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,5120,64,0.0025994665920734406
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,5120,64,0.014570666352907815
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,5120,32,0.0026517334083716077
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,5120,32,0.014661332964897156
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,4096,65536,0.09664639631907145
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,4096,65536,0.08689066569010416
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,6144,65536,0.128109868367513
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,4096,16384,0.029811199506123858
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,4096,16384,0.036423468589782716
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,4096,16384,0.03696320056915283
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,4096,12288,0.024253867069880166
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,4096,12288,0.028986666599909467
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,4096,12288,0.02962239980697632
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,4096,10240,0.018665599822998046
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,4096,10240,0.027533866961797077
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,4096,10240,0.02576213280359904
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,4096,8192,0.014947199821472168
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,4096,8192,0.024745599428812663
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,5120,65536,0.1277450640996297
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,4096,7168,0.013420800367991129
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,4096,7168,0.02490560015042623
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,4096,8192,0.021973333756128945
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,4096,6144,0.01225600043932597
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,4096,6144,0.023265065749486287
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,4096,7168,0.020082134008407592
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,4096,5120,0.010653866330782573
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,4096,5120,0.021603200833002725
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,4096,6144,0.018260266383488974
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,4096,4096,0.008897067109743754
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,4096,4096,0.018891733884811402
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,4096,5120,0.016196266810099284
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,4096,3584,0.008254933357238769
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,4096,3584,0.019143466154734293
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,4096,3584,0.013264000415802002
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,4096,3072,0.007522133489449819
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,4096,4096,0.01409386694431305
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,4096,3072,0.017808000246683754
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,4096,2560,0.0067669332027435304
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,4096,2560,0.017395200332005818
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,4096,2560,0.011306666334470113
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,4096,2048,0.005739733576774597
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,4096,2048,0.016733866930007935
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,4096,3072,0.012141866485277812
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,4096,1536,0.004987733562787374
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,4096,1536,0.016039466857910155
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,4096,1536,0.009296000003814697
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,4096,1024,0.0035487999518712364
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,4096,1024,0.016167466839154564
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,4096,2048,0.010417067011197408
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,4096,768,0.0034506666163603462
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,4096,768,0.01469546655813853
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,4096,768,0.008042666812737782
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,4096,1024,0.008730666836102803
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,4096,512,0.0030048000315825146
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,4096,512,0.01477120021979014
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,4096,256,0.0025792000194390613
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,4096,256,0.014452266693115234
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,4096,256,0.007522133489449819
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,4096,128,0.002598399917284648
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,4096,128,0.014444800217946372
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,4096,512,0.007876266539096833
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,4096,64,0.0025898667673269907
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,4096,64,0.014443733294804893
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,4096,32,0.0029696000119050344
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,4096,128,0.007131733496983846
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,4096,32,0.01453013320763906
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,3584,65536,0.08555946350097657
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,3584,65536,0.07856960296630859
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,3584,16384,0.024984532594680788
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,3584,16384,0.035182933012644454
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,4096,65536,0.1275274674097697
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,3584,12288,0.020381865898768108
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,3584,12288,0.028908799091974895
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,3584,16384,0.03639573256174723
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,3584,10240,0.017923200130462648
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,3584,10240,0.026766933997472125
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,3584,12288,0.02930240035057068
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,3584,8192,0.014536533753077188
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,3584,8192,0.024979199965794882
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,3584,10240,0.02564799984296163
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,3584,7168,0.012905599673589072
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,3584,7168,0.024768000841140746
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,3584,8192,0.021934932470321654
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,3584,6144,0.011665067076683045
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,3584,6144,0.022792534033457438
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,3584,7168,0.019847466548283895
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,3584,5120,0.010027733445167542
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,3584,5120,0.020920532941818237
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,3584,5120,0.01576640009880066
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,3584,4096,0.008738133311271667
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,3584,4096,0.019934932390848793
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,3584,6144,0.01790613333384196
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,3584,3584,0.008017066617806752
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,3584,3584,0.018702934185663857
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,3584,3584,0.013395200173060099
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,3584,3072,0.007132799923419952
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,3584,3072,0.018437333901723228
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,3584,4096,0.014150399963061014
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,3584,2560,0.006268799801667531
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,3584,2560,0.016781866550445557
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,3584,2560,0.011316266655921937
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,3584,2048,0.005456000069777171
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,3584,2048,0.016525866587956746
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,3584,3072,0.012386133273442585
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,3584,1536,0.004695466657479604
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,3584,1536,0.017448532581329345
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,3584,1536,0.009657599528630574
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,3584,1024,0.0038624001046021783
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,3584,1024,0.015459199746449789
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,3584,2048,0.010391466816266378
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,3584,768,0.0034645333886146545
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,3584,768,0.014507733782132468
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,3584,1024,0.00881599982579549
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,3584,512,0.003070933371782303
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,3584,512,0.014570666352907815
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,3584,768,0.007924266656239827
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,3584,256,0.003337600082159042
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,3584,256,0.014871467153231302
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,3584,512,0.0075882668296496075
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,3584,256,0.00746666689713796
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,3584,128,0.014936533570289613
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,3584,128,0.0067562664548556015
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,3584,64,0.0027583998938401537
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,3584,64,0.01456106702486674
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,3584,32,0.0031637333333492277
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,3584,32,0.01456106702486674
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,3072,65536,0.07109119892120361
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,3072,65536,0.07180799643198649
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,5120,768,0.008329600095748901
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,3072,16384,0.024321067333221435
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,3072,16384,0.03450453281402588
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,3584,65536,0.1278986692428589
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,3072,12288,0.018710400660832724
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,3072,12288,0.030557866891225176
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,3072,16384,0.03668800195058187
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,3072,10240,0.01634666621685028
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,3072,10240,0.02714666724205017
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,3072,12288,0.029253333806991577
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,3072,8192,0.013605333367983499
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,3072,8192,0.0249674657980601
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,3072,10240,0.02553706765174866
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,3072,7168,0.012462932864824932
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,3072,7168,0.023861332734425863
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,3072,8192,0.021142399311065672
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,3072,6144,0.011256532867749532
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,3072,6144,0.02129279971122742
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,3072,7168,0.019976532459259032
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,3072,5120,0.010058666268984478
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,3072,5120,0.020427733659744263
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,3072,6144,0.017825067043304443
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,3072,4096,0.008538666367530822
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,3072,4096,0.01955839991569519
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,3072,4096,0.014132266243298849
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,3072,3584,0.0075530668099721265
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,3072,3584,0.019884800910949706
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,3072,5120,0.016203733285268147
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,3072,3072,0.006833066542943318
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,3072,3072,0.018572799364725747
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,3072,3584,0.013274666666984559
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,3072,2560,0.006248533229033152
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,3072,3072,0.012037332852681477
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,3072,2560,0.017986132701237997
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,3072,2048,0.0050687998533248905
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,3072,2048,0.016570666432380678
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,3072,2048,0.01036906639734904
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,3072,1536,0.004628266890843709
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,3072,1536,0.016820265849431356
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,3072,2560,0.011594667037328085
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,3072,1024,0.0038506666819254553
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,3072,1024,0.01629866659641266
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,3072,1024,0.008402132987976074
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,3072,768,0.003505066782236099
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,3072,768,0.014603733023007711
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,3072,1536,0.009525332848230999
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,3072,512,0.0030602666238943735
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,3072,512,0.014535466829935709
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,3072,768,0.008306133250395458
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,3072,256,0.0029909332593282064
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,3072,512,0.00787306676308314
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,3072,256,0.015105066696802774
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,3072,128,0.0026357332865397137
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,3072,128,0.014587733149528503
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,3072,256,0.0074869334697723385
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,3072,64,0.0025973332424958544
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,3072,64,0.014477866888046264
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,3072,32,0.0025802666942278544
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,3072,32,0.014459733168284097
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,2560,65536,0.06323413451512655
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,2560,65536,0.06776959896087646
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,3072,65536,0.12760426998138427
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,2560,16384,0.02024959921836853
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,2560,16384,0.03262400031089783
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,2560,16384,0.036474665006001786
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,2560,12288,0.01698453426361084
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,2560,12288,0.028924800952275592
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,2560,12288,0.02887786626815796
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,2560,10240,0.015382400155067444
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,2560,10240,0.026053333282470705
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,2560,10240,0.025246934096018477
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,2560,8192,0.012870400150616964
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,2560,8192,0.023105067014694215
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,2560,8192,0.021585067113240562
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,2560,7168,0.012301866213480632
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,2560,7168,0.022986666361490885
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,2560,7168,0.01952426632245382
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,2560,6144,0.011533866326014202
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,2560,6144,0.02134079933166504
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,2560,6144,0.017844265699386595
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,2560,5120,0.009733333190282186
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,2560,5120,0.02035413384437561
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,2560,5120,0.016135467092196147
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,2560,4096,0.008152533570925396
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,2560,4096,0.018576000134150186
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,2560,65536,0.1272437334060669
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,2560,3584,0.007527466615041096
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,2560,3584,0.018101332585016887
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,2560,4096,0.014125866691271463
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,2560,3072,0.00664213349421819
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,2560,3072,0.01703146696090698
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,2560,3584,0.013303466637929282
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,2560,2560,0.005874133110046387
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,2560,2560,0.01717653274536133
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,2560,3072,0.012110933661460876
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,2560,2048,0.005155199766159057
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,2560,2048,0.016684800386428833
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,2560,2560,0.011593600114186604
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,2560,1536,0.00425600012143453
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,2560,1536,0.01576533317565918
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,2560,2048,0.010368000467618306
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,2560,1024,0.003822933385769526
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,2560,1024,0.015282133221626281
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,2560,1536,0.009551999966303508
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,2560,768,0.0033930666744709016
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,2560,768,0.014643200238545737
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,2560,1024,0.008355200290679932
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,2560,512,0.0030464000999927522
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,2560,512,0.016247466206550598
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,2560,768,0.0082997332016627
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,2560,512,0.007838933169841767
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,2560,256,0.0025962665677070618
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,2560,256,0.014443733294804893
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,2560,128,0.0025610665480295816
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,2560,128,0.014666666587193808
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,2560,256,0.007506133119265239
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,2560,64,0.002162133405605952
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,2560,128,0.007146666447321574
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,2560,32,0.0025792000194390613
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,2560,64,0.014502400159835815
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,2560,32,0.014473600188891092
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,2048,65536,0.05523413419723511
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,2048,65536,0.06090026696523031
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,2048,16384,0.016545066237449647
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,2048,16384,0.029561599095662434
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,2048,16384,0.03633066813151042
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,2048,12288,0.012865066528320312
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,2048,12288,0.027397332588831584
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,2048,12288,0.028568534056345622
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,2048,10240,0.0127210666735967
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,2048,10240,0.025809067487716674
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,2048,10240,0.02518080075581868
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,2048,8192,0.0129120002190272
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,2048,8192,0.02471253275871277
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,3584,128,0.0026496000587940215
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,2048,7168,0.011684266726175944
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,2048,7168,0.023619200785954794
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,2048,8192,0.021149865786234536
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,2048,6144,0.01018773317337036
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,2048,6144,0.02068693240483602
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,2048,7168,0.01943146586418152
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,2048,5120,0.009156266848246258
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,2048,6144,0.01773653427759806
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,2048,5120,0.020795732736587524
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,2048,4096,0.007635200023651123
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,2048,4096,0.018793600797653198
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,2048,4096,0.014098133643468222
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,2048,3584,0.007169066866238911
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,2048,3584,0.017816533644994102
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,2048,5120,0.015821866194407144
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,2048,3072,0.006377600133419037
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,2048,3072,0.01691626707712809
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,2048,3584,0.013338667154312134
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,2048,3072,0.01209173301855723
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,2048,2560,0.0058378666639328
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,2048,2560,0.017068799336751303
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,2048,2048,0.004638933142026265
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,2048,2048,0.017382399241129557
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,2048,2048,0.01043839951356252
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,2048,1536,0.004232533276081085
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,2048,2560,0.011199999849001567
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,2048,1536,0.01673706571261088
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,2048,1024,0.003409066547950109
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,2048,1024,0.015084800124168397
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,2048,1024,0.00842026670773824
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,2048,768,0.0030271999537944795
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,2048,768,0.015010133385658264
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,2048,1536,0.009597866733868917
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,2048,512,0.0029653333127498626
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,2048,512,0.014636799693107605
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,2048,768,0.008008533219496409
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,2048,512,0.007520000139872233
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,2048,256,0.0025898667673269907
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,2048,256,0.014513066411018372
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,2048,128,0.0022229333718617756
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,2048,128,0.014436266819636025
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,2048,256,0.0074869334697723385
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,2048,64,0.0021674667795499166
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,2048,128,0.007134933272997539
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,2048,64,0.014475733041763306
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,2048,32,0.014476799964904785
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,1536,65536,0.04108800093332927
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,1536,65536,0.055980801582336426
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,2048,65536,0.1267317295074463
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,1536,16384,0.014100266496340432
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,1536,16384,0.028651734193166096
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,1536,16384,0.03590720097223918
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,1536,12288,0.01165013313293457
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,1536,12288,0.025515733162562053
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,1536,12288,0.028941865762074786
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,1536,10240,0.010877866546312969
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,1536,10240,0.025012266635894776
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,1536,10240,0.025250132878621417
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,1536,8192,0.00985706647237142
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,1536,8192,0.023041067520777385
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,1536,8192,0.02155839999516805
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,1536,7168,0.008861866593360902
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,1536,7168,0.020794665813446044
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,1536,7168,0.019488000869750978
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,1536,6144,0.00832533339659373
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,1536,6144,0.02111146648724874
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,1536,65536,0.12612586816151936
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,1536,5120,0.00881706674893697
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,1536,5120,0.018837332725524902
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,1536,6144,0.01758400003115336
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,1536,4096,0.00716480016708374
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,1536,4096,0.018614399433135986
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,1536,5120,0.016119466225306193
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,1536,3584,0.006333866715431213
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,1536,3584,0.01794346570968628
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,1536,4096,0.01366933286190033
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,1536,3072,0.0058559998869895935
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,1536,3072,0.017032533884048462
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,1536,3584,0.012849066654841104
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,1536,3072,0.01202133297920227
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,1536,2560,0.0050453335046768185
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,1536,2560,0.018557866414388023
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,1536,2048,0.004632533093293508
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,1536,2048,0.016731733083724977
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,1536,2560,0.011161599556605022
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,1536,2048,0.010362666845321656
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,1536,1536,0.0042026668787002565
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,1536,1536,0.014574933052062988
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,1536,1024,0.0034485332667827605
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,1536,1024,0.014896000425020853
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,1536,1024,0.008798933029174805
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,1536,768,0.003389866650104523
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,1536,1536,0.009161600470542907
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,1536,768,0.015970133741696677
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,1536,512,0.002995199958483378
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,1536,512,0.014500266313552857
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,1536,768,0.008306133250395458
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,1536,512,0.0075434664885203045
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,1536,256,0.002584533393383026
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,1536,256,0.015200000007947287
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,1536,128,0.002540799975395203
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,1536,128,0.015026133259137472
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,1536,256,0.0070826664566993715
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,1536,64,0.002182399978240331
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,1536,128,0.007038933535416921
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,1536,64,0.014456533392270408
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,1536,32,0.002372266600529353
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,1536,32,0.014477866888046264
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,1024,65536,0.029290666182835896
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,1024,65536,0.04881386756896973
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,1024,16384,0.011198932925860088
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,1024,16384,0.02738453348477681
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,1024,16384,0.03591146469116211
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,1024,12288,0.012493866682052612
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,1024,12288,0.02532373269399007
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,1024,12288,0.028148265679677327
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,1024,10240,0.011319466431935628
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,1024,10240,0.024397865931193034
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,1024,10240,0.024784000714619954
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,1024,8192,0.00997226635615031
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,1024,8192,0.023433599869410196
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,1024,8192,0.02109760046005249
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,1024,7168,0.009225599964459737
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,1024,7168,0.02132800022761027
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,1024,7168,0.019383466243743895
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,1024,6144,0.008371200164159138
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,1024,6144,0.020457599560419717
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,1024,6144,0.017844265699386595
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,1024,5120,0.007569066683451335
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,1024,5120,0.019578667481740315
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,3072,128,0.00714026689529419
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,1024,4096,0.0067775999506314594
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,1024,4096,0.018580265839894614
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,1024,5120,0.016119466225306193
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,1024,3584,0.0062496001521746315
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,1024,3584,0.01699840029080709
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,1024,4096,0.014132266243298849
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,1024,3072,0.005550933380921682
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,1024,3072,0.017874133586883546
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,1024,3584,0.01279146671295166
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,1024,2560,0.005017599960168203
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,1024,2560,0.017130666971206666
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,1024,3072,0.012036266922950744
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,1024,2048,0.0047189335028330484
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,1024,2048,0.017846399545669557
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,1024,2560,0.01123306651910146
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,1024,1536,0.003885866701602936
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,1024,1536,0.016313599546750386
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,1024,2048,0.010445866982142131
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,1024,1024,0.003487999985615412
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,1024,1024,0.015280000368754067
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,1024,65536,0.12477333545684814
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,1024,1536,0.009196799993515015
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,1024,768,0.0030591999491055804
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,1024,768,0.014458666245142618
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,1024,512,0.002593066543340683
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,1024,512,0.014490666985511779
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,1024,1024,0.00876800020535787
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,1024,256,0.0025792000194390613
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,1024,768,0.00798826664686203
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,1024,512,0.007577600081761678
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,1024,256,0.014536533753077188
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,1024,128,0.0021909333765506743
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,1024,128,0.014455466469128927
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,1024,256,0.0070602665344874065
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,1024,64,0.0021727999051411946
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,1024,128,0.0071168000499407455
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,1024,64,0.01390506625175476
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,1024,32,0.0021664001047611236
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,1024,32,0.013358933726946512
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,768,65536,0.02477226654688517
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,768,65536,0.042777601877848306
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,768,16384,0.01199893355369568
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,768,16384,0.026719999313354493
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,768,12288,0.010618666807810467
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,768,12288,0.024756266673405965
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,768,16384,0.03559360106786092
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,768,12288,0.028508800268173217
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,768,10240,0.009194667140642803
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,768,10240,0.024578134218851723
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,768,8192,0.008411733309427898
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,768,8192,0.022949333985646567
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,768,8192,0.021104000012079873
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,768,7168,0.007521066566308339
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,768,7168,0.02208426594734192
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,768,10240,0.024800000588099162
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,768,6144,0.0071381335457166035
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,768,6144,0.020702934265136717
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,768,6144,0.01780479947725932
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,768,7168,0.019425066312154134
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,768,5120,0.0071274667978286745
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,768,5120,0.020804266134897866
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,768,4096,0.0066538666685422255
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,768,4096,0.018744534254074095
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,768,5120,0.016177067160606386
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,768,4096,0.014141866564750671
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,768,3584,0.00591786652803421
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,768,3584,0.018322134017944337
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,768,3072,0.005489066739877065
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,768,3072,0.018131200472513834
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,768,3584,0.012857600053151449
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,768,3072,0.01202133297920227
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,768,2560,0.005020800232887268
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,768,2560,0.01732053359349569
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,768,2048,0.0046847999095916745
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,768,2048,0.017375999689102174
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,768,2560,0.011291733384132386
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,768,1536,0.00383146678407987
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,768,2048,0.010045866171518963
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,768,1536,0.014918399850527444
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,768,1024,0.003419733295838038
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,768,1024,0.015468800067901611
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,768,1536,0.009577600161234538
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,768,1024,0.008303999900817871
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,768,768,0.0030378667016824085
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,768,768,0.014441600441932679
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,768,512,0.0026335999369621276
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,768,512,0.015217066804567973
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,768,768,0.008277333279450735
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,768,256,0.002616533388694127
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,768,512,0.007701333363850911
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,768,256,0.0147189329067866
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,768,128,0.0022485333184401194
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,768,128,0.01464959979057312
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,768,256,0.007150933146476746
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,768,64,0.002205866575241089
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,768,128,0.006743466854095459
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,768,64,0.014451199769973755
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,768,32,0.002254933367172877
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,768,32,0.014467199643452963
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,512,65536,0.01762239933013916
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,512,65536,0.04275733232498169
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,512,16384,0.010020266969998677
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,512,16384,0.024920533100763954
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,512,16384,0.03540800015131633
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,512,12288,0.008819199601809184
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,512,12288,0.02282879948616028
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,512,12288,0.028472532828648884
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,512,10240,0.0075573335091273
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,512,10240,0.02267626722653707
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,768,65536,0.12233920097351074
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,512,8192,0.007146666447321574
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,512,8192,0.022716800371805825
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,512,10240,0.024795732895533242
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,512,7168,0.006677333513895671
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,512,7168,0.021938133239746093
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,512,8192,0.02147946755091349
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,512,6144,0.00631466656923294
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,512,6144,0.020683733622233073
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,512,7168,0.01943040092786153
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,512,5120,0.006687999765078227
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,512,5120,0.01918399930000305
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,512,6144,0.01739733417828878
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,512,4096,0.006730666756629944
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,512,4096,0.018769067525863648
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,512,5120,0.016103466351826988
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,512,3584,0.006076799829800924
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,512,3584,0.01690559983253479
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,512,3584,0.01285653313000997
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,512,4096,0.013708800077438354
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,512,3072,0.0054848000407218935
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,512,3072,0.017874133586883546
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,512,3072,0.012086400389671325
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,512,2560,0.0050016000866889955
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,512,2560,0.01800533334414164
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,512,2560,0.011223466197649638
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,512,2048,0.004340266684691111
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,512,2048,0.015962666273117064
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,512,1536,0.003910399973392487
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,512,1536,0.017412267128626504
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,512,2048,0.01036906639734904
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,512,1024,0.003373866776625315
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,512,1024,0.015260799725850423
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,512,1536,0.00920960009098053
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,512,768,0.003049599876006444
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,512,768,0.015310933192571005
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,512,1024,0.008312533299128216
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,512,512,0.002632533262173335
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,512,768,0.007997866471608479
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,512,512,0.01503679951032003
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,512,256,0.0025802666942278544
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,512,256,0.014858667055765787
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,512,65536,0.12257920106252033
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,512,512,0.007525333265463512
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,512,128,0.0023029332359631857
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,512,256,0.007075199981530507
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,512,128,0.013651200135548911
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,512,64,0.002179199953873952
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,512,32,0.002186666677395503
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,512,64,0.012865066528320312
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,512,128,0.007131733496983846
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,512,32,0.013461333513259888
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,256,65536,0.015622400244077048
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,256,16384,0.006403199831644694
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,256,65536,0.03691946665445964
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,256,16384,0.024380799134572348
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,256,12288,0.006678399940331777
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,256,12288,0.022718934218088786
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,256,12288,0.028479999303817748
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,256,10240,0.006299733122189839
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,256,10240,0.02281600038210551
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,256,16384,0.03578773339589437
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,256,8192,0.005940266450246175
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,256,8192,0.02255893349647522
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,256,10240,0.02489173412322998
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,256,7168,0.005514666438102722
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,256,8192,0.02137920061747233
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,256,7168,0.020807466904322305
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,256,6144,0.005909333129723867
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,256,6144,0.020589866240819297
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,256,7168,0.019783467054367065
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,256,5120,0.006523733337720235
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,256,6144,0.017744000752766928
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,256,5120,0.020990933974583945
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,256,4096,0.006165333092212677
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,256,4096,0.019192532698313395
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,256,5120,0.01577279965082804
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,256,3584,0.00551146666208903
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,256,4096,0.014094932874043783
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,256,3584,0.018692266941070557
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,256,3072,0.005426133175690969
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,256,3072,0.016622933745384216
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,256,3584,0.0129120002190272
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,256,2560,0.005097599824269613
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,256,3072,0.01206933359305064
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,256,2560,0.016659200191497803
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,256,2048,0.004700799783070883
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,256,2048,0.01662506659825643
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,256,2560,0.011179733276367187
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,256,1536,0.0037791999677817024
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,256,2048,0.010033067067464192
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,256,1536,0.01643946667512258
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,256,1024,0.003066666672627131
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,256,1024,0.01513706644376119
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,256,1536,0.009537067015965779
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,256,768,0.003032533327738444
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,256,768,0.015169066190719605
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,256,1024,0.008370133241017659
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,256,512,0.00266239990790685
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,256,768,0.007993599772453308
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,256,512,0.014467199643452963
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,256,256,0.002194133400917053
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,256,256,0.014065066973368326
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,256,512,0.007565866907437642
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,256,128,0.0022122666239738466
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,256,256,0.007089066505432129
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,256,128,0.014546133081118264
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,256,64,0.0022101332743962605
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,256,128,0.007099733253320058
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,256,64,0.014460800091425577
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,256,32,0.0021557333568731946
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,128,65536,0.010770133137702942
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,256,32,0.014571733276049294
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,128,65536,0.032842665910720825
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,128,16384,0.005497600138187409
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,128,16384,0.02410986622174581
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,256,65536,0.12150719960530598
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,128,12288,0.005543466905752818
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,128,12288,0.022775467236836752
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,128,16384,0.03579093217849731
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,128,12288,0.028450133403142293
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,128,10240,0.0054613331953684485
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,128,10240,0.020812799533208214
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,128,10240,0.024717867374420166
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,128,8192,0.005468800167242686
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,128,8192,0.021245867013931274
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,128,7168,0.005643733342488607
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,128,8192,0.021027199427286782
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,128,7168,0.02072426676750183
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,128,6144,0.005515733361244201
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,128,6144,0.020433066288630168
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,128,65536,0.12220160166422527
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,128,5120,0.005052799979845682
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,128,5120,0.020360533396402994
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,128,7168,0.019348265727361043
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,128,4096,0.0054400001962979635
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,128,4096,0.01917653282483419
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,128,6144,0.01747093399365743
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,128,5120,0.015799466768900552
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,128,3584,0.005320533116658529
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,128,4096,0.013651200135548911
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,128,3072,0.005026133358478546
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,128,3584,0.018032000462214152
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,128,3072,0.016661333044370015
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,128,2560,0.005087999999523163
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,128,2560,0.0165802667538325
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,128,3584,0.012837333480517068
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,128,2048,0.005046399931112925
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,128,3072,0.011677866180737812
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,128,2048,0.01615466674168905
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,128,1536,0.0038122666378815973
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,128,2560,0.010759466886520385
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,128,1536,0.01664959987004598
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,128,2048,0.010032000144322713
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,128,1024,0.003017599880695343
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,128,1024,0.01511146624883016
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,128,768,0.0029781334102153777
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,128,1536,0.009105066458384197
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,128,1024,0.008310399949550629
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,128,768,0.015612799922625223
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,128,512,0.0025802666942278544
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,128,768,0.007894399762153625
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,128,512,0.01576533317565918
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,128,256,0.0022304000953833262
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,128,512,0.007462400197982788
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,128,256,0.016450132926305136
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,128,128,0.0021536000072956084
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,128,256,0.007086933155854543
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,128,128,0.014491732915242514
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,128,64,0.0021920000513394673
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,128,128,128,0.006723199784755707
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,128,32,0.002152533332506816
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,128,64,0.014497066537539164
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,64,65536,0.007500799993673961
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,128,32,0.014454399545987448
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,64,16384,0.005883733431498209
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,64,65536,0.03293333252271016
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,64,12288,0.005467733244101206
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,64,10240,0.0055189331372578945
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,64,16384,0.023844265937805177
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,64,12288,0.02328959902127584
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,64,8192,0.005845333139101664
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,64,7168,0.005445333321889242
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,64,10240,0.02081813414891561
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,64,8192,0.020754132668177286
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,64,6144,0.005072000126043955
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,64,7168,0.02028053402900696
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,64,5120,0.005026133358478546
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,64,6144,0.021237333615620933
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,64,5120,0.019475199778874717
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,64,4096,0.01909760038057963
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,64,3584,0.005046399931112925
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,64,3072,0.004694400231043497
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,64,3584,0.01795413295427958
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,64,3072,0.01753386656443278
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,64,2560,0.0050687998533248905
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,64,2048,0.004668800036112467
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,64,2560,0.016778665781021117
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,64,1536,0.0037920000652472176
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,64,2048,0.016681599617004394
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,64,1024,0.003049599876006444
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,64,1536,0.014589866995811463
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,64,768,0.0029898665845394133
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,64,1024,0.015227733055750528
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,64,512,0.002595199892918269
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,64,768,0.014453333616256715
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,64,256,0.002271999915440877
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,64,512,0.01662399967511495
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,64,128,0.0021898667017618816
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,64,256,0.01467626690864563
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,64,64,0.002145066608985265
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,64,128,0.014476799964904785
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,64,32,0.00198186660806338
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,64,64,0.014288000265757241
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,32,65536,0.008710400263468424
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,64,32,0.014413866400718688
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,32,16384,0.005847466488679251
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,32,65536,0.03283626635869344
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,32,16384,0.025037866830825806
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,32,12288,0.02327573299407959
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,32,10240,0.005570133527119955
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,32,10240,0.021873066822687783
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,32,8192,0.005573333303133646
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,32,8192,0.020977065960566203
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,32,7168,0.005514666438102722
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,32,7168,0.020840533574422202
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,32,6144,0.005154133339722951
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,32,6144,0.020411733786265054
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,32,5120,0.005094400048255921
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,32,5120,0.019118932882944743
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,32,4096,0.0047082667549451195
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,32,4096,0.0192138671875
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,32,3584,0.005076266825199127
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,32,3584,0.019246933857599895
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,32,3072,0.004653866589069367
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,32,3072,0.017483733097712197
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,32,2560,0.004791466891765595
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,32,2560,0.016771199305852254
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,32,2048,0.004625066618124644
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,32,2048,0.016546133160591125
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,32,1536,0.003790933390458425
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,32,1536,0.017939200003941856
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,32,1024,0.003065599997838338
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,32,1024,0.014659200112024942
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,32,768,0.002865066627661387
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,32,768,0.017270400126775106
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,32,512,0.002587733417749405
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,32,512,0.01455573340257009
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,32,256,0.0022570667167504626
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,32,256,0.014879999558130899
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,32,128,0.0021183999876181287
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,32,128,0.014653866489728292
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,32,64,0.0019776000330845515
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,32,64,0.014493866761525472
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,32,32,0.001833600054184596
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,128,32,32,0.014502400159835815
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,65536,16384,0.3756437301635742
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,65536,16384,0.20933012962341307
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,2048,32,0.002233600119749705
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,65536,12288,0.27259626388549807
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,65536,12288,0.16200106938680012
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,64,4096,0.004650666813055674
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,65536,10240,0.21087892850240073
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,65536,10240,0.12416213353474934
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,128,32,12288,0.005448533097902933
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,65536,8192,0.1711498737335205
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,65536,8192,0.10230293273925781
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,65536,12288,0.0853109359741211
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,65536,7168,0.14528853098551434
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,65536,7168,0.09305919806162516
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,65536,10240,0.07356053193410238
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,65536,6144,0.12946346600850422
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,65536,6144,0.08324159781138102
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,65536,16384,0.1116106669108073
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,65536,5120,0.1071786642074585
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,65536,5120,0.07282453378041585
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,65536,8192,0.06124693155288696
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,65536,4096,0.08665280342102051
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,65536,4096,0.06172586679458618
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,65536,6144,0.04915519952774048
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,65536,3584,0.07681600252787271
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,65536,3584,0.055785600344340006
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,65536,7168,0.054220799605051676
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,65536,3072,0.06696106592814127
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,65536,3072,0.049659732977549234
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,65536,5120,0.04263039827346802
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,65536,2560,0.057062399387359616
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,65536,2560,0.04417706727981567
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,65536,3584,0.031600000460942586
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,65536,2048,0.04879039923350016
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,65536,2048,0.04003413518269856
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,65536,3072,0.0285045325756073
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,65536,1536,0.03625813325246175
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,65536,1536,0.03441173235575358
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,65536,2560,0.02516053318977356
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,65536,1024,0.025624533494313557
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,65536,1024,0.02728640039761861
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,65536,4096,0.034687999884287515
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,65536,768,0.019978666305541994
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,65536,768,0.02384106715520223
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,65536,768,0.012970667084058127
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,65536,512,0.014106667041778565
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,65536,512,0.02249173323313395
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,65536,1024,0.014231466253598533
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,65536,256,0.008481066425641377
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,65536,256,0.01855573256810506
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,65536,256,0.009215999643007915
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,65536,128,0.00633493314186732
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,65536,128,0.017475199699401856
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,65536,128,0.008778666456540424
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,65536,64,0.005132799843947092
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,65536,64,0.018548266092936198
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,65536,32,0.005901866654555003
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,65536,32,0.018736000855763754
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,16384,65536,0.3263807932535807
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,16384,65536,0.2309823989868164
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,65536,1536,0.01831573247909546
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,16384,16384,0.08742612997690836
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,16384,16384,0.06450133323669434
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,65536,512,0.010915199915568035
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,16384,12288,0.06501226822535197
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,16384,12288,0.05268799861272176
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,65536,2048,0.0221834659576416
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,16384,10240,0.05615893205006918
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,16384,10240,0.04639999866485596
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,16384,10240,0.029344000418980914
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,16384,8192,0.045360000928243
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,16384,8192,0.04006293217341105
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,16384,12288,0.033668267726898196
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,16384,7168,0.039377065499623616
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,16384,7168,0.037147732575734456
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,16384,16384,0.041995731989542644
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,16384,6144,0.03466879924138387
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,16384,6144,0.03352533181508382
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,16384,8192,0.024845866362253825
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,16384,5120,0.029654399553934736
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,16384,5120,0.03141866723696391
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,16384,7168,0.02238933245340983
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,16384,4096,0.024000000953674317
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,16384,4096,0.02717973391215007
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,16384,5120,0.018221867084503175
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,16384,3584,0.021410133441289267
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,16384,3584,0.025491199890772503
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,16384,6144,0.020207999149958293
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,16384,3072,0.0187008003393809
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,16384,3072,0.024782933791478477
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,16384,4096,0.016189866264661155
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,16384,2560,0.016244266430536905
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,16384,2560,0.022486400604248048
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,16384,2560,0.012549333771069846
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,16384,2048,0.0134634663661321
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,16384,3584,0.014748799800872802
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,16384,2048,0.02063360015551249
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,16384,1536,0.01067200005054474
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,16384,1536,0.018565332889556883
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,16384,3072,0.013722667098045349
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,16384,1024,0.007969066500663757
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,16384,1024,0.018387200435002644
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,16384,1024,0.00846613347530365
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,16384,768,0.006674133241176605
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,16384,768,0.016603733102480568
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,16384,1536,0.00977066655953725
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,16384,512,0.004659200211366018
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,16384,512,0.01532373329003652
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,16384,768,0.008017066617806752
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,16384,256,0.0034272000193595886
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,16384,256,0.014573867122332254
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,16384,512,0.007949866851170858
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,16384,128,0.003035733352104823
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,16384,128,0.014545067151387533
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,16384,256,0.007549866537253062
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,16384,128,0.007470933099587758
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,16384,64,0.002643200010061264
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,16384,32,0.003032533327738444
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,16384,64,0.014460800091425577
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,16384,32,0.014566399653752646
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,12288,65536,0.263539203008016
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,12288,65536,0.1687317371368408
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,12288,16384,0.07424533367156982
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,12288,16384,0.05811626513799032
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,16384,2048,0.011220266421635944
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,12288,12288,0.05402026573816935
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,12288,12288,0.047006932894388835
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,12288,12288,0.0313973327477773
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,12288,10240,0.04683306614557902
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,12288,10240,0.04116586844126384
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,12288,16384,0.039496533075968426
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,12288,8192,0.03851093451182048
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,12288,8192,0.03672426541646322
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,12288,10240,0.028008532524108887
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,12288,7168,0.032686932881673174
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,12288,7168,0.03207146724065145
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,12288,8192,0.023155200481414794
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,12288,6144,0.029240532716115313
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,12288,6144,0.03138239979743958
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,12288,7168,0.02152000069618225
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,12288,5120,0.024923733870188394
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,12288,5120,0.027881600459416706
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,12288,6144,0.01954560081164042
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,12288,4096,0.020610133806864418
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,12288,4096,0.026872533559799194
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,12288,5120,0.017576533555984496
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,12288,3584,0.018055466810862224
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,12288,3584,0.02502506573994954
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,12288,4096,0.01535146633783976
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,12288,3072,0.015736533204714458
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,12288,3072,0.022744532426198324
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,12288,3584,0.014243200421333313
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,12288,2560,0.01368106702963511
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,12288,2560,0.023242666323979696
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,12288,3072,0.013333333532015481
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,12288,2048,0.012081066767374676
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,12288,2048,0.021717333793640138
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,12288,2560,0.012065066893895467
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,12288,1536,0.009668266773223877
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,12288,1536,0.019041067361831664
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,12288,2048,0.010507733623186747
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,12288,1024,0.007565866907437642
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,12288,1024,0.016681599617004394
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,12288,1536,0.009581866860389709
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,12288,768,0.0062047998110453285
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,12288,768,0.016664533813794454
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,12288,768,0.008365866541862488
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,12288,512,0.004706133405367533
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,12288,512,0.01593386630217234
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,12288,1024,0.008798933029174805
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,12288,256,0.0034773332377274835
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,12288,256,0.014436266819636025
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,12288,256,0.007499733567237854
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,12288,128,0.002998399982849757
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,12288,128,0.014472533265749613
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,12288,512,0.0075765331586201985
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,12288,64,0.0027434666951497394
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,12288,64,0.014640000462532044
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,12288,32,0.002958933264017105
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,12288,32,0.014547200004259745
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,12288,128,0.006683733562628429
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,10240,65536,0.2193354606628418
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,10240,65536,0.15309972763061525
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,10240,16384,0.06060053507486979
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,10240,16384,0.050910933812459314
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,10240,16384,0.03882559935251872
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,10240,12288,0.044674134254455565
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,10240,12288,0.04138559897740682
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,10240,12288,0.030897066990534468
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,10240,10240,0.0383296012878418
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,10240,10240,0.037801599502563475
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,16384,65536,0.14489706357320148
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,10240,8192,0.031725867589314775
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,10240,8192,0.03299306631088257
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,10240,10240,0.027297067642211913
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,10240,7168,0.027843199173609418
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,10240,7168,0.0308405339717865
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,10240,8192,0.022796799739201866
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,10240,6144,0.02467093269030253
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,10240,6144,0.029015467564264937
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,12288,65536,0.13307200272878011
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,10240,5120,0.02111146648724874
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,10240,5120,0.0252074658870697
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,10240,7168,0.021138133605321248
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,10240,4096,0.017217065890630087
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,10240,4096,0.023370667298634847
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,10240,6144,0.01948053240776062
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,10240,3584,0.014536533753077188
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,10240,3584,0.02276159922281901
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,10240,4096,0.015404799580574035
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,10240,3072,0.013674666484196981
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,10240,3072,0.02173653244972229
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,10240,5120,0.01778986652692159
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,10240,2560,0.011276800433794658
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,10240,2560,0.020692267020543418
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,10240,3584,0.014120533068974813
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,10240,2048,0.009616000453631084
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,10240,2048,0.020552533864974975
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,10240,2560,0.011666133006413778
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,10240,1536,0.008107733229796093
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,10240,1536,0.018093866109848023
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,10240,3072,0.012643200159072877
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,10240,1024,0.0062837332487106325
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,10240,1024,0.01867093245188395
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,10240,1024,0.008813866972923278
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,10240,768,0.005231999854246775
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,10240,768,0.017598932981491087
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,10240,1536,0.00960213343302409
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,10240,512,0.004278400043646494
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,10240,2048,0.010454400380452474
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,10240,512,0.014569600423177084
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,10240,256,0.0034624000390370687
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,10240,256,0.014479999740918478
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,10240,256,0.0075082664688428235
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,10240,128,0.002993066608905792
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,10240,128,0.014468266566594442
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,10240,128,0.006849066913127899
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,10240,64,0.0029898665845394133
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,10240,64,0.014459733168284097
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,10240,32,0.0033813332517941795
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,10240,512,0.0075541332364082335
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,10240,32,0.01446293294429779
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,8192,65536,0.173800532023112
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,8192,65536,0.12759466965993244
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,8192,16384,0.0499295989672343
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,8192,16384,0.044888532161712645
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,10240,768,0.008091733356316884
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,8192,12288,0.03631573518117269
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,8192,12288,0.03724693457285563
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,8192,12288,0.030089600880940752
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,8192,10240,0.03238506714502971
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,8192,10240,0.03321066697438558
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,8192,16384,0.03804266850153605
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,8192,8192,0.025917865832646686
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,8192,8192,0.028894933064778645
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,8192,10240,0.026133332649866742
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,8192,7168,0.0226197342077891
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,8192,7168,0.027436800797780353
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,8192,8192,0.022408533096313476
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,8192,6144,0.02004800041516622
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,8192,6144,0.025891200701395674
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,8192,7168,0.020581332842508952
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,8192,5120,0.01737706661224365
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,8192,5120,0.023468800385793052
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,10240,65536,0.12970026334126789
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,8192,4096,0.014036267002423605
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,8192,4096,0.022804266214370726
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,8192,6144,0.019077332814534505
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,8192,3584,0.012422399719556172
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,8192,3584,0.021194666624069214
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,8192,4096,0.014881066481272378
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,8192,3072,0.011001599828402202
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,8192,3072,0.021290665864944457
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,8192,5120,0.0165503998597463
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,8192,2560,0.009831466277440389
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,8192,2560,0.018802134195963542
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,8192,3584,0.014058666427930197
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,8192,2048,0.008334933718045553
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,8192,2048,0.019382399320602418
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,8192,2560,0.011597866813341778
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,8192,1536,0.0071381335457166035
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,8192,1536,0.016635732849438985
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,8192,3072,0.012494933605194092
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,8192,1024,0.005053866902987162
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,8192,1024,0.01663146714369456
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,8192,2048,0.010468266407648722
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,8192,768,0.004086400071779886
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,8192,768,0.016328533490498863
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,8192,768,0.007968000074227651
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,8192,512,0.0034111998975276947
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,8192,512,0.014672000209490457
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,8192,1024,0.008413867155710856
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,8192,256,0.0030805334448814393
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,8192,256,0.014428800344467163
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,8192,256,0.007493333518505096
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,8192,128,0.0026517334083716077
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,8192,1536,0.00958293378353119
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,8192,128,0.01474240024884542
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,8192,64,0.002619733413060506
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,8192,64,0.014852266510327658
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,8192,32,0.002625066787004471
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,8192,512,0.007918933530648549
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,8192,32,0.014567466576894126
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,8192,128,0.007178666690985362
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,7168,65536,0.14678293863932293
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,7168,65536,0.1139082670211792
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,7168,16384,0.04710719982783
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,7168,16384,0.04193066755930583
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,7168,12288,0.03496426741282145
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,7168,12288,0.03664960066477458
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,7168,12288,0.02969493269920349
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,7168,10240,0.030107732613881426
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,7168,10240,0.0312554657459259
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,7168,16384,0.037123199303944907
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,7168,8192,0.023427200317382813
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,7168,8192,0.028488532702128096
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,7168,10240,0.026054400205612182
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,7168,7168,0.020247467358907065
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,7168,7168,0.027074132363001508
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,7168,8192,0.022699733575185142
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,7168,6144,0.01816426714261373
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,7168,6144,0.024851200977961223
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,7168,7168,0.020862932999928793
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,7168,5120,0.016220800081888833
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,7168,5120,0.023705599705378215
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,7168,6144,0.018680532773335777
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,7168,4096,0.012897066275278726
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,7168,4096,0.02225173314412435
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,7168,5120,0.016964266697565712
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,7168,3584,0.012302933136622111
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,7168,3584,0.02094506621360779
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,7168,4096,0.014504533012708029
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,7168,3072,0.010428800185521444
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,7168,3072,0.020333866278330483
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,7168,3584,0.01332586705684662
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,7168,2560,0.009222400188446046
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,7168,2560,0.020165334145228066
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,7168,3072,0.012167466680208842
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,7168,2048,0.008044800162315369
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,7168,2048,0.01863893270492554
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,7168,2560,0.011708799997965496
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,7168,1536,0.006714666883150737
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,7168,1536,0.016636799772580466
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,7168,1536,0.00921493371327718
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,7168,1024,0.005231999854246775
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,7168,1024,0.01651946703592936
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,7168,2048,0.010079999764760334
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,7168,768,0.0042357335488001505
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,7168,768,0.017462400595347087
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,7168,1024,0.008455466230710347
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,7168,512,0.0038463999827702843
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,7168,512,0.014652799566586813
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,7168,768,0.007971199850241344
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,7168,256,0.0033781332274278007
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,7168,256,0.014495999614397685
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,7168,512,0.007546666761239369
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,7168,128,0.0029994666576385496
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,7168,128,0.0144896000623703
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,7168,128,0.006708266834417979
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,7168,64,0.0026208000878492994
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,7168,64,0.014517333110173544
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,7168,32,0.002946133414904277
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,7168,32,0.014482133587201438
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,6144,65536,0.12842240333557128
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,6144,65536,0.099726931254069
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,8192,65536,0.12860266367594403
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,6144,16384,0.03983893394470215
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,6144,16384,0.04154666662216187
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,6144,16384,0.0377781351407369
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,6144,12288,0.030980267127354938
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,6144,12288,0.03305066625277202
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,6144,12288,0.030108799537022907
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,6144,10240,0.02736639976501465
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,6144,10240,0.031229867537816362
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,7168,65536,0.12798293431599933
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,6144,8192,0.02111999988555908
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,6144,8192,0.027432533105214436
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,6144,10240,0.026405332485834758
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,6144,7168,0.01915839910507202
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,6144,7168,0.02720639904340108
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,6144,7168,0.020754132668177286
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,6144,6144,0.01712533235549927
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,6144,6144,0.02521386742591858
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,6144,8192,0.022341332832972207
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,6144,5120,0.014807466665903726
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,6144,5120,0.023280000686645506
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,6144,5120,0.0162282665570577
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,6144,4096,0.012034133076667786
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,6144,4096,0.021372799078623453
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,6144,6144,0.0191103994846344
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,6144,3584,0.011126400033632914
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,6144,3584,0.020829866329828896
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,6144,3584,0.013331199685732523
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,6144,3072,0.010028800368309021
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,6144,3072,0.020323199033737183
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,6144,4096,0.014133333166440328
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,6144,2560,0.008820266524950663
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,6144,2560,0.018990933895111084
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,6144,2560,0.0113045334815979
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,6144,2048,0.00754559983809789
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,6144,2048,0.01796906590461731
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,6144,3072,0.01209173301855723
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,6144,1536,0.006259199976921081
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,6144,1536,0.0167413334051768
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,6144,1536,0.00965333382288615
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,6144,1024,0.004621866842110952
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,6144,1024,0.016681599617004394
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,6144,2048,0.010441600282986959
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,6144,768,0.0038101332883040107
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,6144,768,0.016523733735084534
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,6144,1024,0.008424533406893413
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,6144,512,0.003470933437347412
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,6144,512,0.014587733149528503
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,6144,768,0.007946666578451793
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,6144,256,0.0029834667841593427
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,6144,256,0.014520532886187234
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,6144,256,0.00729066679875056
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,6144,128,0.0025888000925381976
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,6144,128,0.014526933431625366
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,6144,512,0.007585066556930542
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,6144,64,0.002628266563018163
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,6144,128,0.006659199794133504
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,6144,32,0.002587733417749405
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,6144,64,0.014563199877738953
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,6144,32,0.014497066537539164
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,5120,65536,0.11575573285420734
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,5120,65536,0.09398399988810222
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,5120,16384,0.03527786731719971
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,5120,16384,0.04025919834772746
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,6144,65536,0.12757546901702882
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,5120,12288,0.027695999542872114
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,5120,12288,0.030580266316731768
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,5120,16384,0.03628373146057129
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,5120,10240,0.025330134232838947
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,5120,10240,0.028505599498748778
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,7168,256,0.0074879998962084455
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,5120,8192,0.01755839983622233
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,5120,8192,0.025702399015426636
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,5120,12288,0.029420799016952513
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,5120,7168,0.015355733036994935
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,5120,7168,0.02545493245124817
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,5120,10240,0.02590186595916748
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,5120,6144,0.013804800311724343
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,5120,6144,0.022954666614532472
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,5120,8192,0.02195733388264974
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,5120,5120,0.012051199873288471
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,5120,5120,0.02235520084698995
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,5120,5120,0.01618773341178894
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,5120,4096,0.011653332908948263
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,5120,4096,0.02251840035120646
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,5120,7168,0.02030506730079651
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,5120,3584,0.01050453285376231
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,5120,3584,0.020687999327977498
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,5120,6144,0.018288000424702962
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,5120,3072,0.009360000491142273
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,5120,3072,0.0191648006439209
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,5120,3584,0.012964266538619994
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,5120,2560,0.00831573357184728
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,5120,2560,0.019107200702031455
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,5120,3072,0.012074666221936543
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,5120,2048,0.00732479989528656
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,5120,2048,0.01829119920730591
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,5120,4096,0.014154666662216186
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,5120,1536,0.006260266900062561
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,5120,1536,0.017992534240086875
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,5120,2048,0.010372266173362732
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,5120,1024,0.004640000065167745
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,5120,1024,0.015874133507410685
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,5120,2560,0.011270399888356526
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,5120,768,0.004299733539422353
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,5120,768,0.017026132345199584
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,5120,1536,0.00918293297290802
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,5120,512,0.003807999938726425
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,5120,512,0.014493866761525472
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,5120,768,0.007986133297284443
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,5120,256,0.0030080000559488933
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,5120,256,0.01456106702486674
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,5120,1024,0.008330666522185007
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,5120,128,0.002616533388694127
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,5120,512,0.007880533238252004
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,5120,128,0.01430506706237793
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,5120,64,0.002239999920129776
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,5120,64,0.01440000037352244
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,5120,32,0.0025642665723959604
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,5120,32,0.014452266693115234
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,5120,128,0.006785066425800323
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,5120,256,0.007513600091139476
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,4096,65536,0.093777068456014
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,4096,16384,0.028203733762105304
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,4096,65536,0.08064106305440268
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,4096,16384,0.035239466031392414
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,4096,12288,0.02390186587969462
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,4096,12288,0.030427734057108562
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,4096,12288,0.02964800000190735
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,4096,10240,0.01902079979578654
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,4096,10240,0.026808534065882367
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,4096,16384,0.03651306629180908
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,4096,8192,0.015079466501871744
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,4096,8192,0.023982934157053628
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,4096,10240,0.025814400116602583
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,4096,7168,0.013290666540463767
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,4096,7168,0.022669865687688192
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,4096,8192,0.021541333198547362
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,4096,6144,0.011763200163841248
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,4096,6144,0.021874133745829263
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,4096,7168,0.020310399929682414
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,4096,5120,0.010500267148017883
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,4096,5120,0.020885332425435384
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,4096,6144,0.01788053313891093
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,4096,4096,0.008809600273768108
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,4096,4096,0.020884267489115396
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,4096,5120,0.01574613352616628
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,4096,3584,0.008255999783674877
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,4096,3584,0.01861013372739156
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,4096,4096,0.013792000214258828
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,4096,3072,0.007493333518505096
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,4096,3072,0.018075732390085857
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,4096,3584,0.012945066889127096
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,4096,2560,0.0063274666666984555
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,4096,2560,0.016775466998418174
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,4096,3072,0.012102400263150532
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,4096,2048,0.005512533088525137
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,4096,2048,0.016501333316167197
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,4096,2560,0.011286399761835734
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,4096,1536,0.004324266811211904
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,4096,1536,0.015631999572118124
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,4096,2048,0.010069333513577779
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,4096,1024,0.0036703998843828833
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,4096,1024,0.016453333695729575
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,4096,1536,0.009134933352470398
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,4096,768,0.003390933324893316
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,4096,768,0.01635840038458506
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,4096,768,0.008358400066693623
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,4096,512,0.0029866665601730345
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,4096,512,0.014503467082977294
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,4096,1024,0.008297599852085114
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,4096,256,0.002621866762638092
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,4096,256,0.014453333616256715
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,4096,256,0.007144533097743988
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,4096,128,0.0025568000972270967
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,4096,128,0.013964800039927163
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,4096,512,0.0075082664688428235
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,4096,64,0.0022485333184401194
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,4096,64,0.01418880025545756
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,4096,32,0.0023018665611743926
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,4096,128,0.0070698668559392285
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,4096,32,0.01442346672217051
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,3584,65536,0.07847572962443033
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,3584,65536,0.07366399765014649
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,3584,16384,0.025989333788553875
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,3584,16384,0.03365439971288045
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,5120,65536,0.1278805335362752
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,3584,12288,0.020785067478815714
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,3584,12288,0.03116053342819214
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,3584,16384,0.03629120190938313
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,3584,10240,0.017578667402267455
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,3584,10240,0.02635519901911418
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,3584,12288,0.029604266087214153
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,3584,8192,0.016389333208402
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,3584,8192,0.02477653423945109
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,3584,10240,0.025600000222524004
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,3584,7168,0.01498240033785502
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,3584,7168,0.022936532894770302
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,3584,8192,0.021884800990422566
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,3584,6144,0.011704533298810323
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,3584,6144,0.02290346622467041
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,3584,6144,0.017819732427597046
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,3584,5120,0.010428800185521444
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,3584,5120,0.02061226765314738
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,3584,7168,0.019672532876332603
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,3584,4096,0.008887466788291932
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,3584,4096,0.01949119965235392
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,4096,65536,0.12707839806874593
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,3584,3584,0.008004266520341237
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,3584,3584,0.018902399142583213
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,3584,4096,0.014107732971509298
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,3584,3072,0.00876586635907491
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,3584,3072,0.017692800362904867
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,3584,5120,0.01580586632092794
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,3584,2560,0.0063498665889104204
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,3584,2560,0.017086933056513466
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,3584,2560,0.011276800433794658
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,3584,2048,0.005453866720199585
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,3584,3584,0.013298133015632629
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,3584,2048,0.01625279982884725
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,3584,1536,0.0043594668308893835
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,3584,1536,0.017205333709716795
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,3584,3072,0.012097066640853882
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,3584,1024,0.0035114665826161706
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,3584,1024,0.01662506659825643
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,3584,1024,0.008389332890510559
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,3584,768,0.003398400048414866
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,3584,768,0.015031466881434122
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,3584,1536,0.009636267026265462
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,3584,512,0.0029824001093705496
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,3584,512,0.014878933628400167
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,3584,768,0.008005333443482716
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,3584,2048,0.010332799951235453
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,3584,512,0.007574399809042613
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,3584,256,0.0026208000878492994
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,3584,128,0.002614400039116542
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,3584,256,0.015367466211318969
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,3584,128,0.014614400267601014
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,3584,64,0.002403199921051661
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,3584,64,0.014442666371663412
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,3584,32,0.0025162667036056517
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,3584,128,0.00720000018676122
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,3584,32,0.0144896000623703
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,3584,256,0.007533866663773854
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,3072,65536,0.0704085350036621
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,3072,65536,0.0696938673655192
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,3072,16384,0.022791467110315957
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,3072,16384,0.03104640046755473
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,3072,12288,0.01755839983622233
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,3072,12288,0.028266666332880656
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,3072,12288,0.02925119996070862
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,3072,10240,0.014748799800872802
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,3072,10240,0.026342399915059406
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,3072,16384,0.036266668637593584
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,3072,8192,0.013489066561063131
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,3072,8192,0.02440213362375895
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,3072,10240,0.025115732351938886
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,3072,7168,0.012135466933250428
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,3072,7168,0.023809067408243813
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,3072,8192,0.021219199895858763
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,3072,6144,0.0108842670917511
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,3072,6144,0.022006400426228843
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,3072,7168,0.01951893369356791
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,3072,5120,0.009679999947547913
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,3072,5120,0.020909865697224937
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,3072,6144,0.017463467518488564
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,3072,4096,0.00823466678460439
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,3072,4096,0.019078399737675986
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,3072,5120,0.015819733341534935
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,3072,3584,0.007570133109887441
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,3072,3584,0.018566399812698364
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,3072,4096,0.014123732844988504
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,3072,3072,0.006762666503588359
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,3072,3072,0.0180896004041036
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,3072,3584,0.012937600413958231
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,3072,2560,0.006009600063165029
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,3072,2560,0.0166101336479187
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,3584,65536,0.12742186387379964
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,3072,2048,0.005238399902979533
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,3072,2048,0.016797866423924765
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,3072,3072,0.012063999970753986
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,3072,1536,0.004328533510367076
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,3072,1536,0.016344533363978068
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,3072,1536,0.009611733754475911
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,3072,1024,0.0035061334570248926
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,3072,2560,0.011178666353225708
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,3072,1024,0.01527679959932963
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,3072,768,0.0034048000971476236
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,3072,768,0.014749866724014283
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,3072,2048,0.010356266299883525
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,3072,512,0.003009066730737686
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,3072,512,0.014473600188891092
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,3072,768,0.007957333326339721
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,3072,256,0.0025792000194390613
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,3072,256,0.014646400014559427
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,3072,1024,0.008719999591509502
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,3072,128,0.002242133269707362
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,3072,128,0.014490666985511779
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,3072,512,0.007913599908351897
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,3072,256,0.007480533421039581
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,3072,64,0.002250666668017705
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,3072,64,0.014485333363215128
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,3072,128,0.007074133555094402
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,3072,32,0.002186666677395503
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,3072,32,0.0144405335187912
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,2560,65536,0.05934826532999674
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,2560,65536,0.0648959994316101
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,2560,16384,0.019913599888483683
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,2560,16384,0.030352000395456952
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,2560,12288,0.016294399897257485
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,2560,12288,0.02802346746126811
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,2560,12288,0.028946133454640706
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,2560,10240,0.01458453337351481
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,2560,10240,0.02612373431523641
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,2560,16384,0.03629546562830607
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,2560,8192,0.011713066697120666
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,2560,8192,0.023004800081253052
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,2560,10240,0.025156267484029132
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,2560,7168,0.012138666709264119
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,2560,7168,0.023075199127197264
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,2560,8192,0.021104000012079873
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,2560,6144,0.010924800237019857
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,2560,6144,0.020745599269866945
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,2560,7168,0.019474132855733236
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,2560,5120,0.009258666634559631
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,2560,5120,0.020870399475097657
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,2560,6144,0.018152532974878947
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,2560,4096,0.008241066833337148
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,2560,4096,0.018870399395624796
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,2560,5120,0.01618133286635081
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,2560,3584,0.007582933207352956
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,2560,3584,0.018437333901723228
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,2560,4096,0.013684266805648803
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,2560,3072,0.006675200164318084
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,2560,3072,0.016591999928156534
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,2560,3072,0.01209920048713684
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,2560,2560,0.005894400179386139
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,2560,3584,0.012899200121561686
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,2560,2560,0.017131733894348144
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,2560,2048,0.005310933291912079
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,2560,2048,0.01585599978764852
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,2560,2048,0.010150399804115296
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,2560,1536,0.004369066655635833
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,2560,1536,0.015024000406265258
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,2560,2560,0.011256532867749532
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,2560,1024,0.0035317334036032355
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,2560,1024,0.015898666779200234
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,3072,65536,0.12725226879119872
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,2560,768,0.0030239999294281008
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,2560,768,0.014551466703414917
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,2560,1024,0.00839466651280721
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,2560,512,0.002914133419593175
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,2560,512,0.014501333236694336
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,2560,1536,0.009593600034713745
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,2560,256,0.0025920001169045764
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,2560,256,0.014127999544143677
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,2560,512,0.0079925333460172
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,2560,128,0.0022133332987626392
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,2560,256,0.007143466671307881
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,2560,128,0.014605866869290671
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,2560,64,0.0021749332547187803
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,2560,64,0.014501333236694336
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,2560,32,0.0022463999688625337
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,2560,128,0.007079466680685679
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,2560,32,0.014461867014567056
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,2048,65536,0.048537600040435794
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,2048,65536,0.05764586528142294
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,2048,16384,0.018257067600886027
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,2048,16384,0.030369067192077638
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,2048,16384,0.03634666601816813
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,2048,12288,0.014492799838383993
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,2048,12288,0.027025065819422406
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,2048,12288,0.028521599372227986
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,2048,10240,0.013066666324933371
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,2048,10240,0.027149866024653118
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,2560,65536,0.12689706484476726
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,2048,8192,0.011618133385976155
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,2048,8192,0.024871466557184856
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,2048,10240,0.025293866793314617
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,2048,7168,0.01072106659412384
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,2048,7168,0.02360639969507853
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,2048,7168,0.019450666507085164
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,2048,8192,0.02112213373184204
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,2048,6144,0.009902933239936828
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,2048,6144,0.022430932521820067
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,2048,5120,0.011633066336313884
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,2048,5120,0.02068159977595011
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,2048,5120,0.016229333480199178
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,2048,4096,0.009946667154630025
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,2048,4096,0.020590933163960774
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,2048,6144,0.017496534188588462
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,2048,3584,0.008750933408737182
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,2048,3584,0.0181386669476827
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,2048,4096,0.014089600245157877
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,2048,3584,0.0128330667813619
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,2048,3072,0.007896533111731212
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,2048,3072,0.018743467330932618
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,2048,2560,0.00699946681658427
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,2048,2560,0.016824533541997276
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,2048,3072,0.012029866377512615
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,2048,2048,0.005909333129723867
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,2048,2048,0.017242666085561117
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,2048,2560,0.011271466811498005
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,2048,1536,0.0043136000633239744
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,2048,2048,0.01035520037015279
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,2048,1536,0.01711146632830302
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,2048,1024,0.0035018667578697203
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,2048,1024,0.017068799336751303
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,2048,1024,0.008348799745241801
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,2048,768,0.0033749334514141084
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,2048,768,0.016472533345222473
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,2048,1536,0.009199999769528707
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,2048,512,0.0030250666042168934
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,2048,512,0.014497066537539164
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,2048,512,0.007699200014273326
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,2048,768,0.008291199803352356
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,2048,256,0.0026101333399613695
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,2048,256,0.01479573349157969
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,2048,128,0.0022250667214393617
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,2048,128,0.014536533753077188
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,2048,256,0.007485866546630859
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,2048,64,0.0022069332500298817
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,2048,128,0.00714026689529419
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,2048,32,0.0022613334159056345
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,2048,64,0.014556800325711569
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,2048,32,0.014507733782132468
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,1536,65536,0.040250666936238605
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,1536,65536,0.05257920026779175
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,1536,16384,0.014095999797185264
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,1536,16384,0.028893866141637164
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,2048,65536,0.1266976038614909
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,1536,12288,0.011614933609962463
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,1536,12288,0.026245333751042682
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,1536,16384,0.03598399957021077
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,1536,10240,0.012183466553688049
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,1536,10240,0.024966400861740113
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,1536,10240,0.02511253356933594
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,1536,8192,0.01090666651725769
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,1536,8192,0.02262293299039205
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,1536,12288,0.028409600257873535
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,1536,7168,0.010062932968139648
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,1536,7168,0.020794665813446044
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,1536,7168,0.01955519914627075
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,1536,6144,0.00918506681919098
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,1536,6144,0.018810667594273887
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,1536,8192,0.020716800292332967
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,1536,5120,0.00853013296922048
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,1536,5120,0.0186463991800944
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,1536,6144,0.0178272008895874
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,1536,4096,0.007089066505432129
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,1536,4096,0.017825067043304443
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,1536,5120,0.016190933187802632
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,1536,3584,0.006585599978764851
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,1536,3584,0.018023467063903807
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,1536,3584,0.012862933675448099
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,1536,3072,0.00591893345117569
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,1536,3072,0.018399999539057414
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,1536,4096,0.013719466328620911
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,1536,2560,0.005211733281612396
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,1536,2560,0.01665493349234263
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,1536,2560,0.01134933332602183
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,1536,2048,0.0046186665693918865
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,1536,3072,0.011638399958610535
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,1536,2048,0.016935465733210246
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,1536,1536,0.003866666555404663
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,1536,1536,0.016753067572911583
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,1536,1536,0.009604266285896302
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,1536,1024,0.0034101332227389016
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,1536,1024,0.014478933811187745
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,1536,2048,0.010361599922180175
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,1536,768,0.0030933332939942675
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,1536,768,0.014903466900189719
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,1536,1024,0.008333866794904072
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,1536,512,0.0029909332593282064
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,1536,768,0.007956266899903615
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,1536,512,0.014495999614397685
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,1536,256,0.002600533266862233
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,1536,256,0.013633066415786743
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,1536,512,0.007893333335717519
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,1536,128,0.002232533444960912
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,1536,256,0.007483733197053273
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,1536,128,0.014309333761533103
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,1536,64,0.0022528000175952912
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,1536,128,0.007073066631952922
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,1536,32,0.0021461332837740582
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,1536,64,0.01461120049158732
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,1536,32,0.013413332899411521
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,1024,65536,0.02839360038439433
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,1024,65536,0.04783466657002767
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,1024,16384,0.012562132875124612
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,1024,16384,0.02687679926554362
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,1024,16384,0.03532586495081584
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,1024,12288,0.010858666896820069
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,1024,12288,0.025643734137217204
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,1024,12288,0.028101332982381183
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,1024,10240,0.0100490669409434
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,1024,10240,0.024602667490641276
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,1536,65536,0.12625280221303303
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,1024,8192,0.009215999643007915
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,1024,8192,0.023504000902175904
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,1024,10240,0.024782933791478477
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,1024,7168,0.010365866621335347
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,1024,7168,0.020810665686925252
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,1024,8192,0.021112533410390218
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,1024,6144,0.009554133812586466
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,1024,6144,0.020965333779652914
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,1024,7168,0.019461333751678467
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,1024,5120,0.008727467060089112
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,1024,5120,0.019041067361831664
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,1024,6144,0.017820799350738527
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,1024,4096,0.007964799801508587
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,1024,4096,0.019284266233444213
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,1024,5120,0.01578986644744873
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,1024,3584,0.006259199976921081
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,1024,3584,0.017939200003941856
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,1024,4096,0.013686399658521017
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,1024,3072,0.005827199916044871
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,1024,3584,0.012839466333389282
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,1024,3072,0.01816320021947225
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,1024,2560,0.005081599950790406
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,1024,2560,0.019052799542744955
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,1024,2560,0.010873599847157796
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,1024,2048,0.005584000051021576
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,1024,3072,0.012018133203188579
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,1024,2048,0.0171509325504303
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,1024,1536,0.004881066580613455
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,1024,1536,0.016432000199953715
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,1024,1536,0.00921066701412201
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,1024,1024,0.0038506666819254553
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,1024,2048,0.009946667154630025
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,1024,1024,0.014699733257293702
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,1024,768,0.0034527999659379324
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,1024,768,0.014544000228246054
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,1024,1024,0.008488532900810242
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,1024,768,0.008017066617806752
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,1024,512,0.002993066608905792
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,1024,512,0.01595200002193451
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,1024,256,0.0026506667335828146
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,1024,256,0.014532267053922018
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,1024,512,0.007099733253320058
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,1024,128,0.002251733342806498
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,1024,256,0.007150933146476746
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,1024,128,0.014430933197339377
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,1024,64,0.0021749332547187803
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,1024,128,0.007065600156784058
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,1024,64,0.014465066790580749
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,1024,32,0.002619733413060506
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,1024,32,0.014460800091425577
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,768,65536,0.02461973428726196
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,768,65536,0.03938026825586955
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,768,16384,0.014962133765220643
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,768,16384,0.026971733570098876
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,1024,65536,0.12359466552734374
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,768,12288,0.012457600235939026
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,768,12288,0.024899200598398844
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,768,16384,0.03551466862360637
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,768,10240,0.01093226671218872
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,768,10240,0.022383999824523926
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,768,10240,0.02480213244756063
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,768,8192,0.009588266412417095
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,768,12288,0.028460800647735596
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,768,8192,0.02136533260345459
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,768,7168,0.009187199672063192
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,768,7168,0.020835200945536293
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,768,7168,0.019103999932607016
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,768,6144,0.008295466502507527
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,768,6144,0.0198527991771698
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,768,8192,0.02109760046005249
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,768,5120,0.007925333579381307
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,768,5120,0.020824533700942994
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,768,6144,0.01736853321393331
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,768,5120,0.01612160007158915
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,768,4096,0.006701866785685222
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,768,4096,0.01853440006573995
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,768,3584,0.006300800045331319
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,768,3584,0.01883946657180786
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,768,3584,0.012851199507713318
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,768,3072,0.005851733187834421
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,768,3072,0.017511467138926186
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,768,4096,0.01409066617488861
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,768,2560,0.005095466474692027
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,768,2560,0.01665066679318746
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,768,2560,0.011240532994270325
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,768,2048,0.004636799792448679
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,768,3072,0.011971199512481689
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,768,2048,0.017271467049916587
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,768,1536,0.003798400113979975
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,768,1536,0.01560533344745636
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,768,1536,0.009150933225949604
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,768,2048,0.00997759997844696
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,768,1024,0.003454933315515518
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,768,1024,0.017578667402267455
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,768,768,0.0030154667794704436
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,768,768,0.014586666226387024
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,768,1024,0.008368000388145447
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,768,512,0.002585600068171819
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,768,768,0.00795413355032603
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,768,512,0.014477866888046264
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,768,512,0.007546666761239369
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,768,256,0.0025397333006064097
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,768,256,0.014509866635004679
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,768,256,0.007131733496983846
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,768,128,0.002233600119749705
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,768,128,0.014460800091425577
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,768,128,0.007132799923419952
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,768,64,0.0022367998957633973
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,768,64,0.014167466759681701
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,768,32,0.002253866692384084
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,768,32,0.014548266927401224
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,512,65536,0.017211733261744182
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,512,65536,0.0392138679822286
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,512,16384,0.011043199896812439
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,512,16384,0.02489173412322998
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,512,16384,0.03511999845504761
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,512,12288,0.009211732943852743
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,512,12288,0.022714666525522866
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,768,65536,0.12397119998931885
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,512,10240,0.00872320036093394
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,512,10240,0.02264960010846456
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,512,10240,0.024581333001454674
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,512,8192,0.007956266899903615
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,512,12288,0.0279968003431956
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,512,8192,0.021449599663416544
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,512,7168,0.007470933099587758
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,512,7168,0.021271467208862305
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,512,7168,0.019451733430226645
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,512,6144,0.006324266890684764
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,512,6144,0.020787199338277183
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,512,8192,0.021092265844345093
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,512,5120,0.006701866785685222
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,512,5120,0.01915093262990316
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,512,5120,0.016115199526151022
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,512,6144,0.017375999689102174
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,512,4096,0.006684799989064534
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,512,4096,0.018914133310317993
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,512,3584,0.0068256000677744556
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,512,3584,0.0171509325504303
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,512,4096,0.013277866442998252
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,512,3584,0.012821333607037863
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,512,3072,0.005458133419354757
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,512,3072,0.01730453372001648
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,512,2560,0.005073066552480062
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,512,2560,0.017033600807189943
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,512,3072,0.012034133076667786
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,512,2048,0.004644266764322917
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,512,2560,0.01084266702334086
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,512,2048,0.01653866668542226
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,512,1536,0.0037952000896135964
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,512,1536,0.014856533209482829
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,512,2048,0.010256000359853109
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,512,1024,0.003386666625738144
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,512,1536,0.0091839998960495
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,512,1024,0.016922666629155477
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,512,768,0.003070933371782303
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,512,768,0.016774400075276693
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,512,1024,0.007960533102353413
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,512,512,0.0026346666117509207
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,512,512,0.014468266566594442
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,512,768,0.008152533570925396
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,512,256,0.0025397333006064097
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,512,256,0.014716800053914389
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,512,512,0.007467733323574066
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,512,128,0.002219733347495397
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,512,256,0.007099733253320058
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,512,128,0.0138154665629069
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,512,64,0.0021759999295075734
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,512,128,0.006734933455785115
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,512,32,0.0022272000710169475
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,512,32,0.014510933558146158
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,256,65536,0.015457066893577575
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,256,65536,0.03660586675008138
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,512,65536,0.12228480180104573
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,256,16384,0.007563733557860057
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,256,16384,0.022966400782267252
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,256,16384,0.035454932848612467
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,256,12288,0.006728533407052357
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,256,12288,0.02339413364728292
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,256,12288,0.028025599320729573
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,256,10240,0.006325333317120869
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,256,10240,0.022632533311843873
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,256,10240,0.024741333723068238
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,256,8192,0.005931733548641205
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,256,8192,0.02177066604296366
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,256,65536,0.12116586367289225
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,256,7168,0.005963733295599619
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,256,7168,0.02123519976933797
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,256,8192,0.02108373244603475
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,256,6144,0.006285866598288219
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,256,6144,0.020498132705688475
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,256,7168,0.019001599152882895
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,256,5120,0.0067669332027435304
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,256,6144,0.01774186690648397
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,256,5120,0.019411200284957887
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,256,4096,0.006212266782919565
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,256,4096,0.01889280080795288
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,256,5120,0.01570453345775604
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,256,3584,0.005894400179386139
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,256,4096,0.014014933506647745
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,256,3584,0.018348799149195353
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,256,3584,0.01292693316936493
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,256,3072,0.005498666564623515
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,256,3072,0.01806186636288961
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,256,3072,0.012114133437474568
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,256,2560,0.005471999943256378
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,256,2560,0.016822399695714314
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,256,2560,0.01125973363717397
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,256,2048,0.005054933329423269
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,256,2048,0.015868799885114034
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,256,1536,0.0038965334494908653
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,256,2048,0.010018133123715718
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,256,1536,0.016641066471735636
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,256,1024,0.0034186666210492453
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,256,1024,0.015340800086657206
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,256,1536,0.00956053336461385
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,256,768,0.0029824001093705496
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,256,768,0.015870933731396995
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,256,1024,0.008434133728345235
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,256,512,0.0026922665536403658
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,256,768,0.007877333462238312
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,256,512,0.014541866381963095
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,256,512,0.007567999760309856
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,256,256,0.014421332875887552
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,256,256,0.007098666826883952
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,256,128,0.0022069332500298817
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,256,128,0.014470400412877402
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,256,128,0.006763733426729838
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,256,64,0.002229333420594533
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,256,64,0.014622933665911355
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,256,32,0.0021674667795499166
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,256,32,0.014453333616256715
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,128,65536,0.010382933417956035
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,128,65536,0.0329365332921346
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,128,65536,0.12177706559499105
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,128,16384,0.005861333509286245
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,128,16384,0.022798933585484824
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,128,16384,0.03568533261617025
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,128,12288,0.005464533468087515
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,128,12288,0.021052799622217813
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,128,12288,0.028008532524108887
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,128,10240,0.005868799984455109
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,128,10240,0.020746666193008422
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,128,10240,0.024701867500940958
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,128,8192,0.005542399982611338
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,128,8192,0.020651733875274657
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,128,8192,0.021065600713094077
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,128,7168,0.005529599885145823
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,128,7168,0.020809600750605263
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,128,7168,0.019398399194081626
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,128,6144,0.005086933573087057
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,128,6144,0.02026026646296183
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,128,6144,0.01776319940884908
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,128,5120,0.005026133358478546
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,128,5120,0.019942400852839152
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,128,5120,0.015715199708938598
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,128,4096,0.005457066496213277
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,128,4096,0.018729599316914876
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,128,4096,0.013666133085886637
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,128,3584,0.00505920002857844
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,128,3584,0.0173418660958608
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,128,3584,0.01286186675230662
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,128,3072,0.005030400057633718
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,128,3072,0.017986132701237997
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,128,3072,0.01204266647497813
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,128,2560,0.005094400048255921
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,128,2560,0.017240534226099648
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,128,2560,0.010828800002733866
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,128,2048,0.0050335998336474095
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,128,2048,0.01650773286819458
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,128,2048,0.009947733084360758
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,128,1536,0.003748266647259394
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,128,1536,0.016103466351826988
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,128,1536,0.009317333499590557
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,128,1024,0.003047466774781545
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,128,1024,0.015657599767049155
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,128,1024,0.007975466549396515
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,128,768,0.0030229332546393077
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,128,768,0.014644267161687216
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,128,768,0.007863466441631318
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,128,512,0.002632533262173335
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,128,512,0.014657066265741984
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,128,512,0.007569066683451335
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,128,256,0.002290133386850357
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,128,256,0.014504533012708029
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,128,256,0.007050666709740956
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,128,128,0.0022069332500298817
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,128,128,0.014485333363215128
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,128,128,0.006720000008742015
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,128,64,0.002088533341884613
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,128,64,0.01455573340257009
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,128,32,0.0022240000466505687
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,128,32,0.014469333489735923
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,64,65536,0.009511466821034749
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,64,65536,0.031193600098292036
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,64,16384,0.005926399926344554
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,64,16384,0.023319466908772787
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,64,12288,0.00552106648683548
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,64,12288,0.02253119945526123
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,64,10240,0.005508266886075338
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,64,10240,0.02140373388926188
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,64,8192,0.005577600002288819
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,64,8192,0.0209279994169871
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,64,7168,0.005515733361244201
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,64,7168,0.02002453406651815
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,64,6144,0.005172266562779745
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,64,6144,0.018897066513697304
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,64,5120,0.0050794666012128195
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,64,5120,0.019755733013153077
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,64,4096,0.004704000055789947
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,64,4096,0.01878933310508728
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,64,3584,0.005089066425959269
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,64,3584,0.018413867553075156
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,64,3072,0.004730666677157084
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,64,3072,0.01688213348388672
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,64,2560,0.005018666883309682
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,64,2560,0.016687999169031777
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,64,2048,0.004714666803677877
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,64,2048,0.016479999820391337
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,64,1536,0.003538133452335993
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,64,1536,0.015506133437156677
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,64,1024,0.003092266619205475
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,64,1024,0.015145599842071533
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,64,768,0.0029696000119050344
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,64,768,0.014549332857131957
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,64,512,0.0026464000344276427
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,64,512,0.014512000481287637
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,64,256,0.0022143999735514323
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,64,256,0.015613866845766702
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,64,128,0.002251733342806498
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,64,128,0.014447999993960061
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,64,64,0.002065066620707512
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,64,64,0.014446933070818582
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,64,32,0.002103466788927714
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,64,32,0.014113066593805948
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,32,65536,0.008780800302823384
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,32,65536,0.031139200925827025
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,32,16384,0.005897599955399831
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,32,16384,0.02286400000254313
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,32,12288,0.005142400165398916
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,32,12288,0.021305600802103676
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,32,10240,0.005866666634877523
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,32,10240,0.021319466829299926
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,32,8192,0.005487999816735586
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,32,8192,0.020307199160257975
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,32,7168,0.005494399865468343
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,32,7168,0.020214400688807168
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,32,6144,0.005087999999523163
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,32,6144,0.02016213337580363
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,32,5120,0.005006933212280273
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,32,5120,0.01944533387819926
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,32,4096,0.004622933268547058
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,32,4096,0.018744534254074095
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,32,3584,0.004653866589069367
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,32,3584,0.017027199268341064
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,32,3072,0.004602666695912679
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,32,3072,0.016701867183049522
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,32,2560,0.004702933132648468
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,32,2560,0.016816000143686928
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,32,2048,0.004643199841181437
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,32,2048,0.015402666727701821
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,32,1536,0.003822933385769526
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,32,1536,0.016429866353670754
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,32,1024,0.0029909332593282064
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,32,1024,0.015666133165359496
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,32,768,0.002609066665172577
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,32,768,0.01477226714293162
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,32,512,0.002515200028816859
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,32,512,0.014647466937700906
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,32,256,0.0022005334496498106
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,32,256,0.014388266205787658
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,32,128,0.0022101332743962605
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,32,128,0.015309866269429526
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,32,64,0.002091733366250992
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,32,64,0.014454399545987448
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,32,32,0.0018112000077962875
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,32,32,0.014439466595649719
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,65536,16384,0.3736778577168783
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,65536,16384,0.2017130692799886
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,96,2560,768,0.008269866804281871
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,65536,12288,0.27244908014933267
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,65536,12288,0.15818559328715007
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,96,512,64,0.014732799927393594
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,65536,10240,0.21167680422465004
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,65536,10240,0.12448639869689941
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,96,256,256,0.0023071999351183575
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,65536,8192,0.17165333429972332
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,65536,8192,0.10375786622365316
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,65536,12288,0.08606932957967123
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,65536,7168,0.14800853729248048
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,65536,7168,0.09269333680470784
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,65536,8192,0.0605674664179484
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,65536,6144,0.1284063975016276
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,65536,6144,0.08176320393880208
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,65536,10240,0.07291306654612223
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,65536,5120,0.10823893547058105
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,65536,5120,0.07070293426513671
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,65536,16384,0.11188586552937825
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,65536,4096,0.08821973005930582
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,65536,4096,0.06215466658274332
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,65536,5120,0.04149653514226277
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,65536,3584,0.07696213722229003
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,65536,3584,0.05530666510264078
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,65536,6144,0.048189866542816165
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,65536,3072,0.06665493249893188
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,65536,3072,0.04991573492685954
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,65536,7168,0.0542357325553894
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,65536,2560,0.05642026662826538
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,65536,2560,0.04412480195363362
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,65536,4096,0.03669973214467366
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,65536,2048,0.04655040105183919
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,65536,2048,0.04062826633453369
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,65536,3584,0.031668267647425336
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,65536,1536,0.035656531651814774
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,65536,1536,0.03344853321711223
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,65536,3072,0.028907734155654907
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,65536,1024,0.025496532519658405
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,65536,1024,0.026763733228047686
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,65536,2560,0.026054400205612182
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,65536,768,0.01990613341331482
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,65536,768,0.025187200307846068
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,65536,1536,0.018454400698343913
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,65536,512,0.013644799590110779
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,65536,512,0.020803199211756388
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,65536,1024,0.01421119968096415
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,65536,256,0.008794666330019633
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,65536,256,0.01832853356997172
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,65536,768,0.012989866733551025
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,65536,128,0.005771733323733012
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,65536,128,0.01651413341363271
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,65536,2048,0.021771732966105142
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,65536,64,0.004695466657479604
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,65536,256,0.00920960009098053
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,65536,32,0.005502933263778686
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,65536,64,0.018667733669281004
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,65536,32,0.018685867389043175
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,16384,65536,0.313431453704834
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,16384,65536,0.21112000147501625
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,16384,16384,0.08437120119730632
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,65536,128,0.009164800246556599
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,16384,16384,0.06179093519846598
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,16384,12288,0.06172053416570028
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,16384,12288,0.05111253261566162
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,65536,512,0.010893866419792175
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,16384,10240,0.05312960147857666
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,16384,10240,0.045878398418426516
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,16384,10240,0.029180800914764403
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,16384,8192,0.04301120042800903
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,16384,8192,0.038533333937327066
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,16384,12288,0.03339946667353312
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,16384,7168,0.03776853481928508
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,16384,7168,0.03608746528625488
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,16384,16384,0.041934935251871745
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,16384,6144,0.03326293428738912
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,16384,6144,0.03242133259773254
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,16384,8192,0.024820266167322795
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,16384,5120,0.02845120032628377
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,16384,5120,0.029018666346867877
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,16384,7168,0.022347732384999593
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,16384,4096,0.023130667209625245
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,16384,4096,0.026654932896296186
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,16384,5120,0.01810773412386576
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,16384,3584,0.0206389327843984
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,16384,3584,0.02579306761423747
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,16384,4096,0.01636373301347097
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,16384,3072,0.018195199966430663
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,16384,3072,0.02473919987678528
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,16384,6144,0.019897600015004478
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,16384,2560,0.0153546671072642
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,16384,2560,0.022372267643610635
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,16384,2560,0.012868266304334006
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,16384,2048,0.012519466876983642
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,16384,2048,0.02109439969062805
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,16384,3584,0.014961066842079162
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,16384,1536,0.010012800494829815
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,16384,3072,0.013665067156155905
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,16384,1536,0.01863573392232259
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,16384,1024,0.007670400043328603
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,16384,1024,0.016541866461435954
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,16384,1024,0.008763733506202697
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,16384,768,0.0063274666666984555
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,16384,768,0.015597866972287497
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,16384,1536,0.010599467158317565
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,16384,512,0.004709333181381226
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,16384,512,0.018580265839894614
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,16384,768,0.008398933211962382
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,16384,256,0.0034261333445707956
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,16384,256,0.014487466216087342
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,16384,2048,0.011325866977373759
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,16384,128,0.0030026666820049284
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,16384,128,0.014773333072662353
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,16384,128,0.00718506673971812
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,16384,64,0.0025792000194390613
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,16384,64,0.01455466647942861
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,16384,32,0.0029440000653266907
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,16384,32,0.014475733041763306
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,16384,512,0.00788800021012624
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,16384,256,0.0071733335653940845
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,12288,65536,0.25025067329406736
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,12288,16384,0.07273920377095541
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,12288,65536,0.1659850597381592
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,12288,16384,0.05603839953740438
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,12288,12288,0.053642666339874266
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,12288,12288,0.045499734083811444
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,12288,12288,0.031275733311971025
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,12288,10240,0.04586666822433472
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,12288,10240,0.04159786701202393
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,12288,16384,0.039666132132212324
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,12288,8192,0.037596801916758224
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,12288,8192,0.035359998544057206
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,12288,10240,0.027666133642196656
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,12288,7168,0.03253226677576701
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,12288,7168,0.033048532406489056
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,12288,8192,0.023205333948135377
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,12288,6144,0.028918399413426714
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,12288,6144,0.030779733260472612
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,12288,7168,0.021065600713094077
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,12288,5120,0.024921600023905435
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,12288,5120,0.027061333258946735
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,12288,6144,0.0199072003364563
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,12288,4096,0.019847466548283895
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,12288,4096,0.02573973337809245
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,12288,5120,0.017399466037750243
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,12288,3584,0.017879466215769448
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,12288,3584,0.024540799856185912
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,12288,4096,0.015387733777364096
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,12288,3072,0.015401599804560342
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,12288,3072,0.023372799158096313
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,12288,3584,0.014405333002408347
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,12288,2560,0.013448533415794373
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,12288,2560,0.021597866217295328
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,12288,3072,0.0134634663661321
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,12288,2048,0.01165013313293457
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,12288,2048,0.020669867595036827
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,12288,2560,0.011716266473134358
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,12288,1536,0.009408000111579894
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,12288,1536,0.01861013372739156
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,12288,2048,0.010462933778762817
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,12288,1024,0.007463466624418895
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,12288,1024,0.018221867084503175
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,12288,1536,0.009641599655151368
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,12288,768,0.00592853327592214
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,12288,768,0.01536853313446045
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,12288,1024,0.0088128000497818
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,12288,512,0.0047775998711586
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,12288,512,0.015687466661135355
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,12288,768,0.007965866724650066
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,12288,256,0.003469866762558619
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,12288,256,0.014813866217931113
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,12288,256,0.0071946665644645694
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,12288,128,0.003020799905061722
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,12288,128,0.014451199769973755
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,12288,512,0.007529599964618683
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,12288,64,0.0027029333015282948
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,12288,64,0.014803199966748556
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,12288,32,0.0030239999294281008
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,12288,32,0.014678399761517844
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,10240,65536,0.2152714729309082
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,12288,128,0.006662400066852569
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,10240,65536,0.15016214052836102
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,10240,16384,0.0602570652961731
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,10240,16384,0.05035200119018555
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,10240,16384,0.038726401329040525
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,10240,12288,0.04488319953282674
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,10240,12288,0.041788800557454424
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,10240,12288,0.030345600843429566
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,10240,10240,0.038262399037679036
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,10240,10240,0.037215999762217206
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,16384,65536,0.1464853286743164
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,10240,8192,0.03188479940096538
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,10240,8192,0.032857600847880045
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,10240,10240,0.02642773389816284
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,10240,7168,0.027315199375152588
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,10240,7168,0.029089067379633588
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,10240,8192,0.023010132710138957
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,10240,6144,0.02431679964065552
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,10240,6144,0.028897066911061604
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,12288,65536,0.13229226271311442
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,10240,5120,0.020875734090805054
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,10240,5120,0.02599253257115682
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,10240,7168,0.02072746753692627
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,10240,4096,0.017237333456675212
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,10240,4096,0.024937599897384644
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,10240,4096,0.014962133765220643
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,10240,3584,0.0145578662554423
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,10240,3584,0.022657066583633423
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,10240,5120,0.01746986707051595
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,10240,3072,0.012985600034395852
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,10240,3072,0.020990933974583945
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,10240,6144,0.01918506622314453
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,10240,2560,0.011260799566904704
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,10240,2560,0.021236266692479452
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,10240,3584,0.01416000028451284
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,10240,2048,0.009610666831334432
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,10240,2048,0.018942934274673463
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,10240,3072,0.012906666596730551
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,10240,1536,0.008248533308506011
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,10240,1536,0.017283199230829875
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,10240,2560,0.011705600221951802
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,10240,1024,0.006295466423034668
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,10240,1024,0.019758933782577516
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,10240,1536,0.009550933043162029
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,10240,768,0.005070933202902476
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,10240,768,0.016054399808247886
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,10240,1024,0.008477866649627686
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,10240,512,0.004315733412901561
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,10240,512,0.016661333044370015
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,10240,2048,0.010458667079607646
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,10240,256,0.0035061334570248926
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,10240,256,0.014612266421318054
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,10240,512,0.007534933090209961
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,10240,128,0.0032885332902272543
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,10240,128,0.014605866869290671
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,10240,768,0.00790293316046397
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,10240,64,0.002586666742960612
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,10240,64,0.014654933412869772
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,10240,32,0.00296426663796107
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,10240,32,0.014498133460680643
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,8192,65536,0.17078399658203125
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,10240,128,0.007101866602897644
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,8192,65536,0.12607786655426026
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,8192,16384,0.04805546601613363
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,10240,256,0.007098666826883952
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,8192,16384,0.0431658665339152
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,8192,12288,0.03514986832936605
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,8192,12288,0.03663146495819092
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,8192,12288,0.029702399174372358
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,8192,10240,0.03050559957822164
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,8192,10240,0.03245333234469096
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,8192,16384,0.03792746861775716
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,8192,8192,0.025357866287231447
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,8192,8192,0.029849600791931153
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,8192,10240,0.026395734151204425
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,8192,7168,0.021917865673700968
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,8192,7168,0.0273087998231252
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,8192,8192,0.02251733342806498
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,8192,6144,0.01981653372446696
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,8192,6144,0.025326933463414508
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,8192,7168,0.020730666319529214
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,8192,5120,0.016732800006866454
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,8192,5120,0.02281173268953959
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,10240,65536,0.12932159900665283
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,8192,4096,0.013612799843152366
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,8192,4096,0.02190933426221212
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,8192,6144,0.019067732493082683
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,8192,3584,0.012455466389656066
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,8192,3584,0.02095466653505961
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,8192,5120,0.01698026657104492
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,8192,3072,0.010988799730936687
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,8192,3072,0.02064639925956726
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,8192,4096,0.014614400267601014
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,8192,2560,0.009714133540789286
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,8192,2560,0.018544000387191773
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,8192,3584,0.013878400127092997
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,8192,2048,0.00857919951279958
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,8192,2048,0.01950719952583313
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,8192,2560,0.011229866743087768
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,8192,1536,0.007085866729418437
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,8192,1536,0.01604480048020681
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,8192,2048,0.010392533739407857
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,8192,1024,0.00506986677646637
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,8192,1024,0.016474666198094685
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,8192,3072,0.012537599603335062
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,8192,768,0.004287999868392944
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,8192,768,0.014486400286356607
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,8192,1024,0.008858666817347208
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,8192,512,0.0034986667335033415
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,8192,512,0.015525333086649575
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,8192,768,0.007941333452860515
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,8192,256,0.0029738667110602063
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,8192,256,0.01446399986743927
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,8192,1536,0.009609599908192951
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,8192,128,0.0025973332424958544
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,8192,128,0.014477866888046264
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,8192,256,0.007176533341407776
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,8192,64,0.002605866640806198
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,8192,64,0.014471466342608133
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,8192,32,0.0025578667720158894
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,8192,128,0.00668693333864212
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,8192,32,0.014478933811187745
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,8192,512,0.00761599987745285
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,7168,65536,0.15630933443705242
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,7168,65536,0.10858773390452067
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,7168,16384,0.04559359947840373
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,7168,16384,0.04518933296203613
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,7168,12288,0.03578133185704549
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,7168,12288,0.035537068049112955
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,7168,12288,0.029715200265248615
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,7168,10240,0.02987733284632365
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,7168,10240,0.03290133277575175
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,7168,16384,0.03751893440882365
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,7168,8192,0.028757333755493164
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,7168,8192,0.030419200658798218
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,7168,10240,0.026046933730443318
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,7168,7168,0.024495999018351235
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,7168,7168,0.02654186685880025
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,7168,8192,0.02231360077857971
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,7168,6144,0.021176532904307047
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,7168,6144,0.024753065903981526
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,7168,7168,0.02060906688372294
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,7168,5120,0.018308266003926595
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,7168,5120,0.02421440084775289
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,7168,6144,0.01908586621284485
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,7168,4096,0.01490133305390676
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,7168,4096,0.022490666309992472
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,7168,5120,0.01703146696090698
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,7168,3584,0.011677866180737812
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,7168,3584,0.02062826752662659
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,7168,4096,0.014115200440088908
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,7168,3072,0.010486400127410889
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,7168,3072,0.019396267334620156
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,7168,3584,0.013318399588267008
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,7168,2560,0.00942080020904541
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,7168,2560,0.01883413394292196
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,7168,3072,0.012132267157236736
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,7168,2048,0.008109866579373678
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,7168,2048,0.017244799931844076
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,7168,2560,0.011318399508794149
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,7168,1536,0.006378666559855144
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,7168,1536,0.01637226641178131
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,7168,2048,0.010420266787211101
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,7168,1024,0.004691199958324432
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,7168,1024,0.015173332889874777
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,7168,1536,0.00959999958674113
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,7168,768,0.004144000013669332
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,7168,768,0.016711467504501344
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,7168,1024,0.008385066191355388
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,7168,512,0.0033749334514141084
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,7168,512,0.014458666245142618
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,7168,512,0.007512533167997997
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,7168,256,0.0029866665601730345
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,7168,256,0.014459733168284097
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,7168,768,0.007933866480986278
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,7168,128,0.0026229334374268847
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,7168,128,0.014472533265749613
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,7168,128,0.007088000078996022
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,7168,64,0.002585600068171819
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,7168,64,0.014465066790580749
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,7168,32,0.002963199963172277
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,7168,256,0.0074976002176602675
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,7168,32,0.01444586714108785
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,6144,65536,0.1228874683380127
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,6144,65536,0.09433706601460776
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,6144,16384,0.038262399037679036
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,6144,16384,0.04124159812927246
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,8192,65536,0.12822399934132894
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,6144,12288,0.030129067103068036
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,6144,12288,0.03205440044403076
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,6144,16384,0.03707520167032878
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,6144,10240,0.026999467611312868
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,6144,10240,0.02978666623433431
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,6144,12288,0.030125866333643597
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,6144,8192,0.022696532805760703
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,6144,8192,0.027894399563471478
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,6144,8192,0.022396800915400188
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,6144,7168,0.019077332814534505
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,6144,7168,0.02478613257408142
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,6144,10240,0.02608213424682617
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,6144,6144,0.015944533546765647
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,6144,6144,0.02423680027325948
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,6144,7168,0.02071146567662557
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,6144,5120,0.01546239952246348
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,6144,5120,0.02418773372968038
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,6144,6144,0.018588799238204955
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,6144,4096,0.012055466572443645
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,6144,4096,0.02138026754061381
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,6144,4096,0.0141184002161026
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,6144,3584,0.01099733312924703
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,6144,3584,0.020474666357040407
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,6144,5120,0.016636799772580466
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,6144,3072,0.0105813334385554
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,6144,3072,0.020522665977478028
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,6144,3584,0.012994133432706199
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,6144,2560,0.008814932902654012
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,6144,2560,0.018572799364725747
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,6144,3072,0.012551466623942057
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,6144,2048,0.0075434664885203045
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,6144,2048,0.020014933745066323
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,7168,65536,0.12773226896921794
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,6144,1536,0.006284800171852112
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,6144,1536,0.01762666702270508
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,6144,2048,0.010481066505114238
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,6144,1024,0.004705066482226053
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,6144,1024,0.01646506687005361
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,6144,2560,0.010855467120806376
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,6144,768,0.0038293334345022834
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,6144,768,0.014539733529090881
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,6144,1536,0.009547733267148336
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,6144,512,0.00346666673819224
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,6144,512,0.014573867122332254
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,6144,768,0.007957333326339721
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,6144,256,0.0029653333127498626
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,6144,256,0.014454399545987448
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,6144,512,0.007542400062084198
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,6144,128,0.0025642665723959604
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,6144,128,0.014450132846832275
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,6144,1024,0.008379733562469483
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,6144,64,0.0025578667720158894
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,6144,64,0.014573867122332254
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,6144,32,0.002935466667016347
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,6144,256,0.007108266651630402
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,6144,128,0.007125333448251088
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,6144,32,0.014447999993960061
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,5120,65536,0.10690666834513347
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,5120,65536,0.09017600218454996
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,5120,16384,0.0335317333539327
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,5120,16384,0.03905813296635945
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,5120,12288,0.026280534267425538
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,5120,12288,0.03236693342526754
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,5120,12288,0.02972053289413452
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,5120,10240,0.023487999041875204
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,5120,10240,0.02927680015563965
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,5120,16384,0.0366922656695048
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,5120,8192,0.021320533752441407
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,5120,8192,0.025330134232838947
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,5120,10240,0.02613226572672526
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,5120,7168,0.01843413313229879
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,5120,7168,0.025257599353790284
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,5120,8192,0.021917865673700968
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,5120,6144,0.013833600282669067
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,5120,6144,0.024676267306009927
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,5120,7168,0.02029973268508911
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,5120,5120,0.012035199999809265
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,5120,5120,0.023093332846959434
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,5120,6144,0.017812265952428182
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,5120,4096,0.01016960044701894
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,5120,4096,0.02248426675796509
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,5120,5120,0.016155733664830526
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,5120,3584,0.009547733267148336
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,5120,3584,0.021771732966105142
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,5120,4096,0.013805866241455078
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,5120,3072,0.00960106650988261
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,5120,3072,0.018582399686177573
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,5120,3584,0.012931199868520102
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,5120,2560,0.009698133667310078
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,5120,2560,0.020170666774113975
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,5120,3072,0.012173866232236225
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,5120,2048,0.007079466680685679
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,5120,2048,0.016664533813794454
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,5120,2560,0.011261866490046183
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,5120,1536,0.0063285330931345625
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,5120,1536,0.016616533199946083
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,5120,2048,0.01046399970849355
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,5120,1024,0.004682666560014089
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,6144,65536,0.12797119617462158
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,5120,1024,0.014896000425020853
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,5120,768,0.004223999877770742
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,5120,768,0.014732799927393594
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,5120,768,0.008370133241017659
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,5120,512,0.0034005333979924522
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,5120,512,0.014524799585342408
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,5120,1536,0.009195733070373534
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,5120,256,0.0030432000756263735
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,5120,256,0.014567466576894126
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,5120,1024,0.008363733688990276
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,5120,128,0.0026122666895389557
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,5120,128,0.014990933736165366
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,5120,128,0.007057066758473713
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,5120,64,0.0025781333446502686
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,5120,256,0.007129600147406261
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,5120,512,0.007539199789365132
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,5120,32,0.0026517334083716077
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,5120,32,0.014510933558146158
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,4096,65536,0.09172266324361165
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,4096,65536,0.0759434700012207
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,4096,16384,0.026315732796986895
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,4096,16384,0.03480639855066935
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,4096,16384,0.03644693295160929
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,4096,12288,0.022809600830078124
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,4096,12288,0.030611199140548707
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,4096,12288,0.029286400477091475
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,4096,10240,0.01999359925587972
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,4096,10240,0.026213334004084272
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,4096,10240,0.025590399901072185
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,4096,8192,0.015241600076357522
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,4096,8192,0.024563199281692503
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,5120,65536,0.1268064022064209
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,4096,7168,0.01332373321056366
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,4096,7168,0.02279040018717448
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,4096,8192,0.0215829332669576
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,4096,6144,0.011988266309102377
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,4096,6144,0.021835732460021972
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,4096,7168,0.019891200462977092
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,4096,5120,0.010475732882817586
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,4096,5120,0.020761599143346153
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,4096,6144,0.01781760056813558
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,4096,4096,0.008813866972923278
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,4096,4096,0.019578667481740315
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,4096,5120,0.01618346671263377
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,4096,3584,0.008378666639328004
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,4096,3584,0.018963199853897095
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,4096,4096,0.01411733329296112
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,4096,3072,0.0071733335653940845
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,4096,3072,0.01656000018119812
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,4096,3584,0.013289599617322286
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,4096,2560,0.006431999802589417
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,4096,2560,0.016552533706029257
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,4096,2560,0.01122773289680481
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,4096,2048,0.005580799778302511
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,4096,2048,0.01675093372662862
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,4096,3072,0.012157866358757019
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,4096,1536,0.004885333279768625
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,4096,1536,0.016645333170890807
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,4096,2048,0.010181333621342976
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,4096,1024,0.0036992001036802924
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,4096,1024,0.016749866803487144
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,4096,1536,0.009191466371218364
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,4096,768,0.003446399917205175
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,4096,768,0.01451520025730133
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,4096,1024,0.008374399940172831
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,4096,512,0.003032533327738444
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,4096,768,0.008277333279450735
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,4096,512,0.014486400286356607
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,4096,256,0.002632533262173335
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,4096,256,0.014387200276056925
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,4096,256,0.00709440012772878
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,4096,128,0.002603733291228612
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,4096,512,0.007514666517575581
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,4096,128,0.01451520025730133
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,4096,64,0.0022261333962281544
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,4096,64,0.014482133587201438
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,4096,32,0.0022762666145960493
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,4096,32,0.014419200023015341
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,4096,128,0.007108266651630402
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,3584,65536,0.07420480251312256
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,3584,65536,0.07330559889475505
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,3584,16384,0.024361600478490196
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,3584,16384,0.03293333252271016
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,4096,65536,0.12722880045572918
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,3584,12288,0.01909760038057963
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,3584,12288,0.03107200066248576
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,3584,16384,0.03661226828893026
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,3584,10240,0.017485866943995156
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,3584,10240,0.029156267642974854
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,3584,12288,0.02925119996070862
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,3584,8192,0.015533866484959922
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,3584,8192,0.02406826615333557
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,3584,10240,0.025649066766103106
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,3584,7168,0.014935466647148132
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,3584,7168,0.022773333390553794
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,3584,7168,0.01988053321838379
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,3584,6144,0.012971733013788858
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,3584,6144,0.022424532969792684
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,3584,8192,0.02112320065498352
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,3584,5120,0.01071999967098236
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,3584,5120,0.02135466734568278
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,3584,6144,0.017925333976745606
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,3584,4096,0.008780800302823384
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,3584,4096,0.019862399498621622
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,3584,5120,0.016233600179354348
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,3584,3584,0.008052266637484233
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,3584,3584,0.018693333864212035
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,3584,3584,0.012961066762606301
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,3584,3072,0.007300266623497009
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,3584,3072,0.01797119975090027
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,3584,4096,0.014105600118637086
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,3584,2560,0.006298666695753734
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,3584,2560,0.016551466782887776
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,3584,3072,0.012489599982897441
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,3584,2048,0.005445333321889242
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,3584,2560,0.011238400141398113
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,3584,2048,0.017678932348887125
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,3584,1536,0.004296533266703288
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,3584,1536,0.016548267006874083
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,3584,1536,0.00955839951833089
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,3584,1024,0.0035861333211263022
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,3584,1024,0.015566933155059814
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,3584,2048,0.010055466492970785
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,3584,768,0.0034304000437259674
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,3584,768,0.014607999722162882
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,3584,1024,0.008493866523106892
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,3584,512,0.0030623999734719592
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,3584,512,0.01486186683177948
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,3584,768,0.007969066500663757
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,3584,256,0.002639999985694885
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,3584,256,0.0150709331035614
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,3584,512,0.007516799867153168
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,3584,128,0.002583466718594233
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,3584,256,0.007459199925263722
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,3584,128,0.01458026667435964
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,3584,64,0.0022069332500298817
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,3584,64,0.014602667093276978
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,3584,32,0.002552533398071925
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,3584,32,0.014453333616256715
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,3584,128,0.007130666573842366
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,3072,65536,0.06708479722340902
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,3072,65536,0.06508053143819173
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,3072,16384,0.022038400173187256
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,3072,16384,0.030849067370096843
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,3072,16384,0.036263465881347656
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,3072,12288,0.017245866854985557
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,3072,12288,0.0272597332795461
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,5120,64,0.014473600188891092
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,3072,10240,0.015273599823315939
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,3072,10240,0.024842667579650878
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,3072,12288,0.029046400388081865
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,3072,8192,0.013165866335233053
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,3072,8192,0.022906666994094847
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,3584,65536,0.12694400151570637
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,3072,7168,0.012081066767374676
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,3072,7168,0.02295359969139099
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,3072,10240,0.024810665845870973
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,3072,6144,0.010825600226720173
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,3072,6144,0.021604265769322714
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,3072,8192,0.02119040091832479
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,3072,5120,0.009637332955996196
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,3072,5120,0.020754132668177286
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,3072,7168,0.019222400585810342
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,3072,4096,0.008308266599973042
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,3072,4096,0.020143999656041463
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,3072,5120,0.01588053305943807
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,3072,3584,0.007546666761239369
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,3072,3584,0.01908479928970337
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,3072,6144,0.01745706597963969
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,3072,3072,0.006814933319886525
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,3072,3072,0.016934400796890257
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,3072,4096,0.013738666971524557
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,3072,2560,0.006102400024731954
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,3072,2560,0.016658133268356322
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,3072,3584,0.01297813355922699
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,3072,2048,0.005416533350944519
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,3072,2048,0.017425066232681273
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,3072,3072,0.011689600348472596
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,3072,1536,0.004271999994913737
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,3072,1536,0.015910399953524272
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,3072,2560,0.010889599720637005
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,3072,1024,0.0034282666941483817
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,3072,1024,0.016330666343371072
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,3072,1536,0.009169066945711773
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,3072,768,0.00306986669699351
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,3072,768,0.015422933300336204
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,3072,2048,0.010062932968139648
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,3072,512,0.0030410667260487873
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,3072,512,0.015108266472816467
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,3072,768,0.007919999957084655
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,3072,512,0.007533866663773854
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,3072,256,0.002566399921973546
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,3072,1024,0.008327466746171314
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,3072,256,0.01497066617012024
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,3072,128,0.002203733225663503
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,3072,128,0.01439573367436727
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,3072,64,0.002205866575241089
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,3072,64,0.014453333616256715
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,3072,32,0.0022528000175952912
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,3072,128,0.006677333513895671
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,3072,32,0.014487466216087342
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,3072,256,0.007084799806276958
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,2560,65536,0.056092798709869385
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,2560,16384,0.019853866100311278
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,2560,65536,0.0612607995669047
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,2560,16384,0.029422932863235475
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,2560,12288,0.015742933750152587
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,2560,12288,0.026868265867233277
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,2560,12288,0.028509867191314698
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,2560,10240,0.014177067081133523
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,2560,10240,0.024718934297561647
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,2560,16384,0.03606719970703125
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,2560,8192,0.01251413325468699
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,2560,8192,0.023462400833765665
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,2560,10240,0.024842667579650878
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,2560,7168,0.01162986656030019
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,2560,7168,0.021810134251912437
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,2560,8192,0.020812799533208214
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,2560,6144,0.010821333527565003
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,2560,6144,0.021822933355967203
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,2560,7168,0.019475199778874717
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,2560,5120,0.009492266178131103
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,2560,5120,0.01867093245188395
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,2560,6144,0.017463467518488564
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,2560,4096,0.008203733464082081
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,2560,4096,0.020499199628829956
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,3072,65536,0.12749866644541424
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,2560,3584,0.00751146674156189
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,2560,3584,0.017950934171676636
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,2560,5120,0.015754666924476624
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,2560,3072,0.006673066814740498
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,2560,3072,0.019960532585779824
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,2560,4096,0.013700266679128012
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,2560,2560,0.005909333129723867
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,2560,2560,0.018473599354426065
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,2560,3584,0.012969600160916648
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,2560,2048,0.004735999802748362
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,2560,2048,0.016569599509239197
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,2560,3072,0.012156800429026286
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,2560,1536,0.004278400043646494
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,2560,1536,0.014870400230089823
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,2560,2560,0.01123199959595998
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,2560,1024,0.003504000107447306
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,2560,2048,0.01000853379567464
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,2560,1024,0.015054933230082192
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,2560,768,0.0030645333230495454
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,2560,768,0.014575999975204468
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,2560,1536,0.009591466188430786
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,2560,512,0.0026410666604836782
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,2560,768,0.008318933347860973
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,2560,512,0.014430933197339377
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,2560,256,0.0025770666698614756
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,2560,1024,0.008360532919565837
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,2560,256,0.0145087997118632
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,2560,128,0.002229333420594533
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,2560,256,0.0071370666225751235
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,2560,128,0.014493866761525472
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,2560,512,0.007607466479142506
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,2560,64,0.0022069332500298817
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,2560,64,0.014603733023007711
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,2560,32,0.0021877333521842955
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,2560,128,0.0067221333583196
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,2560,32,0.01446293294429779
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,2048,65536,0.04688533147176106
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,2048,65536,0.05519893169403076
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,2048,16384,0.01613866686820984
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,2048,16384,0.028562132517496747
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,2048,12288,0.014272000392278036
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,2048,12288,0.0254314661026001
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,2048,12288,0.028486400842666626
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,2048,10240,0.012821333607037863
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,2048,10240,0.025830399990081788
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,2048,16384,0.035785599549611406
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,2048,8192,0.011249066392580668
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,2048,8192,0.02416106661160787
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,2048,10240,0.02516053318977356
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,2048,7168,0.01046720047791799
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,2048,7168,0.023753599325815836
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,2048,8192,0.02109439969062805
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,2048,6144,0.00960533320903778
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,2048,6144,0.02129813234011332
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,2048,7168,0.019435733556747437
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,2048,5120,0.008523733417193095
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,2048,5120,0.020881066719690956
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,2048,6144,0.01784320076306661
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,2048,4096,0.009488000472386678
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,2048,4096,0.018465065956115724
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,2048,5120,0.01578879952430725
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,2048,3584,0.009537067015965779
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,2048,3584,0.016716800133387246
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,2048,4096,0.013639466961224875
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,2048,3072,0.008447999755541485
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,2048,3072,0.01690559983253479
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,2048,3072,0.012123733758926392
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,2048,2560,0.007534933090209961
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,2048,2560,0.016578132907549538
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,2048,3584,0.012488533059755962
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,2048,2048,0.004695466657479604
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,2048,2048,0.016474666198094685
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,2048,2048,0.010412800312042236
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,2048,1536,0.004078933348258337
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,2048,2560,0.01111253301302592
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,2048,1536,0.01562666694323222
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,2048,1024,0.003402666747570038
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,2048,1024,0.016099199652671814
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,2048,1024,0.008400000135103862
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,2048,768,0.0030432000756263735
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,2048,768,0.014507733782132468
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,2048,1536,0.009526399771372478
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,2048,512,0.003048533449570338
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,2048,512,0.014857600132624308
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,2048,768,0.007938133180141449
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,2048,512,0.007500799993673961
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,2048,256,0.0025941332181294756
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,2048,256,0.014520532886187234
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,2048,128,0.002639999985694885
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,2048,128,0.014549332857131957
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,2048,256,0.007167999943097432
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,2048,64,0.002216533323129018
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,2048,128,0.007231999933719635
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,2048,32,0.0022570667167504626
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,2048,64,0.014453333616256715
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,2048,32,0.014408533771832785
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,1536,65536,0.039459200700124104
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,1536,65536,0.04912533362706502
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,1536,16384,0.01409173309803009
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,1536,16384,0.028820266326268513
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,2560,65536,0.12654080390930175
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,1536,12288,0.011896533767382304
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,1536,12288,0.02553173303604126
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,1536,12288,0.02823253273963928
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,1536,10240,0.012135466933250428
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,1536,10240,0.024564266204833984
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,1536,16384,0.03592106501261393
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,1536,8192,0.010904533664385478
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,1536,8192,0.022834134101867676
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,1536,10240,0.024817067384719848
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,1536,7168,0.01002346674601237
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,1536,7168,0.021237333615620933
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,1536,8192,0.020725333690643312
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,1536,6144,0.010044800241788228
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,1536,6144,0.02064639925956726
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,1536,7168,0.019118932882944743
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,1536,5120,0.0088128000497818
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,1536,5120,0.018663465976715088
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,1536,6144,0.017794134219487508
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,1536,4096,0.00746666689713796
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,1536,4096,0.019637332359949747
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,1536,5120,0.015786666671435037
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,1536,3584,0.006708266834417979
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,1536,3584,0.016665599743525186
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,1536,4096,0.013763200243314108
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,1536,3072,0.006331733365853627
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,1536,3072,0.016927999258041383
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,1536,3584,0.012868266304334006
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,1536,3072,0.012146133184432983
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,1536,2560,0.005453866720199585
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,1536,2560,0.016548267006874083
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,2048,65536,0.12637440363566083
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,1536,2048,0.004737066725889841
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,1536,2048,0.01653439998626709
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,1536,1536,0.004323199888070424
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,1536,2560,0.011181867122650147
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,1536,1536,0.015481600165367126
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,1536,1024,0.003525333354870478
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,1536,1024,0.014710399508476257
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,1536,1536,0.009253333012262981
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,1536,1024,0.00832533339659373
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,1536,2048,0.010411733388900756
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,1536,768,0.003028266628583272
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,1536,512,0.002881066749493281
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,1536,768,0.014487466216087342
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,1536,512,0.014532267053922018
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,1536,256,0.002629333237806956
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,1536,256,0.014418133099873862
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,1536,512,0.007530666887760162
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,1536,256,0.007113599777221679
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,1536,768,0.007894399762153625
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,1536,128,0.014619732896486918
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,1536,64,0.0022026665508747103
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,1536,64,0.014455466469128927
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,1536,128,0.006888533135255177
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,1536,32,0.00225600004196167
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,1536,32,0.014519466956456503
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,1024,65536,0.027639466524124145
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,1024,65536,0.04328533411026001
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,1024,16384,0.010818133751551311
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,1024,16384,0.026821333169937133
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,1024,16384,0.035445332527160645
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,1024,12288,0.010823466380437215
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,1024,12288,0.024821333090464272
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,1024,12288,0.02841493288675944
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,1024,10240,0.010028800368309021
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,1024,10240,0.024800000588099162
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,1024,10240,0.02482773264249166
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,1024,8192,0.008782933155695598
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,1024,8192,0.02232853372891744
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,1536,65536,0.12612266540527345
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,1024,7168,0.01034879982471466
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,1024,7168,0.01983039975166321
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,1024,8192,0.021126399437586464
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,1024,6144,0.00985599954922994
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,1024,6144,0.01970240076382955
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,1024,7168,0.019483733177185058
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,1024,5120,0.008748799562454224
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,1024,5120,0.018564265966415406
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,1024,6144,0.017458132902781167
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,1024,4096,0.007098666826883952
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,1024,4096,0.01880000034968058
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,1024,5120,0.01585493286450704
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,1024,3584,0.006289066871007283
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,1024,3584,0.018871466318766274
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,1024,4096,0.013730133573214212
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,1024,3072,0.005858133236567179
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,1024,3072,0.01991573373476664
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,1024,3584,0.012893866499265036
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,1024,2560,0.005039999882380167
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,1024,2560,0.01686506668726603
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,1024,3072,0.011694932977358501
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,1024,2048,0.005895466605822245
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,1024,2048,0.01700053413709005
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,1024,2560,0.011220266421635944
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,1024,1536,0.004826666911443075
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,1024,1536,0.01618559956550598
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,1024,2048,0.010338133573532105
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,1024,1024,0.0038783999780813852
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,1024,1024,0.015069866180419922
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,1024,1536,0.009131733576456707
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,1024,768,0.0034730667869249977
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,1024,768,0.01456106702486674
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,1024,1024,0.008284799754619598
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,1024,512,0.003014400104681651
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,1024,768,0.007573333382606506
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,1024,512,0.015309866269429526
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,1024,256,0.0026410666604836782
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,1024,256,0.015290666619936624
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,1024,512,0.0074879998962084455
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,1024,128,0.0025813333690166474
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,1024,256,0.007077333331108093
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,1024,128,0.014166399836540222
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,1024,64,0.002238933245340983
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,1024,128,0.006759466727574666
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,1024,64,0.014510933558146158
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,1024,32,0.0025685332715511323
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,1024,32,0.014454399545987448
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,768,65536,0.023170133431752525
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,768,65536,0.037988265355428055
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,768,16384,0.014573867122332254
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,768,16384,0.02678080002466838
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,1024,65536,0.12532587051391603
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,768,12288,0.012131200234095255
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,768,12288,0.02542613347371419
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,768,12288,0.028109866380691528
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,768,10240,0.011325866977373759
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,768,10240,0.022829866409301756
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,768,16384,0.03505706787109375
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,768,8192,0.010175999999046326
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,768,8192,0.020860799153645835
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,768,10240,0.02483946681022644
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,768,7168,0.008770133058230083
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,768,7168,0.020848000049591066
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,768,8192,0.021086933215459187
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,768,6144,0.008402132987976074
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,768,6144,0.0196341335773468
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,768,7168,0.019399466117223103
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,768,5120,0.007484800120194752
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,768,5120,0.018616533279418944
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,768,6144,0.017538134256998697
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,768,4096,0.006723199784755707
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,768,4096,0.019318399826685588
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,768,5120,0.015731199582417806
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,768,3584,0.006277333199977875
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,768,3584,0.016683733463287352
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,768,4096,0.013730133573214212
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,768,3072,0.00561599979797999
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,768,3584,0.012843733032544455
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,768,3072,0.016567466656366985
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,768,2560,0.005097599824269613
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,768,2560,0.017757866779963175
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,768,3072,0.011636267105738323
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,768,2048,0.004612266520659129
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,768,2048,0.016237866878509522
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,768,2048,0.010357333223025005
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,768,1536,0.0038058665891488397
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,768,1536,0.016008533040682473
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,768,1536,0.009165866176287334
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,768,1024,0.003387733300526937
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,768,1024,0.017089066902796428
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,768,1024,0.007914666831493378
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,768,768,0.0029504001140594482
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,768,768,0.0156960000594457
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,768,768,0.007534933090209961
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,768,512,0.0025898667673269907
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,768,512,0.01444906691710154
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,768,512,0.007480533421039581
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,768,256,0.0024256000916163126
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,768,256,0.014180266857147216
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,768,256,0.0067104001839955645
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,768,128,0.002186666677395503
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,768,128,0.014441600441932679
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,768,128,0.006687999765078227
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,768,64,0.0021727999051411946
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,768,64,0.01448853313922882
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,768,32,0.002160000056028366
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,768,32,0.013182933131853739
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,512,65536,0.017998933792114258
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,512,65536,0.03626773357391357
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,768,65536,0.1226207971572876
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,512,16384,0.010942932963371278
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,512,16384,0.02490239938100179
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,512,16384,0.03541546662648519
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,512,12288,0.009723732868830364
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,512,12288,0.024167466163635253
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,512,12288,0.028126933177312213
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,512,10240,0.0087936004002889
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,512,10240,0.022872533400853476
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,512,10240,0.024817067384719848
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,512,8192,0.007549866537253062
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,512,8192,0.021314134200414024
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,512,65536,0.12181866963704427
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,512,7168,0.007206400235493978
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,512,7168,0.02062506675720215
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,512,8192,0.020812799533208214
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,512,6144,0.006774400174617767
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,512,6144,0.0195360004901886
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,512,7168,0.01903466582298279
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,512,5120,0.006692266464233399
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,512,5120,0.01908586621284485
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,512,6144,0.01736853321393331
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,512,4096,0.006683733562628429
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,512,4096,0.018504534165064493
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,512,5120,0.015756799777348836
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,512,3584,0.00590826670328776
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,512,3584,0.016511999567349753
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,512,4096,0.013267200191815695
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,512,3072,0.005485866467158
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,512,3072,0.019320533672968546
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,512,3584,0.012820266683896384
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,512,2560,0.005031466484069824
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,512,3072,0.012027733524640401
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,512,2560,0.017220266660054526
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,512,2048,0.004758400221665701
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,512,2048,0.01656426688035329
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,512,2560,0.010818133751551311
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,512,1536,0.004271999994913737
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,512,2048,0.00995306670665741
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,512,1536,0.015599999825159708
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,512,1024,0.003491200009981791
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,512,1024,0.015796266992886863
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,512,1536,0.009121066331863404
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,512,768,0.002961066613594691
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,512,1024,0.00829013337691625
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,512,768,0.01451520025730133
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,512,512,0.00264533335963885
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,512,512,0.014478933811187745
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,512,768,0.007942399879296621
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,512,512,0.007468800246715546
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,512,256,0.0026154667139053345
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,512,256,0.014612266421318054
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,512,256,0.007102933526039123
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,512,128,0.002268799891074498
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,512,128,0.014600533246994018
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,512,128,0.007031466563542683
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,512,64,0.0021290667355060577
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,512,64,0.014503467082977294
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,512,32,0.0022570667167504626
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,512,32,0.014519466956456503
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,256,65536,0.017038933436075845
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,256,65536,0.03460586468378703
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,256,16384,0.00809279978275299
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,256,16384,0.022822399934132896
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,256,16384,0.03542399803797404
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,256,12288,0.007181866466999054
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,256,12288,0.021800533930460612
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,256,12288,0.028032000859578448
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,256,10240,0.006299733122189839
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,256,10240,0.021423999468485513
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,256,10240,0.02435413400332133
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,256,8192,0.0059125334024429325
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,256,8192,0.020793600877126058
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,256,8192,0.02101226647694906
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,256,7168,0.005862399935722351
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,256,7168,0.02068480054537455
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,256,65536,0.12120532989501953
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,256,6144,0.006390400230884552
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,256,7168,0.018990933895111084
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,256,6144,0.019734400510787963
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,256,5120,0.006698666512966156
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,256,5120,0.02097919980684916
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,256,6144,0.01733013391494751
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,256,4096,0.00586346685886383
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,256,5120,0.01609493295351664
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,256,4096,0.017537067333857216
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,256,3584,0.005500799914201101
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,256,3584,0.01842133402824402
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,256,4096,0.01365013321240743
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,256,3072,0.005485866467158
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,256,3584,0.012795733412106833
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,256,3072,0.016924800475438435
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,256,2560,0.005100800096988678
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,256,2560,0.01684160033861796
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,256,3072,0.011642666657765706
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,256,2048,0.0042133331298828125
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,256,2560,0.01125866671403249
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,256,2048,0.016160000363985697
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,256,1536,0.0038218667109807336
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,256,1536,0.01448319951693217
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,256,2048,0.010066133737564088
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,256,1024,0.0033973333736260734
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,256,1536,0.009103999535242716
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,256,1024,0.017459199825922648
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,256,768,0.002976000060637792
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,256,768,0.014578133821487427
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,256,1024,0.008297599852085114
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,256,512,0.002586666742960612
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,256,512,0.014531200130780539
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,256,768,0.007940266529719036
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,256,256,0.002234666546185811
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,256,256,0.014477866888046264
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,256,512,0.007147733370463054
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,256,128,0.0021898667017618816
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,256,128,0.01458026667435964
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,256,256,0.007048533360163371
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,256,64,0.002186666677395503
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,256,64,0.01446613371372223
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,256,128,0.0066997334361076355
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,256,32,0.002145066608985265
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,128,65536,0.011945600310961407
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,256,32,0.01344106694062551
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,128,65536,0.034916265805562334
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,128,16384,0.005886933207511902
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,128,16384,0.02273919979731242
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,128,16384,0.035010135173797606
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,128,12288,0.005946666498978933
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,128,12288,0.022550400098164877
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,128,12288,0.02799786726633708
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,128,10240,0.005982933441797892
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,128,10240,0.020778665939966835
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,128,10240,0.02473919987678528
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,128,8192,0.005842133363087972
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,128,8192,0.020619734128316244
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,128,8192,0.021005866924921672
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,128,7168,0.005508266886075338
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,128,65536,0.12081813017527263
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,128,7168,0.019926400979359944
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,128,6144,0.005869866907596588
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,128,6144,0.019666133324305217
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,128,7168,0.01937920053799947
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,128,5120,0.00582826683918635
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,128,5120,0.018984532356262206
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,128,5120,0.015686399737993875
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,128,4096,0.0054954667886098225
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,128,4096,0.018372267484664917
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,128,4096,0.013636266191800436
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,128,3584,0.005226666728655497
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,128,3584,0.01675093372662862
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,128,3584,0.012889599800109864
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,128,3072,0.00514986664056778
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,128,3072,0.016548267006874083
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,128,3072,0.011653332908948263
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,128,2560,0.005479466418425242
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,128,2560,0.016760534048080443
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,128,2560,0.010808533430099488
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,128,2048,0.005055999755859375
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,128,2048,0.015873066584269204
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,128,2048,0.01032533347606659
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,128,1536,0.003804799914360046
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,128,1536,0.0150709331035614
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,128,1536,0.009141332904497783
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,128,1024,0.0030421334008375804
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,128,1024,0.014684800306955972
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,128,1024,0.007949866851170858
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,128,768,0.0029397333661715193
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,128,768,0.014514133334159851
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,128,768,0.007490133245786031
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,128,512,0.0025792000194390613
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,128,512,0.014444800217946372
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,128,512,0.007119999825954437
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,128,256,0.0022005334496498106
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,128,256,0.013724799950917563
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,128,256,0.006762666503588359
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,128,128,0.002162133405605952
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,128,128,0.013513599832852682
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,128,128,0.006651733318964641
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,128,64,0.002133333434661229
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,128,64,0.014406399925549826
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,128,32,0.0021685334543387097
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,128,32,0.013432533542315165
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,64,65536,0.008725333213806152
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,64,65536,0.028516266743342084
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,64,16384,0.005883733431498209
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,64,16384,0.021660800774892172
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,64,12288,0.005103999873002371
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,64,12288,0.020600533485412596
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,64,10240,0.00584853341182073
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,64,10240,0.02044693430264791
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,64,8192,0.005866666634877523
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,64,8192,0.018956800301869713
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,64,7168,0.0054954667886098225
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,64,7168,0.019246933857599895
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,64,6144,0.005098666747411092
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,64,6144,0.01871466636657715
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,64,5120,0.005062399804592133
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,64,5120,0.0185973326365153
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,64,4096,0.00461760014295578
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,64,4096,0.01869759956995646
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,64,3584,0.005041066805521647
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,64,3584,0.016798933347066246
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,64,3072,0.004628266890843709
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,64,3072,0.016806399822235106
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,64,2560,0.005006933212280273
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,64,2560,0.016428800423940022
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,64,2048,0.0046293333172798155
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,64,2048,0.014740266402562461
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,64,1536,0.0038122666378815973
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,64,1536,0.015440000096956888
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,64,1024,0.002946133414904277
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,64,1024,0.015363199512163797
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,64,768,0.0026858667532602947
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,64,768,0.014467199643452963
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,64,512,0.0026687999566396077
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,64,512,0.013851733009020487
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,64,256,0.002194133400917053
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,64,256,0.013117866714795432
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,64,128,0.002205866575241089
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,64,128,0.01439466675122579
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,64,64,0.0019402666638294857
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,64,64,0.014245333274205527
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,64,32,0.0020213333268960317
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,64,32,0.012613333264986672
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,32,65536,0.009031466643015544
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,32,65536,0.028794666131337483
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,32,16384,0.006237866481145223
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,32,16384,0.02244480053583781
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,32,12288,0.005452799797058106
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,32,12288,0.02073919971783956
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,32,10240,0.0054848000407218935
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,32,10240,0.020147200425465903
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,32,8192,0.005438933273156484
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,32,8192,0.01919893423716227
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,32,7168,0.005425066749254862
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,32,7168,0.01880426605542501
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,32,6144,0.005077333251635233
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,32,6144,0.01924053430557251
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,32,5120,0.00505920002857844
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,32,5120,0.01887680093447367
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,32,4096,0.004633600016434988
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,32,4096,0.017964800198872886
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,32,3584,0.005002666513125102
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,32,3584,0.017642666896184288
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,32,3072,0.004682666560014089
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,32,3072,0.016546133160591125
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,32,2560,0.005020800232887268
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,32,2560,0.016701867183049522
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,32,2048,0.004619733492533366
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,32,2048,0.017013333241144814
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,32,1536,0.003533866753180822
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,32,1536,0.016676266988118492
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,32,1024,0.0030933332939942675
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,32,1024,0.014938666423161825
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,32,768,0.0027306665976842242
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,32,768,0.01458346645037333
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,32,512,0.002674133330583572
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,32,512,0.013982933759689332
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,32,256,0.0022976001103719073
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,32,256,0.014247467120488485
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,32,128,0.002204799900452296
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,32,128,0.014545067151387533
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,32,64,0.0021888000269730886
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,32,64,0.014266666769981385
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,32,32,0.002201599876085917
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,80,32,32,0.013338667154312134
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,65536,16384,0.36560640335083006
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,65536,16384,0.20069333712259926
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,80,1536,128,0.002195200075705846
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,65536,12288,0.2703296025594076
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,65536,12288,0.15770667394002277
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,768,2560,0.011185066898663838
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,65536,10240,0.2267786661783854
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,65536,10240,0.12940373420715331
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,80,128,6144,0.017562667528788246
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,65536,8192,0.18626987139383952
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,65536,8192,0.10809280077616375
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,65536,10240,0.07295680046081543
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,65536,7168,0.15932586987813313
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,65536,7168,0.09722879727681479
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,65536,8192,0.06108373403549194
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,65536,6144,0.1367008050282796
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,65536,6144,0.08546133041381836
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,65536,12288,0.08919573624928792
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,65536,5120,0.11360639731089275
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,65536,5120,0.07474559942881266
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,65536,16384,0.11301439603169758
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,65536,4096,0.09287040233612061
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,65536,4096,0.06265279849370321
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,65536,6144,0.052008533477783205
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,65536,3584,0.08308053016662598
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,65536,3584,0.056109865506490074
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,65536,5120,0.04210026661554973
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,65536,3072,0.07105279763539632
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,65536,3072,0.053559466203053796
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,65536,7168,0.056753067175547275
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,65536,2560,0.059197866916656496
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,65536,2560,0.04638826847076416
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,65536,4096,0.03654293219248454
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,65536,2048,0.04809066851933797
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,65536,2048,0.0401962677637736
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,65536,3584,0.03217599987983703
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,65536,1536,0.03829760154088338
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,65536,1536,0.03475199937820435
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,65536,2560,0.025571199258168538
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,65536,1024,0.027910399436950683
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,65536,1024,0.026745599508285523
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,65536,3072,0.03049280047416687
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,65536,768,0.02144533395767212
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,65536,768,0.024754132827123007
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,65536,2048,0.021938133239746093
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,65536,512,0.01418773333231608
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,65536,512,0.021294933557510377
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,65536,768,0.012959999839464822
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,65536,256,0.008449066678682964
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,65536,256,0.016458666324615477
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,65536,512,0.011174399654070537
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,65536,128,0.005051733553409576
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,65536,128,0.016294399897257485
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,65536,1024,0.014777599771817525
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,65536,64,0.004193066557248434
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,65536,64,0.016450132926305136
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,65536,32,0.004515199859937032
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,65536,32,0.01639573375384013
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,16384,65536,0.3407669385274251
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,16384,65536,0.21698667208353678
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,65536,256,0.00915839970111847
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,16384,16384,0.08408426443735759
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,16384,16384,0.061248000462849936
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,65536,128,0.008785067001978557
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,16384,12288,0.061096533139546716
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,16384,12288,0.04902506669362386
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,65536,1536,0.01904426614443461
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,16384,10240,0.05392640034357706
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,16384,10240,0.04456640084584554
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,16384,10240,0.029179733991622925
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,16384,8192,0.04268693526585897
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,16384,8192,0.03770240147908528
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,16384,12288,0.03327359954516093
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,16384,7168,0.03747626543045044
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,16384,7168,0.03512533505757649
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,16384,16384,0.04148693482081096
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,16384,6144,0.03333333333333333
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,16384,6144,0.03219733238220215
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,16384,8192,0.02434879938761393
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,16384,5120,0.028497066100438433
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,16384,5120,0.03020586570103963
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,16384,7168,0.022341332832972207
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,16384,4096,0.022282665967941283
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,16384,4096,0.025711999336878462
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,16384,5120,0.0182261327902476
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,16384,3584,0.020641066630681357
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,16384,3584,0.024685867627461753
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,16384,6144,0.020032000541687012
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,16384,3072,0.018080000082651773
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,16384,3072,0.024090667565663658
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,16384,4096,0.015828266739845276
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,16384,2560,0.015338666240374246
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,16384,2560,0.02097599903742472
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,16384,2560,0.012485333283742269
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,16384,2048,0.012452266613642375
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,16384,2048,0.02063039938608805
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,16384,3584,0.014502400159835815
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,16384,1536,0.010082133611043294
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,16384,1536,0.018733867009480796
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,16384,1536,0.009993599851926167
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,16384,1024,0.007558399935563405
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,16384,1024,0.01665493349234263
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,16384,2048,0.011228799819946289
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,16384,768,0.006286933521429698
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,16384,768,0.015084800124168397
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,16384,1024,0.008506666620572407
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,16384,512,0.00415040006240209
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,16384,512,0.014811733365058899
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,16384,768,0.008027733365694682
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,16384,256,0.0034111998975276947
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,16384,256,0.015549866358439126
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,16384,512,0.007593599955240886
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,16384,128,0.0031093334158261614
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,16384,128,0.01455573340257009
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,16384,128,0.006714666883150737
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,16384,64,0.003294933338960012
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,16384,256,0.007143466671307881
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,16384,32,0.002963199963172277
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,16384,64,0.014669866363207499
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,16384,32,0.014471466342608133
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,12288,65536,0.24891200065612792
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,12288,65536,0.15900266965230306
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,12288,16384,0.06963093280792236
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,12288,16384,0.050246401627858484
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,12288,16384,0.039452799161275226
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,12288,12288,0.04676160017649333
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,12288,12288,0.04195733467737834
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,12288,12288,0.03147733410199483
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,12288,10240,0.04125119845072429
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,12288,10240,0.037196799119313555
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,16384,3072,0.013221333424250284
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,12288,8192,0.033190399408340454
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,12288,8192,0.03272106647491455
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,12288,10240,0.027284266551335652
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,12288,7168,0.028838400046030683
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,12288,7168,0.03022506634394328
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,12288,8192,0.02281493345896403
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,12288,6144,0.025964800516764325
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,12288,6144,0.029681066672007244
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,12288,7168,0.02118933399518331
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,12288,5120,0.022321067253748574
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,12288,5120,0.026474666595458985
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,12288,6144,0.019038933515548705
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,12288,4096,0.01769066651662191
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,12288,4096,0.024012800057729086
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,12288,5120,0.017796266078948974
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,12288,3584,0.015929599603017174
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,12288,3584,0.023291732867558798
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,12288,4096,0.015044266978899637
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,12288,3072,0.014751999576886495
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,12288,3072,0.023400533199310302
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,12288,3584,0.014212266604105631
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,12288,2560,0.01276693344116211
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,12288,2560,0.020687999327977498
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,12288,2560,0.011672533551851908
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,12288,2048,0.010863999525705974
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,12288,2048,0.021271467208862305
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,12288,3072,0.01325759987036387
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,12288,1536,0.008746666709582011
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,12288,1536,0.018252799908320107
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,12288,1536,0.009171199798583985
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,12288,1024,0.00710399995247523
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,12288,1024,0.015783466895421348
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,12288,2048,0.010447999835014344
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,12288,768,0.005520000060399374
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,12288,768,0.015448533495267234
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,12288,1024,0.008424533406893413
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,12288,512,0.0035904000202814737
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,12288,512,0.014455466469128927
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,12288,768,0.008360532919565837
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,12288,256,0.0030432000756263735
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,12288,256,0.014563199877738953
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,12288,256,0.007142400244871776
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,12288,128,0.002963199963172277
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,12288,128,0.014516266187032065
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,12288,128,0.007102933526039123
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,12288,64,0.0026101333399613695
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,12288,512,0.007501866420110066
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,12288,32,0.002995199958483378
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,12288,64,0.014500266313552857
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,12288,32,0.014468266566594442
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,10240,65536,0.21394880612691244
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,10240,65536,0.1449461301167806
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,10240,16384,0.053651201725006106
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,10240,16384,0.045459198951721194
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,16384,65536,0.14696426391601564
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,10240,12288,0.039833601315816244
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,10240,12288,0.03785706758499145
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,10240,16384,0.03862826824188233
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,10240,10240,0.034622931480407716
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,10240,10240,0.03424319823582967
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,12288,65536,0.13172586758931476
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,10240,8192,0.027990400791168213
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,10240,8192,0.02950506607691447
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,10240,12288,0.03051626682281494
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,10240,7168,0.024682666858037314
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,10240,7168,0.028218666712443035
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,10240,10240,0.027335466941197713
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,10240,6144,0.021900800863901775
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,10240,6144,0.026276266574859618
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,10240,7168,0.02090346614519755
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,10240,5120,0.018769067525863648
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,10240,5120,0.024396799008051553
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,10240,8192,0.023116799195607503
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,10240,4096,0.014972800016403198
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,10240,4096,0.02353386680285136
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,10240,6144,0.019013333320617675
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,10240,3584,0.013683199882507324
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,10240,3584,0.02063573400179545
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,10240,5120,0.01753386656443278
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,10240,3072,0.01192639966805776
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,10240,3072,0.020958934227625528
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,10240,4096,0.015190399686495461
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,10240,2560,0.01049066682656606
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,10240,2560,0.020632533232371013
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,10240,3584,0.013745066523551942
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,10240,2048,0.008833066622416178
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,10240,2048,0.01845973332722982
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,10240,3072,0.012885333100954691
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,10240,1536,0.007584000130494435
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,10240,1536,0.017121066649754844
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,10240,2560,0.011668266852696736
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,10240,1024,0.005448533097902933
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,10240,1024,0.015824000040690102
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,10240,2048,0.010434133807818095
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,10240,768,0.004113066693147024
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,10240,768,0.01456106702486674
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,10240,1536,0.009267200032869976
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,10240,512,0.0035071998834609987
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,10240,512,0.014625066518783569
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,10240,1024,0.008427733182907104
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,10240,256,0.003094399968783061
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,10240,256,0.014761599898338317
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,10240,512,0.007559466858704884
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,10240,128,0.002679466704527537
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,10240,128,0.014603733023007711
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,10240,768,0.008334933718045553
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,10240,64,0.0025898667673269907
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,10240,256,0.007109333574771881
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,10240,32,0.002657066782315572
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,10240,64,0.014510933558146158
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,10240,32,0.014531200130780539
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,10240,128,0.006786133348941803
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,8192,65536,0.18528532981872559
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,8192,65536,0.12171413103739422
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,8192,16384,0.045737600326538085
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,8192,16384,0.039903998374938965
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,8192,12288,0.031699200471242264
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,8192,12288,0.03431253433227539
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,8192,12288,0.029870933294296263
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,8192,10240,0.027610667546590167
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,8192,10240,0.0312991996606191
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,8192,16384,0.03787200053532918
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,8192,8192,0.022216532627741495
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,8192,8192,0.026973867416381837
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,8192,10240,0.026055467128753663
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,8192,7168,0.01986133257548014
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,8192,7168,0.026895999908447266
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,8192,8192,0.022151466210683188
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,8192,6144,0.017790933450063072
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,8192,6144,0.02566933234532674
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,8192,7168,0.021092265844345093
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,8192,5120,0.015346133708953857
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,8192,5120,0.02563520073890686
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,8192,6144,0.019033600886662803
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,8192,4096,0.012291199962298075
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,8192,4096,0.020886399348576865
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,8192,5120,0.017132800817489625
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,8192,3584,0.012334932883580525
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,8192,3584,0.020968532562255858
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,8192,4096,0.014205867052078247
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,8192,3072,0.010794666409492493
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,8192,3072,0.022078933318456014
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,8192,3584,0.013710932930310568
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,8192,2560,0.009639466802279156
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,8192,2560,0.01883093317349752
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,8192,3072,0.012103466192881267
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,8192,2048,0.008330666522185007
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,8192,2048,0.017939200003941856
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,8192,2560,0.011307733257611592
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,8192,1536,0.007134933272997539
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,8192,1536,0.016590933005015053
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,8192,1536,0.009274666508038838
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,8192,2048,0.010007466872533162
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,8192,1024,0.004857600231965383
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,8192,1024,0.014710399508476257
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,8192,768,0.004013866682847341
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,8192,768,0.0165802667538325
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,8192,1024,0.00841813286145528
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,8192,512,0.00311253344019254
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,8192,512,0.014805333813031516
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,8192,768,0.008222933113574981
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,8192,256,0.0028181334336598715
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,8192,256,0.014747732877731323
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,8192,256,0.0071829333901405334
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,8192,128,0.002629333237806956
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,8192,512,0.007130666573842366
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,8192,128,0.014542933305104574
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,8192,64,0.002598399917284648
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,8192,64,0.014577066898345948
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,8192,32,0.0029728000362714132
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,8192,32,0.014641066392262777
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,8192,128,0.007090133428573608
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,7168,65536,0.1617311954498291
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,7168,65536,0.10380266507466633
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,7168,16384,0.04584533373514811
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,7168,16384,0.04165120124816894
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,10240,65536,0.12898879845937092
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,7168,12288,0.03541866540908813
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,7168,12288,0.0327455997467041
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,7168,16384,0.03752853473027547
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,7168,10240,0.026740266879399614
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,7168,10240,0.02912213404973348
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,7168,12288,0.02981013258298238
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,7168,8192,0.021548799673716226
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,7168,8192,0.025573333104451496
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,7168,10240,0.02586666742960612
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,7168,7168,0.018990933895111084
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,7168,7168,0.024411733945210776
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,7168,8192,0.022010666131973267
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,7168,6144,0.016800000270207723
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,7168,6144,0.02309760053952535
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,7168,7168,0.020651733875274657
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,7168,5120,0.014361600081125895
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,7168,5120,0.023030400276184082
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,7168,6144,0.018845866123835243
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,7168,4096,0.011640533804893494
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,7168,4096,0.02060906688372294
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,7168,5120,0.01699840029080709
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,7168,3584,0.010831999778747558
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,7168,3584,0.019942400852839152
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,8192,65536,0.12756160100301106
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,7168,3072,0.009664000074068705
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,7168,3072,0.020197333892186482
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,7168,3584,0.013299199938774108
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,7168,2560,0.008515200018882752
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,7168,2560,0.01862506667772929
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,7168,4096,0.014194132884343466
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,7168,2048,0.007517866790294647
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,7168,2048,0.01660160024960836
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,7168,3072,0.012446932991345723
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,7168,1536,0.005965866645177205
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,7168,2560,0.011206400394439698
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,7168,1536,0.016692266861597697
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,7168,1024,0.004366933306058248
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,7168,1024,0.01472106675306956
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,7168,2048,0.010079999764760334
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,7168,768,0.0038389332592487337
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,7168,768,0.014536533753077188
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,7168,768,0.008071466783682505
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,7168,512,0.0030826665461063385
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,7168,512,0.016217600305875143
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,7168,1024,0.008313600222269695
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,7168,256,0.0029813334345817565
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,7168,256,0.014737066626548768
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,7168,256,0.0071274667978286745
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,7168,128,0.0026026666164398193
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,7168,128,0.014665599664052328
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,7168,1536,0.009253333012262981
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,7168,64,0.0026133333643277483
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,7168,64,0.0145087997118632
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,7168,32,0.002566399921973546
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,7168,128,0.0071168000499407455
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,7168,32,0.0144405335187912
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,7168,512,0.007539199789365132
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,6144,65536,0.09016746679941813
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,6144,65536,0.12406720320383709
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,6144,16384,0.038286932309468585
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,6144,16384,0.04133653243382772
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,6144,12288,0.030369067192077638
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,6144,12288,0.03513386646906535
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,6144,12288,0.029148799180984498
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,6144,10240,0.02651519974072774
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,6144,10240,0.028523733218510944
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,6144,16384,0.03751573165257772
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,6144,8192,0.024279467264811196
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,6144,8192,0.025594667593638105
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,6144,10240,0.025603200991948443
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,6144,7168,0.02174826661745707
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,6144,7168,0.02493013342221578
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,6144,8192,0.02199466625849406
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,6144,6144,0.015318399667739869
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,6144,6144,0.022524799903233847
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,6144,7168,0.02034986615180969
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,6144,5120,0.013299199938774108
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,6144,5120,0.022683733701705934
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,6144,6144,0.018253866831461588
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,6144,4096,0.011183999975522359
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,6144,4096,0.020757333437601725
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,6144,5120,0.016619732975959776
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,6144,3584,0.009987200299898785
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,6144,3584,0.019350399573644005
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,6144,4096,0.013691733280817667
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,6144,3072,0.008830933769543966
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,6144,3072,0.01864853302637736
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,6144,3584,0.013339733084042868
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,6144,2560,0.007926400005817413
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,6144,2560,0.017103999853134155
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,7168,65536,0.12763306299845378
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,6144,2048,0.006757333377997081
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,6144,2048,0.016530133287111917
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,6144,2560,0.011306666334470113
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,6144,1536,0.005134933193524678
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,6144,1536,0.016694400707880655
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,6144,3072,0.012056533495585125
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,6144,1024,0.0038560000558694207
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,6144,1024,0.016157866517702738
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,6144,2048,0.010708266496658325
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,6144,768,0.0034602666894594826
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,6144,768,0.01458346645037333
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,6144,1024,0.008327466746171314
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,6144,512,0.002974933385848999
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,6144,512,0.014481066664059957
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,6144,1536,0.009532800316810608
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,6144,256,0.002632533262173335
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,6144,256,0.014666666587193808
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,6144,256,0.0071168000499407455
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,6144,128,0.002629333237806956
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,6144,128,0.0144896000623703
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,6144,512,0.007502933343251546
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,6144,64,0.0021909333765506743
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,6144,64,0.014428800344467163
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,6144,128,0.007092266778151194
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,6144,32,0.0025781333446502686
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,6144,32,0.014520532886187234
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,6144,768,0.007916800181070964
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,5120,65536,0.10444053014119466
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,5120,65536,0.08648426532745361
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,5120,16384,0.032500267028808594
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,5120,16384,0.0370410680770874
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,5120,12288,0.025561600923538208
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,5120,12288,0.031319467226664226
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,5120,12288,0.029233066240946452
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,5120,10240,0.02272426684697469
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,5120,10240,0.031352533896764116
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,5120,16384,0.03668053150177002
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,5120,8192,0.019282132387161255
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,5120,8192,0.02474773327509562
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,5120,10240,0.025574400027592974
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,5120,7168,0.019082667430241902
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,5120,7168,0.023119999965031942
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,5120,8192,0.02216426730155945
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,5120,6144,0.01723626653353373
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,5120,6144,0.022553600867589316
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,5120,7168,0.019959467649459838
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,5120,5120,0.012174933155377706
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,5120,5120,0.022244266668955483
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,5120,6144,0.017836799224217735
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,5120,4096,0.01004266639550527
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,5120,4096,0.020334933201471964
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,5120,5120,0.016173866391181946
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,5120,3584,0.009398399790128072
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,5120,3584,0.019205333789189656
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,5120,4096,0.014115200440088908
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,5120,3072,0.00839573343594869
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,5120,3072,0.0165173331896464
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,5120,3584,0.01318826675415039
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,5120,2560,0.007156266768773396
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,5120,2560,0.018601600329081217
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,5120,3072,0.012062933047612507
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,5120,2048,0.005913599828879039
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,5120,2048,0.01581013302008311
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,5120,2560,0.011203199625015259
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,5120,1536,0.004897066454092661
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,5120,1536,0.016800000270207723
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,5120,1536,0.009313066800435383
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,5120,1024,0.004285866518815359
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,5120,1024,0.01737066706021627
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,5120,2048,0.010014933347702027
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,5120,768,0.003437866767247518
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,5120,768,0.015359999736150107
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,5120,1024,0.008309333523114523
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,5120,512,0.003048533449570338
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,5120,768,0.007515733440717061
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,5120,512,0.01446613371372223
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,5120,256,0.002598399917284648
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,5120,256,0.014753066500027976
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,5120,256,0.007136000196139018
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,5120,128,0.002609066665172577
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,5120,128,0.014563199877738953
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,5120,512,0.007534933090209961
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,5120,64,0.0021664001047611236
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,5120,64,0.013868799805641175
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,5120,128,0.007052800059318543
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,5120,32,0.002570666621128718
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,5120,32,0.01447466711203257
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,4096,65536,0.08938240210215251
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,4096,65536,0.07425706386566162
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,4096,16384,0.02616320053736369
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,4096,16384,0.030982400973637896
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,6144,65536,0.1273685296376546
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,4096,12288,0.02000746726989746
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,4096,12288,0.026962133248647054
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,4096,12288,0.02908160090446472
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,4096,10240,0.01812160015106201
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,4096,10240,0.024922666947046916
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,4096,16384,0.03640106519063314
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,4096,8192,0.014590932925542196
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,4096,8192,0.02278719941775004
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,4096,8192,0.02157013416290283
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,4096,7168,0.013142399986584983
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,4096,7168,0.022936532894770302
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,4096,10240,0.025906133651733398
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,4096,6144,0.01188373366991679
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,4096,6144,0.0207370658715566
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,4096,7168,0.01954560081164042
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,4096,5120,0.01042133371035258
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,4096,6144,0.01741546591122945
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,4096,5120,0.01965973377227783
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,4096,4096,0.008774399757385254
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,4096,4096,0.01813439925511678
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,4096,4096,0.014108799894650779
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,4096,3584,0.00798399994770686
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,4096,3584,0.019244800011316933
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,4096,5120,0.01618133286635081
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,4096,3072,0.007341866691907247
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,4096,3072,0.016561067104339598
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,4096,3584,0.012893866499265036
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,4096,2560,0.0063509335120519
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,4096,2560,0.01704853375752767
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,4096,3072,0.01204266647497813
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,4096,2048,0.005133866767088572
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,4096,2048,0.014994133512179056
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,4096,2048,0.010444800059000652
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,4096,1536,0.004229333500067393
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,4096,1536,0.017299199104309083
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,4096,2560,0.011226666967074077
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,4096,1024,0.0034506666163603462
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,4096,1024,0.01730560064315796
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,5120,65536,0.12726399898529053
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,4096,768,0.0030741333961486817
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,4096,768,0.016764799753824867
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,4096,1024,0.008451199531555176
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,4096,512,0.002985599885384242
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,4096,512,0.015014400084813436
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,4096,1536,0.009642666578292847
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,4096,256,0.0025568000972270967
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,4096,256,0.014834133783976236
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,4096,768,0.008006399869918824
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,4096,128,0.002311466634273529
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,4096,512,0.0074976002176602675
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,4096,256,0.007121066749095917
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,4096,128,0.01444906691710154
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,4096,64,0.002152533332506816
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,4096,32,0.0021514666577180227
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,4096,64,0.014458666245142618
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,4096,32,0.013867732882499696
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,4096,128,0.007079466680685679
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,3584,65536,0.07571732997894287
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,3584,65536,0.06477333307266235
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,3584,16384,0.02434026598930359
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,3584,16384,0.03169493277867635
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,3584,12288,0.01882879932721456
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,3584,12288,0.02704213261604309
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,3584,12288,0.028904533386230467
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,3584,10240,0.01618880033493042
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,3584,10240,0.025464532772699992
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,3584,16384,0.036449066797892254
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,3584,8192,0.015041066209475198
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,3584,8192,0.023034665981928507
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,3584,10240,0.025643734137217204
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,3584,7168,0.013594667116800943
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,3584,7168,0.021477333704630532
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,3584,8192,0.02114560008049011
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,3584,6144,0.011508267124493916
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,3584,6144,0.020907733837763467
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,3584,7168,0.0194815993309021
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,3584,5120,0.009941333532333374
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,3584,5120,0.019938133160273232
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,3584,6144,0.017800533771514894
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,3584,4096,0.008476799726486206
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,3584,4096,0.018615466356277467
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,3584,5120,0.016152532895406087
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,3584,3584,0.007597866654396057
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,3584,3584,0.017014400164286295
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,3584,4096,0.013680000106493631
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,3584,3072,0.007222400108973186
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,3584,3072,0.016662399967511496
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,4096,65536,0.12670400142669677
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,3584,2560,0.005913599828879039
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,3584,2560,0.01649173299471537
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,3584,3584,0.01290773351987203
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,3584,2048,0.005156266689300537
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,3584,2560,0.011290666460990906
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,3584,2048,0.016160000363985697
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,3584,1536,0.0043594668308893835
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,3584,1536,0.014806399742762247
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,3584,3072,0.011982933680216471
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,3584,1024,0.0034261333445707956
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,3584,1024,0.01479039986928304
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,3584,1024,0.008327466746171314
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,3584,768,0.003124266614516576
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,3584,768,0.014863999684651694
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,3584,1536,0.009202133615811665
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,3584,512,0.002844800055027008
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,3584,512,0.014493866761525472
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,3584,2048,0.010071466366449993
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,3584,256,0.002648533384005229
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,3584,256,0.01446613371372223
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,3584,768,0.007762133578459422
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,3584,128,0.0025781333446502686
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,3584,256,0.0071263998746871945
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,3584,128,0.01454080045223236
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,3584,512,0.007470933099587758
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,3584,64,0.0026208000878492994
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,3584,32,0.002583466718594233
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,3584,64,0.014495999614397685
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,3584,32,0.014553599556287131
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,3584,128,0.006704000135262807
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,3072,65536,0.06663573185602824
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,3072,16384,0.022886399428049722
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,3072,65536,0.062371198336283365
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,3072,16384,0.030854399998982745
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,3072,12288,0.01794559955596924
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,3072,12288,0.027083732684453327
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,3072,12288,0.028889600435892743
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,3072,10240,0.016166399916013083
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,3072,10240,0.024422399202982583
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,3072,16384,0.036296534538269046
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,3072,8192,0.014072533448537192
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,3072,8192,0.022835199038187662
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,3072,10240,0.025269333521525068
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,3072,7168,0.012793599565823873
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,3072,7168,0.021401600042978922
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,3072,8192,0.021527467171351115
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,3072,6144,0.011699199676513672
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,3072,6144,0.020682666699091592
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,3072,7168,0.01946880022684733
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,3072,5120,0.009646933277448018
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,3072,5120,0.018655999501546224
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,3072,6144,0.01752426624298096
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,3072,4096,0.008323200047016144
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,3072,4096,0.01886826753616333
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,3072,5120,0.01579093337059021
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,3072,3584,0.0073845331867535904
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,3072,3584,0.016616533199946083
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,3072,4096,0.013742933670679728
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,3072,3072,0.006392533580462138
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,3072,3072,0.01665173371632894
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,3072,3584,0.012875733772913614
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,3072,2560,0.0058794667323430385
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,3072,2560,0.016692266861597697
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,3072,3072,0.012088533242543538
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,3072,2048,0.0047189335028330484
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,3072,2048,0.016612266500790916
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,3072,2560,0.011220266421635944
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,3072,1536,0.004264533519744873
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,3072,1536,0.014654933412869772
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,3072,2048,0.010056533416112264
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,3072,1024,0.00347626656293869
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,3072,1024,0.016127999623616537
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,3584,65536,0.12708266576131183
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,3072,768,0.0031285333136717476
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,3072,1536,0.009193600217501322
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,3072,768,0.014891733725865683
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,3072,512,0.002993066608905792
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,3072,512,0.014506666858990987
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,3072,512,0.0075103998184204105
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,3072,256,0.0025802666942278544
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,3072,256,0.014432000120480857
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,3072,768,0.007985066870848339
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,3072,128,0.002518400053183238
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,3072,128,0.014512000481287637
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,3072,1024,0.008378666639328004
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,3072,64,0.002217599997917811
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,3072,64,0.014451199769973755
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,3072,32,0.002586666742960612
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,3072,256,0.007088000078996022
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,3072,32,0.014546133081118264
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,3072,128,0.006673066814740498
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,2560,65536,0.05635840098063151
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,2560,65536,0.057461333274841306
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,2560,16384,0.021002666155497233
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,2560,16384,0.029578665892283123
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,2560,12288,0.01704960068066915
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,2560,12288,0.027309866746266682
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,2560,12288,0.029235200087229414
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,2560,10240,0.015340800086657206
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,2560,10240,0.025240532557169598
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,2560,16384,0.03588693141937256
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,2560,8192,0.013008000453313193
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,2560,8192,0.022102399667104086
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,2560,10240,0.02483946681022644
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,2560,7168,0.012523733576138816
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,2560,7168,0.02127893368403117
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,2560,8192,0.021160533030827842
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,2560,6144,0.011361066500345867
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,2560,6144,0.019358932971954346
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,2560,7168,0.01914880077044169
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,2560,5120,0.00960106650988261
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,2560,5120,0.020385066668192543
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,2560,6144,0.017773866653442383
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,2560,4096,0.007942399879296621
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,2560,4096,0.017790933450063072
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,2560,4096,0.01333440045515696
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,2560,3584,0.006871466835339864
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,2560,3584,0.016804265975952148
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,2560,5120,0.01578133304913839
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,2560,3072,0.006330666442712148
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,2560,3072,0.018592000007629395
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,2560,3072,0.01202880044778188
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,2560,2560,0.005184000233809153
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,2560,2560,0.016696532567342125
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,2560,3584,0.012900267044703165
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,2560,2048,0.004640000065167745
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,2560,2048,0.014684800306955972
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,3072,65536,0.1264906644821167
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,2560,1536,0.004265599946180979
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,2560,1536,0.016344533363978068
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,2560,1536,0.00916266640027364
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,2560,1024,0.0034805332620938623
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,2560,1024,0.01823893388112386
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,2560,2048,0.00999786655108134
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,2560,768,0.0030858665704727173
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,2560,2560,0.011215999722480774
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,2560,768,0.01552959978580475
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,2560,512,0.002974933385848999
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,2560,512,0.014461867014567056
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,2560,512,0.007492266595363617
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,2560,256,0.002621866762638092
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,2560,1024,0.008332799871762593
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,2560,768,0.007997866471608479
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,2560,128,0.002493866781393687
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,2560,256,0.013990400234858194
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,2560,128,0.014532267053922018
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,2560,64,0.002221866697072983
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,2560,64,0.014475733041763306
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,2560,32,0.002611200014750163
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,2560,128,0.007089066505432129
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,2560,32,0.014435199896494546
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,2560,256,0.007528533538182576
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,2048,65536,0.04688320159912109
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,2048,16384,0.017303466796875
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,2048,65536,0.05085973342259725
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,2048,16384,0.02688746651013692
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,2048,12288,0.013688533504803976
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,2048,12288,0.0247925341129303
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,2048,12288,0.02839253346125285
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,2048,10240,0.012513066331545511
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,2048,10240,0.023206400871276855
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,2048,16384,0.03592746655146281
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,2048,8192,0.010820266604423524
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,2048,8192,0.022091732422510783
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,2048,10240,0.024881066878636678
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,2048,7168,0.01028053363164266
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,2048,7168,0.020592000087102255
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,2048,8192,0.020786132415135702
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,2048,6144,0.009448533256848652
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,2048,6144,0.019223467508951823
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,2048,7168,0.019509333372116088
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,2048,5120,0.008373333017031352
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,2048,5120,0.01819733381271362
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,2048,6144,0.017806933323542277
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,2048,4096,0.007441066702206929
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,2048,4096,0.018811732530593872
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,2048,5120,0.01585599978764852
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,2048,3584,0.006715733309586842
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,2048,3584,0.016574933131535848
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,2048,4096,0.013741866747538248
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,2048,3072,0.005880533158779145
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,2048,3072,0.018743467330932618
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,2048,3584,0.012852266430854797
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,2048,2560,0.00510506679614385
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,2048,2560,0.017463467518488564
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,2048,3072,0.01167680025100708
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,2048,2048,0.004693333307902018
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,2048,2048,0.01623146633307139
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,2048,2560,0.011248000462849935
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,2048,1536,0.0038560000558694207
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,2048,1536,0.015217066804567973
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,2048,2048,0.010081066687901815
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,2048,1024,0.003443199892838796
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,2048,1024,0.014954666296641031
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,2048,1536,0.009187199672063192
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,2048,768,0.0030613332986831666
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,2048,768,0.014686933159828186
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,2048,768,0.008022400240103405
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,2048,1024,0.008345599969228108
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,2048,512,0.002647466709216436
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,2048,512,0.014840533336003622
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,2048,256,0.0026496000587940215
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,2048,256,0.015566933155059814
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,2560,65536,0.12680319945017499
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,2048,256,0.00719893326361974
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,2048,128,0.0022485333184401194
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,2048,128,0.01446399986743927
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,2048,128,0.006760533154010773
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,2048,64,0.0021856000026067098
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,2048,64,0.014546133081118264
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,2048,32,0.0021802666286627453
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,2048,32,0.014497066537539164
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,1536,65536,0.035716267426808675
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,1536,65536,0.045661866664886475
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,1536,16384,0.013805866241455078
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,1536,16384,0.02542080084482829
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,1536,16384,0.03558613459269206
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,1536,12288,0.011316266655921937
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,1536,12288,0.023740800221761067
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,1536,12288,0.028540800015131634
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,1536,10240,0.010780800382296245
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,1536,10240,0.02301759918530782
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,2048,65536,0.12636160055796306
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,1536,8192,0.009645866354306538
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,1536,8192,0.021079466740290324
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,1536,10240,0.02486720085144043
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,1536,7168,0.008609066406885784
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,1536,7168,0.020806399981180827
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,1536,8192,0.020810665686925252
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,1536,6144,0.007975466549396515
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,1536,6144,0.01977919936180115
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,1536,7168,0.019488000869750978
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,1536,5120,0.007536000013351441
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,1536,5120,0.018269866704940796
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,1536,5120,0.0157258669535319
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,1536,4096,0.006729599833488464
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,1536,4096,0.01844266653060913
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,1536,6144,0.017460266749064125
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,1536,3584,0.006284800171852112
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,1536,3584,0.01729493339856466
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,1536,3584,0.012890666723251343
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,1536,3072,0.005584000051021576
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,1536,3072,0.016578132907549538
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,1536,4096,0.013686399658521017
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,1536,2560,0.005066666503747304
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,1536,2560,0.017356799046198527
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,1536,2560,0.01121493379275004
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,1536,2048,0.004634666442871094
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,1536,2048,0.014878933628400167
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,1536,3072,0.011587199568748475
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,1536,1536,0.003852800031503042
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,1536,1536,0.016054399808247886
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,1536,1536,0.009131733576456707
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,1536,1024,0.003370666752258936
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,1536,1024,0.01593066652615865
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,1536,2048,0.009947733084360758
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,1536,768,0.002998399982849757
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,1536,768,0.014452266693115234
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,1536,1024,0.008354133367538452
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,1536,768,0.007901866734027863
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,1536,512,0.0026517334083716077
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,1536,512,0.01453439990679423
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,1536,256,0.002570666621128718
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,1536,256,0.01444586714108785
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,1536,512,0.007149866720040639
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,1536,256,0.007143466671307881
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,1536,128,0.0022463999688625337
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,1536,128,0.014456533392270408
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,1536,64,0.002198400100072225
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,1536,128,0.006657066444555919
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,1536,64,0.014513066411018372
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,1536,32,0.002203733225663503
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,1536,32,0.014436266819636025
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,1024,65536,0.029392000039418536
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,1024,65536,0.039389868577321366
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,1024,16384,0.010806399583816528
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,1024,16384,0.023306665817896526
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,1536,65536,0.12598933378855387
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,1024,12288,0.009543466567993163
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,1024,12288,0.022717867294947305
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,1024,16384,0.03549439907073974
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,1024,10240,0.00848426620165507
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,1024,10240,0.020859734217325846
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,1024,10240,0.024488532543182374
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,1024,12288,0.028089600801467895
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,1024,8192,0.007835733393828075
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,1024,8192,0.020760534207026164
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,1024,7168,0.006814933319886525
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,1024,7168,0.020510933796564736
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,1024,7168,0.019435733556747437
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,1024,6144,0.006309333443641663
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,1024,6144,0.020042665799458823
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,1024,8192,0.020721065998077392
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,1024,5120,0.006758399804433187
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,1024,5120,0.0177130659421285
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,1024,6144,0.017411200205485027
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,1024,4096,0.006746666630109151
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,1024,5120,0.01542080044746399
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,1024,4096,0.01859626571337382
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,1024,3584,0.006296533346176148
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,1024,3584,0.016587733229001363
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,1024,3584,0.012558933099110922
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,1024,3072,0.005542399982611338
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,1024,3072,0.016934400796890257
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,1024,4096,0.013699199755986533
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,1024,3072,0.011732266346613566
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,1024,2560,0.005055999755859375
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,1024,2560,0.016723199685414632
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,1024,2048,0.004786133269468943
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,1024,2048,0.01646080017089844
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,1024,2560,0.010813867052396137
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,1024,1536,0.003802666564782461
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,1024,1536,0.016454399625460307
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,1024,2048,0.01002346674601237
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,1024,1024,0.0033952000240484873
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,1024,1024,0.015686399737993875
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,1024,1536,0.009107200304667155
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,1024,768,0.0029728000362714132
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,1024,768,0.01495039959748586
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,1024,1024,0.00836906631787618
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,1024,512,0.0026208000878492994
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,1024,512,0.01474133332570394
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,1024,768,0.007948799928029378
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,1024,256,0.0023786666492621104
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,1024,256,0.014522666732470194
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,1024,512,0.007433600227038066
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,1024,128,0.0022463999688625337
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,1024,128,0.014459733168284097
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,1024,256,0.007107200225194295
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,1024,64,0.002183466653029124
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,1024,64,0.014428800344467163
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,1024,128,0.00631039987007777
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,1024,32,0.0021344001094500227
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,768,65536,0.02368320027987162
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,1024,32,0.01447466711203257
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,768,65536,0.035382401943206784
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,768,16384,0.00993386705716451
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,768,16384,0.02205973267555237
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,768,16384,0.03509226640065511
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,768,12288,0.008036266764005024
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,768,12288,0.021305600802103676
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,1024,65536,0.12405227025349934
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,768,10240,0.007575466732184092
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,768,10240,0.021643733978271483
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,768,12288,0.02849493424097697
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,768,8192,0.006358399987220764
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,768,8192,0.020866133769353232
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,768,10240,0.024371200799942018
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,768,7168,0.006302933394908905
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,768,7168,0.02099306583404541
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,768,8192,0.02071146567662557
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,768,6144,0.006266666452089946
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,768,6144,0.01895786722501119
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,768,7168,0.019129600127538046
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,768,5120,0.006763733426729838
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,768,5120,0.01888213356335958
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,768,6144,0.01772586703300476
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,768,4096,0.006295466423034668
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,768,4096,0.01679146687189738
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,768,5120,0.015363199512163797
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,768,3584,0.005919999877611796
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,768,3584,0.016638933618863424
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,768,4096,0.01369599997997284
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,768,3072,0.005510400235652924
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,768,3072,0.01653333306312561
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,768,3584,0.012440533439318339
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,768,2560,0.0054624001185099285
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,768,2560,0.016682666540145875
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,768,3072,0.012009599804878235
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,768,2048,0.004557866851488749
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,768,2560,0.010600533088048298
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,768,2048,0.016353066762288412
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,768,1536,0.0038560000558694207
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,768,1536,0.014987732966740927
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,768,1536,0.00913813312848409
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,768,2048,0.010021332899729412
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,768,1024,0.0033962666988372804
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,768,1024,0.016056533654530844
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,768,768,0.0030378667016824085
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,768,768,0.01576746702194214
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,768,1024,0.008319999774297078
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,768,512,0.002631466587384542
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,768,768,0.0075434664885203045
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,768,512,0.014460800091425577
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,768,256,0.0023189333577950796
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,768,256,0.014614400267601014
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,768,512,0.007485866546630859
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,768,128,0.0022698665658632917
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,768,256,0.007146666447321574
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,768,128,0.014428800344467163
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,768,64,0.002133333434661229
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,768,128,0.006735999882221222
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,768,64,0.014549332857131957
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,768,32,0.0021898667017618816
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,512,65536,0.017516799767812095
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,512,65536,0.030990932385126752
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,768,65536,0.12162026564280193
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,512,16384,0.007124266525109608
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,512,16384,0.023398399353027344
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,512,16384,0.035306668281555174
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,512,12288,0.006748799979686737
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,512,12288,0.022105600436528525
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,512,12288,0.028099199136098225
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,512,10240,0.006559999783833821
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,512,10240,0.022125866015752158
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,512,10240,0.02478826642036438
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,512,65536,0.12123520374298095
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,512,8192,0.005888000130653381
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,512,8192,0.020517333348592123
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,512,7168,0.005852800110975901
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,512,7168,0.01964799960454305
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,512,7168,0.018978132804234823
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,512,6144,0.005925333499908448
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,512,6144,0.020422399044036865
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,512,8192,0.02101226647694906
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,512,5120,0.006668800115585327
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,512,5120,0.019961599508921304
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,512,5120,0.01569919983545939
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,512,6144,0.016927999258041383
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,512,4096,0.005929600199063619
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,512,4096,0.01836693286895752
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,512,3584,0.005784533421198527
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,512,3584,0.01797440052032471
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,512,3584,0.012402133146921793
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,512,3072,0.00544106662273407
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,512,3072,0.016571733355522155
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,512,3072,0.011603200435638427
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,512,2560,0.005092266698678335
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,512,2560,0.016569599509239197
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,512,2560,0.010777599612871806
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,512,2048,0.0050005331635475155
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,512,2048,0.014617600043614707
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,512,2048,0.010002133250236512
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,512,1536,0.0038101332883040107
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,512,1536,0.01639573375384013
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,512,1536,0.008921600381533305
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,512,1024,0.003385599950949351
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,512,1024,0.0166485329469045
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,512,1024,0.008286933104197185
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,512,768,0.0029909332593282064
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,512,768,0.014514133334159851
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,512,768,0.007914666831493378
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,512,512,0.0025813333690166474
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,512,512,0.014473600188891092
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,512,512,0.007542400062084198
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,512,256,0.002550400048494339
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,512,256,0.014647466937700906
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,512,256,0.007089066505432129
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,512,128,0.002147199958562851
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,512,128,0.013076266646385193
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,512,128,0.007066666583220164
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,512,64,0.002169599880774816
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,512,64,0.01444586714108785
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,512,32,0.002201599876085917
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,512,32,0.013717333475748697
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,256,65536,0.01202453374862671
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,256,65536,0.026948267221450807
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,256,65536,0.12077013651529948
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,256,16384,0.0063274666666984555
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,256,16384,0.021316266059875487
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,256,16384,0.03496426741282145
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,256,12288,0.005467733244101206
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,256,12288,0.020564266045888267
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,256,12288,0.0280021329720815
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,256,10240,0.005930666625499725
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,256,10240,0.0188810666402181
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,256,10240,0.02433919906616211
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,256,8192,0.0061152001221974695
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,256,8192,0.01950506567955017
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,256,8192,0.02062186598777771
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,256,7168,0.005497600138187409
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,256,7168,0.018722132841746012
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,256,7168,0.018936532735824584
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,256,6144,0.005433600147565206
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,256,6144,0.018601600329081217
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,256,6144,0.017350399494171144
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,256,5120,0.005089066425959269
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,256,5120,0.019323732455571493
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,256,5120,0.015252266327540079
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,256,4096,0.005426133175690969
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,256,4096,0.018631466229756675
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,256,4096,0.013633066415786743
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,256,3584,0.005418666700522105
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,256,3584,0.016500266393025716
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,256,3584,0.012390399972597759
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,256,3072,0.005108266572157542
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,256,3072,0.0165503998597463
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,256,3072,0.011627733707427979
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,256,2560,0.005432533224423727
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,256,2560,0.01662613352139791
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,256,2560,0.011159466703732808
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,256,2048,0.005029333134492239
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,256,2048,0.014500266313552857
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,256,2048,0.010373333096504211
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,256,1536,0.003790933390458425
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,256,1536,0.015466666221618653
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,256,1536,0.009124267101287841
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,256,1024,0.0032117334504922234
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,256,1024,0.014513066411018372
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,256,1024,0.008365866541862488
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,256,768,0.002994133283694585
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,256,768,0.014428800344467163
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,256,768,0.0075434664885203045
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,256,512,0.0025909334421157838
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,256,512,0.014567466576894126
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,256,512,0.007609599828720092
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,256,256,0.00230826660990715
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,256,256,0.01588266690572103
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,256,256,0.00673173318306605
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,256,128,0.0022079999248186748
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,256,128,0.013555199901262919
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,256,128,0.006748799979686737
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,256,64,0.0022026665508747103
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,256,64,0.014512000481287637
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,256,32,0.0021738665799299877
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,256,32,0.01309440036614736
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,128,65536,0.008359466989835102
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,128,65536,0.02763413389523824
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,128,65536,0.1207306702931722
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,128,16384,0.005931733548641205
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,128,16384,0.021660800774892172
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,128,16384,0.034986666838328045
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,128,12288,0.005509333312511444
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,128,12288,0.020824533700942994
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,128,12288,0.027992532650629683
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,128,10240,0.0058890665570894875
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,128,10240,0.019012266397476198
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,128,10240,0.024353067080179848
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,128,8192,0.0058442667126655575
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,128,8192,0.01958079934120178
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,128,8192,0.020705066124598184
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,128,7168,0.005482666691144307
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,128,7168,0.018943999210993448
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,128,7168,0.019033600886662803
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,128,6144,0.0054506664474805195
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,128,6144,0.019552000363667808
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,128,6144,0.017362133661905924
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,128,5120,0.005083733300367991
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,128,5120,0.018609066804250084
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,128,5120,0.01564906636873881
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,128,4096,0.004634666442871094
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,128,4096,0.018242132663726807
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,128,4096,0.013459199666976928
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,128,3584,0.0050357331832249965
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,128,3584,0.016489600141843162
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,128,3584,0.012469333410263062
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,128,3072,0.00480320006608963
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,128,3072,0.01649173299471537
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,128,3072,0.01162559986114502
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,128,2560,0.005017599960168203
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,128,2560,0.017004799842834473
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,128,2560,0.010743467013041179
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,128,2048,0.004645333190759023
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,128,2048,0.015220266580581666
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,128,2048,0.009616000453631084
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,128,1536,0.0038272000849246977
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,128,1536,0.015954132874806723
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,128,1536,0.009136000275611877
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,128,1024,0.0030026666820049284
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,128,1024,0.015516799688339234
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,128,1024,0.00809279978275299
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,128,768,0.0029728000362714132
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,128,768,0.014793599645296732
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,128,768,0.00751146674156189
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,128,512,0.0025770666698614756
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,128,512,0.013809067010879517
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,128,512,0.0074976002176602675
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,128,256,0.0021759999295075734
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,128,256,0.013607466220855713
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,128,256,0.007092266778151194
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,128,128,0.002162133405605952
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,128,128,0.012945066889127096
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,128,128,0.0066549330949783325
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,128,64,0.0021301334102948504
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,128,64,0.01309866706530253
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,128,32,0.0021738665799299877
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,128,32,0.012757333119710288
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,64,65536,0.00751146674156189
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,64,65536,0.02690560022989909
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,64,16384,0.005862399935722351
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,64,16384,0.0225983997186025
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,64,12288,0.005432533224423727
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,64,12288,0.020881066719690956
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,64,10240,0.005064533154169718
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,64,10240,0.01885973413785299
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,64,8192,0.00544213354587555
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,64,8192,0.018895999590555827
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,64,7168,0.005070933202902476
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,64,7168,0.019000534216562906
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,64,6144,0.005065600077311197
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,64,6144,0.01857066750526428
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,64,5120,0.004690133531888326
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,64,5120,0.018845866123835243
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,64,4096,0.004600533346335093
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,64,4096,0.018186666568120322
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,64,3584,0.0046517332394917805
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,64,3584,0.018028799692789713
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,64,3072,0.0046079998215039575
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,64,3072,0.01686720053354899
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,64,2560,0.0050453335046768185
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,64,2560,0.016437333822250367
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,64,2048,0.004634666442871094
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,64,2048,0.01639253298441569
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,64,1536,0.0034634667138258614
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,64,1536,0.014801067113876343
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,64,1024,0.003081600119670232
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,64,1024,0.014686933159828186
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,64,768,0.0027424000203609467
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,64,768,0.01477120021979014
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,64,512,0.0026677332818508146
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,64,512,0.013768532872200012
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,64,256,0.0022698665658632917
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,64,256,0.014085333546002707
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,64,128,0.0022111999491850535
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,64,128,0.013225600123405457
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,64,64,0.0018304000298182172
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,64,64,0.012547199924786886
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,64,32,0.0018304000298182172
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,64,32,0.014443733294804893
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,32,65536,0.0091839998960495
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,32,65536,0.026796799898147584
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,32,16384,0.005866666634877523
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,32,16384,0.020779732863108316
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,32,12288,0.005133866767088572
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,32,12288,0.020760534207026164
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,32,10240,0.005076266825199127
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,32,10240,0.019197867314020792
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,32,8192,0.005446400245030721
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,32,8192,0.018601600329081217
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,32,7168,0.005128533144791921
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,32,7168,0.019427200158437095
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,32,6144,0.005038933455944061
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,32,6144,0.018542933464050292
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,32,5120,0.004645333190759023
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,32,5120,0.01845653255780538
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,32,4096,0.004265599946180979
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,32,4096,0.018052266041437785
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,32,3584,0.004669866462548574
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,32,3584,0.017361066738764443
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,32,3072,0.0046165332198143
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,32,3072,0.016509866714477538
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,32,2560,0.004634666442871094
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,32,2560,0.01709653337796529
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,32,2048,0.004273066421349844
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,32,2048,0.016267733772595723
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,32,1536,0.003398400048414866
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,32,1536,0.015501866738001505
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,32,1024,0.002998399982849757
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,32,1024,0.014427733421325684
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,32,768,0.0025343999266624452
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,32,768,0.014458666245142618
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,32,512,0.0025813333690166474
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,32,512,0.015084800124168397
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,32,256,0.002182399978240331
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,32,256,0.01448853313922882
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,32,128,0.002170666555563609
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,32,128,0.014378666877746582
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,32,64,0.0017279999951521556
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,32,64,0.013901866475741067
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,64,32,32,0.001803733284274737
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,32,32,0.012781866391499839
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,65536,16384,0.36473067601521814
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,65536,16384,0.20215360323588052
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,2048,512,0.007527466615041096
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,65536,12288,0.266211207707723
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,65536,12288,0.15665067036946614
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,64,768,32,0.014485333363215128
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,65536,10240,0.22668800354003907
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,65536,10240,0.1270965337753296
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,64,512,4096,0.013659733533859252
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,65536,8192,0.18546346028645833
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,65536,8192,0.10486079851786297
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,65536,8192,0.062244268258412686
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,65536,7168,0.15549440383911134
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,65536,7168,0.094651730855306
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,65536,10240,0.07453760306040445
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,65536,6144,0.13929707209269207
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,65536,6144,0.08444159825642904
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,65536,12288,0.09431040287017822
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,65536,5120,0.11251520315806071
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,65536,5120,0.07355413436889649
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,65536,16384,0.11538026332855225
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,65536,4096,0.09620587031046549
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,65536,4096,0.06563733418782552
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,65536,6144,0.04867946704228719
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,65536,3584,0.08050666650136312
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,65536,3584,0.05704106489817301
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,65536,5120,0.04597973426183065
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,65536,3072,0.07076586882273356
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,65536,3072,0.05241599877675375
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,65536,7168,0.06069013277689615
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,65536,2560,0.05918613274892172
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,65536,2560,0.04620159864425659
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,65536,4096,0.035358933607737224
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,65536,2048,0.05054080088933309
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,65536,2048,0.04003200133641561
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,65536,3072,0.028962133328119914
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,65536,1536,0.03804693222045898
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,65536,1536,0.033662935098012284
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,65536,3584,0.03222613334655762
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,65536,1024,0.02717439929644267
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,65536,1024,0.027289599180221558
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,65536,2560,0.025681066513061523
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,65536,768,0.020703999201456706
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,65536,768,0.02463573416074117
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,65536,1024,0.014586666226387024
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,65536,512,0.014457600315411887
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,65536,512,0.02076373298962911
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,65536,768,0.013326932986577352
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,65536,256,0.008636800448099773
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,65536,256,0.018710400660832724
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,65536,1536,0.018623999754587808
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,65536,128,0.005864533285299936
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,65536,128,0.016563199957211814
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,65536,2048,0.022083199024200438
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,65536,64,0.005449600021044413
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,65536,64,0.016536532839139303
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,65536,32,0.005449600021044413
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,65536,32,0.018645334243774413
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,16384,65536,0.3433237393697103
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,16384,65536,0.20773332913716636
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,65536,128,0.008753066261609394
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,16384,16384,0.08287893136342367
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,16384,16384,0.05752213398615519
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,65536,256,0.009592533111572266
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,16384,12288,0.0608245333035787
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,16384,12288,0.048028798898061116
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,65536,512,0.010937600334485372
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,16384,10240,0.05224639972050985
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,16384,10240,0.04386026859283447
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,16384,10240,0.029292800029118854
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,16384,8192,0.04194453159968058
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,16384,8192,0.03741013209025065
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,16384,12288,0.0334826668103536
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,16384,7168,0.037140266100565596
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,16384,7168,0.035181868076324466
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,16384,16384,0.04157973527908325
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,16384,6144,0.03216960032780965
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,16384,6144,0.030958932638168336
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,16384,8192,0.024551467100779215
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,16384,5120,0.028191999594370527
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,16384,5120,0.02996586759885152
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,16384,7168,0.022062933444976805
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,16384,4096,0.02247040073076884
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,16384,4096,0.026177066564559936
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,16384,5120,0.018145066499710084
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,16384,3584,0.020515199502309164
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,16384,3584,0.025199999411900837
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,16384,6144,0.020257065693537392
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,16384,3072,0.017619200547536216
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,16384,3072,0.022737065951029457
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,16384,4096,0.015787733594576518
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,16384,2560,0.015086932977040609
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,16384,2560,0.022256000836690267
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,16384,2560,0.012525866429011026
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,16384,2048,0.01241386632124583
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,16384,2048,0.021197867393493653
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,16384,3584,0.014884266257286071
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,16384,1536,0.009956266482671101
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,16384,1536,0.01964799960454305
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,16384,3072,0.013969066739082336
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,16384,1024,0.007520000139872233
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,16384,1024,0.015250133474667868
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,16384,1024,0.008433066805203756
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,16384,768,0.006270933151245117
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,16384,768,0.014491732915242514
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,16384,1536,0.009708799918492635
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,16384,512,0.004228266576925913
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,16384,512,0.015836800138155617
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,16384,512,0.00766186664501826
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,16384,256,0.0030346666773160298
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,16384,256,0.014500266313552857
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,16384,2048,0.010892800490061442
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,16384,128,0.0027072000006834666
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,16384,128,0.014658133188883463
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,16384,768,0.008398933211962382
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,16384,64,0.00257493332028389
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,16384,64,0.0144405335187912
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,16384,32,0.002586666742960612
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,16384,32,0.014507733782132468
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,12288,65536,0.2577621301015218
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,16384,128,0.007156266768773396
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,16384,256,0.007493333518505096
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,12288,16384,0.07140586376190186
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,12288,65536,0.15920106569925946
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,12288,16384,0.052128001054128015
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,12288,12288,0.05030293464660644
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,12288,12288,0.042957866191864015
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,12288,12288,0.03104640046755473
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,12288,10240,0.04327573378880818
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,12288,10240,0.03889919916788737
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,12288,16384,0.03919253349304199
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,12288,8192,0.03548906644185384
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,12288,8192,0.03290773431460063
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,12288,10240,0.02735466758410136
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,12288,7168,0.03069973389307658
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,12288,7168,0.03107733329137166
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,12288,8192,0.022806400060653688
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,12288,6144,0.027463465929031372
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,12288,6144,0.02885226607322693
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,12288,7168,0.02118826707204183
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,12288,5120,0.0231221338113149
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,12288,5120,0.026985599597295122
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,12288,6144,0.019030400117238364
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,12288,4096,0.018835200866063436
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,12288,4096,0.02505600055058797
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,12288,5120,0.017285333077112833
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,12288,3584,0.01683733264605204
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,12288,3584,0.024691200256347655
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,12288,4096,0.01508799990018209
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,12288,3072,0.015362133582433065
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,12288,3072,0.022533333301544188
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,12288,3584,0.014221866925557455
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,12288,2560,0.013315199812253316
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,12288,2560,0.020756266514460244
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,12288,2560,0.012060800194740295
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,12288,2048,0.011250133315722149
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,12288,2048,0.021747199694315593
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,12288,3072,0.012842667102813721
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,12288,1536,0.009164800246556599
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,12288,1536,0.017468800147374473
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,12288,1536,0.00960213343302409
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,12288,1024,0.007108266651630402
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,12288,1024,0.017588265736897788
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,12288,2048,0.010486400127410889
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,12288,768,0.005942399799823761
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,12288,768,0.01624853312969208
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,12288,768,0.007985066870848339
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,12288,512,0.004345599810282389
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,12288,512,0.015658666690190635
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,12288,1024,0.008313600222269695
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,12288,256,0.0034517332911491393
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,12288,256,0.014468266566594442
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,12288,256,0.007144533097743988
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,12288,128,0.00297173336148262
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,12288,128,0.014593066771825156
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,12288,128,0.006807466844717662
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,12288,64,0.002570666621128718
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,12288,64,0.014497066537539164
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,12288,32,0.0027253332237402597
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,12288,32,0.01444906691710154
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,12288,512,0.0075445334116617845
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,10240,65536,0.21098453203837075
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,10240,65536,0.14476799964904785
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,10240,16384,0.06101653178532919
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,10240,16384,0.045348266760508224
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,10240,16384,0.03854506810506185
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,10240,12288,0.03997653325398763
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,10240,12288,0.0373418649037679
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,10240,12288,0.03038613398869832
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,10240,10240,0.03465919891993205
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,10240,10240,0.03423253297805786
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,16384,65536,0.1437109311421712
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,10240,8192,0.02791999975840251
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,10240,8192,0.02917226751645406
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,10240,10240,0.026246400674184163
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,10240,7168,0.02458239992459615
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,10240,7168,0.028388265768686933
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,10240,8192,0.022346667448679605
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,10240,6144,0.02196906606356303
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,10240,6144,0.02765866716702779
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,12288,65536,0.1323743979136149
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,10240,5120,0.018900267283121743
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,10240,5120,0.024681599934895833
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,10240,7168,0.020887466271718343
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,10240,4096,0.015218133727709452
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,10240,4096,0.02294720013936361
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,10240,4096,0.01495573321978251
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,10240,3584,0.013511466979980468
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,10240,3584,0.020830933252970377
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,10240,5120,0.017450666427612303
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,10240,3072,0.01200213332970937
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,10240,3072,0.022005333503087362
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,10240,6144,0.019323732455571493
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,10240,2560,0.010797866185506185
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,10240,2560,0.02015893260637919
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,10240,2560,0.011267200112342834
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,10240,2048,0.009224533041318258
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,10240,2048,0.01851200064023336
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,10240,3072,0.01241386632124583
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,10240,1536,0.007480533421039581
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,10240,1536,0.0165173331896464
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,10240,1536,0.00960106650988261
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,10240,1024,0.00544213354587555
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,10240,1024,0.017038933436075845
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,10240,2048,0.010097066561381023
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,10240,768,0.0038975998759269714
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,10240,768,0.0145578662554423
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,10240,1024,0.008358400066693623
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,10240,512,0.0036330667634805044
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,10240,512,0.01606826682885488
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,10240,768,0.008025600016117096
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,10240,256,0.002974933385848999
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,10240,256,0.014546133081118264
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,10240,512,0.007492266595363617
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,10240,128,0.002993066608905792
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,10240,128,0.014539733529090881
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,10240,128,0.006734933455785115
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,10240,64,0.002643200010061264
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,10240,64,0.014493866761525472
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,10240,256,0.0075103998184204105
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,10240,32,0.0026367999613285064
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,10240,32,0.014473600188891092
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,8192,65536,0.18063039779663087
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,8192,65536,0.11757439772288006
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,8192,16384,0.04614400068918864
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,8192,16384,0.04178453286488851
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,8192,16384,0.03835626840591431
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,8192,12288,0.036500267187754315
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,8192,12288,0.03360319932301839
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,10240,65536,0.13025600115458172
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,8192,10240,0.03038826584815979
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,8192,10240,0.029077333211898804
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,8192,12288,0.029781333605448407
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,8192,8192,0.022456532716751097
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,8192,8192,0.027239465713500978
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,8192,8192,0.022380799055099487
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,8192,7168,0.01994880040486654
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,8192,7168,0.02513386607170105
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,8192,10240,0.02603413263956706
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,8192,6144,0.01775146722793579
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,8192,6144,0.024064000447591147
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,8192,6144,0.018631466229756675
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,8192,5120,0.015219199657440185
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,8192,5120,0.023731199900309245
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,8192,7168,0.020719999074935914
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,8192,4096,0.012302933136622111
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,8192,4096,0.021970132986704506
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,8192,5120,0.01699626644452413
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,8192,3584,0.011220266421635944
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,8192,3584,0.020719999074935914
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,8192,4096,0.014647466937700906
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,8192,3072,0.01088533302148183
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,8192,3072,0.020547199249267577
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,8192,3072,0.012121599912643433
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,8192,2560,0.009616000453631084
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,8192,2560,0.018782933553059898
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,8192,3584,0.012967466314633688
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,8192,2048,0.008400000135103862
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,8192,2048,0.018147200345993042
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,8192,2048,0.009986133376757304
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,8192,1536,0.0068341334660847975
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,8192,1536,0.016212266683578492
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,8192,2560,0.01130986710389455
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,8192,1024,0.005102933446566264
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,8192,1024,0.015896532932917276
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,8192,1024,0.008402132987976074
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,8192,768,0.003987200061480204
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,8192,768,0.016429866353670754
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,8192,1536,0.009165866176287334
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,8192,512,0.0033919999996821085
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,8192,512,0.014792533715566
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,8192,768,0.007982933521270752
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,8192,256,0.002980266759792964
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,8192,256,0.014663466811180114
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,8192,512,0.00747626672188441
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,8192,128,0.002585600068171819
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,8192,128,0.014456533392270408
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,8192,128,0.006669866542021434
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,8192,64,0.00257493332028389
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,8192,64,0.014461867014567056
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,8192,32,0.0025973332424958544
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,8192,256,0.007102933526039123
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,8192,32,0.014436266819636025
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,7168,65536,0.15199893315633137
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,7168,65536,0.1018229325612386
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,7168,16384,0.04364800055821737
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,7168,16384,0.04166613419850667
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,10240,3584,0.01409173309803009
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,7168,12288,0.03436479965845744
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,7168,12288,0.031896533568700154
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,7168,12288,0.029421865940093994
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,7168,10240,0.030136533578236896
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,7168,10240,0.028860799471537274
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,7168,16384,0.0374570647875468
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,7168,8192,0.02070080041885376
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,7168,8192,0.02504533330599467
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,7168,10240,0.025998934110005693
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,7168,7168,0.018246400356292724
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,7168,7168,0.024711465835571288
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,8192,65536,0.12760960261027018
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,7168,6144,0.016160000363985697
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,7168,6144,0.022677334149678548
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,7168,8192,0.021996800104777017
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,7168,5120,0.014019200205802917
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,7168,5120,0.023471999168395995
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,7168,7168,0.020725333690643312
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,7168,4096,0.011706667145093282
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,7168,5120,0.016614400347073875
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,7168,4096,0.020770132541656494
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,7168,3584,0.01050453285376231
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,7168,3584,0.0207370658715566
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,7168,6144,0.018668800592422485
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,7168,3072,0.009365333120028178
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,7168,3072,0.018939733505249023
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,7168,3584,0.012922666470209756
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,7168,2560,0.009198932846387228
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,7168,2560,0.018579200903574625
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,7168,3072,0.012074666221936543
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,7168,2048,0.007946666578451793
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,7168,2048,0.016531200210253397
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,7168,4096,0.014227199554443359
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,7168,1536,0.006274133423964183
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,7168,1536,0.017543466885884602
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,7168,2560,0.010868266224861145
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,7168,1024,0.005016533533732096
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,7168,1024,0.016034133235613503
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,7168,2048,0.010043733318646749
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,7168,768,0.00391146664818128
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,7168,768,0.014598400394121806
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,7168,1536,0.00925973355770111
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,7168,512,0.003291733314593633
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,7168,512,0.014494933684666953
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,7168,1024,0.008477866649627686
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,7168,256,0.0030378667016824085
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,7168,256,0.015465600291887918
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,7168,512,0.007590400179227193
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,7168,128,0.002569599946339925
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,7168,128,0.014473600188891092
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,7168,256,0.007130666573842366
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,7168,64,0.0026378666361172995
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,7168,768,0.00802346666653951
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,7168,64,0.014528000354766845
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,7168,32,0.002467199911673864
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,7168,32,0.014538666605949402
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,7168,128,0.0066890666882197065
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,6144,65536,0.12077333132425945
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,6144,16384,0.03678186734517415
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,6144,65536,0.09753920237223307
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,6144,16384,0.03734826644261678
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,6144,12288,0.02824746568997701
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,6144,12288,0.031163734197616578
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,6144,12288,0.029307732979456585
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,6144,10240,0.025126399596532185
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,6144,10240,0.029011199871699016
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,6144,16384,0.03664000034332275
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,6144,8192,0.021358933051427206
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,6144,8192,0.025464532772699992
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,6144,10240,0.026097067197163898
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,6144,7168,0.018253866831461588
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,6144,7168,0.024756266673405965
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,6144,8192,0.021986132860183714
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,6144,6144,0.017308799425760905
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,6144,6144,0.022770132621129355
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,6144,7168,0.020719999074935914
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,6144,5120,0.014805333813031516
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,6144,5120,0.0212991992632548
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,6144,6144,0.01858133276303609
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,6144,4096,0.012277332941691081
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,6144,4096,0.02231360077857971
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,6144,5120,0.01657600005467733
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,6144,3584,0.011190399527549744
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,6144,3584,0.02029866576194763
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,6144,4096,0.01416106621424357
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,6144,3072,0.009995733698209126
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,6144,3072,0.018481065829594932
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,6144,3584,0.012940800189971924
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,6144,2560,0.008593066533406576
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,6144,2560,0.016590933005015053
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,6144,3072,0.012051199873288471
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,6144,2048,0.006776533524195353
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,6144,2048,0.016691199938456216
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,6144,2560,0.011283199985822041
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,6144,1536,0.005179733534653981
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,6144,1536,0.017299199104309083
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,6144,2048,0.009991466999053955
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,6144,1024,0.004237866898377737
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,6144,1024,0.018119466304779053
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,6144,1536,0.009301333626111349
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,6144,768,0.003416533271471659
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,6144,768,0.014454399545987448
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,6144,1024,0.008356266220410665
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,6144,512,0.0030400000512599947
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,6144,512,0.01458560029665629
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,6144,512,0.007202133536338806
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,6144,256,0.0029845332105954488
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,6144,256,0.01488746702671051
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,6144,768,0.007935999830563863
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,6144,128,0.002550400048494339
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,6144,128,0.013740799824396768
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,6144,128,0.0070602665344874065
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,6144,64,0.0024330665667851763
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,6144,64,0.013763200243314108
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,6144,32,0.0022229333718617756
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,6144,256,0.0070783997575442
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,6144,32,0.014437333742777506
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,5120,65536,0.10532053311665852
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,5120,65536,0.08690346876780192
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,5120,16384,0.033532798290252686
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,5120,16384,0.03337279955546062
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,7168,65536,0.1277728001276652
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,5120,12288,0.02434239983558655
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,5120,12288,0.030101333061854047
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,5120,16384,0.03639359871546428
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,5120,10240,0.021447465817133585
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,5120,10240,0.027745066086451213
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,5120,12288,0.029422932863235475
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,5120,8192,0.018132267395655315
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,5120,8192,0.02483946681022644
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,5120,10240,0.025590399901072185
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,5120,7168,0.01620586713155111
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,5120,7168,0.023489065965016685
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,5120,8192,0.0220960001150767
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,5120,6144,0.014890666802724203
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,5120,6144,0.02482453385988871
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,5120,7168,0.020310399929682414
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,5120,5120,0.01359999974568685
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,5120,5120,0.022029866774876915
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,5120,6144,0.018256000677744546
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,5120,4096,0.01120746632417043
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,5120,4096,0.018910932540893554
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,5120,5120,0.01617173353830973
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,5120,3584,0.010134399930636088
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,5120,3584,0.018600533405939736
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,5120,4096,0.01409066617488861
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,5120,3072,0.009141332904497783
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,5120,3072,0.019035732746124266
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,5120,3584,0.012873599926630655
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,5120,2560,0.008232533435026805
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,5120,2560,0.017428267002105712
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,5120,3072,0.012029866377512615
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,5120,2048,0.00592853327592214
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,5120,2048,0.01690559983253479
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,6144,65536,0.12685333093007406
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,5120,1536,0.004680533210436503
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,5120,1536,0.014703999956448874
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,5120,2560,0.011318399508794149
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,5120,1024,0.0038218667109807336
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,5120,1024,0.015390933553377787
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,5120,2048,0.010075733065605164
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,5120,768,0.0034677334129810332
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,5120,768,0.015442132949829102
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,5120,1536,0.009194667140642803
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,5120,512,0.0030250666042168934
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,5120,512,0.014501333236694336
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,5120,512,0.00716480016708374
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,5120,1024,0.008386133114496867
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,5120,256,0.002595199892918269
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,5120,256,0.013959466417630514
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,5120,768,0.007939200103282928
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,5120,128,0.002480000009139379
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,5120,128,0.01458560029665629
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,5120,64,0.002268799891074498
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,5120,64,0.014629333217938741
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,5120,32,0.0025653332471847535
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,5120,32,0.014497066537539164
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,5120,128,0.007124266525109608
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,5120,256,0.007132799923419952
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,4096,65536,0.07089493274688721
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,4096,16384,0.02759679953257243
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,4096,16384,0.030247465769449873
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,4096,16384,0.036761601765950516
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,4096,12288,0.020140800873438516
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,4096,12288,0.027374933163324993
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,4096,12288,0.02927253246307373
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,4096,10240,0.01740586757659912
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,4096,10240,0.024297600984573363
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,5120,65536,0.12709333101908366
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,4096,8192,0.014500266313552857
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,4096,8192,0.023011199633280435
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,4096,10240,0.025634133815765382
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,4096,7168,0.01330560048421224
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,4096,7168,0.02190613349278768
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,4096,7168,0.019900800784428914
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,4096,6144,0.012476799885431926
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,4096,6144,0.02065066695213318
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,4096,8192,0.021585067113240562
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,4096,5120,0.010472533106803895
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,4096,5120,0.020412800709406535
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,4096,5120,0.015819733341534935
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,4096,4096,0.008779733379681905
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,4096,4096,0.019777067502339683
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,4096,6144,0.01788053313891093
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,4096,3584,0.008124800026416778
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,4096,3584,0.017214934031168617
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,4096,3584,0.012915199995040894
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,4096,3072,0.007237333556016285
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,4096,3072,0.016523733735084534
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,4096,4096,0.014134400089581809
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,4096,2560,0.006341333190600078
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,4096,2560,0.018133334318796792
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,4096,3072,0.012078932921091716
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,4096,2048,0.005459199845790863
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,4096,2048,0.0164874662955602
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,4096,2560,0.011178666353225708
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,4096,1536,0.00425493319829305
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,4096,1536,0.016316800316174825
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,4096,1536,0.009622400005658466
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,4096,1024,0.003454933315515518
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,4096,1024,0.014679466684659323
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,4096,2048,0.010067199667294819
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,4096,768,0.0032490665713946023
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,4096,768,0.015252266327540079
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,4096,1024,0.008364799618721008
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,4096,512,0.0030517332255840302
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,4096,512,0.014472533265749613
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,4096,768,0.007937066753705342
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,4096,256,0.00258240004380544
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,4096,256,0.014204800128936768
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,4096,256,0.00710399995247523
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,4096,128,0.002232533444960912
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,4096,128,0.014900267124176025
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,4096,512,0.007530666887760162
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,4096,64,0.0021877333521842955
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,4096,128,0.007073066631952922
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,4096,64,0.014484266440073649
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,4096,32,0.00257493332028389
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,4096,32,0.01447466711203257
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,3584,65536,0.07330880165100098
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,3584,65536,0.06474453210830688
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,3584,16384,0.023822933435440063
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,3584,16384,0.029367466767628986
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,4096,65536,0.12711679935455322
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,3584,12288,0.01817493240038554
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,3584,12288,0.026844799518585205
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,3584,16384,0.036348799864451095
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,3584,10240,0.015803733468055726
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,3584,10240,0.024618667364120484
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,3584,10240,0.025447465976079303
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,3584,8192,0.013767466942469279
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,3584,8192,0.022985599438349404
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,3584,12288,0.028914133707682293
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,3584,7168,0.012652800480524699
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,3584,7168,0.020768000682195028
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,3584,7168,0.01944640080134074
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,3584,6144,0.011935999989509583
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,3584,6144,0.020407466093699138
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,3584,8192,0.02142613331476847
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,3584,5120,0.010006399949391682
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,3584,5120,0.01907520095507304
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,3584,5120,0.015796266992886863
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,3584,4096,0.00842133363087972
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,3584,4096,0.019288533926010133
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,3584,6144,0.017836799224217735
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,3584,3584,0.007500799993673961
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,3584,3584,0.016535466909408568
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,3584,4096,0.013778133193651834
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,3584,3072,0.006996266543865204
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,3584,3072,0.016639999548594155
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,3584,3584,0.012892799576123557
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,3584,2560,0.005994666616121928
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,3584,2560,0.01585706671079
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,3584,2560,0.010904533664385478
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,3584,2048,0.005067733426888784
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,3584,2048,0.016661333044370015
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,3584,3072,0.012139733632405598
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,3584,1536,0.004265599946180979
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,3584,1536,0.01543786625067393
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,3584,1536,0.009231999516487122
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,3584,1024,0.003554133325815201
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,3584,1024,0.01607039968172709
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,3584,2048,0.009987200299898785
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,3584,768,0.0030271999537944795
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,3584,768,0.01448319951693217
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,3584,768,0.007982933521270752
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,3584,512,0.002997333308060964
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,3584,512,0.014663466811180114
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,3584,1024,0.008345599969228108
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,3584,256,0.0026367999613285064
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,3584,256,0.014457600315411887
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,3584,256,0.007503999769687653
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,3584,128,0.0025429333249727885
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,3584,128,0.014632532993952433
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,3584,512,0.007541333138942718
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,3584,64,0.0021759999295075734
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,3584,64,0.013753599921862283
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,3584,32,0.0021685334543387097
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,3584,32,0.013165866335233053
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,3584,128,0.0067210664351781205
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,3072,65536,0.06651413440704346
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,3072,65536,0.06234986782073974
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,3072,16384,0.020524799823760986
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,3072,16384,0.028495999177296956
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,3072,16384,0.036195198694864913
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,3072,12288,0.015736533204714458
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,3072,12288,0.026405332485834758
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,3072,12288,0.0285589337348938
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,3072,10240,0.01416213313738505
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,3072,10240,0.024386133750279745
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,4096,65536,0.0866709311803182
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,3072,8192,0.012258133292198181
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,3072,8192,0.023346134026845298
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,3584,65536,0.12693866888682048
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,3072,7168,0.011945600310961407
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,3072,7168,0.02094399929046631
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,3072,7168,0.01911999980608622
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,3072,6144,0.010855467120806376
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,3072,6144,0.020670932531356812
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,3072,8192,0.021127466360727945
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,3072,5120,0.009612799684206644
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,3072,5120,0.019747199614842732
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,3072,10240,0.025242666403452556
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,3072,4096,0.008172800143559773
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,3072,4096,0.01890346606572469
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,3072,6144,0.017417599757512413
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,3072,3584,0.007551999886830647
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,3072,3584,0.016769067446390788
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,3072,5120,0.015713066856066386
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,3072,3072,0.006811733543872833
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,3072,3072,0.017411200205485027
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,3072,4096,0.013755733768145243
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,3072,2560,0.00594346672296524
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,3072,2560,0.016740266482035318
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,3072,2560,0.011293866237004598
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,3072,3072,0.012121599912643433
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,3072,2048,0.004839466512203216
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,3072,2048,0.016521599888801575
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,3072,3584,0.012897066275278726
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,3072,1536,0.004231466849644979
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,3072,1536,0.01569386621316274
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,3072,1024,0.0035445332527160645
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,3072,1024,0.014844800035158793
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,3072,1536,0.009143466750780743
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,3072,768,0.003005866706371307
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,3072,768,0.015030399958292643
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,3072,2048,0.010368000467618306
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,3072,512,0.002995199958483378
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,3072,512,0.014919466773668923
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,3072,1024,0.007956266899903615
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,3072,256,0.0026154667139053345
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,3072,256,0.014504533012708029
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,3072,768,0.007978666822115581
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,3072,128,0.002170666555563609
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,3072,128,0.014418133099873862
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,3072,256,0.007143466671307881
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,3072,128,0.006667733192443848
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,3072,64,0.002120533337195714
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,3072,32,0.002205866575241089
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,3072,64,0.014458666245142618
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,3072,512,0.007493333518505096
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,3072,32,0.014454399545987448
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,2560,65536,0.05882773399353027
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,2560,65536,0.05714240074157715
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,2560,16384,0.018184532721837364
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,2560,16384,0.02691626747449239
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,2560,12288,0.01390506625175476
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,2560,12288,0.02629973292350769
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,2560,12288,0.02858346700668335
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,2560,10240,0.0129120002190272
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,2560,10240,0.024046933650970458
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,2560,16384,0.03619946638743083
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,2560,8192,0.011252267162005107
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,2560,8192,0.02304853399594625
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,2560,10240,0.024785067637761435
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,2560,7168,0.0121888001759847
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,2560,7168,0.021109332640965782
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,2560,8192,0.021144533157348634
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,2560,6144,0.010441600282986959
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,2560,6144,0.01962239940961202
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,2560,7168,0.019093332688013713
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,2560,5120,0.009218133489290873
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,2560,5120,0.020204800367355346
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,2560,6144,0.017485866943995156
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,2560,4096,0.008066133161385854
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,2560,4096,0.0182751993338267
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,3072,65536,0.12686293125152587
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,2560,3584,0.007204266885916392
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,2560,3584,0.016549332936604818
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,2560,5120,0.015822933117548624
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,2560,3072,0.006454400221506755
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,2560,3072,0.01823893388112386
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,2560,3584,0.012828800082206725
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,2560,2560,0.0055071999629338585
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,2560,2560,0.015619200468063355
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,2560,4096,0.013741866747538248
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,2560,2048,0.004645333190759023
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,2560,2048,0.014467199643452963
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,2560,3072,0.01170133352279663
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,2560,1536,0.004046933352947235
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,2560,1536,0.017042134205500284
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,2560,2560,0.010845866799354554
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,2560,1024,0.00340693344672521
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,2560,1024,0.015285332997639975
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,2560,2048,0.010406399766604107
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,2560,768,0.002998399982849757
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,2560,768,0.015427199999491372
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,2560,1024,0.007980800171693166
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,2560,768,0.007915733257929485
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,2560,512,0.002629333237806956
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,2560,512,0.015315199891726175
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,2560,256,0.0025600001215934755
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,2560,1536,0.0091839998960495
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,2560,256,0.013547733426094055
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,2560,128,0.0022111999491850535
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,2560,128,0.014467199643452963
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,2560,256,0.007099733253320058
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,2560,128,0.007085866729418437
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,2560,64,0.0022069332500298817
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,2560,32,0.0022357332209746042
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,2560,64,0.01443839967250824
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,2560,512,0.007502933343251546
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,2560,32,0.0145087997118632
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,2048,65536,0.04650346835454305
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,2048,65536,0.0505791982014974
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,2048,16384,0.015054933230082192
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,2048,16384,0.026574933528900148
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,2048,12288,0.013854933778444925
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,2048,12288,0.024650667111078897
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,2048,12288,0.028471465905507403
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,2048,10240,0.012546133001645407
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,2048,10240,0.02323946754137675
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,2048,16384,0.03592853148778279
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,2048,8192,0.010824533303578694
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,2048,8192,0.021590399742126464
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,2048,10240,0.024833067258199056
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,2048,7168,0.010462933778762817
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,2048,7168,0.019598933060963948
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,2048,8192,0.021129600207010903
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,2048,6144,0.00955839951833089
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,2048,6144,0.019157334168752035
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,2048,7168,0.01949653426806132
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,2048,5120,0.008474666873613994
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,2048,5120,0.018595200777053834
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,2048,6144,0.01744106610616048
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,2048,4096,0.007480533421039581
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,2048,4096,0.016775466998418174
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,2048,5120,0.01577173372109731
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,2048,3584,0.00668693333864212
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,2048,3584,0.017217065890630087
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,2048,3584,0.012564266721407572
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,2048,3072,0.005876266459623972
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,2048,3072,0.01664426624774933
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,2048,4096,0.013649066289265951
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,2048,2560,0.005133866767088572
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,2048,2560,0.016445866227149962
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,2048,2560,0.01083626647790273
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,2048,3072,0.012010666728019714
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,2048,2048,0.004615466793378194
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,2048,2048,0.01628159979979197
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,2048,1536,0.0038431999584039056
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,2048,1536,0.015698132912317912
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,2048,1536,0.0087936004002889
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,2048,1024,0.0034282666941483817
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,2048,1024,0.014591999848683677
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,2048,2048,0.01050453285376231
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,2048,768,0.003018666555484136
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,2048,768,0.014856533209482829
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,2048,1024,0.008390399813652038
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,2048,512,0.002700799951950709
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,2048,512,0.014567466576894126
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,2048,768,0.007925333579381307
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,2048,256,0.0026079999903837843
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,2048,256,0.014802133043607077
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,2048,512,0.007531733314196269
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,2048,256,0.0071946665644645694
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,2048,128,0.0022229333718617756
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,2048,128,0.014425599575042724
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,2048,64,0.002181333303451538
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,2048,64,0.014411733547846476
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,2048,128,0.007088000078996022
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,2048,32,0.002196266750494639
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,2560,65536,0.12660480340321859
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,2048,32,0.013442132870356241
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,1536,65536,0.03531519969304402
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,1536,16384,0.013618133465449014
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,1536,65536,0.04510613282521565
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,1536,16384,0.024363734324773154
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,1536,12288,0.011168000102043153
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,1536,12288,0.024430932601292928
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,1536,12288,0.028149332602818804
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,1536,10240,0.01046720047791799
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,1536,10240,0.02375040054321289
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,1536,16384,0.03533013264338176
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,1536,8192,0.00963200032711029
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,1536,8192,0.021427200237909953
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,1536,10240,0.02483839988708496
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,1536,7168,0.010036266843477885
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,1536,7168,0.0194432000319163
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,1536,8192,0.020745599269866945
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,1536,6144,0.009190400441487629
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,1536,6144,0.018787199258804323
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,1536,7168,0.019437867403030395
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,1536,5120,0.008436266581217449
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,1536,5120,0.018224000930786133
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,1536,6144,0.017072000106175742
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,1536,4096,0.007400533556938172
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,1536,4096,0.016950400670369466
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,1536,5120,0.015862400333086647
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,1536,3584,0.00673173318306605
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,1536,3584,0.016552533706029257
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,1536,4096,0.013758933544158936
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,1536,3072,0.00547626664241155
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,1536,3072,0.016782933473587038
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,1536,3584,0.012942933042844138
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,1536,2560,0.005066666503747304
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,1536,2560,0.016406400005022685
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,1536,3072,0.011705600221951802
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,1536,2048,0.004689066608746847
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,1536,2048,0.01675093372662862
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,2048,65536,0.12627519766489664
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,1536,1536,0.004279466470082601
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,1536,1536,0.019054933389027914
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,1536,2560,0.010856533050537109
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,1536,1024,0.003398400048414866
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,1536,1024,0.015341867009798685
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,1536,2048,0.010389332969983418
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,1536,768,0.0030421334008375804
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,1536,768,0.014541866381963095
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,1536,1536,0.009091200431187947
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,1536,512,0.002962133288383484
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,1536,512,0.01448319951693217
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,1536,1024,0.007990399996439617
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,1536,256,0.0025621332228183747
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,1536,256,0.014487466216087342
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,1536,768,0.007909333209196727
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,1536,512,0.007456000149250031
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,1536,128,0.00222080002228419
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,1536,128,0.014504533012708029
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,1536,256,0.007122133175532024
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,1536,64,0.0021375998854637145
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,1536,128,0.007124266525109608
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,1536,32,0.0025994665920734406
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,1536,32,0.014459733168284097
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,1024,65536,0.025016532341639204
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,1024,65536,0.04016213417053223
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,1024,16384,0.010717866818110149
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,1024,16384,0.024503467480341594
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,1024,16384,0.0350816011428833
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,1024,12288,0.010558933019638062
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,1024,12288,0.021078399817148843
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,1024,12288,0.028066132465998334
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,1024,10240,0.009894399841626485
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,1024,10240,0.021064533789952596
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,1024,10240,0.024472532669703166
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,1024,8192,0.008890666564305623
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,1024,8192,0.01988053321838379
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,1536,65536,0.1264245351155599
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,1024,7168,0.007934933404127757
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,1024,7168,0.018850133816401164
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,1024,8192,0.02108479936917623
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,1024,6144,0.0076000000039736434
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,1024,6144,0.018637865781784058
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,1024,7168,0.019090133905410766
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,1024,5120,0.006748799979686737
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,1024,5120,0.020088533560434975
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,1024,6144,0.017433599630991618
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,1024,4096,0.006332799792289734
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,1024,4096,0.01715946594874064
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,1024,5120,0.015703466534614564
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,1024,3584,0.00591786652803421
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,1024,3584,0.01669013301531474
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,1024,3584,0.01284160017967224
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,1024,3072,0.005445333321889242
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,1024,3072,0.01667520006497701
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,1024,4096,0.013309866189956665
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,1024,2560,0.00505920002857844
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,1024,2560,0.016174933314323424
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,1024,2560,0.010874666770299276
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,1024,3072,0.011603200435638427
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,1024,2048,0.004642133414745331
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,1024,2048,0.016453333695729575
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,1024,1536,0.0038122666378815973
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,1024,1536,0.01448319951693217
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,1024,1536,0.009144533673922222
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,1024,1024,0.0034175999462604523
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,1024,2048,0.009948800007502239
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,1024,1024,0.015358933806419372
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,1024,768,0.002977066735426585
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,1024,768,0.014481066664059957
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,1024,1024,0.008328533172607422
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,1024,768,0.007460266848405202
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,1024,512,0.0026101333399613695
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,1024,512,0.014476799964904785
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,1024,256,0.002549333373705546
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,1024,256,0.014510933558146158
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,1024,512,0.007286400099595388
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,1024,128,0.0021759999295075734
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,1024,256,0.007067733506361644
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,1024,128,0.014542933305104574
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,1024,64,0.0021546666820844015
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,1024,128,0.0066655998428662615
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,1024,64,0.013817600409189858
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,1024,32,0.002182399978240331
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,1024,32,0.014450132846832275
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,768,65536,0.02313279906908671
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,768,65536,0.034962133566538496
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,768,16384,0.010361599922180175
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,768,16384,0.023616000016530355
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,1024,65536,0.12316373189290364
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,768,12288,0.00913279950618744
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,768,12288,0.020827732483545938
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,768,16384,0.03548053503036499
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,768,10240,0.008341333270072937
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,768,10240,0.021411200364430748
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,768,10240,0.0247871994972229
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,768,8192,0.007547733187675476
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,768,8192,0.02041920026143392
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,768,8192,0.020709333817164104
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,768,7168,0.007203199962774913
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,768,7168,0.019817600647608437
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,768,7168,0.019016534090042114
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,768,6144,0.006764799853165944
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,768,6144,0.01909760038057963
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,768,6144,0.017427200078964235
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,768,5120,0.006706133484840393
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,768,5120,0.017783466974894205
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,768,5120,0.015490133563677469
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,768,4096,0.006664533416430156
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,768,4096,0.01781760056813558
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,768,4096,0.01332266628742218
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,768,3584,0.006251733501752217
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,768,3584,0.016851200660069784
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,768,3584,0.012410666545232136
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,768,3072,0.005432533224423727
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,768,3072,0.01662613352139791
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,768,65536,0.12271467049916585
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,768,2560,0.005117866893609365
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,768,2560,0.017189333836237587
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,768,3072,0.01169599990049998
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,768,2048,0.004268800218900045
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,768,2048,0.0164000004529953
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,768,2560,0.010780800382296245
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,768,1536,0.0038506666819254553
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,768,1536,0.014453333616256715
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,768,2048,0.00999786655108134
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,768,1024,0.003402666747570038
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,768,1024,0.015869866808255514
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,768,1536,0.009105066458384197
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,768,1024,0.007922133306662242
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,768,768,0.0029728000362714132
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,768,768,0.01504746675491333
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,768,512,0.002571733295917511
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,768,512,0.01458346645037333
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,768,768,0.0077344000339508055
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,768,256,0.002570666621128718
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,768,512,0.007490133245786031
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,768,256,0.014518400033315023
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,768,128,0.0021674667795499166
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,768,128,0.013756799697875976
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,768,128,0.0067210664351781205
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,768,64,0.002152533332506816
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,768,64,0.013392000397046407
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,768,32,0.0021557333568731946
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,768,32,0.013449600338935852
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,512,65536,0.02007360061009725
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,512,65536,0.033581864833831784
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,512,65536,0.12148693402608235
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,512,16384,0.009117866555849712
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,512,16384,0.02148373325665792
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,512,16384,0.03537600040435791
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,512,12288,0.007323733468850453
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,512,12288,0.02038080096244812
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,512,12288,0.02803093393643697
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,512,10240,0.0062943999965985615
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,512,10240,0.02060906688372294
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,512,10240,0.024289067586263022
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,512,8192,0.005907199780146281
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,512,8192,0.01973973313967387
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,512,8192,0.020679465929667153
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,512,7168,0.005861333509286245
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,512,7168,0.02029119928677877
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,512,7168,0.0189301331837972
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,512,6144,0.0058677335580190025
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,512,6144,0.01958079934120178
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,512,6144,0.017164800564448038
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,512,5120,0.006337066491444905
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,512,5120,0.01841493248939514
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,512,5120,0.01530346671740214
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,512,4096,0.005843199789524078
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,512,4096,0.017923200130462648
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,512,4096,0.013653332988421122
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,512,3584,0.005524266759554545
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,512,3584,0.01656000018119812
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,512,3584,0.012795733412106833
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,512,3072,0.005427200098832448
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,512,3072,0.01627840002377828
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,512,3072,0.011621333161989848
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,512,2560,0.0050346667567888895
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,512,2560,0.01492693324883779
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,512,2560,0.010769066214561463
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,512,2048,0.004248533149560293
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,512,2048,0.017072000106175742
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,512,2048,0.009944533308347065
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,512,1536,0.0037791999677817024
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,512,1536,0.016242133577664693
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,512,1536,0.009131733576456707
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,512,1024,0.003044266750415166
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,512,1024,0.014468266566594442
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,512,1024,0.00793280005455017
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,512,768,0.0029845332105954488
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,512,768,0.014497066537539164
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,512,768,0.007843199868996937
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,512,512,0.0026208000878492994
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,512,512,0.01474453310171763
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,512,512,0.007490133245786031
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,512,256,0.0025920001169045764
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,512,256,0.014419200023015341
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,512,256,0.007077333331108093
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,512,128,0.002182399978240331
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,512,128,0.01444906691710154
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,512,128,0.006331733365853627
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,512,64,0.0021173333128293352
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,512,64,0.014428800344467163
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,512,32,0.002165333429972331
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,512,32,0.014473600188891092
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,256,65536,0.013684266805648803
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,256,65536,0.02992960015932719
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,256,65536,0.12130239804585774
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,256,16384,0.006363733112812043
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,256,16384,0.02073600093523661
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,256,16384,0.03501226504643758
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,256,12288,0.005913599828879039
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,256,12288,0.020482132832209267
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,256,12288,0.027944533030192058
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,256,10240,0.0059690664211908976
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,256,10240,0.020227199792861937
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,256,10240,0.024701867500940958
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,256,8192,0.005832533538341522
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,256,8192,0.018894932667414346
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,256,8192,0.02105600039164225
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,256,7168,0.005448533097902933
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,256,7168,0.020009599129358926
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,256,7168,0.01899519960085551
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,256,6144,0.005939200023810069
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,256,6144,0.019141334295272826
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,256,6144,0.017335466543833413
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,256,5120,0.005864533285299936
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,256,5120,0.019143466154734293
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,256,5120,0.015659733613332113
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,256,4096,0.0054517333706219995
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,256,4096,0.018956800301869713
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,256,4096,0.013710932930310568
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,256,3584,0.005038933455944061
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,256,3584,0.016761600971221924
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,256,3584,0.012847999731699625
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,256,3072,0.005448533097902933
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,256,3072,0.016549332936604818
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,256,3072,0.011637333035469054
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,256,2560,0.0050346667567888895
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,256,2560,0.016295466820398966
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,256,2560,0.010423466563224792
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,256,2048,0.004955733319123586
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,256,2048,0.015719466408093772
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,256,2048,0.00995733340581258
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,256,1536,0.0038101332883040107
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,256,1536,0.016289066274960837
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,256,1536,0.00916373332341512
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,256,1024,0.003031466652949651
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,256,1024,0.017002665996551515
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,256,1024,0.008307200173536937
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,256,768,0.0029866665601730345
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,256,768,0.014473600188891092
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,256,768,0.007890133559703827
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,256,512,0.002595199892918269
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,256,512,0.014924800395965577
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,256,512,0.007097599903742473
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,256,256,0.0022261333962281544
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,256,256,0.013577600320180258
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,256,256,0.007073066631952922
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,256,128,0.002181333303451538
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,256,128,0.014349866906801859
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,256,128,0.006675200164318084
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,256,64,0.0021546666820844015
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,256,64,0.014330666263898215
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,256,32,0.002148266633351644
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,256,32,0.014344533284505209
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,128,65536,0.011549866199493409
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,128,65536,0.028361600637435914
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,128,65536,0.12072106997172039
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,128,16384,0.005910400052865346
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,128,16384,0.020772266387939452
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,128,16384,0.0350325345993042
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,128,12288,0.0054666668176651
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,128,12288,0.0190175990263621
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,128,12288,0.02803093393643697
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,128,10240,0.005485866467158
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,128,10240,0.018563199043273925
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,128,10240,0.024276266495386757
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,128,8192,0.0054848000407218935
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,128,8192,0.018623999754587808
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,128,8192,0.020558933417002358
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,128,7168,0.005436799923578898
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,128,7168,0.01928000052769979
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,128,7168,0.01894506613413493
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,128,6144,0.0050911997755368555
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,128,6144,0.019337600469589232
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,128,6144,0.01734293301900228
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,128,5120,0.00462719996770223
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,128,5120,0.019090133905410766
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,128,5120,0.01528320014476776
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,128,4096,0.00508480022350947
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,128,4096,0.017237333456675212
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,128,4096,0.013615999619166055
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,128,3584,0.005006933212280273
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,128,3584,0.01648533344268799
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,128,3584,0.012817066907882691
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,128,3072,0.004666666686534882
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,128,3072,0.01546986699104309
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,128,3072,0.012006400028864543
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,128,2560,0.004646400113900503
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,128,2560,0.01586560010910034
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,128,2560,0.010757333040237427
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,128,2048,0.0046079998215039575
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,128,2048,0.014831999937693277
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,128,2048,0.009965866804122925
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,128,1536,0.0034517332911491393
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,128,1536,0.014453333616256715
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,128,1536,0.009145599603652955
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,128,1024,0.0030229332546393077
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,128,1024,0.014478933811187745
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,128,1024,0.008317866424719492
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,128,768,0.002631466587384542
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,128,768,0.01444906691710154
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,128,768,0.007482666770617168
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,128,512,0.0028149334092934927
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,128,512,0.014250666896502177
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,128,512,0.007203199962774913
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,128,256,0.0023370665808518727
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,128,256,0.013069867094357809
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,128,256,0.007108266651630402
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,128,128,0.0022250667214393617
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,128,128,0.014335999886194864
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,128,128,0.006343466540177663
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,128,64,0.0018239999810854596
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,128,64,0.014247467120488485
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,128,32,0.001842133328318596
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,128,32,0.013788800438245138
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,64,65536,0.007947733501593272
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,64,65536,0.026580266157786053
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,64,16384,0.005491200089454651
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,64,16384,0.020966400702794395
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,64,12288,0.005085866649945577
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,64,12288,0.01885546644528707
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,64,10240,0.00547626664241155
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,64,10240,0.019127466281255088
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,64,8192,0.00510506679614385
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,64,8192,0.01856213410695394
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,64,7168,0.005043200155099233
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,64,7168,0.01880853374799093
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,64,6144,0.0050357331832249965
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,64,6144,0.016668800512949625
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,64,5120,0.004642133414745331
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,64,5120,0.018351999918619792
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,64,4096,0.004264533519744873
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,64,4096,0.01655359963575999
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,64,3584,0.004604800045490265
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,64,3584,0.017001599073410034
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,64,3072,0.004261333247025808
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,64,3072,0.016671999295552572
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,64,2560,0.004660266637802124
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,64,2560,0.016131200393040977
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,64,2048,0.004277333120505015
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,64,2048,0.01470186710357666
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,64,1536,0.0037962667644023894
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,64,1536,0.014437333742777506
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,64,1024,0.0029663999875386557
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,64,1024,0.014433067043622336
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,64,768,0.002616533388694127
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,64,768,0.014456533392270408
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,64,512,0.0025685332715511323
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,64,512,0.014728533228238425
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,64,256,0.0021685334543387097
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,64,256,0.014582399527231851
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,64,128,0.0023999998966852823
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,64,128,0.012602667013804117
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,64,64,0.001791999985774358
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,64,64,0.01416213313738505
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,64,32,0.0018101333330074946
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,64,32,0.014051199952761332
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,32,65536,0.008342400193214417
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,32,65536,0.026964267094930012
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,32,16384,0.0058378666639328
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,32,16384,0.02142613331476847
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,32,12288,0.005085866649945577
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,32,12288,0.018678400913874307
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,32,10240,0.0054282665252685545
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,32,10240,0.01856213410695394
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,32,8192,0.0054282665252685545
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,32,8192,0.017853866020838417
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,32,7168,0.00505920002857844
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,32,7168,0.01920106609662374
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,32,6144,0.004996266464392344
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,32,6144,0.017245866854985557
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,32,5120,0.004605866471926371
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,32,5120,0.019435733556747437
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,32,4096,0.004194133480389913
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,32,4096,0.016556800405184428
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,32,3584,0.004634666442871094
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,32,3584,0.016962132851282754
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,32,3072,0.004252799848715464
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,32,3072,0.016568533579508462
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,32,2560,0.004636799792448679
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,32,2560,0.015561599532763162
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,32,2048,0.004231466849644979
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,32,2048,0.014759467045466105
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,32,1536,0.0033845332761605583
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,32,1536,0.014665599664052328
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,32,1024,0.002993066608905792
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,32,1024,0.014975999792416891
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,32,768,0.0025973332424958544
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,32,768,0.015641599893569946
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,32,512,0.0022261333962281544
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,32,512,0.013874133427937826
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,32,256,0.002152533332506816
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,32,256,0.0131221334139506
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,32,128,0.002147199958562851
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,32,128,0.012713600198427835
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,32,64,0.0018677332748969395
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,32,64,0.014376533031463624
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,48,32,32,0.001802666609485944
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,32,32,0.013671466708183288
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,65536,16384,0.36902828216552735
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,65536,16384,0.19685014088948566
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,48,1536,64,0.014494933684666953
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,65536,12288,0.2649322668711344
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,65536,12288,0.15134399731953938
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,768,12288,0.028075732787450153
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,65536,10240,0.2241546630859375
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,65536,10240,0.12767893473307293
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,48,768,256,0.006708266834417979
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,65536,8192,0.18297386169433594
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,65536,8192,0.1083946704864502
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,65536,8192,0.06613866488138834
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,65536,7168,0.15505919456481934
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,65536,7168,0.09566293557484945
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,65536,10240,0.08116587003072104
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,65536,6144,0.14031786918640138
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,65536,6144,0.08226239681243896
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,65536,12288,0.08717119693756104
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,65536,5120,0.11294293403625488
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,65536,5120,0.07156159877777099
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,65536,16384,0.11996479829152425
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,65536,4096,0.0945472002029419
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,65536,4096,0.06234879891077677
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,65536,6144,0.05049920082092285
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,65536,3584,0.0800661325454712
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,65536,3584,0.05404266516367594
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,65536,7168,0.058474667867024745
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,65536,3072,0.06939413547515869
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,65536,3072,0.050139733155568446
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,65536,5120,0.044536534945170084
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,65536,2560,0.05840426683425903
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,65536,2560,0.04388800064722697
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,65536,4096,0.03509653409322103
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,65536,2048,0.04928106864293416
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,65536,2048,0.03803413311640422
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,65536,3584,0.03430399894714355
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,65536,1536,0.03681066830952962
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,65536,1536,0.032765867312749226
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,65536,2560,0.02540053327878316
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,65536,1024,0.025656533241271973
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,65536,1024,0.026020266612370807
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,65536,3072,0.028043733040491743
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,65536,768,0.019423999389012656
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,65536,768,0.022768000761667885
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,65536,1536,0.019001599152882895
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,65536,512,0.013672533631324767
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,65536,512,0.020806399981180827
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,65536,1024,0.015176533659299215
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,65536,256,0.007499733567237854
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,65536,256,0.01809813380241394
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,65536,768,0.013748266299565635
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,65536,128,0.0051018665234247845
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,65536,128,0.016563199957211814
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,65536,128,0.008413867155710856
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,65536,64,0.004267733295758565
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,65536,64,0.016541866461435954
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,65536,32,0.004666666686534882
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,65536,32,0.016566399733225504
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,16384,65536,0.37408854166666666
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,16384,65536,0.19943359692891438
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,65536,256,0.009225599964459737
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,16384,16384,0.08069760004679362
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,16384,16384,0.056986665725708006
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,65536,512,0.011273599664370219
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,16384,12288,0.05894720156987508
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,16384,12288,0.047890134652455646
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,65536,2048,0.021780266364415487
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,16384,10240,0.05029759804407755
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,16384,10240,0.041118931770324704
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,16384,10240,0.02924906611442566
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,16384,8192,0.04074879884719849
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,16384,8192,0.03501760164896647
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,16384,12288,0.033045333623886106
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,16384,7168,0.03604160149892171
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,16384,7168,0.03320533235867818
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,16384,16384,0.04177813529968262
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,16384,6144,0.031429332494735715
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,16384,6144,0.03104426662127177
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,16384,8192,0.023982934157053628
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,16384,5120,0.026789333422978717
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,16384,5120,0.028279467423756914
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,16384,7168,0.022054400046666464
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,16384,4096,0.021641600131988525
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,16384,4096,0.025035732984542848
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,16384,6144,0.01989439924558004
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,16384,3584,0.019550933440526327
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,16384,3584,0.023906133572260537
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,16384,4096,0.015786666671435037
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,16384,3072,0.017126399278640746
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,16384,5120,0.018348799149195353
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,16384,3072,0.023188267151514688
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,16384,2560,0.014523733655611673
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,16384,2560,0.02099840044975281
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,16384,3584,0.014596266547838846
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,16384,2048,0.012088533242543538
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,16384,2048,0.02122453252474467
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,16384,2560,0.012471466263135274
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,16384,1536,0.009800533453623455
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,16384,1536,0.01874133348464966
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,16384,3072,0.014102400342623393
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,16384,1024,0.007565866907437642
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,16384,1024,0.01667413314183553
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,16384,1536,0.009609599908192951
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,16384,1024,0.008710400263468424
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,16384,768,0.005868799984455109
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,16384,768,0.014491732915242514
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,16384,512,0.0038805333276589714
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,16384,512,0.01525973379611969
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,16384,2048,0.01128426690896352
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,16384,256,0.0030464000999927522
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,16384,256,0.014478933811187745
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,16384,512,0.007618133227030437
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,16384,128,0.0026410666604836782
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,16384,128,0.014626133441925048
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,16384,256,0.007295999924341838
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,16384,768,0.008303999900817871
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,16384,64,0.002548266698916753
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,16384,32,0.002639999985694885
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,16384,64,0.014391466975212097
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,16384,128,0.0067775999506314594
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,16384,32,0.014565333724021912
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,12288,65536,0.24774293899536132
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,12288,16384,0.0688266674677531
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,12288,16384,0.05284800132115682
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,12288,16384,0.03876693248748779
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,12288,12288,0.05051519870758057
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,12288,12288,0.04383999903996785
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,12288,12288,0.03095253308614095
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,12288,10240,0.04400533437728882
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,12288,10240,0.041015466054280594
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,12288,10240,0.02722666660944621
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,12288,8192,0.03710399866104126
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,12288,8192,0.03188479940096538
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,12288,65536,0.15274880727132162
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,12288,7168,0.03277973333994548
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,12288,7168,0.030434133609135945
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,12288,7168,0.02092693249384562
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,12288,6144,0.029014400641123456
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,12288,6144,0.026946133375167845
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,12288,8192,0.022705066204071044
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,12288,5120,0.02185813387235006
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,12288,5120,0.02574186722437541
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,12288,6144,0.019194666544596353
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,12288,4096,0.017398399114608765
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,12288,4096,0.02434879938761393
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,12288,5120,0.017411200205485027
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,12288,3584,0.01573013365268707
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,12288,3584,0.024037333329518636
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,12288,3584,0.014526933431625366
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,12288,3072,0.013852799932161966
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,12288,3072,0.022272000710169472
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,12288,4096,0.015495466192563376
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,12288,2560,0.012616533041000366
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,12288,2560,0.02090453306833903
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,12288,3072,0.012875733772913614
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,12288,2048,0.010910933216412861
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,12288,2048,0.01899306575457255
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,12288,2560,0.011447466413180033
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,12288,1536,0.008774399757385254
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,12288,1536,0.017797333002090455
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,12288,1536,0.009205333391825358
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,12288,1024,0.00709440012772878
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,12288,1024,0.015863466262817382
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,12288,2048,0.010456533233324686
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,12288,768,0.005478399991989136
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,12288,768,0.014491732915242514
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,12288,1024,0.008387200037638346
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,12288,512,0.0039327998956044516
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,12288,512,0.01548906664053599
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,12288,768,0.007970133423805236
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,12288,256,0.00311253344019254
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,12288,256,0.015020799636840821
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,12288,512,0.007613866527875264
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,12288,128,0.0028319999575614927
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,12288,128,0.014532267053922018
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,12288,128,0.006795733173688252
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,12288,64,0.002621866762638092
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,12288,256,0.007102933526039123
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,12288,32,0.002571733295917511
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,12288,64,0.014539733529090881
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,12288,32,0.014453333616256715
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,10240,65536,0.13887359301249186
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,16384,65536,0.14575573603312175
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,10240,16384,0.05981760025024414
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,10240,16384,0.04937493403752645
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,12288,65536,0.1315232038497925
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,10240,12288,0.04490240017573039
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,10240,12288,0.04141120115915935
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,10240,16384,0.038430933157602945
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,10240,10240,0.03805546760559082
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,10240,10240,0.03434133529663086
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,10240,12288,0.0305567999680837
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,10240,8192,0.032339199384053545
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,10240,8192,0.029373866319656373
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,10240,10240,0.026745599508285523
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,10240,7168,0.025388799111048382
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,10240,7168,0.027086933453877766
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,10240,8192,0.022708266973495483
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,10240,6144,0.022294400135676067
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,10240,6144,0.026616533597310383
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,10240,7168,0.02077440023422241
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,10240,5120,0.01907306710879008
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,10240,5120,0.025589332977930708
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,10240,6144,0.019132800896962485
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,10240,4096,0.015471999843915304
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,10240,4096,0.0228000005086263
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,10240,65536,0.22187199592590331
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,10240,3584,0.014404267072677612
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,10240,3584,0.023272534211476646
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,10240,5120,0.017076265811920167
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,10240,3072,0.012356266379356384
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,10240,3072,0.02097919980684916
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,10240,4096,0.015335466464360556
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,10240,2560,0.01086293359597524
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,10240,2560,0.02081386645634969
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,10240,3584,0.013754666845003764
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,10240,2048,0.010396800438563029
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,10240,2048,0.018718934059143065
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,10240,3072,0.012503467003504434
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,10240,1536,0.0083146666487058
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,10240,1536,0.016645333170890807
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,10240,2048,0.010411733388900756
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,10240,1024,0.006391466657320659
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,10240,1024,0.015797332922617594
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,10240,2560,0.01153600017229716
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,10240,768,0.005087999999523163
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,10240,768,0.016886399189631144
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,10240,1536,0.009152000149091084
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,10240,512,0.004103466620047887
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,10240,512,0.014469333489735923
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,10240,512,0.007623466849327088
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,10240,256,0.003065599997838338
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,10240,256,0.015068800250689188
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,10240,768,0.008273066580295562
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,10240,128,0.002639999985694885
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,10240,128,0.014201600352923074
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,10240,256,0.007175466914971669
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,10240,64,0.0026186667382717133
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,10240,64,0.01439786652723948
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,10240,32,0.002569599946339925
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,10240,32,0.014460800091425577
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,10240,1024,0.008356266220410665
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,10240,128,0.006740266581376393
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,8192,65536,0.16079786618550618
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,8192,65536,0.10939199924468994
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,8192,16384,0.0499125321706136
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,8192,16384,0.037172265847524005
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,8192,12288,0.03133119940757752
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,8192,12288,0.03208426634470622
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,8192,12288,0.029824000597000123
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,8192,10240,0.02728853424390157
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,8192,10240,0.030239999294281006
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,8192,16384,0.03744853337605794
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,8192,8192,0.021977599461873373
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,8192,8192,0.025884799162546795
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,8192,10240,0.02614506681760152
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,8192,7168,0.019857066869735717
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,8192,7168,0.024830933411916098
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,8192,8192,0.022326399882634483
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,8192,6144,0.01733760039011637
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,8192,6144,0.025415466229120894
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,8192,7168,0.02103253404299418
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,8192,5120,0.014978133638699851
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,8192,5120,0.02224853237469991
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,10240,65536,0.1290719985961914
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,8192,4096,0.012268799543380737
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,8192,4096,0.02066133419672648
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,8192,6144,0.018684800465901694
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,8192,3584,0.011197866996129353
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,8192,3584,0.02220266660054525
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,8192,5120,0.0172437330087026
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,8192,3072,0.010044800241788228
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,8192,3072,0.01980479955673218
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,8192,3584,0.013262933492660523
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,8192,2560,0.008724266290664673
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,8192,2560,0.018360533316930137
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,8192,4096,0.014497066537539164
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,8192,2048,0.007970133423805236
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,8192,2048,0.016574933131535848
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,8192,3072,0.012105600039164225
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,8192,1536,0.006695466736952464
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,8192,1536,0.0164682666460673
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,8192,2048,0.010075733065605164
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,8192,1024,0.00461760014295578
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,8192,1024,0.01607360045115153
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,8192,2560,0.011222400267918905
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,8192,768,0.003853866706291834
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,8192,768,0.014460800091425577
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,8192,1536,0.009245866537094116
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,8192,512,0.0031615999837716425
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,8192,768,0.007995733122030894
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,8192,512,0.014958932995796204
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,8192,256,0.0026186667382717133
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,8192,256,0.014447999993960061
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,8192,1024,0.008343467116355896
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,8192,128,0.002234666546185811
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,8192,128,0.014479999740918478
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,8192,256,0.007134933272997539
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,8192,64,0.0022624000906944276
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,8192,64,0.014401066303253173
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,8192,32,0.002197333425283432
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,8192,32,0.014486400286356607
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,8192,128,0.00673173318306605
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,7168,65536,0.14989333152770995
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,7168,65536,0.09719573656717936
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,8192,512,0.007516799867153168
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,7168,16384,0.04122666517893474
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,7168,16384,0.037715200583140054
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,7168,12288,0.03101973334948222
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,7168,12288,0.029101866483688354
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,7168,12288,0.029445334275563555
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,7168,10240,0.026820266246795656
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,7168,10240,0.026833067337671917
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,7168,16384,0.036848000685373944
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,7168,8192,0.022265599171320597
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,7168,8192,0.024701867500940958
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,7168,10240,0.026001065969467163
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,7168,7168,0.020152533054351808
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,7168,7168,0.022719999154408775
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,7168,8192,0.022032000621159873
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,7168,6144,0.015685333808263143
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,7168,6144,0.02315946618715922
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,7168,7168,0.02031146685282389
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,7168,5120,0.013828266660372415
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,7168,5120,0.020719999074935914
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,7168,6144,0.018266665935516357
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,7168,4096,0.011260799566904704
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,7168,4096,0.020336000124613444
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,7168,5120,0.016990933815638223
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,7168,3584,0.010382933417956035
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,7168,3584,0.019901865720748903
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,7168,4096,0.014144000411033631
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,7168,3072,0.009240532914797466
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,7168,3072,0.01830186645189921
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,7168,3584,0.013313066959381104
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,7168,2560,0.008354133367538452
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,7168,2560,0.01729493339856466
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,7168,3072,0.01209493378798167
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,7168,2048,0.0070613334576288865
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,7168,2048,0.015554133057594299
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,7168,2560,0.01128000020980835
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,7168,1536,0.005892266829808554
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,7168,1536,0.015184000134468079
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,7168,2048,0.010439466436703999
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,7168,1024,0.004366933306058248
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,7168,1024,0.014479999740918478
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,7168,1536,0.009240532914797466
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,7168,768,0.003786666691303253
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,7168,768,0.01454080045223236
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,7168,1024,0.008321066697438557
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,7168,512,0.0030645333230495454
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,7168,512,0.014537599682807923
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,7168,768,0.007949866851170858
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,7168,256,0.0025941332181294756
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,7168,256,0.014485333363215128
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,7168,256,0.007211733361085255
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,7168,128,0.002600533266862233
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,7168,128,0.014418133099873862
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,7168,512,0.00751146674156189
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,7168,64,0.002160000056028366
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,7168,64,0.01446399986743927
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,7168,32,0.0021856000026067098
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,7168,32,0.013860266407330832
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,7168,128,0.0070602665344874065
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,6144,65536,0.11615253289540609
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,6144,65536,0.09148693084716797
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,6144,16384,0.035529601573944095
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,6144,16384,0.03528106609980265
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,8192,65536,0.12751039663950603
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,6144,12288,0.027482666571935016
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,6144,12288,0.031086933612823487
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,6144,12288,0.029395200808842975
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,6144,10240,0.02439360022544861
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,6144,10240,0.028974932432174683
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,6144,16384,0.03664746681849162
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,6144,8192,0.020222934087117513
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,6144,8192,0.024710400899251302
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,6144,10240,0.025880533456802367
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,6144,7168,0.017870932817459106
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,6144,7168,0.022772266467412313
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,6144,8192,0.022026666005452476
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,6144,6144,0.015821866194407144
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,6144,6144,0.02337706685066223
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,6144,6144,0.01829119920730591
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,6144,5120,0.014054399728775025
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,6144,5120,0.02101759910583496
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,6144,7168,0.020363734165827433
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,6144,4096,0.010424533486366272
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,6144,4096,0.018819200992584228
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,6144,4096,0.014179199934005737
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,6144,3584,0.009643733501434326
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,6144,3584,0.019477333625157675
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,6144,5120,0.01604480048020681
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,6144,3072,0.00876693328221639
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,6144,3072,0.019432532787322997
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,6144,3072,0.012125866611798604
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,6144,2560,0.007739733159542084
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,6144,2560,0.016747732957204185
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,6144,3584,0.013347199559211731
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,6144,2048,0.006449066599210103
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,6144,2048,0.016671999295552572
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,6144,2048,0.010442666212717692
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,6144,1536,0.005960533519585928
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,6144,2560,0.010893866419792175
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,6144,1536,0.014634666840235391
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,7168,65536,0.12741973400115966
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,6144,1024,0.0039711999396483105
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,6144,1024,0.01603626708189646
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,6144,768,0.003621333340803782
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,6144,768,0.015684266885121666
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,6144,1536,0.009270399808883667
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,6144,512,0.0030261332790056865
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,6144,512,0.014548266927401224
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,6144,768,0.008332799871762593
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,6144,256,0.0026389333109060925
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,6144,256,0.014601600170135499
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,6144,1024,0.008255999783674877
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,6144,128,0.002251733342806498
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,6144,128,0.014547200004259745
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,6144,256,0.007122133175532024
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,6144,64,0.0022122666239738466
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,6144,64,0.014471466342608133
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,6144,32,0.0022805333137512207
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,6144,32,0.01446293294429779
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,6144,512,0.0071370666225751235
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,5120,65536,0.1040277322133382
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,5120,65536,0.08118506272633871
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,5120,16384,0.03234240015347799
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,5120,16384,0.033056000868479415
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,5120,16384,0.03624320030212402
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,5120,12288,0.024684800704320272
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,5120,12288,0.028843732674916585
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,5120,12288,0.02895466685295105
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,5120,10240,0.021496532360712688
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,5120,10240,0.027799467245737713
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,5120,10240,0.025621332724889118
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,5120,8192,0.018236800034840902
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,5120,8192,0.02310933272043864
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,6144,65536,0.1270645300547282
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,5120,7168,0.016157866517702738
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,5120,7168,0.02273599902788798
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,5120,8192,0.021617066860198975
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,5120,6144,0.014615466197331747
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,5120,6144,0.021705599625905354
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,5120,7168,0.019921066363652547
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,5120,5120,0.012874666849772134
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,5120,5120,0.02071146567662557
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,5120,5120,0.015745066603024802
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,5120,4096,0.009436800082524618
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,5120,6144,0.01775253415107727
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,5120,4096,0.018184532721837364
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,5120,3584,0.008789333701133727
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,5120,3584,0.018565332889556883
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,5120,3584,0.012941867113113403
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,5120,3072,0.007964799801508587
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,5120,3072,0.016897066434224447
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,5120,4096,0.013683199882507324
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,5120,2560,0.007214933137098948
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,5120,2560,0.016523733735084534
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,5120,3072,0.012072533369064331
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,5120,2048,0.0055754666527112326
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,5120,2048,0.016272000471750894
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,5120,2560,0.010822400450706482
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,5120,1536,0.005468800167242686
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,5120,1536,0.014923733472824097
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,5120,1536,0.009160533547401428
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,5120,1024,0.004233600199222564
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,5120,1024,0.014523733655611673
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,5120,2048,0.010044800241788228
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,5120,768,0.003815466662247976
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,5120,768,0.014501333236694336
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,5120,1024,0.008396800359090168
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,5120,512,0.0030901332696278887
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,5120,512,0.014465066790580749
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,5120,768,0.007959466675917308
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,5120,256,0.00258240004380544
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,5120,256,0.01443839967250824
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,5120,256,0.007110400001207988
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,5120,128,0.002595199892918269
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,5120,128,0.012983466188112894
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,5120,512,0.007333333293596904
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,5120,64,0.0021909333765506743
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,5120,64,0.014186666409174601
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,5120,32,0.0021802666286627453
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,5120,128,0.006679466863473256
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,5120,32,0.013960533340771995
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,4096,65536,0.08573973178863525
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,4096,65536,0.06609706481297811
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,4096,16384,0.026998400688171387
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,4096,16384,0.02886613408724467
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,5120,65536,0.1267690658569336
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,4096,12288,0.019700266917546592
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,4096,12288,0.026554665962855023
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,4096,16384,0.03625813325246175
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,4096,10240,0.016979199647903443
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,4096,10240,0.02600533366203308
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,4096,10240,0.025544534126917522
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,4096,8192,0.014088533322016396
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,4096,8192,0.023283199469248454
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,4096,12288,0.0289792001247406
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,4096,7168,0.013235200444857279
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,4096,7168,0.022589866320292154
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,4096,8192,0.021473066012064616
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,4096,7168,0.019411200284957887
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,4096,6144,0.012131200234095255
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,4096,6144,0.023612799247105916
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,4096,5120,0.011450666189193725
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,4096,5120,0.020230400562286376
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,4096,5120,0.015793066223462424
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,4096,4096,0.009891200065612792
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,4096,4096,0.01737706661224365
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,4096,6144,0.017499732971191406
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,4096,3584,0.009092266360918682
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,4096,3584,0.016694400707880655
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,4096,3584,0.012890666723251343
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,4096,3072,0.008359466989835102
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,4096,3072,0.016786134243011473
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,4096,4096,0.013738666971524557
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,4096,2560,0.007145600020885467
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,4096,2560,0.015758933623631795
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,4096,3072,0.012125866611798604
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,4096,2048,0.004863999783992767
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,4096,2048,0.016158933440844216
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,4096,2560,0.010833066701889039
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,4096,1536,0.003966933240493139
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,4096,1536,0.014471466342608133
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,6144,128,0.006743466854095459
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,4096,1024,0.0035360001027584078
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,4096,1024,0.014597333470980325
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,4096,1536,0.00918293297290802
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,4096,768,0.0030229332546393077
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,4096,768,0.014456533392270408
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,4096,2048,0.010107733805974324
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,4096,512,0.002980266759792964
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,4096,512,0.014733866850535075
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,4096,1024,0.007980800171693166
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,4096,256,0.0025802666942278544
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,4096,256,0.014086400469144186
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,4096,768,0.007896533111731212
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,4096,128,0.0021685334543387097
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,4096,128,0.014476799964904785
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,4096,256,0.007122133175532024
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,4096,64,0.00222080002228419
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,4096,64,0.014436266819636025
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,4096,32,0.002203733225663503
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,4096,128,0.006684799989064534
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,4096,32,0.01439786652723948
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,4096,512,0.0074890668193499255
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,3584,65536,0.07135039965311686
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,3584,16384,0.023332266012827556
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,3584,65536,0.061292799313863125
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,3584,16384,0.028500266869862872
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,3584,12288,0.01791680057843526
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,3584,12288,0.025972266991933186
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,3584,12288,0.028937600056330365
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,3584,10240,0.015639467040697734
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,3584,10240,0.02505706747372945
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,3584,16384,0.036142933368682864
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,3584,8192,0.013313066959381104
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,3584,8192,0.024089600642522177
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,3584,10240,0.02553173303604126
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,3584,7168,0.01241919994354248
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,3584,7168,0.020806399981180827
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,3584,8192,0.021156267325083414
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,3584,6144,0.011397332946459452
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,3584,6144,0.022299732764561972
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,3584,7168,0.01946453253428141
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,3584,5120,0.011573333541552227
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,3584,5120,0.019961599508921304
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,4096,65536,0.12664852937062582
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,3584,4096,0.009757866462071735
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,3584,4096,0.016787199179331462
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,3584,6144,0.017454934120178223
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,3584,3584,0.009169066945711773
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,3584,3584,0.016636799772580466
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,3584,5120,0.01573973298072815
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,3584,3072,0.007950933277606964
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,3584,3072,0.01726079980532328
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,3584,3584,0.01269760032494863
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,3584,2560,0.007175466914971669
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,3584,2560,0.01731200019518534
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,3584,4096,0.01369599997997284
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,3584,2048,0.004822400212287903
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,3584,2048,0.014679466684659323
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,3584,2048,0.010046933094660442
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,3584,1536,0.003934933245182038
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,3584,3072,0.012039466698964437
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,3584,1536,0.014578133821487427
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,3584,1024,0.0034495999415715536
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,3584,1024,0.014975999792416891
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,3584,2560,0.010845866799354554
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,3584,768,0.0030389333764712016
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,3584,768,0.014667733510335287
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,3584,768,0.008002133170763651
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,3584,512,0.0026133333643277483
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,3584,512,0.0144896000623703
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,3584,1024,0.00796693315108617
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,3584,256,0.002603733291228612
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,3584,256,0.014324266711870828
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,3584,1536,0.008941866954167684
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,3584,128,0.0022314667701721193
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,3584,256,0.007091199855009715
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,3584,512,0.007144533097743988
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,3584,128,0.01447466711203257
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,3584,64,0.002179199953873952
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,3584,64,0.013563733299573263
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,3584,32,0.0022645334402720133
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,3584,128,0.006659199794133504
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,3584,32,0.01383039951324463
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,3072,65536,0.06288853486378988
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,3072,65536,0.05625706513722738
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,3072,16384,0.02065599958101908
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,3072,16384,0.027025065819422406
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,3072,12288,0.015520000457763672
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,3072,12288,0.025917865832646686
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,3072,12288,0.028990934292475384
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,3072,10240,0.014133333166440328
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,3072,10240,0.0249834676583608
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,3072,16384,0.03615573247273763
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,3072,8192,0.012054399649302164
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,3072,8192,0.02064853310585022
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,3072,10240,0.025225599606831867
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,3072,7168,0.011462400356928509
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,3072,7168,0.0208512008190155
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,3072,8192,0.020713599522908528
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,3072,6144,0.010388267040252686
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,3072,6144,0.02120959957440694
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,3072,7168,0.019382399320602418
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,3072,5120,0.009655466675758362
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,3072,5120,0.01877760092417399
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,3072,6144,0.01744106610616048
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,3072,4096,0.009179733196894328
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,3072,4096,0.016544000307718913
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,3072,5120,0.016156799594561257
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,3072,3584,0.008357333143552144
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,3072,3584,0.016883200407028197
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,3072,4096,0.013768532872200012
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,3072,3072,0.007196799914042156
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,3072,3072,0.016505600015322367
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,3072,3584,0.012868266304334006
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,3072,2560,0.006617600222428639
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,3072,2560,0.01751040021578471
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,3072,3072,0.011655466755231221
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,3072,2048,0.004680533210436503
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,3072,2048,0.014493866761525472
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,3072,2048,0.010436266660690308
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,3072,1536,0.003828266759713491
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,3072,1536,0.01452906628449758
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,3072,2560,0.011206400394439698
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,3072,1024,0.003489066660404205
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,3072,1024,0.016035200158754984
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,3072,1024,0.008425600330034892
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,3072,768,0.0030591999491055804
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,3072,768,0.014591999848683677
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,3072,1536,0.009224533041318258
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,3072,512,0.0026464000344276427
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,3072,512,0.013937066992123923
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,3072,512,0.007522133489449819
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,3072,768,0.007949866851170858
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,3072,256,0.002549333373705546
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,3072,256,0.014509866635004679
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,3072,128,0.0023103999594847363
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,3072,128,0.014523733655611673
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,3584,65536,0.12637653350830078
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,3072,256,0.007202133536338806
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,3072,128,0.006741333504517872
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,3072,64,0.0022229333718617756
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,3072,64,0.01439466675122579
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,3072,32,0.0022240000466505687
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,3072,32,0.014506666858990987
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,2560,65536,0.055035734176635744
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,2560,65536,0.05305920044581095
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,2560,16384,0.018593066930770875
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,2560,16384,0.026900267601013182
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,2560,12288,0.013677866260210673
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,2560,12288,0.024150399367014568
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,2560,12288,0.028645332654317218
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,2560,10240,0.012882133324941
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,2560,10240,0.02382826606432597
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,2560,16384,0.03604480028152466
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,2560,8192,0.011241599917411804
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,2560,8192,0.021517866849899293
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,2560,10240,0.024830933411916098
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,2560,7168,0.010354133447011311
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,2560,7168,0.021977599461873373
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,2560,8192,0.0207370658715566
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,2560,6144,0.009619200229644775
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,2560,6144,0.019363200664520262
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,2560,7168,0.01949866612752279
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,2560,5120,0.008805333574612936
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,2560,5120,0.019371734062830607
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,2560,6144,0.01753600041071574
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,2560,4096,0.00840106705824534
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,2560,4096,0.016824533541997276
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,2560,5120,0.015829333662986757
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,2560,3584,0.00786240001519521
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,2560,3584,0.016609066724777223
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,2560,4096,0.013803733388582864
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,2560,3072,0.006792533397674561
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,2560,3072,0.01548906664053599
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,2560,3584,0.012885333100954691
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,2560,2560,0.0061258668700853985
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,2560,2560,0.017334399620691936
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,2560,3072,0.011729066570599873
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,3072,65536,0.12655680179595946
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,2560,2048,0.004633600016434988
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,2560,2048,0.0145578662554423
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,2560,1536,0.003976533313592275
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,2560,1536,0.014711466431617738
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,2560,1536,0.009211732943852743
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,2560,1024,0.0034154665966828666
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,2560,1024,0.014775466918945313
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,2560,2560,0.011199999849001567
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,2560,2048,0.010362666845321656
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,2560,768,0.0030250666042168934
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,2560,768,0.014459733168284097
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,2560,512,0.002583466718594233
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,2560,512,0.014470400412877402
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,2560,512,0.007130666573842366
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,2560,256,0.002566399921973546
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,2560,768,0.007513600091139476
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,2560,256,0.014667733510335287
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,2560,128,0.002251733342806498
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,2560,128,0.014497066537539164
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,2560,1024,0.00874773363272349
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,2560,64,0.002239999920129776
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,2560,256,0.007495466868082683
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,2560,64,0.014062933127085366
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,2560,32,0.002143999934196472
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,2560,128,0.006753066678841908
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,2560,32,0.014475733041763306
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,2048,65536,0.04516373475392659
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,2048,16384,0.015523200233777365
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,2048,65536,0.04689706563949585
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,2048,16384,0.025594667593638105
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,2048,12288,0.011949867010116577
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,2048,12288,0.02577386697133382
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,2048,12288,0.028152533372243244
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,2048,10240,0.01087679962317149
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,2048,10240,0.023118933041890465
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,2048,16384,0.0358570655186971
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,2048,8192,0.011642666657765706
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,2048,8192,0.02081813414891561
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,2048,10240,0.02487466732660929
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,2048,7168,0.010549333691596986
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,2048,7168,0.018692266941070557
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,2048,8192,0.020794665813446044
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,2048,6144,0.009899733463923137
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,2048,6144,0.018593066930770875
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,2048,7168,0.01913386583328247
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,2048,5120,0.008937600255012512
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,2048,5120,0.01668693423271179
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,2048,6144,0.01744106610616048
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,2048,4096,0.007974400122960409
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,2048,4096,0.01660693287849426
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,2048,5120,0.015740799903869628
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,2048,3584,0.006748799979686737
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,2048,3584,0.016541866461435954
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,2048,4096,0.013783466815948487
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,2048,3072,0.00590826670328776
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,2048,3072,0.015955199797948204
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,2048,3584,0.012918399771054587
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,2048,2560,0.005096533397833506
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,2048,2560,0.01606613298257192
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,2048,3072,0.011904000242551168
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,2048,2048,0.0046517332394917805
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,2048,2048,0.014986667037010192
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,2560,65536,0.12647679646809895
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,2048,1536,0.003941333293914795
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,2048,1536,0.017069866259892784
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,2048,2560,0.010838400324185688
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,2048,1024,0.0031306666632493338
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,2048,1024,0.01446399986743927
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,2048,1536,0.009186133742332458
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,2048,768,0.003019733230272929
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,2048,1024,0.008382933338483174
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,2048,768,0.014798933267593383
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,2048,512,0.00266239990790685
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,2048,2048,0.01000426709651947
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,2048,512,0.014636799693107605
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,2048,256,0.0025653332471847535
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,2048,256,0.014501333236694336
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,2048,256,0.0071168000499407455
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,2048,512,0.007197866837183635
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,2048,128,0.0022154666483402254
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,2048,128,0.014379733800888061
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,2048,64,0.002161066730817159
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,2048,768,0.007569066683451335
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,2048,32,0.0021461332837740582
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,2048,64,0.013235200444857279
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,2048,128,0.006663466493288676
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,2048,32,0.014153599739074707
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,1536,65536,0.03545920054117839
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,1536,16384,0.014944000045458474
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,1536,65536,0.04125119845072429
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,1536,16384,0.02528426647186279
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,1536,12288,0.012166399757067363
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,1536,12288,0.02263573408126831
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,1536,12288,0.028154667218526202
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,1536,10240,0.011239467064539592
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,1536,10240,0.021504000822703043
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,1536,16384,0.03553386529286702
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,1536,8192,0.010001066327095031
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,1536,8192,0.022475733359654745
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,1536,10240,0.024493867158889772
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,1536,7168,0.009292800227801006
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,1536,7168,0.020478934049606323
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,1536,8192,0.021022933721542358
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,1536,6144,0.008831999699274699
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,1536,6144,0.019310933351516724
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,1536,7168,0.01906986633936564
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,1536,5120,0.007906133433183034
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,1536,5120,0.016812799374262492
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,1536,6144,0.017422932386398315
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,1536,4096,0.007145600020885467
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,1536,5120,0.01574186682701111
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,1536,4096,0.01766293247540792
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,1536,3584,0.006358399987220764
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,1536,3584,0.017846399545669557
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,1536,3584,0.012852266430854797
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,1536,3072,0.005718400080998739
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,1536,3072,0.017211733261744182
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,1536,4096,0.01372053325176239
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,1536,2560,0.005060266455014547
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,1536,2560,0.014825600385665893
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,1536,3072,0.012151466806729634
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,1536,2048,0.005196799834569296
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,1536,2048,0.014833066860834757
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,1536,2560,0.01120746632417043
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,1536,1536,0.0038090666135152185
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,1536,1536,0.014542933305104574
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,2048,65536,0.12610560258229572
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,1536,1024,0.0034400001168251038
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,1536,1024,0.015187199910481772
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,1536,1536,0.009206400314966837
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,1536,768,0.002994133283694585
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,1536,1024,0.007987200220425924
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,1536,768,0.014885333180427552
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,1536,512,0.00258240004380544
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,1536,512,0.014416000247001648
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,1536,2048,0.009981866677602131
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,1536,256,0.0025813333690166474
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,1536,256,0.014565333724021912
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,1536,768,0.00787199983994166
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,1536,256,0.007152000069618225
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,1536,512,0.007157333195209503
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,1536,128,0.00222080002228419
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,1536,64,0.0021322667598724365
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,1536,128,0.014485333363215128
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,1536,32,0.0021418665846188864
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,1536,64,0.014346667130788169
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,1536,128,0.006662400066852569
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,1536,32,0.014407466848691305
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,1024,65536,0.025333333015441894
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,1024,16384,0.011092266440391541
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,1024,65536,0.033243733644485476
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,1024,16384,0.022705066204071044
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,1024,12288,0.009549867113431294
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,1024,12288,0.02132479945818583
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,1024,12288,0.02776319980621338
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,1024,10240,0.0086325337489446
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,1024,10240,0.019307732582092285
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,1024,16384,0.03539520104726156
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,1024,8192,0.007642666498819987
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,1024,8192,0.018862932920455933
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,1024,10240,0.02449493408203125
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,1024,7168,0.007245866457621257
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,1024,7168,0.02115839918454488
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,1024,8192,0.02069973349571228
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,1024,6144,0.006316799918810527
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,1024,6144,0.01902186671892802
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,1024,7168,0.01909439961115519
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,1024,5120,0.006109866499900818
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,1024,5120,0.018579200903574625
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,1024,6144,0.016990933815638223
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,1024,4096,0.005936000247796377
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,1024,4096,0.01651946703592936
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,1024,5120,0.015871999661127727
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,1024,3584,0.006309333443641663
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,1024,3584,0.017977599302927652
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,1024,4096,0.013310933113098144
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,1024,3072,0.005467733244101206
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,1024,3072,0.016657066345214844
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,1024,3584,0.012513066331545511
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,1024,2560,0.005047466854254405
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,1024,2560,0.016581333676973977
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,1024,3072,0.011596799890200297
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,1024,2048,0.005264000097910563
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,1024,2048,0.01499626636505127
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,1024,2560,0.011256532867749532
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,1024,1536,0.00468800018231074
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,1024,1536,0.014969600240389505
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,1024,2048,0.010090667009353637
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,1024,1024,0.003538133452335993
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,1024,1024,0.014516266187032065
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,1024,1536,0.009249066313107807
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,1024,768,0.003382399926582972
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,1024,768,0.015180800358454385
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,1024,1024,0.007977599898974102
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,1024,512,0.00276053324341774
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,1024,512,0.014520532886187234
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,1536,65536,0.12508800029754638
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,1024,768,0.007942399879296621
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,1024,256,0.0026101333399613695
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,1024,256,0.014504533012708029
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,1024,512,0.007097599903742473
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,1024,128,0.002218666672706604
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,1024,128,0.013102933764457703
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,1024,256,0.006769066552321117
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,1024,64,0.0021877333521842955
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,1024,128,0.006738133231798808
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,1024,64,0.014444800217946372
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,1024,32,0.002169599880774816
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,768,65536,0.020796799659729005
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,1024,32,0.013608533143997192
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,768,65536,0.033973332246144614
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,768,16384,0.010230400164922078
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,768,16384,0.02119999925295512
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,768,12288,0.008782933155695598
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,768,12288,0.020642133553822835
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,768,16384,0.0351583997408549
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,768,10240,0.0075093333919843035
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,768,10240,0.020510933796564736
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,768,12288,0.02800533374150594
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,768,8192,0.006728533407052357
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,768,8192,0.021132800976435342
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,768,8192,0.020708266894022623
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,768,7168,0.005977599819501241
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,768,7168,0.019476266702016194
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,768,10240,0.024819199244181314
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,768,6144,0.006326400240262349
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,768,6144,0.01882773240407308
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,768,7168,0.019157334168752035
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,768,5120,0.005914666752020518
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,768,5120,0.01870186726252238
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,768,6144,0.017406932512919106
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,768,4096,0.006250666578610738
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,768,4096,0.017374932765960693
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,1024,65536,0.12227413654327393
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,768,3584,0.005932799975077311
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,768,3584,0.01669013301531474
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,768,4096,0.013682132959365845
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,768,3072,0.005513600011666616
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,768,3584,0.012468266487121581
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,768,3072,0.016381866733233132
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,768,5120,0.015737600127855935
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,768,2560,0.005538133283456167
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,768,2560,0.016119466225306193
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,768,2048,0.004378666480382284
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,768,2048,0.01530880033969879
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,768,2560,0.010790399710337321
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,768,1536,0.0042303999265035
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,768,2048,0.009586133559544881
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,768,3072,0.011609599987665812
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,768,1024,0.0034901333351929987
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,768,1536,0.01639253298441569
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,768,1024,0.015838932991027833
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,768,768,0.0031637333333492277
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,768,768,0.014442666371663412
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,768,1024,0.007978666822115581
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,768,512,0.002682666728893916
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,768,1536,0.009148800373077392
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,768,768,0.007897600034872691
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,768,512,0.014385066429773965
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,768,256,0.0025727999707063037
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,768,256,0.013086932897567748
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,768,128,0.002160000056028366
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,768,128,0.013674666484196981
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,768,512,0.007155199845631917
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,768,256,0.007106133302052816
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,768,64,0.002229333420594533
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,768,32,0.0022240000466505687
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,768,64,0.014337066809336343
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,768,128,0.006311466793219249
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,768,32,0.014447999993960061
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,512,65536,0.030164267619450885
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,512,16384,0.007567999760309856
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,512,16384,0.021015467246373494
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,512,16384,0.035076268513997394
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,512,12288,0.007153066496054332
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,512,12288,0.02064639925956726
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,512,12288,0.027705599864323933
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,512,10240,0.006541866560777028
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,512,10240,0.018631466229756675
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,512,10240,0.024394667148590087
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,512,8192,0.006001066664854685
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,512,8192,0.019074134031931558
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,768,65536,0.12182719707489013
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,512,7168,0.005577600002288819
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,512,7168,0.01874879995981852
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,512,7168,0.019045333067576088
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,512,8192,0.020618667205174766
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,512,6144,0.005891199906667074
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,512,6144,0.018677333990732826
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,512,5120,0.005843199789524078
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,512,5120,0.0185589333375295
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,512,5120,0.015363199512163797
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,512,4096,0.005915733178456625
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,512,6144,0.017348267634709678
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,512,4096,0.018687999248504637
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,512,3584,0.005737600227197012
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,512,3584,0.01665279964605967
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,512,3584,0.01242026686668396
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,512,4096,0.01362666686375936
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,512,3072,0.0054730668663978575
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,512,3072,0.01808533271153768
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,512,3072,0.011603200435638427
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,512,2560,0.005051733553409576
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,512,2560,0.016695467631022136
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,512,2560,0.010811733206113179
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,512,2048,0.005055999755859375
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,512,2048,0.014595199624697366
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,512,1536,0.003973333289225897
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,512,1536,0.01530239979426066
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,512,2048,0.009576533238093059
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,512,1024,0.002994133283694585
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,512,1024,0.014839466412862143
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,512,65536,0.12205013434092205
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,512,768,0.0030730667213598887
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,512,768,0.014536533753077188
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,512,1536,0.00888320008913676
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,512,1024,0.00795306662718455
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,512,512,0.0026346666117509207
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,512,512,0.014300800363222756
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,512,256,0.0025589334468046824
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,512,768,0.007956266899903615
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,512,256,0.014490666985511779
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,512,128,0.0021749332547187803
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,512,512,0.007150933146476746
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,512,128,0.013617066542307535
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,512,64,0.0021546666820844015
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,512,256,0.00713919997215271
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,512,128,0.006673066814740498
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,512,32,0.002182399978240331
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,512,64,0.013398399949073792
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,256,65536,0.01652479966481527
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,512,32,0.014472533265749613
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,256,16384,0.006675200164318084
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,256,65536,0.026744532585144042
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,256,16384,0.02049386699994405
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,256,12288,0.005505066613356272
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,256,12288,0.01930239995320638
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,256,12288,0.027704532941182452
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,256,10240,0.005859200159708659
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,256,10240,0.018680532773335777
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,256,16384,0.03497386773427327
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,256,8192,0.0058773333827654515
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,256,8192,0.02021119991938273
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,256,10240,0.02432533303896586
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,256,7168,0.005471999943256378
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,256,8192,0.020637865861256918
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,256,7168,0.01858773430188497
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,256,6144,0.00544106662273407
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,256,6144,0.01843093236287435
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,256,7168,0.019351466496785482
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,256,5120,0.005052799979845682
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,256,5120,0.018542933464050292
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,256,6144,0.01698453426361084
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,256,4096,0.0054666668176651
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,256,4096,0.01658986707528432
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,256,5120,0.01569386621316274
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,256,3584,0.005043200155099233
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,256,4096,0.013641599814097086
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,256,3584,0.016658133268356322
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,256,3072,0.004664533336957296
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,256,3072,0.0162581334511439
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,256,3584,0.012474667032559712
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,256,2560,0.005094400048255921
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,256,2560,0.01551253298918406
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,256,3072,0.011593600114186604
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,256,2048,0.00462719996770223
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,256,2048,0.016497066617012023
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,256,2560,0.010762666662534077
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,256,1536,0.0037930667400360107
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,256,1536,0.014760532975196838
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,256,1536,0.009161600470542907
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,256,1024,0.003032533327738444
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,256,1024,0.014579199751218162
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,256,1024,0.007893333335717519
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,256,768,0.0029887999097506206
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,256,768,0.014498133460680643
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,256,768,0.007910400132338206
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,256,512,0.002624000112215678
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,256,512,0.014642133315404256
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,256,512,0.007473066449165344
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,256,256,0.0022111999491850535
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,256,256,0.014455466469128927
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,256,256,0.007084799806276958
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,256,128,0.0021461332837740582
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,256,128,0.014366933703422546
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,256,128,0.006646400193373363
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,256,64,0.0021888000269730886
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,256,65536,0.12168640295664471
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,256,64,0.013238400220870972
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,256,32,0.0022101332743962605
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,128,65536,0.009156266848246258
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,256,32,0.013698133826255798
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,128,65536,0.025006934007008867
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,128,16384,0.005870933334032694
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,128,16384,0.02094399929046631
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,128,16384,0.0353333314259847
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,128,12288,0.005448533097902933
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,128,12288,0.01918399930000305
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,128,65536,0.1203114668528239
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,128,10240,0.00561706672112147
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,128,10240,0.018678400913874307
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,128,12288,0.027806933720906573
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,128,8192,0.005469866593678792
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,128,8192,0.019411200284957887
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,128,10240,0.023939200242360435
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,128,7168,0.005463466544946035
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,128,7168,0.01824959913889567
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,128,8192,0.020642133553822835
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,128,6144,0.005073066552480062
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,128,6144,0.018576000134150186
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,128,7168,0.019011199474334717
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,128,5120,0.005043200155099233
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,128,5120,0.018083200852076212
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,128,6144,0.017318399747212727
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,128,4096,0.004610133171081543
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,128,4096,0.01739306648572286
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,128,5120,0.01566933294137319
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,128,3584,0.005050666630268097
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,128,3584,0.016521599888801575
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,128,4096,0.013309866189956665
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,128,3072,0.004664533336957296
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,128,3072,0.017423999309539796
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,128,3584,0.012408533692359924
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,128,2560,0.005062399804592133
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,128,2560,0.014958932995796204
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,128,3072,0.011658666531244914
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,128,2048,0.004329599936803182
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,128,2048,0.014989866813023885
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,128,2560,0.010538666447003683
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,128,1536,0.003802666564782461
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,128,1536,0.01588053305943807
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,128,2048,0.009970133503278095
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,128,1024,0.0030037333567937215
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,128,1024,0.01474453310171763
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,128,1536,0.009126399954160053
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,128,768,0.0029909332593282064
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,128,768,0.014472533265749613
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,128,1024,0.0079434668024381
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,128,512,0.0026186667382717133
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,128,512,0.014782933394114175
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,128,768,0.007748266557852428
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,128,256,0.0021418665846188864
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,128,256,0.014705066879590353
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,128,256,0.007091199855009715
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,128,128,0.002183466653029124
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,128,128,0.014164266983668008
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,128,128,0.006295466423034668
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,128,64,0.0021344001094500227
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,128,64,0.013335466384887695
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,128,32,0.0021888000269730886
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,128,32,0.013372799754142762
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,64,65536,0.009155199925104777
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,64,65536,0.024903466304143272
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,64,16384,0.0058677335580190025
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,64,16384,0.01951573292414347
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,64,12288,0.005036800106366476
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,64,12288,0.018614399433135986
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,64,10240,0.005032533407211303
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,64,10240,0.019139200448989868
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,64,8192,0.0054400001962979635
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,64,8192,0.01827626625696818
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,64,7168,0.005086933573087057
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,64,7168,0.018013866742451985
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,64,6144,0.005058133105436961
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,64,6144,0.017258665959040322
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,64,5120,0.0046847999095916745
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,64,5120,0.016615466276804606
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,64,4096,0.0045525332291920986
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,64,4096,0.016657066345214844
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,64,3584,0.004756266872088114
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,64,3584,0.016455466548601784
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,64,3072,0.00462719996770223
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,64,3072,0.017038933436075845
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,64,2560,0.004615466793378194
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,64,2560,0.015285332997639975
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,64,2048,0.004257066547870636
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,64,2048,0.014520532886187234
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,64,1536,0.0033887999753157297
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,64,1536,0.014513066411018372
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,64,1024,0.002961066613594691
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,64,1024,0.01550933321317037
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,64,768,0.002570666621128718
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,64,768,0.014311466614405313
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,64,512,0.0025546667476495106
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,64,512,0.014604799946149192
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,64,256,0.0021717332303524016
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,64,256,0.014170666535695394
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,64,128,0.002125866711139679
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,64,128,0.01402453382809957
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,64,64,0.001803733284274737
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,64,64,0.012683733304341634
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,64,32,0.001817600056529045
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,64,32,0.012492799758911132
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,32,65536,0.007493333518505096
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,32,65536,0.024430932601292928
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,32,16384,0.005921066800753275
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,32,16384,0.018913066387176512
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,32,12288,0.005003733436266581
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,32,12288,0.01837973395983378
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,32,10240,0.005172266562779745
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,32,10240,0.01883093317349752
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,32,8192,0.004461866617202759
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,32,8192,0.01752426624298096
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,32,7168,0.004724266628424326
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,32,7168,0.01839253306388855
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,32,6144,0.004786133269468943
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,32,6144,0.01761173407236735
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,32,5120,0.004655999938646952
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,32,5120,0.01754666765530904
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,32,4096,0.004293333490689596
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,32,4096,0.01845759948094686
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,32,3584,0.004692266881465912
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,32,3584,0.016537599762280784
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,32,3072,0.004242133100827535
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,32,3072,0.016520532965660095
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,32,2560,0.004622933268547058
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,32,2560,0.015658666690190635
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,32,2048,0.004274133344491323
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,32,2048,0.01558080017566681
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,32,1536,0.0034965333839257562
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,32,1536,0.015209600329399109
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,32,1024,0.003045333425203959
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,32,1024,0.015057067076365152
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,32,768,0.00262719988822937
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,32,768,0.014484266440073649
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,32,512,0.0022453332940737406
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,32,512,0.01452906628449758
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,32,256,0.002242133269707362
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,32,256,0.014535466829935709
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,32,128,0.0022026665508747103
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,32,128,0.014217600226402283
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,32,64,0.0017984000345071156
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,32,64,0.012625066439310708
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,32,32,0.0018144000321626663
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,32,32,32,0.01246506671110789
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,65536,16384,0.3329375902811686
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,65536,16384,0.19614613850911458
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,32,512,65536,0.027987200021743774
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,65536,12288,0.21888213157653807
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,65536,12288,0.1269706646601359
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,256,2048,0.009949866930643718
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,65536,10240,0.18376213709513348
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,65536,10240,0.10887253284454346
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,32,128,512,0.007133866846561432
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,65536,8192,0.1474837303161621
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,65536,8192,0.08928213119506836
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,65536,8192,0.06918400128682455
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,65536,7168,0.12983253002166747
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,65536,7168,0.08008320331573486
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,65536,10240,0.07889066537221273
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,65536,6144,0.11214400132497152
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,65536,6144,0.07067840099334717
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,65536,16384,0.11635626951853435
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,65536,5120,0.09440213044484456
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,65536,5120,0.06199359893798828
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,65536,12288,0.08921493689219157
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,65536,4096,0.07570772965749105
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,65536,4096,0.053149867057800296
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,65536,6144,0.05460160175959269
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,65536,3584,0.06680533091227213
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,65536,3584,0.04766186475753784
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,65536,7168,0.06212053298950195
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,65536,3072,0.05792106787363688
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,65536,3072,0.04356373151143392
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,65536,4096,0.035778133074442546
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,65536,2560,0.0489408016204834
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,65536,2560,0.03926080067952474
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,65536,5120,0.04463466803232829
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,65536,2048,0.03993279933929443
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,65536,2048,0.03420053323109944
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,65536,2560,0.02553386688232422
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,65536,1536,0.0310208002726237
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,65536,1536,0.02983466585477193
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,65536,3584,0.03337386846542358
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,65536,1024,0.022153600056966146
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,65536,1024,0.02473919987678528
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,65536,3072,0.030048000812530517
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,65536,768,0.017458132902781167
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,65536,768,0.022796799739201866
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,65536,2048,0.022012799978256226
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,65536,512,0.012482133507728577
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,65536,512,0.019879466295242308
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,65536,768,0.013800533612569174
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,65536,256,0.007250133156776428
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,65536,256,0.016637866695721946
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,65536,512,0.011283199985822041
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,65536,128,0.005036800106366476
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,65536,128,0.01605013310909271
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,65536,128,0.009084799885749817
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,65536,64,0.004281599819660187
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,65536,64,0.016525866587956746
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,65536,32,0.004287999868392944
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,65536,32,0.015537066260973611
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,16384,65536,0.3257301330566406
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,16384,65536,0.18531840642293293
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,65536,256,0.009619200229644775
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,16384,16384,0.09365440209706624
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,16384,16384,0.056423465410868325
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,65536,1024,0.015602133671442666
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,16384,12288,0.06834987004597982
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,16384,12288,0.047310932477315264
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,65536,1536,0.01897066632906596
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,16384,10240,0.05936213334401449
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,16384,10240,0.042148268222808837
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,16384,10240,0.02983466585477193
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,16384,8192,0.04213653405507405
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,16384,8192,0.03548266490300496
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,16384,12288,0.03359253406524658
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,16384,7168,0.03667200009028117
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,16384,7168,0.032961066563924155
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,16384,16384,0.045177598794301346
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,16384,6144,0.032280532519022624
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,16384,6144,0.031541333595911665
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,16384,8192,0.02558293342590332
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,16384,5120,0.02763413389523824
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,16384,5120,0.028756266832351683
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,16384,7168,0.021985065937042237
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,16384,4096,0.02241386572519938
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,16384,4096,0.02483200033505758
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,16384,6144,0.020584533611933388
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,16384,3584,0.01995413303375244
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,16384,4096,0.015732266505559287
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,16384,3584,0.02269759972890218
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,16384,3072,0.01735466718673706
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,16384,3072,0.02285760045051575
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,16384,5120,0.018987733125686645
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,16384,2560,0.014944000045458474
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,16384,2560,0.020771199464797975
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,16384,2560,0.012989866733551025
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,16384,2048,0.012432000041007996
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,16384,2048,0.019486933946609497
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,16384,3072,0.014563199877738953
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,16384,1536,0.00965226689974467
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,16384,1536,0.018549333016077675
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,16384,3584,0.014934399724006652
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,16384,1024,0.007478400071461995
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,16384,1024,0.01606613298257192
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,16384,1024,0.0088128000497818
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,16384,768,0.006289066871007283
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,16384,768,0.014673067132631936
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,16384,1536,0.010034132997194927
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,16384,512,0.003420799970626831
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,16384,512,0.014493866761525472
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,16384,2048,0.011987200379371643
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,16384,256,0.002631466587384542
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,16384,256,0.014552533626556396
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,16384,256,0.0072543998559316
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,16384,128,0.0025631998976071674
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,16384,128,0.014544000228246054
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,16384,512,0.007514666517575581
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,16384,64,0.0022495999932289124
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,16384,64,0.014469333489735923
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,16384,128,0.006746666630109151
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,16384,32,0.014447999993960061
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,12288,65536,0.2406602700551351
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,16384,768,0.008311466872692108
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,12288,65536,0.14818026224772135
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,12288,16384,0.06697920163472494
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,12288,16384,0.052324267228444424
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,12288,16384,0.039070932070414226
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,12288,12288,0.0494485338528951
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,12288,12288,0.041509334246317545
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,12288,12288,0.03102186719576518
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,12288,10240,0.042498131593068436
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,12288,10240,0.03784746726353963
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,12288,10240,0.02683200041453044
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,12288,8192,0.03568533261617025
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,12288,8192,0.03443839947382609
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,16384,32,0.0021909333765506743
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,12288,7168,0.030896000067392987
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,12288,7168,0.030989867448806763
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,12288,7168,0.020592000087102255
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,12288,6144,0.027192533016204834
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,12288,6144,0.028904533386230467
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,12288,8192,0.02280319929122925
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,12288,5120,0.023986132939656575
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,12288,5120,0.02490239938100179
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,12288,5120,0.017413334051767985
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,12288,4096,0.02034453352292379
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,12288,4096,0.022940800587336222
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,12288,6144,0.019030400117238364
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,12288,3584,0.017809067169825235
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,12288,3584,0.02146240075429281
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,12288,3584,0.014139733711878457
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,12288,3072,0.01574399967988332
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,12288,3072,0.02138026754061381
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,12288,4096,0.01530880033969879
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,12288,2560,0.011667199929555257
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,12288,2560,0.020662399133046468
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,12288,3072,0.012897066275278726
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,12288,2048,0.009971200426419576
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,12288,2048,0.018583466609319053
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,12288,2560,0.011657599608103435
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,12288,1536,0.008772266904513042
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,12288,1536,0.01829973260561625
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,12288,1536,0.009204266468683879
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,12288,1024,0.006897066533565521
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,12288,1024,0.016787199179331462
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,12288,2048,0.010466133554776508
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,12288,768,0.0052373334765434265
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,12288,768,0.014513066411018372
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,12288,768,0.007977599898974102
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,12288,512,0.003937066594759623
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,12288,512,0.014526933431625366
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,12288,1024,0.008757332960764568
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,12288,256,0.0030229332546393077
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,12288,256,0.014622933665911355
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,12288,256,0.0071936001380284624
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,12288,128,0.002665599932273229
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,12288,128,0.014404267072677612
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,12288,512,0.0074879998962084455
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,12288,64,0.002209066599607468
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,12288,64,0.01446399986743927
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,12288,32,0.002573866645495097
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,12288,32,0.014447999993960061
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,12288,128,0.007100800176461537
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,10240,65536,0.20417386690775552
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,10240,65536,0.12871253490447998
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,10240,16384,0.06131413380304972
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,10240,16384,0.045390931765238444
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,16384,65536,0.15444906552632648
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,10240,12288,0.044121599197387694
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,10240,12288,0.03930346568425496
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,10240,16384,0.038211198647816975
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,10240,10240,0.03745280106862386
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,10240,10240,0.035523200035095216
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,12288,65536,0.13140587011973065
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,10240,8192,0.02991360028584798
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,10240,8192,0.030766934156417847
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,10240,12288,0.030104533831278486
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,10240,7168,0.027319467067718504
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,10240,7168,0.030078933636347456
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,10240,10240,0.026473599672317504
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,10240,6144,0.02445546587308248
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,10240,6144,0.02812160054842631
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,10240,8192,0.022669865687688192
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,10240,7168,0.020687999327977498
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,10240,5120,0.025949867566426595
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,10240,6144,0.019076265891393027
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,10240,4096,0.017156267166137697
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,10240,4096,0.023126399517059325
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,10240,5120,0.017230933904647826
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,10240,3584,0.01580586632092794
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,10240,3584,0.02123840053876241
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,10240,4096,0.014966400464375815
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,10240,3072,0.014260266224543253
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,10240,3072,0.0217685341835022
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,10240,3584,0.013823999961217245
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,10240,2560,0.011193600296974183
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,10240,2560,0.019077332814534505
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,10240,3072,0.012507733702659608
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,10240,2048,0.009900800387064616
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,10240,2048,0.018629332383473717
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,10240,2560,0.011213866869608562
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,10240,1536,0.008358400066693623
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,10240,1536,0.01651946703592936
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,10240,2048,0.01086186667283376
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,10240,1024,0.006331733365853627
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,10240,1024,0.016607999801635742
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,10240,1536,0.009212799866994222
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,10240,768,0.004675200084845225
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,10240,768,0.014487466216087342
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,10240,1024,0.008414933085441589
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,10240,512,0.00413973331451416
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,10240,512,0.014553599556287131
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,10240,768,0.007922133306662242
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,10240,256,0.003052799900372823
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,10240,256,0.014478933811187745
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,10240,512,0.007547733187675476
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,10240,128,0.0025994665920734406
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,10240,128,0.014467199643452963
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,10240,256,0.0070816000302632645
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,10240,64,0.002423466742038727
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,10240,64,0.01374613344669342
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,10240,32,0.002186666677395503
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,10240,32,0.014059733351071677
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,10240,128,0.007077333331108093
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,8192,65536,0.15733760197957355
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,8192,65536,0.104258131980896
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,8192,16384,0.048527999718983965
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,8192,16384,0.03647893269856771
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,8192,16384,0.03745813369750976
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,8192,12288,0.034983468055725095
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,8192,12288,0.031632000207901
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,8192,12288,0.02932159900665283
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,8192,10240,0.030436267455418903
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,8192,10240,0.028172800938288372
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,8192,10240,0.02603413263956706
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,8192,8192,0.021972266832987468
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,8192,8192,0.025142399470011394
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,10240,5120,0.02109439969062805
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,8192,7168,0.019474132855733236
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,8192,7168,0.023589332898457847
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,10240,65536,0.12860480149586995
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,8192,6144,0.017172267039616905
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,8192,6144,0.022824533780415854
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,8192,8192,0.021939200162887574
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,8192,5120,0.015011200308799743
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,8192,5120,0.021154133478800456
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,8192,6144,0.0190175990263621
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,8192,4096,0.012044800321261089
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,8192,4096,0.01960853338241577
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,8192,7168,0.020708266894022623
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,8192,3584,0.011241599917411804
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,8192,3584,0.019421867529551187
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,8192,5120,0.016978132724761962
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,8192,3072,0.009979733824729919
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,8192,3072,0.01981226603190104
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,8192,3072,0.01209920048713684
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,8192,2560,0.008764800429344178
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,8192,2560,0.017875200510025023
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,8192,3584,0.013433600465456644
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,8192,2048,0.007570133109887441
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,8192,2048,0.01658986707528432
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,8192,4096,0.014525866508483887
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,8192,1536,0.006274133423964183
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,8192,1536,0.014921599626541137
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,8192,1536,0.009275733431180318
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,8192,1024,0.004178133110205332
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,8192,1024,0.014476799964904785
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,8192,2048,0.01042133371035258
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,8192,768,0.0034400001168251038
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,8192,768,0.014579199751218162
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,8192,2560,0.011218133568763732
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,8192,512,0.0030069333811601003
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,8192,512,0.014429866274197897
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,8192,512,0.007978666822115581
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,8192,256,0.0026186667382717133
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,8192,256,0.0144405335187912
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,8192,768,0.007968000074227651
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,8192,128,0.0022240000466505687
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,8192,128,0.014467199643452963
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,8192,256,0.0070720002055168155
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,8192,64,0.0021749332547187803
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,8192,128,0.006718933085600535
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,8192,32,0.0022154666483402254
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,8192,64,0.014427733421325684
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,8192,32,0.01444906691710154
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,7168,65536,0.1480021317799886
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,7168,65536,0.09152639706929525
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,7168,16384,0.03988053401311238
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,7168,16384,0.0362773338953654
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,7168,16384,0.03725866476694743
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,7168,12288,0.030024532477060956
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,7168,12288,0.028907734155654907
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,8192,65536,0.12796160380045574
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,7168,10240,0.02604159911473592
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,7168,10240,0.02640639940897624
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,7168,12288,0.029707733790079755
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,7168,8192,0.02170133392016093
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,7168,8192,0.024604799350102742
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,7168,10240,0.02600533366203308
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,7168,7168,0.019452800353368126
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,7168,7168,0.024420267343521117
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,7168,8192,0.02232746680577596
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,7168,6144,0.01567359964052836
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,7168,6144,0.0226090669631958
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,7168,7168,0.020553600788116456
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,7168,5120,0.013642666737238565
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,7168,5120,0.022232532501220703
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,7168,6144,0.01909439961115519
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,7168,4096,0.011012267072995503
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,7168,4096,0.019899733861287437
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,7168,5120,0.017015467087427773
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,7168,3584,0.010317867000897724
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,7168,3584,0.020310399929682414
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,7168,3584,0.01336853305498759
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,7168,3072,0.009269332885742188
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,7168,3072,0.018372267484664917
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,7168,4096,0.014165332913398743
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,7168,2560,0.008291199803352356
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,7168,2560,0.016536532839139303
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,7168,2560,0.011300266782442728
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,7168,2048,0.007111466427644093
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,7168,2048,0.016960000991821288
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,7168,3072,0.012087466319402058
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,7168,1536,0.005487999816735586
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,7168,1536,0.01451520025730133
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,7168,1536,0.009628799557685853
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,7168,1024,0.003703466554482778
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,7168,1024,0.01477120021979014
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,7168,2048,0.0100000003973643
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,7168,768,0.0032042667269706728
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,7168,768,0.0145578662554423
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,7168,1024,0.008360532919565837
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,7168,512,0.0029706666866938275
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,7168,512,0.01448319951693217
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,7168,768,0.0079434668024381
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,7168,256,0.002573866645495097
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,7168,256,0.014613333344459533
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,7168,256,0.007176533341407776
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,7168,128,0.0022528000175952912
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,7168,128,0.01446293294429779
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,7168,512,0.007533866663773854
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,7168,64,0.002148266633351644
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,7168,64,0.014443733294804893
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,7168,32,0.0021727999051411946
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,7168,128,0.006792533397674561
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,7168,32,0.013774933417638144
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,6144,65536,0.12569493452707928
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,6144,65536,0.08458027044932047
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,6144,16384,0.035867734750111895
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,6144,16384,0.03624213139216105
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,6144,16384,0.03629333178202311
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,6144,12288,0.02645439902941386
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,6144,12288,0.029475200176239013
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,8192,1024,0.00839573343594869
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,6144,10240,0.02313813368479411
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,6144,10240,0.028262400627136232
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,7168,65536,0.127293864885966
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,6144,8192,0.01941759983698527
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,6144,8192,0.023436800638834635
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,6144,12288,0.029241599639256793
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,6144,7168,0.017460266749064125
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,6144,7168,0.02381333311398824
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,6144,10240,0.02609279950459798
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,6144,6144,0.015453867117563882
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,6144,6144,0.020920532941818237
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,6144,8192,0.022275199492772423
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,6144,5120,0.013750400145848593
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,6144,5120,0.020657066504160562
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,6144,6144,0.018175999323527016
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,6144,4096,0.010430933038393656
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,6144,4096,0.020060799519220986
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,6144,7168,0.020770132541656494
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,6144,3584,0.00959999958674113
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,6144,3584,0.017927465836207072
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,6144,3584,0.01292693316936493
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,6144,3072,0.008763733506202697
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,6144,3072,0.01906026601791382
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,6144,4096,0.013801599542299906
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,6144,2560,0.007548800110816956
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,6144,2560,0.015871999661127727
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,6144,5120,0.016571733355522155
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,6144,2048,0.006734933455785115
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,6144,2048,0.01572053333123525
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,6144,2560,0.010905599594116211
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,6144,1536,0.0050784001747767125
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,6144,1536,0.015782399972279867
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,6144,2048,0.010486400127410889
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,6144,1024,0.003554133325815201
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,6144,3072,0.01164906620979309
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,6144,1024,0.0165173331896464
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,6144,768,0.0030602666238943735
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,6144,768,0.015131733814875283
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,6144,768,0.007994666695594788
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,6144,512,0.0029205332199732465
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,6144,512,0.014865066607793173
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,6144,1024,0.008377599716186523
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,6144,256,0.002639999985694885
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,6144,256,0.014492799838383993
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,6144,512,0.007570133109887441
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,6144,128,0.0021856000026067098
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,6144,1536,0.009192533294359843
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,6144,256,0.007118933399518331
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,6144,64,0.0022357332209746042
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,6144,128,0.014520532886187234
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,6144,32,0.0021930667261282604
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,6144,64,0.01441493332386017
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,6144,32,0.01446399986743927
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,5120,65536,0.10333973566691082
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,6144,128,0.006674133241176605
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,5120,65536,0.075708802541097
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,5120,16384,0.03250666658083598
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,5120,16384,0.032237867514292404
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,5120,12288,0.024281599124272666
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,5120,12288,0.027730133136113482
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,5120,12288,0.029307732979456585
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,5120,10240,0.02062293291091919
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,5120,10240,0.026633600393931072
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,5120,16384,0.03619946638743083
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,5120,8192,0.017748266458511353
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,5120,8192,0.022081067164738975
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,5120,10240,0.025605332851409913
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,5120,7168,0.01602133313814799
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,5120,7168,0.023814400037129722
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,5120,8192,0.02194026708602905
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,5120,6144,0.014075733224550881
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,5120,6144,0.020967467625935873
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,5120,7168,0.019927465915679933
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,5120,5120,0.012388267119725545
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,5120,5120,0.01970240076382955
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,5120,6144,0.01783039967219035
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,5120,4096,0.009527466694513957
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,5120,4096,0.017108267545700072
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,5120,5120,0.015820800264676412
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,5120,3584,0.00878613293170929
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,5120,3584,0.01872319976488749
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,5120,4096,0.013705600301424661
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,5120,3072,0.007939200103282928
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,5120,3072,0.016007467110951742
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,5120,3584,0.01297599971294403
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,5120,2560,0.007110400001207988
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,5120,2560,0.016461867094039916
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,5120,3072,0.011715199549992878
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,5120,2048,0.005514666438102722
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,5120,2048,0.015209600329399109
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,5120,2560,0.010955733060836793
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,5120,1536,0.004732800026734671
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,5120,1536,0.015982932845751443
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,5120,2048,0.010050132870674133
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,5120,1024,0.0033802665770053864
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,5120,1024,0.014553599556287131
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,6144,65536,0.1275978644688924
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,5120,768,0.0030400000512599947
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,5120,768,0.014486400286356607
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,5120,1536,0.009245866537094116
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,5120,512,0.0026410666604836782
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,5120,512,0.014432000120480857
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,5120,1024,0.008317866424719492
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,5120,256,0.002567466596762339
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,5120,256,0.013077333569526672
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,5120,768,0.007898666461308797
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,5120,128,0.002163200080394745
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,5120,128,0.012849066654841104
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,5120,256,0.00721919983625412
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,5120,64,0.002145066608985265
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,5120,128,0.0066655998428662615
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,5120,64,0.014052266875902811
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,5120,32,0.0021759999295075734
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,5120,512,0.007459199925263722
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,5120,32,0.01399679978688558
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,4096,65536,0.0835103988647461
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,4096,65536,0.06284799973169962
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,4096,16384,0.025671466191609697
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,4096,16384,0.028338134288787842
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,4096,12288,0.01919999917348226
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,4096,12288,0.025684267282485962
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,4096,12288,0.028787199656168622
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,4096,10240,0.01698026657104492
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,4096,10240,0.024937599897384644
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,4096,16384,0.03617386817932129
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,4096,8192,0.01400213340918223
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,4096,8192,0.024244266748428344
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,4096,10240,0.025337600708007814
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,4096,7168,0.012846933801968894
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,4096,7168,0.023176532983779908
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,4096,8192,0.0212991992632548
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,4096,6144,0.011748266220092774
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,4096,6144,0.02347946763038635
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,4096,7168,0.01952426632245382
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,4096,5120,0.010829866925875346
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,4096,5120,0.020026665925979615
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,4096,6144,0.01778879960378011
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,4096,4096,0.011760000387827556
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,4096,4096,0.01662506659825643
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,4096,5120,0.015761066476504007
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,4096,3584,0.010644267002741497
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,4096,3584,0.016662399967511496
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,4096,4096,0.013729066650072733
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,4096,3072,0.009493333101272584
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,4096,3072,0.015067733327547708
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,4096,3584,0.01288106640179952
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,4096,2560,0.00595413347085317
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,4096,2560,0.016703999042510985
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,4096,3072,0.011708799997965496
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,4096,2048,0.0050784001747767125
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,4096,2048,0.014533332983652749
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,4096,2560,0.01087679962317149
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,4096,1536,0.004653866589069367
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,4096,1536,0.014602667093276978
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,4096,1536,0.009226666887601216
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,4096,1024,0.003437866767247518
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,4096,1024,0.014850133657455444
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,4096,2048,0.010052266716957092
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,4096,768,0.0030421334008375804
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,4096,768,0.014484266440073649
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,4096,768,0.007674666742483775
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,4096,512,0.0026634665826956432
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,4096,512,0.014455466469128927
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,4096,1024,0.007972266773382823
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,4096,256,0.0026208000878492994
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,4096,256,0.014390400052070618
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,4096,256,0.00716480016708374
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,4096,128,0.0022410665949185687
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,4096,128,0.014465066790580749
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,4096,512,0.007348266740640004
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,4096,64,0.0021546666820844015
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,4096,64,0.013719466328620911
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,4096,32,0.002149333308140437
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,4096,128,0.006723199784755707
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,4096,32,0.013702399532000222
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,3584,65536,0.07250133355458578
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,3584,65536,0.0566271980603536
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,3584,16384,0.02267199953397115
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,3584,16384,0.026921600103378296
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,5120,65536,0.1272447983423869
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,3584,12288,0.0173962672551473
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,3584,12288,0.025009065866470337
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,3584,12288,0.029240532716115313
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,3584,10240,0.015226667126019796
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,3584,10240,0.025193599859873454
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,3584,16384,0.03597013155619304
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,3584,8192,0.013027200102806091
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,3584,8192,0.02418880065282186
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,3584,10240,0.025605332851409913
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,3584,7168,0.012072533369064331
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,3584,7168,0.02269866665204366
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,3584,8192,0.021121066808700562
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,3584,6144,0.01097813347975413
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,3584,6144,0.020891733964284263
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,3584,6144,0.01745599905649821
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,3584,5120,0.009988266229629516
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,3584,5120,0.018553600708643595
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,3584,7168,0.019537067413330077
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,3584,4096,0.011702400445938111
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,3584,4096,0.016614400347073875
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,3584,4096,0.014109866817792258
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,3584,3584,0.01084266702334086
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,3584,3584,0.017084799210230508
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,3584,5120,0.015748266379038492
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,3584,3072,0.009367466966311137
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,3584,3072,0.015621333320935567
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,4096,65536,0.12646400133768718
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,3584,2560,0.0058773333827654515
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,3584,3072,0.011612799763679505
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,3584,2560,0.0153546671072642
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,3584,2048,0.004633600016434988
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,3584,2048,0.014459733168284097
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,3584,3584,0.012871467073758445
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,3584,1536,0.003961600114901861
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,3584,1536,0.014549332857131957
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,3584,1536,0.00920746624469757
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,3584,1024,0.0033759998778502146
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,3584,1024,0.01478506624698639
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,3584,2048,0.01004266639550527
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,3584,768,0.0030016000072161358
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,3584,768,0.01446613371372223
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,3584,2560,0.010900266965230306
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,3584,512,0.002600533266862233
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,3584,512,0.014473600188891092
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,3584,1024,0.00869866708914439
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,3584,256,0.002611200014750163
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,3584,256,0.014453333616256715
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,3584,512,0.007157333195209503
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,3584,768,0.007915733257929485
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,3584,128,0.002182399978240331
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,3584,128,0.014491732915242514
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,3584,64,0.0021578667064507803
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,3584,64,0.014428800344467163
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,3584,32,0.002165333429972331
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,3584,256,0.0070720002055168155
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,3584,128,0.007079466680685679
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,3584,32,0.014481066664059957
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,3072,65536,0.06360960006713867
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,3072,16384,0.020122667153676353
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,3072,65536,0.05442986488342285
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,3072,16384,0.026426666975021364
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,3072,12288,0.015291733543078103
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,3072,12288,0.02376746733983358
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,3072,12288,0.02893120050430298
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,3072,10240,0.013773866494496665
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,3072,10240,0.024306132396062215
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,3072,16384,0.03603839874267578
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,3072,8192,0.012036266922950744
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,3072,8192,0.022036266326904298
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,3072,10240,0.02531306743621826
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,3072,7168,0.011228799819946289
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,3072,7168,0.021516799926757812
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,3072,8192,0.021040000518163047
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,3072,6144,0.010233599940935771
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,3072,6144,0.020885332425435384
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,3072,7168,0.019499733050664266
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,3072,5120,0.009661866227785747
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,3072,5120,0.020473599433898926
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,3072,6144,0.017480534315109254
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,3072,4096,0.010391466816266378
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,3072,4096,0.016563199957211814
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,3072,5120,0.015848533312479655
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,3072,3584,0.009412266810735067
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,3072,3584,0.0165173331896464
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,3072,4096,0.013713066776593527
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,3072,3072,0.008441600203514098
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,3072,3072,0.016364799936612447
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,3072,3584,0.01295360028743744
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,3072,2560,0.005571199953556061
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,3072,2560,0.015879467129707336
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,3584,65536,0.12655146916707355
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,3072,2048,0.004689066608746847
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,3072,2048,0.016311466693878174
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,3072,3072,0.012083199620246888
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,3072,1536,0.003908266623814901
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,3072,1536,0.014501333236694336
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,3072,2560,0.01129919985930125
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,3072,1024,0.003453866640726725
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,3072,1536,0.008753066261609394
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,3072,1024,0.014777599771817525
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,3072,768,0.0029813334345817565
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,3072,768,0.014597333470980325
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,3072,2048,0.009940266609191895
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,3072,512,0.002603733291228612
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,3072,512,0.014456533392270408
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,3072,512,0.007477333148320516
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,3072,256,0.0025418666501839954
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,3072,256,0.0144405335187912
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,3072,768,0.007897600034872691
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,3072,128,0.0022122666239738466
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,3072,128,0.01446399986743927
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,3072,1024,0.008365866541862488
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,3072,256,0.007062399884064992
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,3072,64,0.0021749332547187803
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,3072,128,0.0067221333583196
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,3072,64,0.01446613371372223
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,3072,32,0.0021738665799299877
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,3072,32,0.014409599701563516
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,2560,65536,0.05389440059661865
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,2560,65536,0.04923839966456096
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,2560,16384,0.018154666821161906
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,2560,16384,0.02550613284111023
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,2560,12288,0.013502933581670127
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,2560,12288,0.022884267568588256
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,2560,12288,0.028693334261576338
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,2560,10240,0.01249066690603892
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,2560,10240,0.023450666666030885
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,2560,16384,0.03600000143051148
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,2560,8192,0.010855467120806376
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,2560,8192,0.022453333934148154
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,2560,10240,0.024835199117660522
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,2560,7168,0.010085333387056987
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,2560,7168,0.02329279979070028
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,2560,8192,0.021449599663416544
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,2560,6144,0.009219200412432352
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,2560,6144,0.020460800329844157
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,2560,7168,0.019578667481740315
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,2560,5120,0.008357333143552144
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,2560,5120,0.019885865847269694
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,2560,6144,0.017797333002090455
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,2560,4096,0.009611733754475911
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,2560,4096,0.016695467631022136
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,2560,5120,0.015782399972279867
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,2560,4096,0.013682132959365845
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,2560,3584,0.008453333377838134
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,2560,3584,0.016682666540145875
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,2560,3072,0.008041599889596303
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,2560,3072,0.01641706625620524
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,2560,3072,0.012082133690516155
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,2560,2560,0.005509333312511444
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,2560,2560,0.016641066471735636
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,2560,3584,0.012942933042844138
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,2560,2048,0.004584533472855886
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,2560,2048,0.014629333217938741
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,2560,2048,0.010136533776919048
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,2560,1536,0.0038794666528701783
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,2560,1536,0.015227733055750528
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,2560,2560,0.010824533303578694
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,2560,1024,0.0033973333736260734
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,2560,1024,0.01453013320763906
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,3072,65536,0.12660693327585856
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,2560,768,0.002993066608905792
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,2560,768,0.014551466703414917
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,2560,1024,0.008312533299128216
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,2560,512,0.0026346666117509207
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,2560,512,0.01446293294429779
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,2560,1536,0.009218133489290873
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,2560,256,0.0025685332715511323
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,2560,256,0.014562132954597472
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,2560,768,0.007893333335717519
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,2560,128,0.0022005334496498106
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,2560,128,0.012985600034395852
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,2560,512,0.007291733225186665
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,2560,64,0.002221866697072983
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,2560,256,0.0070720002055168155
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,2560,128,0.006718933085600535
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,2560,32,0.0022474666436513263
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,2560,64,0.014432000120480857
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,2560,32,0.013639466961224875
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,2048,65536,0.0440341313680013
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,2048,16384,0.015871999661127727
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,2048,65536,0.04375466505686442
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,2048,16384,0.022804266214370726
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,2048,12288,0.011550933122634888
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,2048,12288,0.025333333015441894
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,2048,12288,0.028189865748087566
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,2048,10240,0.010411733388900756
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,2048,10240,0.024796799818674723
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,2048,16384,0.03583253224690755
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,2048,8192,0.009272533655166625
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,2048,8192,0.02234986623128255
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,2048,10240,0.024918399254480996
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,2048,7168,0.008795733253161114
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,2048,7168,0.01909760038057963
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,2048,8192,0.021163733800252278
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,2048,6144,0.011270399888356526
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,2048,6144,0.0187882661819458
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,2048,7168,0.019554134209950766
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,2048,5120,0.00999679962793986
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,2048,5120,0.017966934045155845
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,2048,6144,0.01754666765530904
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,2048,4096,0.00876800020535787
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,2048,4096,0.01801813244819641
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,2048,5120,0.01581760048866272
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,2048,3584,0.0066997334361076355
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,2048,3584,0.016546133160591125
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,2048,4096,0.013745066523551942
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,2048,3072,0.006001066664854685
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,2048,3072,0.016697599490483602
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,2048,3584,0.012915199995040894
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,2048,2560,0.00506986677646637
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,2048,3072,0.012102400263150532
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,2048,2560,0.016769067446390788
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,2560,65536,0.12632213433583578
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,2048,2048,0.004638933142026265
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,2048,2048,0.014756266276041666
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,2048,1536,0.003973333289225897
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,2048,1536,0.014699733257293702
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,2048,2560,0.01122773289680481
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,2048,1024,0.0033952000240484873
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,2048,1024,0.014588800072669984
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,2048,1536,0.009181867043177288
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,2048,768,0.003014400104681651
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,2048,768,0.014479999740918478
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,2048,1024,0.008404266834259034
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,2048,512,0.0026357332865397137
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,2048,2048,0.010385066270828247
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,2048,512,0.01436906655629476
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,2048,256,0.0026208000878492994
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,2048,256,0.014413866400718688
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,2048,768,0.008004266520341237
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,2048,128,0.0022570667167504626
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,2048,256,0.006779733300209046
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,2048,512,0.007159466544787089
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,2048,128,0.014217600226402283
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,2048,64,0.0021888000269730886
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,2048,32,0.002128000060717265
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,2048,64,0.014446933070818582
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,2048,128,0.006793599824110668
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,2048,32,0.014225066701571146
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,1536,65536,0.03454933166503906
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,1536,16384,0.0181877334912618
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,1536,65536,0.03747520049413045
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,1536,16384,0.02701866626739502
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,1536,12288,0.014672000209490457
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,1536,12288,0.025703465938568114
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,1536,12288,0.028217599789301558
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,1536,10240,0.012962133685747782
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,1536,10240,0.02354773283004761
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,1536,16384,0.03554986715316773
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,1536,8192,0.011508267124493916
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,1536,8192,0.02067413330078125
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,1536,10240,0.024808533986409507
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,1536,7168,0.010643200079600016
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,1536,7168,0.019410133361816406
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,1536,8192,0.020810665686925252
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,1536,6144,0.009807999928792317
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,1536,6144,0.018729599316914876
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,1536,7168,0.019476266702016194
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,1536,5120,0.008843732873598735
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,1536,5120,0.01812373399734497
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,1536,6144,0.017436800400416057
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,1536,4096,0.007166933516661327
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,1536,4096,0.01833066741625468
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,1536,5120,0.015812266866366068
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,1536,3584,0.0063178668419520065
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,1536,3584,0.01688533425331116
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,1536,4096,0.013659733533859252
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,1536,3072,0.005779199798901876
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,1536,3072,0.017427200078964235
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,1536,3584,0.012869333227475485
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,1536,2560,0.005195733408133189
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,1536,2560,0.015795200069745382
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,1536,3072,0.011655466755231221
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,1536,2048,0.005757866799831391
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,1536,2048,0.014923733472824097
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,1536,2560,0.011262933413187664
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,1536,1536,0.003866666555404663
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,1536,1536,0.014628266294797262
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,1536,2048,0.010055466492970785
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,1536,1024,0.0034400001168251038
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,2048,65536,0.12557333310445148
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,1536,1024,0.014844800035158793
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,1536,768,0.003018666555484136
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,1536,768,0.014595199624697366
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,1536,768,0.007640533149242401
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,1536,1536,0.009112532933553059
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,1536,512,0.002917333443959554
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,1536,1024,0.007956266899903615
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,1536,256,0.002532266577084859
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,1536,512,0.014498133460680643
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,1536,256,0.014455466469128927
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,1536,128,0.002179199953873952
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,1536,128,0.013713066776593527
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,1536,256,0.007141333321730297
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,1536,512,0.007132799923419952
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,1536,64,0.00222080002228419
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,1536,128,0.006705066561698914
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,1536,32,0.002165333429972331
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,1536,64,0.014357333381970724
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,1536,32,0.014407466848691305
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,1024,65536,0.024755199750264488
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,1024,16384,0.01316373348236084
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,1024,65536,0.03204480012257894
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,1024,16384,0.02285760045051575
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,1024,12288,0.01092800001303355
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,1024,12288,0.019885865847269694
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,1024,12288,0.028100266059239702
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,1024,10240,0.009591466188430786
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,1024,10240,0.020049067338307698
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,1024,16384,0.035496532917022705
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,1024,8192,0.009453866879145305
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,1024,8192,0.01911466717720032
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,1024,10240,0.02486506700515747
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,1024,7168,0.007587199906508128
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,1024,7168,0.018680532773335777
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,1024,7168,0.019006933768590292
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,1024,6144,0.007068799932797749
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,1024,6144,0.018548266092936198
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,1024,8192,0.0211136003335317
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,1024,5120,0.006321066617965698
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,1024,5120,0.018403200308481853
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,1024,5120,0.015381333231925965
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,1024,4096,0.006049066781997681
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,1024,4096,0.018617600202560425
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,1024,6144,0.017384533087412515
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,1024,3584,0.006281599899133046
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,1024,3584,0.016807466745376587
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,1024,3584,0.012883200248082479
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,1024,3072,0.0055285334587097164
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,1024,3072,0.01726079980532328
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,1024,4096,0.01362559994061788
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,1024,2560,0.00508480022350947
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,1024,2560,0.016919465859731038
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,1024,2560,0.01092906693617503
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,1024,3072,0.011657599608103435
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,1024,2048,0.006276266773541768
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,1024,2048,0.016249600052833556
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,1024,1536,0.0051125332713127134
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,1024,1536,0.015004799763361613
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,1024,1536,0.008940800031026205
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,1024,1024,0.004232533276081085
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,1024,2048,0.010020266969998677
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,1024,1024,0.015592533349990844
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,1024,768,0.0034272000193595886
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,1024,768,0.014630400141080222
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,1536,65536,0.12521493434906006
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,1024,512,0.0029845332105954488
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,1024,512,0.01446293294429779
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,1024,1024,0.00798826664686203
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,1024,256,0.002629333237806956
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,1024,768,0.007537066439787547
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,1024,256,0.014475733041763306
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,1024,128,0.002313599983851115
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,1024,512,0.007146666447321574
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,1024,128,0.014551466703414917
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,1024,256,0.006717866659164429
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,1024,64,0.002201599876085917
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,1024,32,0.002234666546185811
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,1024,64,0.014539733529090881
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,1024,128,0.006646400193373363
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,1024,32,0.014439466595649719
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,768,65536,0.01994880040486654
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,768,16384,0.011592533191045125
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,768,65536,0.03498133420944214
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,768,16384,0.02139413356781006
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,768,12288,0.009666132926940917
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,768,12288,0.020218666394551596
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,768,12288,0.0284768005212148
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,768,10240,0.008391466736793519
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,768,10240,0.019008000691731773
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,768,16384,0.03505813280741374
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,768,8192,0.007143466671307881
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,768,8192,0.019746132691701255
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,768,10240,0.024785067637761435
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,768,7168,0.0067114666104316715
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,768,7168,0.020061866442362467
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,768,8192,0.02071573336919149
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,768,6144,0.006690133114655812
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,768,6144,0.01864853302637736
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,768,7168,0.01922453244527181
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,768,5120,0.0063509335120519
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,768,5120,0.01872106591860453
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,768,6144,0.017037866512934367
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,768,4096,0.006247466802597046
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,768,4096,0.016664533813794454
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,1024,65536,0.12156480153401691
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,768,3584,0.005925333499908448
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,768,3584,0.016571733355522155
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,768,5120,0.015692800283432007
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,768,3072,0.005509333312511444
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,768,3072,0.016586666305859886
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,768,4096,0.01325440009435018
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,768,2560,0.005480533341566721
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,768,2560,0.01530880033969879
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,768,3584,0.012382933497428894
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,768,2048,0.004273066421349844
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,768,2048,0.014507733782132468
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,768,3072,0.011613866686820984
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,768,1536,0.004496000210444133
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,768,1536,0.015178666512171427
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,768,2560,0.010780800382296245
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,768,1024,0.0038581334054470064
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,768,1024,0.014841600259145101
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,768,1536,0.008758399883906047
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,768,2048,0.009964799880981446
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,768,768,0.0034741332133611047
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,768,768,0.014470400412877402
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,768,512,0.002994133283694585
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,768,1024,0.007913599908351897
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,768,512,0.013848533233006796
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,768,768,0.007901866734027863
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,768,256,0.0025557334224383037
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,768,256,0.014638933539390563
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,768,256,0.00676800012588501
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,768,128,0.0022026665508747103
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,768,128,0.014121599992116294
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,768,128,0.0067445332805315655
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,768,64,0.002143999934196472
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,768,64,0.014478933811187745
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,768,32,0.002217599997917811
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,768,32,0.014456533392270408
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,512,65536,0.02849919994672139
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,512,65536,0.030115199089050294
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,512,16384,0.008428800106048583
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,512,16384,0.020307199160257975
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,512,16384,0.03476906617482503
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,512,12288,0.00757120003302892
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,512,12288,0.01993280053138733
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,512,12288,0.027743999163309736
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,512,10240,0.007099733253320058
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,512,10240,0.01874666611353556
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,768,65536,0.12152000268300374
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,512,8192,0.006712533533573151
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,512,8192,0.018549333016077675
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,512,10240,0.024375466505686443
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,512,7168,0.00552106648683548
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,512,7168,0.018807466824849448
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,512,8192,0.02068693240483602
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,512,6144,0.006285866598288219
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,512,6144,0.018973867098490395
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,512,7168,0.019030400117238364
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,512,5120,0.005861333509286245
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,512,5120,0.018550399939219156
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,512,6144,0.01742186745007833
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,512,4096,0.005883733431498209
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,512,4096,0.017875200510025023
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,512,5120,0.015313067038853965
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,512,3584,0.005447466671466827
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,512,3584,0.01668693423271179
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,512,4096,0.013612799843152366
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,512,3072,0.005385600030422211
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,512,3072,0.01764906644821167
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,512,3584,0.01242026686668396
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,512,2560,0.005048533280690512
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,512,2560,0.014487466216087342
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,512,3072,0.011641599734624227
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,512,2048,0.004683733483155568
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,512,2048,0.014890666802724203
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,512,2560,0.010819199681282043
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,512,1536,0.0037994667887687682
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,512,1536,0.014477866888046264
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,512,2048,0.009987200299898785
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,512,1024,0.003052799900372823
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,512,1024,0.014968533317248026
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,512,1536,0.009121066331863404
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,512,768,0.00296426663796107
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,512,768,0.01443839967250824
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,512,1024,0.008037333190441132
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,512,512,0.0026357332865397137
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,512,768,0.0075541332364082335
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,512,512,0.01455893317858378
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,512,256,0.0026069333155949908
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,512,512,0.007114666700363159
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,512,256,0.013915733496348063
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,512,128,0.0024864000578721365
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,512,256,0.007098666826883952
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,512,128,0.014472533265749613
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,512,65536,0.1212725321451823
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,512,64,0.002131200085083644
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,512,32,0.0021877333521842955
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,512,128,0.006690133114655812
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,512,64,0.014502400159835815
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,512,32,0.014424533645311991
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,256,65536,0.01655893325805664
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,256,16384,0.007029333213965099
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,256,65536,0.025172267357508344
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,256,16384,0.01881706714630127
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,256,12288,0.0054730668663978575
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,256,12288,0.019048533837000527
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,256,12288,0.028050132592519122
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,256,10240,0.0058442667126655575
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,256,10240,0.01690559983253479
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,256,16384,0.03540053367614746
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,256,8192,0.005490133166313171
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,256,8192,0.01885120073954264
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,256,10240,0.024753065903981526
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,256,7168,0.005430399874846141
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,256,8192,0.020983467499415077
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,256,7168,0.016502400239308677
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,256,6144,0.005151999990145365
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,256,6144,0.01849173307418823
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,256,7168,0.019326933224995933
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,256,5120,0.004999466737111409
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,256,6144,0.01734293301900228
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,256,5120,0.017990400393803917
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,256,4096,0.0050901333491007485
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,256,4096,0.01727786660194397
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,256,5120,0.015748266379038492
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,256,3584,0.005049600203831991
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,256,4096,0.013313066959381104
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,256,3584,0.01562346617380778
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,256,3584,0.01255466639995575
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,256,3072,0.004774400095144907
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,256,3072,0.017089066902796428
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,256,3072,0.011698133746782939
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,256,2560,0.005055999755859375
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,256,2560,0.014706133802731832
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,256,2560,0.010900266965230306
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,256,2048,0.004678399860858917
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,256,2048,0.01649173299471537
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,256,2048,0.010002133250236512
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,256,1536,0.003804799914360046
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,256,1536,0.014479999740918478
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,256,1024,0.0029898665845394133
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,256,1024,0.014492799838383993
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,256,1536,0.00920746624469757
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,256,768,0.003050666550795237
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,256,1024,0.008309333523114523
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,256,768,0.01469013293584188
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,256,512,0.002799999962250392
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,256,768,0.007899733384450276
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,256,512,0.014513066411018372
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,256,256,0.002254933367172877
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,256,512,0.007493333518505096
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,256,256,0.01453013320763906
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,256,128,0.002221866697072983
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,256,65536,0.12073280016581218
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,256,256,0.006659199794133504
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,256,128,0.014505599935849508
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,256,64,0.0021375998854637145
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,256,32,0.002149333308140437
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,256,64,0.013857066631317139
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,256,128,0.006727466483910878
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,256,32,0.013725866874059042
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,128,65536,0.009992532928784688
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,128,16384,0.005884799857934316
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,128,65536,0.022988800207773843
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,128,16384,0.01872533361117045
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,128,12288,0.005458133419354757
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,128,12288,0.018613332509994508
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,128,12288,0.028435200452804565
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,128,10240,0.005514666438102722
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,128,16384,0.03462506532669067
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,128,10240,0.01849386692047119
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,128,8192,0.005453866720199585
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,128,8192,0.017528533935546875
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,128,10240,0.02433919906616211
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,128,7168,0.005107200145721436
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,128,8192,0.02089386582374573
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,128,7168,0.017382399241129557
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,128,6144,0.005123200019200643
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,128,6144,0.01740480065345764
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,128,7168,0.019368533293406168
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,128,5120,0.004668800036112467
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,128,6144,0.01692906618118286
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,128,5120,0.018657066424687705
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,128,4096,0.004262400170167288
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,128,4096,0.0165994664033254
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,128,5120,0.015666133165359496
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,128,3584,0.004657066861788432
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,128,4096,0.013661866386731466
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,128,3584,0.01623040040334066
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,128,3072,0.004277333120505015
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,128,3072,0.017573332786560057
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,128,3584,0.012814933061599731
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,128,2560,0.004650666813055674
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,128,2560,0.014740266402562461
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,128,3072,0.011555199821790058
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,128,2048,0.004287999868392944
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,128,2048,0.014596266547838846
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,128,2560,0.010753066341082255
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,128,1536,0.003475199888149897
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,128,65536,0.12097386519114177
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,128,1536,0.016385066509246825
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,128,2048,0.009937066833178203
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,128,1024,0.003035733352104823
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,128,1024,0.014403200149536133
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,128,768,0.0026506667335828146
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,128,1536,0.00913706620534261
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,128,1024,0.008312533299128216
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,128,768,0.014966400464375815
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,128,512,0.0025898667673269907
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,128,512,0.014363732933998109
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,128,768,0.007541333138942718
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,128,256,0.0022154666483402254
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,128,512,0.007242666681607564
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,128,256,0.012584533294041952
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,128,128,0.0021216000119845075
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,128,256,0.00673173318306605
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,128,128,0.012513066331545511
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,128,64,0.002143999934196472
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,128,128,0.006694399813810985
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,128,32,0.0021087999145189924
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,128,64,0.01406613290309906
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,64,65536,0.009192533294359843
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,128,32,0.014387200276056925
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,64,16384,0.005875200033187866
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,64,65536,0.022789333264033
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,64,12288,0.005030400057633718
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,64,10240,0.005083733300367991
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,64,16384,0.018736000855763754
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,64,12288,0.017754666010538735
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,64,8192,0.005452799797058106
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,64,10240,0.018557866414388023
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,64,7168,0.005085866649945577
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,64,8192,0.0169706662495931
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,64,6144,0.005058133105436961
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,64,7168,0.018268799781799315
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,64,5120,0.004283733169237773
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,64,6144,0.016924800475438435
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,64,4096,0.004242133100827535
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,64,5120,0.01738133430480957
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,64,3584,0.004737066725889841
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,64,4096,0.016484266519546507
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,64,3072,0.004275199770927429
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,64,3072,0.016235733032226564
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,64,2560,0.0046079998215039575
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,64,2560,0.01590933303038279
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,64,2048,0.004283733169237773
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,64,2048,0.01546986699104309
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,64,1536,0.0034602666894594826
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,64,1536,0.014839466412862143
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,64,1024,0.002980266759792964
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,64,1024,0.014491732915242514
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,64,768,0.002586666742960612
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,64,768,0.013700266679128012
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,64,512,0.00222080002228419
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,64,512,0.01423893372217814
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,64,256,0.002169599880774816
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,64,256,0.01251413325468699
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,64,128,0.0021802666286627453
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,64,128,0.013889066378275552
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,64,64,0.0018634666999181111
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,64,64,0.013422933220863343
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,64,32,0.001803733284274737
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,64,32,0.014051199952761332
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,32,65536,0.0074986666440963745
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,32,65536,0.02186773419380188
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,32,16384,0.005039999882380167
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,32,16384,0.017063466707865398
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,32,12288,0.005267199873924255
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,32,10240,0.005064533154169718
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,32,12288,0.018212266763051353
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,32,8192,0.004233600199222564
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,32,10240,0.017808000246683754
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,32,7168,0.0046186665693918865
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,32,7168,0.016668800512949625
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,32,6144,0.0045952002207438145
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,32,6144,0.018256000677744546
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,32,5120,0.004223999877770742
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,32,5120,0.01934400002161662
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,32,4096,0.004218666752179464
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,32,4096,0.016689066092173258
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,32,3584,0.004604800045490265
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,32,3584,0.01824959913889567
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,32,3072,0.004318933188915253
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,32,3072,0.016288000345230102
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,32,2560,0.004600533346335093
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,32,2560,0.014879999558130899
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,32,2048,0.004250666499137879
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,32,2048,0.016966400543848674
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,32,1536,0.003500800083080927
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,32,1536,0.014528000354766845
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,32,1024,0.0030570665995279947
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,32,1024,0.014579199751218162
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,32,768,0.0026464000344276427
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,32,768,0.014497066537539164
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,32,512,0.002548266698916753
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,32,512,0.014455466469128927
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,32,256,0.002198400100072225
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,32,256,0.01443839967250824
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,32,128,0.002178133279085159
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,32,128,0.013773866494496665
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,32,64,0.001836799954374631
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,32,64,0.013804800311724343
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,16,32,32,0.0021066665649414062
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,32,32,0.013765333096186319
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,65536,16384,0.3292938550313314
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,65536,16384,0.17971946398417155
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,16,768,512,0.007495466868082683
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,65536,12288,0.23946026166280113
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,65536,12288,0.14257280031840008
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,64,3584,0.016551466782887776
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,65536,10240,0.20344746907552086
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,65536,10240,0.10699093341827393
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,16,32,8192,0.01793280045191447
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,65536,8192,0.14672746658325195
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,65536,8192,0.08771093686421713
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,65536,8192,0.06789226531982422
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,65536,7168,0.12872320016225178
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,65536,7168,0.08060800234476725
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,65536,10240,0.08006933530171713
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,65536,6144,0.11136000156402588
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,65536,6144,0.07000106970469157
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,65536,16384,0.12644800345102947
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,65536,5120,0.09373546441396077
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,65536,5120,0.0627455989519755
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,65536,12288,0.08998719851175943
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,65536,4096,0.0748693307240804
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,65536,4096,0.052298665046691895
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,65536,6144,0.05386666854222616
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,65536,3584,0.06612053314844767
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,65536,3584,0.04960319995880127
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,65536,4096,0.03827093442281087
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,65536,3072,0.057241598765055336
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,65536,3072,0.04376853307088216
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,65536,7168,0.06256320079167685
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,65536,2560,0.04853546619415283
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,65536,2560,0.03832319974899292
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,65536,5120,0.04394986629486084
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,65536,2048,0.03986666599909465
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,65536,2048,0.0347978671391805
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,65536,2560,0.02684906721115112
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,65536,1536,0.030636799335479737
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,65536,1536,0.028973867495854694
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,65536,3584,0.03461440006891887
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,65536,1024,0.02189013361930847
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,65536,1024,0.02481386661529541
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,65536,2048,0.02235520084698995
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,65536,768,0.01741973360379537
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,65536,768,0.02259946664174398
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,65536,3072,0.030293333530426025
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,65536,512,0.012454400459925335
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,65536,512,0.019698133071263633
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,65536,768,0.013576533397038779
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,65536,256,0.007419733206431071
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,65536,256,0.016642133394877114
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,65536,512,0.011121066411336263
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,65536,128,0.004689066608746847
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,65536,128,0.016110933820406594
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,65536,256,0.009348266323407491
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,65536,64,0.004275199770927429
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,65536,64,0.0164682666460673
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,65536,32,0.004297600189844767
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,65536,32,0.016458666324615477
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,16384,65536,0.32597440083821616
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,65536,128,0.008332799871762593
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,16384,65536,0.1827071984608968
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,16384,16384,0.09423573017120361
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,16384,16384,0.0618229349454244
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,65536,1024,0.016194132963816325
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,16384,12288,0.06713706652323405
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,16384,12288,0.043671464920043944
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,65536,1536,0.019859200716018675
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,16384,10240,0.05785599946975708
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,16384,10240,0.03920746644337972
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,16384,10240,0.03194239934285482
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,16384,8192,0.04809066851933797
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,16384,8192,0.03440426588058472
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,16384,12288,0.03419626553853353
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,16384,7168,0.03539520104726156
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,16384,7168,0.03292160034179688
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,16384,16384,0.046470399697621664
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,16384,6144,0.03127999901771546
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,16384,6144,0.030563199520111085
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,16384,7168,0.02201813260714213
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,16384,5120,0.026495999097824095
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,16384,5120,0.026985599597295122
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,16384,8192,0.02693973382314046
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,16384,4096,0.021677867571512858
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,16384,4096,0.026850134134292603
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,16384,6144,0.021486934026082358
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,16384,3584,0.01938026746114095
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,16384,3584,0.022766933838526408
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,16384,5120,0.018230400482813516
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,16384,3072,0.01688533425331116
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,16384,3072,0.022808533906936646
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,16384,4096,0.017348267634709678
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,16384,2560,0.0145087997118632
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,16384,2560,0.02102400064468384
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,16384,2560,0.01325759987036387
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,16384,2048,0.012051199873288471
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,16384,2048,0.019010132551193236
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,16384,3584,0.01580586632092794
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,16384,1536,0.00977066655953725
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,16384,1536,0.01920106609662374
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,16384,3072,0.01416853368282318
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,16384,1024,0.007282133400440216
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,16384,1024,0.015864533185958863
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,16384,1024,0.008851200342178345
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,16384,768,0.0062496001521746315
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,16384,768,0.014590932925542196
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,16384,1536,0.010043733318646749
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,16384,512,0.0034485332667827605
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,16384,512,0.01458346645037333
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,16384,512,0.007607466479142506
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,16384,256,0.0026858667532602947
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,16384,256,0.014513066411018372
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,16384,768,0.008372267087300617
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,16384,128,0.0025546667476495106
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,16384,128,0.014492799838383993
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,16384,128,0.007100800176461537
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,16384,64,0.002219733347495397
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,16384,64,0.014479999740918478
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,16384,32,0.0023210667073726653
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,16384,2048,0.01199893355369568
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,16384,32,0.014475733041763306
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,16384,256,0.007134933272997539
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,12288,65536,0.14429866472880046
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,12288,16384,0.0702122688293457
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,12288,16384,0.049651201566060385
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,12288,16384,0.03887146711349487
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,12288,12288,0.04859093427658081
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,12288,12288,0.04039146502812703
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,12288,12288,0.03092693289120992
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,12288,10240,0.04230719804763794
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,12288,10240,0.03829226493835449
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,12288,65536,0.234878937403361
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,12288,8192,0.03510293165842692
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,12288,8192,0.03423893451690674
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,12288,10240,0.026921600103378296
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,12288,7168,0.030997333923975627
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,12288,7168,0.030961066484451294
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,12288,7168,0.021132800976435342
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,12288,6144,0.02695786754290263
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,12288,6144,0.028955733776092528
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,12288,8192,0.022732800245285033
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,12288,5120,0.02365760008494059
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,12288,5120,0.026341332991917925
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,12288,6144,0.01913706660270691
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,12288,4096,0.019477333625157675
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,12288,4096,0.02256853381792704
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,12288,5120,0.017434666554133095
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,12288,3584,0.01762346625328064
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,12288,3584,0.020744532346725464
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,12288,3584,0.014115200440088908
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,12288,3072,0.01548906664053599
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,12288,3072,0.019619200627009073
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,12288,4096,0.015211733182271323
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,12288,2560,0.013733333349227906
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,12288,2560,0.02064639925956726
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,12288,2560,0.011666133006413778
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,12288,2048,0.009991466999053955
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,12288,2048,0.018595200777053834
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,12288,3072,0.012932266791661581
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,12288,1536,0.008036266764005024
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,12288,1536,0.016458666324615477
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,12288,1536,0.009609599908192951
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,12288,1024,0.006266666452089946
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,12288,1024,0.014802133043607077
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,12288,2048,0.010733866691589355
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,12288,768,0.004295466840267182
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,12288,768,0.014883200327555338
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,12288,1024,0.008380800485610962
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,12288,512,0.0031658666829268134
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,12288,512,0.013797332843144735
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,12288,768,0.008026666442553202
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,12288,256,0.0026229334374268847
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,12288,256,0.014507733782132468
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,12288,512,0.0075434664885203045
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,12288,128,0.0022069332500298817
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,12288,128,0.013298133015632629
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,12288,128,0.007115733126799266
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,12288,64,0.002182399978240331
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,12288,256,0.007154133419195812
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,12288,64,0.013484799861907959
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,12288,32,0.002221866697072983
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,12288,32,0.013484799861907959
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,10240,65536,0.1999285380045573
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,10240,65536,0.12416000366210937
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,10240,16384,0.06017706791559855
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,10240,16384,0.044369065761566163
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,16384,65536,0.1535744031270345
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,10240,12288,0.04144853353500366
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,10240,12288,0.037604268391927084
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,10240,16384,0.03789973258972168
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,10240,10240,0.03621546824773152
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,10240,10240,0.034492798646291095
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,12288,65536,0.13084906737009686
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,10240,8192,0.028884265820185346
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,10240,8192,0.02966826756795247
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,10240,12288,0.030114134152730305
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,10240,7168,0.026794666051864625
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,10240,7168,0.02893120050430298
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,10240,10240,0.026184533039728803
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,10240,6144,0.02395626703898112
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,10240,6144,0.02676266630490621
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,10240,8192,0.022379734118779502
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,10240,5120,0.020521599054336547
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,10240,5120,0.026650667190551758
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,10240,7168,0.02068480054537455
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,10240,4096,0.016522666811943053
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,10240,4096,0.02290560007095337
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,10240,6144,0.019024000565210978
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,10240,3584,0.015688533584276833
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,10240,3584,0.022013866901397706
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,10240,4096,0.015371732910474143
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,10240,3072,0.013783466815948487
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,10240,3072,0.02065920035044352
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,10240,5120,0.017343999942143758
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,10240,2560,0.01259519954522451
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,10240,2560,0.020248534282048543
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,10240,2560,0.01135040024916331
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,10240,2048,0.009607467055320739
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,10240,2048,0.018565332889556883
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,10240,3072,0.012748799721399941
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,10240,1536,0.00788266658782959
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,10240,1536,0.01653333306312561
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,10240,3584,0.013760000467300415
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,10240,1024,0.0058783998092015585
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,10240,1024,0.01469013293584188
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,10240,1536,0.009265066186587016
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,10240,768,0.005026133358478546
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,10240,768,0.014482133587201438
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,10240,1024,0.008798933029174805
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,10240,2048,0.010434133807818095
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,10240,512,0.003833599885304769
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,10240,512,0.014381866653760275
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,10240,256,0.0026378666361172995
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,10240,256,0.014579199751218162
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,10240,256,0.0070709332823753355
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,10240,128,0.002271999915440877
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,10240,128,0.01418880025545756
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,10240,128,0.006762666503588359
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,10240,64,0.002274133265018463
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,10240,64,0.014481066664059957
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,10240,32,0.0022101332743962605
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,10240,512,0.007482666770617168
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,10240,32,0.014392532904942832
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,8192,65536,0.15897493362426757
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,8192,65536,0.10013439655303955
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,8192,16384,0.04677120049794515
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,10240,768,0.007973333199818928
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,8192,16384,0.03750826517740886
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,8192,12288,0.03407680193583171
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,8192,12288,0.030407466491063434
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,8192,12288,0.02972266674041748
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,8192,10240,0.02959679961204529
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,8192,10240,0.028104533751805622
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,8192,16384,0.03750400145848592
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,8192,8192,0.021882667144139608
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,8192,8192,0.023962666591008507
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,8192,10240,0.026428800821304322
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,8192,7168,0.01957546671231588
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,8192,7168,0.025655466318130492
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,8192,8192,0.02209706703821818
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,8192,6144,0.01712426741917928
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,8192,6144,0.02328746716181437
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,8192,7168,0.020323199033737183
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,8192,5120,0.014929067095120749
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,8192,5120,0.023497599363327026
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,8192,6144,0.018710400660832724
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,8192,4096,0.01209173301855723
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,8192,4096,0.01878933310508728
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,8192,5120,0.017014400164286295
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,8192,3584,0.011029332876205444
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,8192,3584,0.01880106727282206
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,8192,4096,0.014525866508483887
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,8192,3072,0.00993280013402303
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,8192,3072,0.01904319922129313
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,8192,3584,0.013693867127100625
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,8192,2560,0.00879039963086446
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,8192,2560,0.016809600591659545
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,8192,3072,0.012077866991360982
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,8192,2048,0.007516799867153168
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,8192,2048,0.015901866555213928
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,8192,2560,0.01128000020980835
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,8192,1536,0.006343466540177663
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,8192,1536,0.016125866770744325
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,8192,1536,0.009611733754475911
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,8192,1024,0.0042357335488001505
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,8192,1024,0.014570666352907815
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,8192,2048,0.010431999961535137
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,8192,768,0.003345066557327906
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,8192,768,0.01444906691710154
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,8192,1024,0.008405333757400513
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,8192,512,0.0030048000315825146
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,8192,512,0.014469333489735923
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,8192,768,0.008001066744327545
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,8192,256,0.0025653332471847535
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,8192,256,0.01344000001748403
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,8192,256,0.007115733126799266
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,8192,128,0.002229333420594533
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,8192,128,0.014505599935849508
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,8192,512,0.007547733187675476
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,8192,64,0.0022005334496498106
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,8192,64,0.014447999993960061
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,8192,32,0.002209066599607468
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,8192,32,0.014486400286356607
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,8192,128,0.007091199855009715
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,7168,65536,0.13901012738545734
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,7168,65536,0.08946773211161295
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,7168,16384,0.03984853426615397
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,7168,16384,0.03531306584676107
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,10240,65536,0.1286837339401245
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,7168,12288,0.029423999786376952
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,7168,12288,0.02882133324940999
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,7168,16384,0.03711680173873901
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,7168,10240,0.025434666872024538
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,7168,10240,0.027427200476328534
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,7168,12288,0.029727999369303388
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,7168,8192,0.021283199389775596
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,7168,8192,0.02434773246447245
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,7168,10240,0.026040534178415935
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,7168,7168,0.019035732746124266
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,7168,7168,0.022658133506774904
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,7168,8192,0.022344533602396646
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,7168,6144,0.01575040022532145
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,7168,6144,0.022822399934132896
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,7168,7168,0.02035413384437561
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,7168,5120,0.01381119986375173
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,7168,5120,0.0206058661142985
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,7168,6144,0.018662399053573607
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,7168,4096,0.011169067025184632
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,7168,4096,0.020917334159215293
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,7168,5120,0.016597333550453185
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,7168,3584,0.01016533374786377
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,7168,3584,0.019233065843582153
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,8192,65536,0.12739306290944416
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,7168,3072,0.009195733070373534
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,7168,3584,0.012818132837613424
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,7168,3072,0.01848640044530233
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,7168,2560,0.008252800007661184
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,7168,2560,0.01692906618118286
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,7168,4096,0.01409386694431305
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,7168,2048,0.00709440012772878
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,7168,2048,0.014825600385665893
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,7168,2560,0.011250133315722149
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,7168,1536,0.005524266759554545
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,7168,1536,0.014517333110173544
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,7168,2048,0.010449066758155823
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,7168,1024,0.003583999971548716
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,7168,1024,0.014579199751218162
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,7168,3072,0.012100266416867574
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,7168,768,0.0030464000999927522
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,7168,768,0.01455573340257009
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,7168,1024,0.008433066805203756
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,7168,512,0.0027477333943049112
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,7168,512,0.014486400286356607
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,7168,768,0.007569066683451335
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,7168,256,0.0025631998976071674
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,7168,256,0.01418773333231608
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,7168,1536,0.009166933099428813
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,7168,128,0.002219733347495397
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,7168,128,0.013587199648221335
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,7168,512,0.007620266576608021
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,7168,64,0.0022709332406520844
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,7168,128,0.006684799989064534
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,7168,64,0.014446933070818582
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,7168,32,0.002182399978240331
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,7168,32,0.014426666498184203
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,6144,65536,0.12074453035990398
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,6144,65536,0.08132373491923015
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,6144,16384,0.03789226611455281
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,6144,16384,0.03277759949366252
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,6144,16384,0.036415998140970865
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,6144,12288,0.02757226626078288
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,6144,12288,0.029564799865086873
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,6144,12288,0.0297760009765625
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,6144,10240,0.022730666399002075
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,6144,10240,0.02694293260574341
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,7168,65536,0.12702720165252684
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,6144,8192,0.018999467293421425
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,6144,8192,0.022770132621129355
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,6144,10240,0.02568746606508891
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,6144,7168,0.01725333333015442
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,6144,7168,0.0223306675752004
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,6144,8192,0.021970132986704506
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,6144,6144,0.015335466464360556
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,6144,6144,0.021884800990422566
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,6144,7168,0.020272000630696615
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,6144,5120,0.013645866513252258
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,6144,5120,0.021138133605321248
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,6144,6144,0.01827733318010966
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,6144,4096,0.010281599561373393
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,6144,4096,0.01935466726620992
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,6144,5120,0.016203733285268147
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,6144,3584,0.009493333101272584
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,6144,3584,0.01862186590830485
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,6144,3584,0.01297599971294403
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,6144,3072,0.00848426620165507
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,6144,3072,0.01663253307342529
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,6144,4096,0.013758933544158936
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,6144,2560,0.007570133109887441
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,6144,2560,0.01681493322054545
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,6144,2560,0.010892800490061442
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,6144,2048,0.00652159998814265
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,6144,2048,0.014860799908638
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,6144,3072,0.012392533818880717
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,6144,1536,0.0047541335225105286
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,6144,1536,0.01516480048497518
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,6144,1536,0.009239466985066731
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,6144,1024,0.0034495999415715536
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,6144,1024,0.014588800072669984
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,6144,2048,0.010058666268984478
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,6144,768,0.0030602666238943735
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,6144,768,0.01463573376337687
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,6144,1024,0.008313600222269695
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,6144,768,0.007900799810886382
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,6144,512,0.0026133333643277483
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,6144,512,0.014619732896486918
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,6144,256,0.0026122666895389557
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,6144,256,0.014411733547846476
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,6144,256,0.0071370666225751235
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,6144,128,0.002243199944496155
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,6144,128,0.014467199643452963
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,6144,512,0.007538133362929027
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,6144,64,0.002196266750494639
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,6144,64,0.014510933558146158
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,6144,32,0.002131200085083644
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,6144,128,0.006664533416430156
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,6144,32,0.014460800091425577
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,5120,65536,0.10183786551157634
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,5120,65536,0.07235413392384847
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,5120,16384,0.03191146651903788
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,5120,16384,0.032306132713953654
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,5120,16384,0.03657066822052002
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,5120,12288,0.023883734146753946
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,5120,12288,0.02683839996655782
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,6144,65536,0.12673813501993816
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,5120,10240,0.020281600952148437
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,5120,10240,0.026553599039713542
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,7168,256,0.00710399995247523
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,5120,8192,0.017384533087412515
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,5120,8192,0.022331732511520385
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,5120,12288,0.029263999064763385
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,5120,7168,0.015619200468063355
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,5120,7168,0.022460800409317017
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,5120,10240,0.02576853235562642
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,5120,6144,0.013664000233014426
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,5120,6144,0.02100373307863871
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,5120,7168,0.020245333512624107
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,5120,5120,0.012072533369064331
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,5120,5120,0.02085439960161845
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,5120,6144,0.017870932817459106
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,5120,4096,0.009616000453631084
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,5120,4096,0.017899733781814576
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,5120,8192,0.02183893322944641
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,5120,3584,0.008758399883906047
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,5120,3584,0.01651093363761902
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,5120,3584,0.012498133381207784
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,5120,3072,0.007922133306662242
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,5120,3072,0.016597333550453185
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,5120,5120,0.01580586632092794
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,5120,2560,0.0072522665063540145
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,5120,4096,0.013857066631317139
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,5120,2560,0.015738667050997416
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,5120,2048,0.005549866457780202
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,5120,2048,0.014722133676211039
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,5120,2048,0.010086400310198466
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,5120,1536,0.004377600053946177
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,5120,1536,0.014897066354751586
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,5120,2560,0.01086293359597524
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,5120,1024,0.0034453332424163817
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,5120,1024,0.014520532886187234
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,5120,3072,0.01211840013662974
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,5120,768,0.002980266759792964
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,5120,768,0.014500266313552857
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,5120,1024,0.008348799745241801
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,5120,512,0.002644266684850057
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,5120,768,0.007523199915885926
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,5120,512,0.014468266566594442
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,5120,256,0.002569599946339925
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,5120,256,0.014484266440073649
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,5120,1536,0.009239466985066731
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,5120,128,0.002243199944496155
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,5120,128,0.014500266313552857
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,5120,128,0.006797866523265838
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,5120,256,0.006674133241176605
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,5120,64,0.0022101332743962605
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,5120,32,0.0022463999688625337
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,5120,64,0.013714133699735006
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,5120,32,0.014467199643452963
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,4096,65536,0.08081920146942138
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,5120,512,0.007506133119265239
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,4096,65536,0.059843198458353675
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,4096,16384,0.02510720094045003
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,4096,16384,0.02861119906107585
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,4096,12288,0.01945919990539551
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,4096,12288,0.02609386642773946
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,4096,12288,0.029252266883850096
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,4096,10240,0.01657919983069102
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,4096,10240,0.024331732590993246
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,4096,16384,0.03631360133488973
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,4096,8192,0.013698133826255798
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,4096,8192,0.02225066622098287
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,4096,10240,0.02563520073890686
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,4096,7168,0.012923733393351237
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,4096,7168,0.022846933205922446
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,4096,8192,0.02177706758181254
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,4096,6144,0.011645866433779399
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,4096,6144,0.022434133291244506
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,4096,7168,0.01983893314997355
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,4096,5120,0.010494933525721232
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,4096,5120,0.020691200097401937
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,4096,6144,0.01750826636950175
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,4096,4096,0.009244799613952637
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,4096,4096,0.01920213301976522
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,4096,5120,0.016260266304016113
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,4096,3584,0.010412800312042236
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,4096,3584,0.019103999932607016
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,4096,4096,0.01406826674938202
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,4096,3072,0.009403733412424724
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,4096,3072,0.01735360026359558
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,4096,3584,0.013182933131853739
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,4096,2560,0.008273066580295562
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,4096,2560,0.01747200091679891
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,5120,65536,0.12656319936116536
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,4096,2048,0.0068341334660847975
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,4096,2048,0.016683733463287352
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,4096,3072,0.012056533495585125
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,4096,2560,0.01086186667283376
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,4096,1536,0.005734399954477946
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,4096,1536,0.016634666919708253
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,4096,1024,0.004380799829959869
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,4096,1024,0.014724266529083253
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,4096,1024,0.008378666639328004
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,4096,2048,0.010058666268984478
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,4096,768,0.0038592000802357995
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,4096,768,0.014492799838383993
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,4096,512,0.0034282666941483817
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,4096,512,0.014441600441932679
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,4096,1536,0.00918826659520467
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,4096,256,0.0022730665902296705
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,4096,256,0.014452266693115234
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,4096,256,0.007051733136177063
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,4096,512,0.0074986666440963745
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,4096,128,0.002251733342806498
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,4096,128,0.013223466277122498
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,4096,64,0.002162133405605952
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,4096,768,0.007549866537253062
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,4096,64,0.013637333114941915
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,4096,32,0.002205866575241089
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,4096,32,0.013092266519864401
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,4096,128,0.006706133484840393
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,3584,65536,0.07149973710378012
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,3584,16384,0.02263039946556091
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,3584,65536,0.05443413257598877
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,3584,16384,0.027497599522272747
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,3584,12288,0.016956800222396852
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,3584,12288,0.02478826642036438
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,3584,12288,0.028899200757344562
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,3584,10240,0.014843733112017313
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,3584,10240,0.023319466908772787
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,3584,16384,0.03594133456548055
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,3584,8192,0.012774399916330972
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,3584,8192,0.021653334299723305
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,3584,10240,0.025195733706156416
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,3584,7168,0.011826133728027344
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,3584,7168,0.02232960065205892
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,3584,8192,0.021208532651265464
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,3584,6144,0.010843732953071594
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,3584,6144,0.020669867595036827
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,3584,7168,0.01942080060640971
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,3584,5120,0.010018133123715718
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,3584,5120,0.020618667205174766
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,3584,6144,0.017491199572881064
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,3584,4096,0.008742400010426839
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,3584,4096,0.017781333128611247
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,3584,5120,0.015753600001335143
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,3584,3584,0.01037440001964569
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,3584,3584,0.018081067005793254
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,3584,4096,0.013379200299580892
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,3584,3072,0.009172266721725464
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,3584,3072,0.016528000434239708
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,3584,3584,0.012594133615493774
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,3584,2560,0.008198399841785432
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,3584,2560,0.01667413314183553
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,3584,3072,0.011947733163833619
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,3584,2048,0.006697600086530049
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,3584,2048,0.015783466895421348
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,3584,2560,0.01087679962317149
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,3584,1536,0.005621333420276642
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,3584,1536,0.016157866517702738
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,4096,65536,0.1266698678334554
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,3584,1024,0.004229333500067393
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,3584,1024,0.016291200121243795
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,3584,2048,0.00999786655108134
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,3584,768,0.0038058665891488397
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,3584,768,0.014470400412877402
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,3584,1536,0.00920746624469757
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,3584,512,0.0038058665891488397
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,3584,512,0.014476799964904785
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,3584,1024,0.008344533046086629
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,3584,256,0.0022122666239738466
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,3584,512,0.007454933226108551
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,3584,256,0.0143477330605189
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,3584,768,0.007493333518505096
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,3584,128,0.0021568000316619872
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,3584,128,0.013550933202107748
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,3584,64,0.0021183999876181287
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,3584,64,0.01402453382809957
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,3584,32,0.002169599880774816
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,3584,128,0.006667733192443848
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,3584,32,0.013250133395195008
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,3584,256,0.007083733379840851
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,3072,65536,0.0620746652285258
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,3072,65536,0.05083733399709066
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,3072,16384,0.019367466370264687
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,3072,16384,0.024945066372553507
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,3072,12288,0.014862933754920959
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,3072,12288,0.022677334149678548
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,3072,12288,0.02850773334503174
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,3072,10240,0.013677866260210673
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,3072,10240,0.02288853327433268
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,3072,16384,0.036321067810058595
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,3072,8192,0.011638399958610535
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,3072,8192,0.02323626677195231
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,3072,10240,0.025076266129811602
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,3072,7168,0.010872532924016316
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,3072,7168,0.020786132415135702
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,3072,8192,0.021116799116134642
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,3072,6144,0.010086400310198466
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,3072,6144,0.020680532852808634
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,3072,7168,0.01908906698226929
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,3072,5120,0.009148800373077392
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,3072,5120,0.018942934274673463
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,3072,6144,0.017838933070500693
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,3072,4096,0.007986133297284443
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,3072,4096,0.01660266617933909
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,3072,4096,0.013741866747538248
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,3072,3584,0.009544533491134644
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,3072,3584,0.016667733589808144
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,3072,5120,0.015727999806404113
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,3072,3072,0.007970133423805236
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,3072,3072,0.016562133034070333
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,3584,65536,0.12695573170979818
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,3072,2560,0.007000533243020375
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,3072,2560,0.017614932854970296
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,3072,3072,0.011693867047627766
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,3072,2048,0.006289066871007283
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,3072,3584,0.012872533003489176
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,3072,2048,0.016874667008717856
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,3072,1536,0.005690666536490122
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,3072,1536,0.015970133741696677
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,3072,1536,0.009179733196894328
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,3072,1024,0.004603733122348785
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,3072,1024,0.015381333231925965
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,3072,2048,0.010052266716957092
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,3072,768,0.003818666686614355
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,3072,2560,0.011286399761835734
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,3072,768,0.014469333489735923
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,3072,512,0.00340693344672521
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,3072,512,0.014588800072669984
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,3072,1024,0.008354133367538452
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,3072,256,0.0025941332181294756
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,3072,256,0.01297706663608551
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,3072,512,0.007514666517575581
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,3072,128,0.002165333429972331
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,3072,128,0.01406719982624054
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,3072,768,0.008004266520341237
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,3072,64,0.002147199958562851
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,3072,64,0.014443733294804893
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,3072,32,0.0021461332837740582
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,3072,256,0.007080533107121785
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,3072,128,0.006675200164318084
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,2560,65536,0.05305920044581095
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,2560,65536,0.046572800477345785
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,2560,16384,0.01773653427759806
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,2560,16384,0.02296746571858724
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,2560,16384,0.03590293327967326
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,2560,12288,0.013332266608874002
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,2560,12288,0.0226474662621816
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,2560,12288,0.028331732749938963
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,2560,10240,0.012289067109425861
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,2560,10240,0.020806399981180827
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,3072,65536,0.12636693318684894
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,2560,8192,0.010680533448855082
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,2560,8192,0.02140586574872335
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,2560,10240,0.024807467063268026
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,2560,7168,0.009687466422716777
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,2560,7168,0.019782400131225585
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,2560,7168,0.019014400243759156
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,2560,6144,0.009177600344022114
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,2560,6144,0.018669867515563966
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,2560,8192,0.02144320011138916
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,2560,5120,0.008332799871762593
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,2560,5120,0.020177066326141357
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,2560,6144,0.017500799894332886
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,2560,4096,0.007119999825954437
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,2560,5120,0.01568000018596649
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,2560,4096,0.017668267091115318
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,2560,3584,0.006713599960009258
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,2560,3584,0.01653439998626709
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,2560,3584,0.012923733393351237
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,2560,3072,0.0076341331005096436
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,2560,3072,0.01658560037612915
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,2560,4096,0.013751467068990072
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,2560,2560,0.006735999882221222
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,2560,2560,0.016718933979670204
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,2560,3072,0.011773866415023804
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,2560,2048,0.006361599763234456
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,2560,2048,0.014906666676203408
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,2560,2560,0.010845866799354554
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,2560,1536,0.005852800110975901
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,2560,1536,0.014807466665903726
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,2560,1536,0.009212799866994222
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,2560,1024,0.004693333307902018
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,2560,1024,0.014628266294797262
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,2560,2048,0.009970133503278095
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,2560,768,0.0038176000118255613
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,2560,768,0.014504533012708029
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,2560,1024,0.008339200417200725
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,2560,512,0.0034954667091369627
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,2560,512,0.014444800217946372
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,2560,768,0.007885866860548655
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,2560,256,0.0027242665489514667
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,2560,256,0.01344106694062551
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,2560,256,0.006962133447329204
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,2560,128,0.002195200075705846
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,2560,512,0.007463466624418895
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,2560,128,0.01458346645037333
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,2560,128,0.006779733300209046
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,2560,64,0.0021898667017618816
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,2560,64,0.014493866761525472
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,2560,32,0.0021290667355060577
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,2560,32,0.014328533411026
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,2048,65536,0.0432426651318868
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,2048,65536,0.04096853335698446
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,2048,16384,0.015524267156918844
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,2048,16384,0.022706133127212525
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,2560,65536,0.12603306770324707
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,2048,12288,0.011223466197649638
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,2048,12288,0.020894932746887206
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,2048,16384,0.03579946756362915
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,2048,10240,0.010222933689753215
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,2048,10240,0.020653865734736123
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,2048,12288,0.028492800394694012
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,2048,8192,0.00962666670481364
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,2048,8192,0.01885333259900411
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,2048,10240,0.024732800324757893
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,2048,7168,0.008850133419036866
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,2048,7168,0.020082134008407592
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,2048,7168,0.01912320057551066
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,2048,6144,0.007946666578451793
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,2048,6144,0.017378133535385133
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,2048,8192,0.02077546715736389
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,2048,5120,0.006774400174617767
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,2048,5120,0.01660693287849426
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,2048,5120,0.01580586632092794
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,2048,6144,0.017372800906499227
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,2048,4096,0.00633493314186732
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,2048,4096,0.017375999689102174
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,2048,3584,0.005905066430568695
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,2048,3584,0.018278400103251137
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,2048,3584,0.012558933099110922
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,2048,3072,0.005905066430568695
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,2048,3072,0.017223467429478966
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,2048,4096,0.01340053379535675
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,2048,2560,0.005110399921735128
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,2048,2560,0.017518933614095053
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,2048,3072,0.01165013313293457
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,2048,2048,0.006233599781990051
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,2048,2048,0.016567466656366985
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,2048,2560,0.010844799876213073
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,2048,1536,0.005386666456858317
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,2048,1536,0.015277866522471109
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,2048,1536,0.009173333644866943
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,2048,1024,0.004216533402601878
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,2048,1024,0.014565333724021912
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,2048,2048,0.009990400075912476
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,2048,768,0.003542399903138479
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,2048,768,0.014536533753077188
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,2048,1024,0.008003200093905132
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,2048,512,0.0033962666988372804
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,2048,768,0.007526400188604991
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,2048,512,0.014453333616256715
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,2048,256,0.002625066787004471
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,2048,256,0.014478933811187745
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,2048,256,0.007150933146476746
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,2048,128,0.002233600119749705
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,2048,512,0.007777066528797149
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,2048,128,0.01328000028928121
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,2048,64,0.002165333429972331
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,2048,64,0.013822933038075766
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,2048,128,0.006682666639486949
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,2048,32,0.0022133332987626392
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,2048,32,0.013347199559211731
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,1536,65536,0.033921066919962564
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,1536,65536,0.03624213139216105
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,1536,16384,0.01202239990234375
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,1536,16384,0.024702932437260947
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,1536,16384,0.03543573220570882
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,1536,12288,0.014913066228230795
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,1536,12288,0.023331199089686075
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,2048,65536,0.12440000375111897
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,1536,10240,0.012958932916323343
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,1536,10240,0.019921066363652547
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,1536,12288,0.027770666281382243
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,1536,8192,0.011339733004570007
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,1536,8192,0.018609066804250084
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,1536,8192,0.02072640061378479
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,1536,7168,0.010663466652234395
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,1536,7168,0.019262933731079103
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,1536,10240,0.02435520092646281
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,1536,6144,0.009726933638254802
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,1536,6144,0.018860799074172974
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,1536,7168,0.01908906698226929
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,1536,5120,0.006306133170922597
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,1536,5120,0.01687893271446228
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,1536,6144,0.017002665996551515
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,1536,4096,0.006357333560784657
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,1536,4096,0.019924267133076986
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,1536,4096,0.013296000162760415
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,1536,3584,0.005937066674232483
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,1536,3584,0.016613333423932394
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,1536,5120,0.015727999806404113
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,1536,3072,0.005514666438102722
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,1536,3072,0.016523733735084534
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,1536,3584,0.012504532933235168
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,1536,3072,0.011892267068227132
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,1536,2560,0.005049600203831991
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,1536,2560,0.015068800250689188
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,1536,2560,0.010913067062695821
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,1536,2048,0.00537066658337911
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,1536,2048,0.01578986644744873
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,1536,1536,0.0048096001148223875
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,1536,1536,0.014672000209490457
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,1536,1536,0.009153067072232565
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,1536,1024,0.0038560000558694207
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,1536,1024,0.014596266547838846
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,1536,2048,0.009613866607348125
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,1536,768,0.003453866640726725
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,1536,768,0.01446293294429779
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,1536,1024,0.008012799918651581
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,1536,768,0.007477333148320516
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,1536,512,0.003044266750415166
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,1536,512,0.014060800274213156
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,1536,256,0.0026208000878492994
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,1536,256,0.014413866400718688
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,1536,512,0.007166933516661327
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,1536,256,0.00709440012772878
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,1536,128,0.002251733342806498
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,1536,128,0.013852799932161966
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,1536,64,0.0021589333812395734
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,1536,128,0.006715733309586842
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,1536,64,0.014461867014567056
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,1536,32,0.0022805333137512207
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,1536,32,0.013319466511408487
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,1024,65536,0.024286933739980063
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,1024,65536,0.02908586661020915
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,1024,16384,0.0130431999762853
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,1024,16384,0.020802134275436403
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,1024,16384,0.03505386511484782
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,1024,12288,0.010779733459154766
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,1024,12288,0.018961066007614137
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,1536,65536,0.1240554650624593
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,1024,10240,0.009593600034713745
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,1024,10240,0.018812799453735353
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,1024,10240,0.024482132991154988
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,1024,8192,0.008801066875457763
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,1024,12288,0.027611732482910156
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,1024,8192,0.018718934059143065
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,1024,7168,0.00772266685962677
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,1024,7168,0.01858133276303609
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,1024,7168,0.018961066007614137
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,1024,6144,0.006141866743564606
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,1024,6144,0.01734293301900228
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,1024,8192,0.021061333020528157
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,1024,5120,0.005885866781075796
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,1024,5120,0.016518400112787882
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,1024,6144,0.017382399241129557
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,1024,4096,0.005859200159708659
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,1024,5120,0.01576746702194214
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,1024,4096,0.018837332725524902
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,1024,3584,0.005513600011666616
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,1024,3584,0.017550933361053466
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,1024,3584,0.012524799505869547
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,1024,3072,0.005497600138187409
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,1024,3072,0.01648533344268799
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,1024,4096,0.013662933309872945
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,1024,2560,0.005038933455944061
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,1024,2560,0.016595199704170227
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,1024,3072,0.011685333649317424
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,1024,2560,0.010830932855606079
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,1024,2048,0.006236800054709116
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,1024,2048,0.014472533265749613
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,1024,1536,0.005072000126043955
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,1024,1536,0.014817066987355552
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,1024,1536,0.009143466750780743
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,1024,1024,0.003853866706291834
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,1024,2048,0.00992746651172638
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,1024,1024,0.014442666371663412
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,1024,768,0.0034986667335033415
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,1024,768,0.014553599556287131
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,1024,1024,0.007896533111731212
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,1024,768,0.007547733187675476
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,1024,512,0.00306986669699351
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,1024,512,0.014111999670664468
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,1024,256,0.0025642665723959604
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,1024,256,0.014428800344467163
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,1024,512,0.0071285332242647815
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,1024,128,0.00230826660990715
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,1024,256,0.006669866542021434
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,1024,128,0.014065066973368326
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,1024,64,0.002160000056028366
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,1024,64,0.013749333222707114
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,1024,128,0.006270933151245117
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,1024,32,0.0021589333812395734
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,1024,32,0.013749333222707114
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,768,65536,0.020577067136764528
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,768,65536,0.03334933519363403
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,768,16384,0.011848533153533935
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,768,16384,0.019118932882944743
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,768,16384,0.03546239932378133
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,768,12288,0.009904000163078307
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,768,12288,0.018879999717076622
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,1024,65536,0.12238186995188396
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,768,10240,0.0083146666487058
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,768,10240,0.018768000602722167
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,768,12288,0.028150399525960285
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,768,8192,0.007358933488527934
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,768,8192,0.019129600127538046
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,768,8192,0.02032853364944458
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,768,7168,0.006007466713587443
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,768,7168,0.01853440006573995
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,768,10240,0.024317866563796996
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,768,6144,0.005899733304977417
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,768,6144,0.018351999918619792
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,768,6144,0.017383466164271034
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,768,7168,0.019036799669265747
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,768,5120,0.005497600138187409
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,768,5120,0.018339200814565023
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,768,4096,0.006123733520507812
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,768,4096,0.016492799917856852
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,768,4096,0.013251200318336487
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,768,5120,0.015777066349983215
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,768,3584,0.005861333509286245
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,768,3584,0.016687999169031777
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,768,3072,0.005487999816735586
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,768,3072,0.01602133313814799
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,768,3584,0.012811733285586038
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,768,2560,0.0054506664474805195
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,768,3072,0.011603200435638427
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,768,2560,0.016936532656351724
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,768,2048,0.004612266520659129
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,768,2048,0.015421866377194723
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,768,2560,0.010831999778747558
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,768,1536,0.004286933441956838
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,768,2048,0.009595732887585957
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,768,1536,0.014502400159835815
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,768,1024,0.0036821333070596062
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,768,1024,0.01446613371372223
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,768,1536,0.008904533584912618
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,768,1024,0.00793280005455017
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,768,768,0.003385599950949351
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,768,768,0.014486400286356607
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,768,512,0.00297173336148262
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,768,512,0.013959466417630514
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,768,768,0.007949866851170858
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,768,256,0.0026229334374268847
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,768,256,0.014426666498184203
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,768,512,0.007506133119265239
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,768,128,0.002170666555563609
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,768,128,0.012854400277137756
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,768,256,0.007129600147406261
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,768,64,0.0021322667598724365
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,768,128,0.0064416001240412395
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,768,64,0.014498133460680643
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,768,32,0.01420906682809194
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,512,65536,0.027560534079869588
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,512,65536,0.02876586715380351
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,768,65536,0.1213759978612264
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,512,16384,0.008365866541862488
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,512,16384,0.01843199928601583
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,512,16384,0.03511039813359578
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,512,12288,0.0075093333919843035
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,512,12288,0.01857173244158427
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,512,12288,0.027677865823109944
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,512,10240,0.007142400244871776
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,512,10240,0.01897066632906596
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,512,10240,0.02434773246447245
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,512,8192,0.005629866818586985
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,512,8192,0.018634666999181114
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,512,8192,0.020677334070205687
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,512,7168,0.0054400001962979635
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,512,7168,0.018487467368443807
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,512,65536,0.12078293164571126
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,512,6144,0.0059114664793014525
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,512,6144,0.018551466862360637
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,512,7168,0.019035732746124266
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,512,5120,0.00544213354587555
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,512,5120,0.016935465733210246
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,512,5120,0.015313067038853965
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,512,4096,0.005841066439946493
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,512,6144,0.01691306630770365
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,512,4096,0.01795626680056254
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,512,3584,0.0054666668176651
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,512,3584,0.016540799538294473
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,512,3584,0.012457600235939026
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,512,4096,0.013317333658536276
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,512,3072,0.005422933399677277
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,512,3072,0.0164682666460673
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,512,2560,0.005041066805521647
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,512,2560,0.015547733505566916
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,512,3072,0.011606400211652119
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,512,2048,0.00469760000705719
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,512,2560,0.010798933108647664
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,512,2048,0.01478506624698639
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,512,1536,0.003798400113979975
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,512,1536,0.014433067043622336
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,512,2048,0.009611733754475911
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,512,1024,0.003324799984693527
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,512,1536,0.009129599730173747
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,512,1024,0.014551466703414917
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,512,768,0.0030378667016824085
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,512,768,0.014486400286356607
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,512,1024,0.007923200229803721
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,512,512,0.0026133333643277483
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,512,512,0.013673599561055502
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,512,512,0.007335466643174489
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,512,256,0.002595199892918269
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,512,256,0.013829333583513894
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,512,256,0.00713919997215271
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,512,128,0.0022143999735514323
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,512,128,0.014500266313552857
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,512,128,0.0064064001043637585
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,512,64,0.0022026665508747103
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,512,64,0.013723733027776084
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,512,32,0.0022069332500298817
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,512,32,0.013779200116793313
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,256,65536,0.016401066382726034
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,256,65536,0.024718934297561647
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,256,65536,0.12025279998779297
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,256,16384,0.006215466558933258
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,256,16384,0.018691200017929076
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,256,16384,0.03459626833597819
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,256,12288,0.005443199972311656
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,256,12288,0.017062399784723917
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,256,12288,0.02759893337885539
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,256,10240,0.005864533285299936
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,256,10240,0.018422400951385497
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,256,10240,0.02431466579437256
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,256,8192,0.005500799914201101
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,256,8192,0.018152532974878947
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,256,8192,0.020579200983047486
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,256,7168,0.005354666709899902
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,256,7168,0.017173333962758382
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,256,7168,0.018973867098490395
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,256,6144,0.00506986677646637
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,256,6144,0.016598400473594666
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,256,6144,0.016915200153986613
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,256,5120,0.005028266708056132
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,256,5120,0.016774400075276693
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,256,5120,0.015452800194422403
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,256,4096,0.005093333125114441
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,256,4096,0.016531200210253397
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,256,4096,0.013209600249926248
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,256,3584,0.005021866659323374
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,256,3584,0.01681173245112101
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,256,3584,0.012079999844233195
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,256,3072,0.004626133541266123
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,256,3072,0.01635840038458506
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,256,3072,0.011623467008272808
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,256,2560,0.005074133475621542
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,256,2560,0.014513066411018372
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,256,2560,0.010737066467603047
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,256,2048,0.004646400113900503
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,256,2048,0.015010133385658264
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,256,2048,0.009613866607348125
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,256,1536,0.0037952000896135964
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,256,1536,0.014471466342608133
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,256,1536,0.009129599730173747
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,256,1024,0.0030218665798505146
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,256,1024,0.014654933412869772
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,256,1024,0.007879466811815897
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,256,768,0.0029525332152843474
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,256,768,0.014481066664059957
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,256,768,0.007500799993673961
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,256,512,0.002603733291228612
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,256,512,0.014430933197339377
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,256,512,0.0070826664566993715
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,256,256,0.002201599876085917
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,256,256,0.01328426698843638
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,256,256,0.006713599960009258
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,256,128,0.0021674667795499166
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,256,128,0.013614933689435324
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,256,128,0.006660266717274983
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,256,64,0.002170666555563609
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,256,64,0.013264000415802002
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,256,32,0.002195200075705846
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,256,32,0.0137855996688207
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,128,65536,0.009608532985051472
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,128,65536,0.020525866746902467
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,128,65536,0.12061973412831624
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,128,16384,0.00553599993387858
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,128,16384,0.017670400937398276
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,3072,32,0.014365866780281067
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,128,12288,0.005132799843947092
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,128,12288,0.018219733238220216
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,128,16384,0.0349397341410319
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,128,10240,0.0054282665252685545
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,128,10240,0.016523733735084534
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,128,12288,0.027591466903686523
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,128,8192,0.0050784001747767125
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,128,10240,0.024308266242345174
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,128,8192,0.018547199169794717
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,128,7168,0.005092266698678335
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,128,7168,0.016642133394877114
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,128,8192,0.020259199539820354
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,128,6144,0.004996266464392344
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,128,6144,0.016562133034070333
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,128,7168,0.019382399320602418
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,128,5120,0.00462719996770223
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,128,5120,0.01664959987004598
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,128,6144,0.01694719990094503
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,128,4096,0.004198400179545084
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,128,5120,0.015733333428700765
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,128,4096,0.016966400543848674
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,128,3584,0.0046741331617037455
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,128,4096,0.01367573340733846
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,128,3584,0.01653333306312561
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,128,3584,0.012637866536776224
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,128,3072,0.0042303999265035
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,128,3072,0.016808533668518068
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,128,3072,0.011982933680216471
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,128,2560,0.004622933268547058
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,128,2560,0.015281066298484802
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,128,2560,0.01116373340288798
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,128,2048,0.004237866898377737
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,128,2048,0.014500266313552857
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,128,2048,0.010029866298039754
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,128,1536,0.003752533346414566
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,128,1536,0.014499200383822122
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,128,1536,0.009449600179990133
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,128,1024,0.0032085334261258446
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,128,1024,0.014506666858990987
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,128,768,0.002656000107526779
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,128,768,0.01458346645037333
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,128,1024,0.008105599880218506
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,128,512,0.0026133333643277483
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,128,768,0.007502933343251546
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,128,512,0.014521599809328715
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,128,256,0.0022815999885400133
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,128,512,0.0075093333919843035
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,128,256,0.013041067123413085
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,128,128,0.0022698665658632917
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,128,256,0.007100800176461537
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,128,128,0.013630933562914529
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,128,64,0.002147199958562851
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,128,128,0.006678399940331777
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,128,32,0.0017792000124851864
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,128,64,0.01378986636797587
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,128,32,0.014365866780281067
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,64,65536,0.007492266595363617
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,64,16384,0.005449600021044413
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,64,65536,0.021450666586558025
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,64,12288,0.005076266825199127
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,64,16384,0.018346667289733887
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,64,10240,0.0051242664456367494
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,64,12288,0.018599466482798258
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,64,8192,0.00552106648683548
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,64,10240,0.016616533199946083
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,64,7168,0.005015466610590617
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,64,7168,0.016682666540145875
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,64,6144,0.004669866462548574
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,64,6144,0.01654293338457743
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,64,5120,0.004649599889914194
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,64,5120,0.016539733608563742
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,64,4096,0.004603733122348785
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,64,4096,0.016471466422080992
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,64,3584,0.004263466596603394
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,64,3584,0.014739200472831726
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,64,3072,0.004199466605981191
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,64,3072,0.014629333217938741
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,64,2560,0.004196266829967499
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,64,2560,0.01469013293584188
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,64,2048,0.0038133333126703895
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,64,2048,0.014516266187032065
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,64,1536,0.0033791999022165934
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,64,1536,0.014808533589045205
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,64,1024,0.002962133288383484
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,64,1024,0.013540266950925191
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,64,768,0.0026101333399613695
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,64,768,0.013773866494496665
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,64,512,0.0021759999295075734
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,64,512,0.01381333371003469
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,64,256,0.002152533332506816
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,64,256,0.014457600315411887
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,64,128,0.0022079999248186748
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,64,128,0.013970133662223817
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,64,64,0.001781333362062772
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,64,64,0.013817600409189858
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,64,32,0.0017759999881188076
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,64,32,0.013324800133705138
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,32,65536,0.007141333321730297
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,32,65536,0.02034453352292379
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,32,16384,0.005430399874846141
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,32,16384,0.016545066237449647
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,32,12288,0.005446400245030721
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,32,12288,0.01657386620839437
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,32,10240,0.0050346667567888895
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,32,10240,0.01774079998334249
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,32,8192,0.004211199780305227
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,32,8192,0.016472533345222473
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,32,7168,0.00462719996770223
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,32,7168,0.017122133572896322
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,32,6144,0.004589866598447164
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,32,6144,0.016530133287111917
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,32,5120,0.0042453333735466
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,32,5120,0.01653866668542226
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,32,4096,0.003789866715669632
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,32,4096,0.017130666971206666
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,32,3584,0.00421973317861557
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,32,3584,0.016023466984430947
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,32,3072,0.0038389332592487337
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,32,3072,0.014551466703414917
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,32,2560,0.00421013335386912
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,32,2560,0.014566399653752646
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,32,2048,0.003857066730658213
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,32,2048,0.014774399995803832
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,32,1536,0.0035861333211263022
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,32,1536,0.014551466703414917
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,32,1024,0.003032533327738444
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,32,1024,0.014646400014559427
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,32,768,0.0027072000006834666
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,32,768,0.014494933684666953
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,32,512,0.0022154666483402254
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,32,512,0.01397333343823751
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,32,256,0.002277333289384842
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,32,256,0.012657066186269125
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,32,128,0.002243199944496155
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,32,128,0.012555733323097229
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,32,64,0.0018858666221300761
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,32,64,0.012565333644549051
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,32,32,0.0018879999717076618
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,32,32,0.012922666470209756
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,65536,16384,0.32255681355794275
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,65536,16384,0.17757867177327474
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,8,768,32,0.0021749332547187803
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,65536,12288,0.23652373949686684
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,65536,12288,0.1396992047627767
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,8,512,768,0.007529599964618683
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,65536,10240,0.19947412808736165
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,65536,10240,0.11949439843495686
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,8,64,8192,0.018603734175364175
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,65536,8192,0.16309439341227214
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,65536,8192,0.09448106288909912
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,65536,10240,0.07791146437327066
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,65536,7168,0.14326400756835939
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,65536,7168,0.08225706418355307
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,65536,8192,0.07181546688079835
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,65536,6144,0.12491199970245362
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,65536,6144,0.0746122678120931
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,65536,12288,0.10344746907552083
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,65536,5120,0.10559893449147542
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,65536,5120,0.064574933052063
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,65536,16384,0.11575360298156738
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,65536,4096,0.07988906701405843
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,65536,4096,0.05381333430608114
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,65536,6144,0.05503253142038981
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,65536,3584,0.07012373606363932
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,65536,3584,0.04944640000661214
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,65536,4096,0.03820586601893107
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,65536,3072,0.06053546667098999
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,65536,3072,0.044078934192657473
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,65536,7168,0.05485973358154297
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,65536,2560,0.050945067405700685
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,65536,2560,0.04071573416392009
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,65536,5120,0.043221334616343185
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,65536,2048,0.042156799634297686
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,65536,2048,0.03476373354593913
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,65536,3584,0.03620693286259969
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,65536,1536,0.032150399684906
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,65536,1536,0.030946133534113567
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,65536,2560,0.024820266167322795
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,65536,1024,0.021918932596842446
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,65536,1024,0.025172267357508344
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,65536,2048,0.023906133572260537
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,65536,768,0.017612799008687337
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,65536,768,0.02272319992383321
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,65536,3072,0.030742400884628297
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,65536,512,0.012607999642690024
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,65536,512,0.019268266359965005
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,65536,512,0.011222400267918905
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,65536,256,0.007684266567230225
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,65536,256,0.016520532965660095
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,65536,1024,0.015819733341534935
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,65536,128,0.0048885335524876915
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,65536,128,0.016548267006874083
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,65536,128,0.00836906631787618
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,65536,64,0.0042463997999827065
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,65536,64,0.016551466782887776
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,65536,32,0.004238933324813843
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,65536,32,0.016530133287111917
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,16384,65536,0.32358293533325194
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,65536,768,0.014146133263905843
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,16384,65536,0.17972373962402344
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,16384,16384,0.0914090633392334
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,16384,16384,0.061178668340047204
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,65536,1536,0.020089600483576456
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,16384,12288,0.06679999828338623
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,16384,12288,0.043532800674438474
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,65536,256,0.00962453285853068
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,16384,10240,0.05670933326085409
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,16384,10240,0.03932160139083862
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,16384,10240,0.03234453399976094
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,16384,8192,0.047278932730356854
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,16384,8192,0.033478399117787674
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,16384,12288,0.035179734230041504
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,16384,7168,0.03554666837056478
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,16384,7168,0.03317013382911682
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,16384,16384,0.04953066507975261
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,16384,6144,0.031062400341033934
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,16384,6144,0.02961066762606303
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,16384,8192,0.025707733631134034
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,16384,5120,0.026898133754730224
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,16384,5120,0.027618134021759035
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,16384,7168,0.022379734118779502
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,16384,4096,0.021756800015767415
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,16384,4096,0.025254400571187337
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,16384,5120,0.019156267245610557
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,16384,3584,0.019385600090026857
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,16384,3584,0.022670932610829673
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,16384,6144,0.02143893241882324
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,16384,3072,0.01692053278287252
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,16384,3072,0.02220053275426229
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,16384,4096,0.01622719963391622
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,16384,2560,0.014513066411018372
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,16384,2560,0.020937599738438926
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,16384,2560,0.012820266683896384
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,16384,2048,0.01211199959119161
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,16384,2048,0.02065920035044352
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,16384,3072,0.014899200201034546
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,16384,1536,0.01002239982287089
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,16384,1536,0.018604799111684164
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,16384,3584,0.015757866700490317
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,16384,1024,0.007490133245786031
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,16384,1024,0.014535466829935709
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,16384,1024,0.008805333574612936
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,16384,768,0.0058783998092015585
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,16384,768,0.014453333616256715
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,16384,1536,0.010389332969983418
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,16384,512,0.0035274667044480645
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,16384,512,0.014552533626556396
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,16384,768,0.008404266834259034
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,16384,256,0.002701866626739502
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,16384,256,0.014509866635004679
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,16384,512,0.00795413355032603
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,16384,128,0.0026015999416510267
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,16384,128,0.014541866381963095
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,16384,2048,0.01125333309173584
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,16384,64,0.0022282667458057405
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,16384,64,0.014403200149536133
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,16384,32,0.002293333411216736
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,16384,32,0.014424533645311991
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,16384,128,0.006872533261775971
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,12288,65536,0.23414719899495445
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,16384,256,0.00709440012772878
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,12288,65536,0.1417280038197835
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,12288,16384,0.04966400067011516
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,12288,16384,0.038747731844584146
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,12288,12288,0.05045440196990967
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,12288,12288,0.04199999968210856
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,12288,12288,0.030586665868759154
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,12288,10240,0.04326186577479045
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,12288,10240,0.03772906859715779
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,12288,10240,0.026778666178385417
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,12288,8192,0.03627306620279948
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,12288,8192,0.03263253370920817
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,12288,16384,0.06890880266825358
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,12288,7168,0.031640533606211343
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,12288,7168,0.02983573277791341
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,12288,7168,0.02109439969062805
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,12288,6144,0.027540266513824463
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,12288,6144,0.028548266490300494
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,12288,8192,0.022692267100016275
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,12288,5120,0.0242741326491038
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,12288,5120,0.027501867214838667
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,12288,6144,0.019051732619603474
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,12288,4096,0.0200000007947286
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,12288,4096,0.022312533855438233
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,12288,5120,0.017399466037750243
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,12288,3584,0.018156800667444864
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,12288,3584,0.020680532852808634
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,12288,3584,0.01423893372217814
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,12288,3072,0.01621333360671997
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,12288,3072,0.02065066695213318
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,12288,4096,0.014969600240389505
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,12288,2560,0.014030933380126953
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,12288,2560,0.018874667088190713
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,12288,3072,0.013061333696047464
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,12288,2048,0.009960533181826273
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,12288,2048,0.01856213410695394
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,12288,2560,0.011724799871444702
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,12288,1536,0.008235733211040496
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,12288,1536,0.016516266266504924
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,12288,1536,0.009679999947547913
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,12288,1024,0.006298666695753734
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,12288,1024,0.015719466408093772
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,12288,2048,0.010838400324185688
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,12288,768,0.004613333443800608
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,12288,768,0.014575999975204468
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,12288,1024,0.008796800176302593
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,12288,768,0.007962666451931
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,12288,512,0.003788800040880839
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,12288,512,0.014435199896494546
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,12288,256,0.0029919999341169994
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,12288,256,0.01432319978872935
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,12288,256,0.007228800157705943
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,12288,128,0.002571733295917511
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,12288,128,0.014501333236694336
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,12288,128,0.007180800040562947
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,12288,64,0.0022143999735514323
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,12288,64,0.014362667004267374
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,12288,32,0.002232533444960912
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,12288,512,0.007551999886830647
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,12288,32,0.01440000037352244
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,10240,65536,0.19484586715698243
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,10240,65536,0.12019626299540202
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,10240,16384,0.05810453494389852
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,10240,16384,0.04321493307749431
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,16384,65536,0.16322986284891766
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,10240,12288,0.04191679954528808
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,10240,12288,0.03745919863382975
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,10240,16384,0.03784426848093669
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,10240,10240,0.036456533273061115
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,10240,10240,0.0352021336555481
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,12288,65536,0.135262934366862
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,10240,8192,0.02977493405342102
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,10240,8192,0.029870933294296263
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,10240,12288,0.029281065861384077
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,10240,7168,0.026884265740712482
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,10240,7168,0.028572799762090047
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,10240,10240,0.026265599330266315
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,10240,6144,0.02412160038948059
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,10240,6144,0.02690560022989909
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,10240,8192,0.02232746680577596
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,10240,5120,0.02078933318456014
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,10240,5120,0.027345067262649535
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,10240,7168,0.02077440023422241
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,10240,4096,0.016852267583211265
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,10240,4096,0.020881066719690956
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,10240,6144,0.01898026665051778
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,10240,3584,0.016350932916005454
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,10240,3584,0.022257065773010253
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,10240,5120,0.0173418660958608
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,10240,3072,0.013967999815940857
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,10240,3072,0.020733867088953653
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,10240,4096,0.014939733346303306
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,10240,2560,0.012872533003489176
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,10240,2560,0.018645334243774413
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,10240,3584,0.013808000087738036
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,10240,2048,0.009698133667310078
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,10240,2048,0.01862506667772929
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,10240,2560,0.010963199536005656
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,10240,1536,0.007921066880226136
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,10240,1536,0.016567466656366985
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,10240,3072,0.012333866953849793
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,10240,1024,0.006679466863473256
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,10240,1024,0.016586666305859886
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,10240,2048,0.010474666953086853
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,10240,768,0.005037866532802582
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,10240,768,0.014587733149528503
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,10240,768,0.007938133180141449
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,10240,512,0.004312533140182495
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,10240,512,0.014485333363215128
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,10240,1024,0.008403199911117553
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,10240,256,0.002647466709216436
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,10240,256,0.014495999614397685
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,10240,256,0.007221333185831706
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,10240,128,0.0022624000906944276
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,10240,128,0.014476799964904785
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,10240,1536,0.009141332904497783
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,10240,512,0.007513600091139476
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,10240,64,0.0021738665799299877
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,10240,32,0.0021536000072956084
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,10240,64,0.014476799964904785
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,10240,32,0.013598933815956116
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,8192,65536,0.1579861323038737
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,8192,65536,0.09804159800211588
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,10240,128,0.006863999863465626
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,8192,16384,0.046494932969411214
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,8192,16384,0.036134401957194015
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,8192,12288,0.03592106501261393
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,8192,12288,0.03025706609090169
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,8192,12288,0.029443200429280596
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,8192,10240,0.030756266911824544
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,8192,10240,0.028344533840815228
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,8192,16384,0.03668053150177002
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,8192,8192,0.02224640051523844
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,8192,8192,0.0249781330426534
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,8192,10240,0.026475733518600462
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,8192,7168,0.01986133257548014
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,8192,7168,0.0243285338083903
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,8192,8192,0.021927465995152794
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,8192,6144,0.01739413340886434
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,8192,6144,0.022949333985646567
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,8192,7168,0.02104533314704895
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,8192,5120,0.015059199929237366
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,8192,5120,0.021121066808700562
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,8192,6144,0.018679465850194296
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,8192,4096,0.012378666798273723
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,8192,4096,0.0208512008190155
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,8192,5120,0.016999467213948568
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,8192,3584,0.011274666587511698
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,8192,3584,0.01899306575457255
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,8192,4096,0.014215466380119324
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,8192,3072,0.01014400025208791
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,8192,3072,0.019451733430226645
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,8192,3584,0.013664000233014426
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,8192,2560,0.008846933643023174
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,8192,2560,0.018651733795801796
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,8192,3072,0.012077866991360982
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,8192,2048,0.007490133245786031
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,8192,2048,0.014734933773676554
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,8192,2560,0.011283199985822041
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,8192,1536,0.0063391998410224915
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,8192,1536,0.015599999825159708
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,8192,2048,0.01016426682472229
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,8192,1536,0.00921493371327718
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,8192,1024,0.00425600012143453
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,8192,1024,0.014941866199175516
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,8192,768,0.0035264000296592714
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,8192,768,0.014567466576894126
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,8192,1024,0.008405333757400513
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,8192,512,0.003031466652949651
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,8192,512,0.014556800325711569
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,8192,768,0.00793280005455017
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,8192,256,0.002605866640806198
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,8192,256,0.014476799964904785
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,8192,256,0.007190399865309398
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,8192,512,0.007472000022729237
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,8192,128,0.0021717332303524016
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,8192,128,0.014493866761525472
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,8192,64,0.002268799891074498
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,8192,64,0.014025599757830302
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,8192,32,0.002194133400917053
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,8192,32,0.014467199643452963
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,8192,128,0.006677333513895671
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,7168,65536,0.13831574122111004
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,7168,65536,0.08728426297505697
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,7168,16384,0.04023573398590088
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,7168,16384,0.03515199820200603
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,10240,65536,0.12831786473592122
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,7168,12288,0.030923734108606975
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,7168,12288,0.029078400135040282
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,7168,16384,0.036665598551432296
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,7168,10240,0.026662399371465046
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,7168,10240,0.02681386669476827
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,7168,12288,0.029049599170684816
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,7168,8192,0.02226346731185913
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,7168,8192,0.024490666389465333
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,7168,10240,0.02643520037333171
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,7168,7168,0.01972800095876058
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,7168,7168,0.022380799055099487
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,7168,8192,0.022426666816075642
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,7168,6144,0.01567466656366984
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,7168,6144,0.02067306637763977
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,7168,7168,0.02023893396059672
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,7168,5120,0.013653332988421122
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,7168,5120,0.02062079906463623
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,7168,6144,0.018633600076039633
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,7168,4096,0.011224533120791119
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,7168,4096,0.019203199942906698
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,7168,5120,0.01655893325805664
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,7168,3584,0.010387200117111205
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,7168,3584,0.018542933464050292
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,7168,4096,0.014103466272354126
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,7168,3072,0.009161600470542907
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,7168,3072,0.018103466431299845
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,8192,65536,0.12741866906483967
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,7168,2560,0.008338133494059246
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,7168,2560,0.016609066724777223
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,7168,3584,0.0132832000652949
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,7168,2048,0.007085866729418437
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,7168,2048,0.014577066898345948
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,7168,2560,0.010883200168609618
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,7168,1536,0.0059445331494013464
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,7168,1536,0.014593066771825156
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,7168,3072,0.012132267157236736
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,7168,1024,0.0038389332592487337
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,7168,1024,0.014487466216087342
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,7168,2048,0.010024533669153849
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,7168,768,0.003387733300526937
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,7168,768,0.01446613371372223
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,7168,1536,0.009155199925104777
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,7168,512,0.003031466652949651
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,7168,512,0.014473600188891092
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,7168,1024,0.008354133367538452
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,7168,256,0.0026122666895389557
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,7168,256,0.013757866621017457
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,7168,768,0.007559466858704884
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,7168,128,0.0022656001150608064
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,7168,128,0.014062933127085366
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,7168,256,0.006774400174617767
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,7168,64,0.0022122666239738466
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,7168,64,0.013693867127100625
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,7168,32,0.0021418665846188864
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,7168,32,0.014114133516947427
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,7168,128,0.006696533163388569
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,7168,512,0.007563733557860057
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,6144,65536,0.08098666667938233
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,6144,65536,0.12015679677327473
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,6144,16384,0.03718719879786174
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,6144,16384,0.03238720099131266
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,6144,12288,0.02821333408355713
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,6144,12288,0.02986133297284444
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,6144,12288,0.02897599935531616
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,6144,10240,0.023367466529210408
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,6144,10240,0.02730773289998372
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,6144,16384,0.03650453488032023
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,6144,8192,0.019038933515548705
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,6144,8192,0.02267199953397115
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,6144,10240,0.026024534304936724
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,6144,7168,0.017540266116460167
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,6144,7168,0.022703999280929567
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,6144,8192,0.02222613294919332
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,6144,6144,0.01532906691233317
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,6144,6144,0.02060479919115702
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,6144,7168,0.020373332500457763
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,6144,5120,0.013659733533859252
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,6144,5120,0.020696532726287842
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,6144,6144,0.018297600746154784
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,6144,4096,0.010612266262372334
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,6144,4096,0.018986666202545167
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,6144,5120,0.01613866686820984
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,6144,3584,0.009498666723569233
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,6144,3584,0.016703999042510985
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,6144,4096,0.013745066523551942
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,6144,3072,0.008717866738637288
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,6144,3072,0.016531200210253397
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,6144,3584,0.012900267044703165
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,6144,2560,0.00763626645008723
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,6144,2560,0.01478506624698639
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,7168,65536,0.12725866635640462
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,6144,2048,0.006738133231798808
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,6144,2048,0.016411733627319337
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,6144,3072,0.01172160009543101
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,6144,1536,0.0049792001644770306
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,6144,1536,0.014492799838383993
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,6144,2560,0.010872532924016316
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,6144,1024,0.003841066608826319
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,6144,1024,0.014491732915242514
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,6144,1024,0.008010666569073994
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,6144,768,0.0034005333979924522
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,6144,768,0.014459733168284097
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,6144,1536,0.009224533041318258
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,6144,512,0.003013333429892858
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,6144,512,0.014114133516947427
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,6144,2048,0.01002239982287089
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,6144,256,0.002593066543340683
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,6144,256,0.01453013320763906
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,6144,256,0.006725333134333293
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,6144,128,0.0022485333184401194
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,6144,128,0.014478933811187745
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,6144,768,0.007916800181070964
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,6144,64,0.002179199953873952
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,6144,512,0.007494399944941203
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,6144,32,0.002181333303451538
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,6144,64,0.014158933361371358
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,6144,128,0.006742399930953979
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,5120,65536,0.10119679768880208
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,5120,65536,0.07195733388264974
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,5120,16384,0.032048000892003374
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,5120,16384,0.032305065790812174
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,5120,16384,0.03634133338928223
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,5120,12288,0.024503467480341594
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,5120,12288,0.02885013421376546
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,5120,12288,0.029339732726414998
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,5120,10240,0.020966400702794395
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,5120,10240,0.02583786646525065
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,5120,10240,0.026187733809153242
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,5120,8192,0.018129066626230875
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,5120,8192,0.021735467513402305
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,6144,65536,0.12671573162078859
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,5120,7168,0.01598186691602071
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,5120,7168,0.02073919971783956
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,5120,8192,0.021947733561197915
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,5120,6144,0.014396799604098
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,5120,6144,0.020639999707539877
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,5120,7168,0.019938133160273232
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,5120,5120,0.012505599856376648
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,5120,5120,0.019475199778874717
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,5120,6144,0.01834986607233683
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,5120,4096,0.009637332955996196
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,5120,4096,0.018167465925216675
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,5120,5120,0.016196266810099284
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,5120,3584,0.00886400043964386
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,5120,3584,0.018333866198857626
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,5120,3584,0.012964266538619994
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,5120,3072,0.00796693315108617
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,5120,3072,0.0165994664033254
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,5120,4096,0.013751467068990072
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,5120,2560,0.007086933155854543
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,5120,2560,0.01636373301347097
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,5120,2560,0.010898133118947346
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,5120,2048,0.00555626650651296
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,5120,2048,0.015877333283424378
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,5120,3072,0.012087466319402058
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,5120,1536,0.0046741331617037455
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,5120,1536,0.014625066518783569
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,5120,1536,0.009219200412432352
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,5120,1024,0.003435733417669932
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,5120,1024,0.014528000354766845
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,5120,2048,0.010063999891281128
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,5120,768,0.0030730667213598887
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,5120,768,0.014542933305104574
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,5120,1024,0.008393599589665731
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,5120,512,0.002682666728893916
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,5120,768,0.00758186678091685
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,5120,512,0.01440000037352244
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,5120,256,0.0025429333249727885
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,5120,256,0.014331733187039694
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,5120,256,0.007145600020885467
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,5120,128,0.0022261333962281544
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,5120,128,0.014492799838383993
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,5120,512,0.007496533294518788
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,5120,64,0.002136533210674922
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,5120,64,0.014446933070818582
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,5120,128,0.00655680000782013
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,5120,32,0.0022367998957633973
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,5120,32,0.01456000010172526
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,4096,65536,0.08385279973347982
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,4096,65536,0.06391253471374511
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,4096,16384,0.02611626585324605
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,4096,16384,0.028560000658035278
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,5120,65536,0.12675519784291583
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,4096,12288,0.02018986741701762
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,4096,12288,0.02480640014012655
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,4096,16384,0.036355201403299967
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,4096,10240,0.016746666034062704
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,4096,10240,0.02386773427327474
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,4096,10240,0.0256223996480306
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,4096,8192,0.014190933108329773
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,4096,8192,0.02359786629676819
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,4096,12288,0.02892373402913411
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,4096,7168,0.013095466295878091
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,4096,7168,0.02291413346926371
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,4096,7168,0.019501866896947224
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,4096,6144,0.012045866250991822
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,4096,6144,0.021205333868662517
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,4096,8192,0.021549866596857706
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,4096,5120,0.0107914666334788
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,4096,5120,0.020568533738454183
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,4096,5120,0.015742933750152587
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,4096,4096,0.009559466441472372
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,4096,4096,0.01853653391202291
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,4096,6144,0.017867734034856163
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,4096,3584,0.010471466183662414
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,4096,3584,0.01739733417828878
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,4096,3584,0.012534399827321371
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,4096,3072,0.009590400258700053
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,4096,3072,0.016716800133387246
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,4096,4096,0.013725866874059042
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,4096,2560,0.008400000135103862
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,4096,2560,0.017044266064961754
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,4096,2560,0.01123520036538442
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,4096,2048,0.008036266764005024
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,4096,2048,0.016621866822242738
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,4096,3072,0.011642666657765706
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,4096,1536,0.005504000186920166
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,4096,1536,0.015710933009783427
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,4096,1536,0.0091839998960495
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,4096,1024,0.004403199752171834
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,4096,1024,0.01448319951693217
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,6144,32,0.012882133324941
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,4096,768,0.003787733366092046
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,4096,768,0.014472533265749613
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,4096,2048,0.01011306643486023
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,4096,512,0.0034677334129810332
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,4096,512,0.014425599575042724
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,4096,768,0.007674666742483775
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,4096,256,0.0025920001169045764
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,4096,256,0.0132341335217158
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,4096,1024,0.007938133180141449
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,4096,128,0.0021461332837740582
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,4096,512,0.007514666517575581
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,4096,128,0.014152533809343972
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,4096,64,0.0021354667842388153
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,4096,64,0.01325440009435018
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,4096,256,0.007088000078996022
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,4096,32,0.013806933164596557
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,4096,128,0.006720000008742015
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,3584,65536,0.07463040351867675
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,3584,65536,0.055130668481191004
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,3584,16384,0.023141332467397056
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,3584,16384,0.02654293378194173
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,3584,16384,0.03634026845296224
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,3584,12288,0.01736533244450887
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,3584,12288,0.024197334051132204
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,4096,65536,0.12679786682128907
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,3584,10240,0.014958932995796204
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,3584,10240,0.0228383998076121
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,3584,12288,0.029178667068481445
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,3584,8192,0.012981333335240684
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,3584,8192,0.02211093306541443
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,3584,8192,0.02121493419011434
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,3584,7168,0.012083199620246888
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,3584,7168,0.022721066077550253
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,3584,10240,0.02545493245124817
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,3584,6144,0.010806399583816528
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,3584,6144,0.020682666699091592
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,3584,6144,0.01771199901898702
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,3584,5120,0.009760000308354696
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,3584,5120,0.01880426605542501
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,3584,7168,0.01953279972076416
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,3584,4096,0.00890239973862966
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,3584,4096,0.018248534202575682
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,3584,5120,0.01574613352616628
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,3584,3584,0.010520533720652262
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,3584,3584,0.016835200786590575
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,3584,4096,0.013805866241455078
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,3584,3072,0.009266133109728496
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,3584,3072,0.016721065839131674
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,3584,3072,0.011733333269755047
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,3584,2560,0.00832533339659373
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,3584,2560,0.016645333170890807
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,3584,3584,0.012923733393351237
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,3584,2048,0.007473066449165344
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,3584,2048,0.016662399967511496
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,3584,2560,0.01115733285744985
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,3584,1536,0.00553173323472341
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,3584,1536,0.016525866587956746
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,3584,2048,0.01011306643486023
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,3584,1024,0.004513066510359446
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,3584,1024,0.016938666502634682
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,3584,1536,0.009113599856694538
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,3584,768,0.003849600007136663
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,3584,768,0.01460693379243215
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,3584,1024,0.008281599978605907
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,3584,768,0.007901866734027863
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,3584,512,0.005087999999523163
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,3584,512,0.01451520025730133
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,3584,256,0.0025621332228183747
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,3584,256,0.014409599701563516
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,3584,256,0.0070592001080513
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,3584,128,0.0021930667261282604
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,3584,128,0.01416106621424357
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,3584,512,0.007123200098673503
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,3584,64,0.0022282667458057405
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,3584,64,0.014339199662208557
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,3584,32,0.002196266750494639
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,3584,128,0.006720000008742015
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,3584,32,0.01446399986743927
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,3072,65536,0.06421653429667154
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,3072,65536,0.05110613505045573
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,3072,16384,0.019410133361816406
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,3072,16384,0.02487466732660929
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,3072,16384,0.036023465792338054
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,3072,12288,0.015482667088508605
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,3072,12288,0.0228000005086263
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,3584,65536,0.12623146375020344
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,3072,10240,0.013506133357683817
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,3072,10240,0.022290132443110146
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,3072,12288,0.02892586588859558
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,3072,8192,0.012106666962305706
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,3072,8192,0.022073600689570108
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,3072,10240,0.02531733314196269
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,3072,7168,0.010912000139554342
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,3072,7168,0.021619200706481934
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,3072,8192,0.021126399437586464
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,3072,6144,0.010292266805966694
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,3072,6144,0.02037013371785482
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,3072,7168,0.019154133399327596
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,3072,5120,0.009202133615811665
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,3072,5120,0.018665599822998046
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,3072,5120,0.015829333662986757
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,3072,4096,0.007971199850241344
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,3072,4096,0.016635732849438985
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,3072,6144,0.017837866147359212
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,3072,3584,0.009236266215642292
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,3072,3584,0.0179584006468455
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,3072,3584,0.01286720037460327
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,3072,3072,0.008507733543713886
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,3072,3072,0.018232532342274985
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,3072,4096,0.013758933544158936
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,3072,2560,0.006806399921576183
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,3072,2560,0.01654293338457743
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,3072,2560,0.01093226671218872
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,3072,2048,0.006679466863473256
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,3072,2048,0.01652906636397044
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,3072,3072,0.012005333105723064
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,3072,1536,0.0055061335364977515
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,3072,1536,0.016155733664830526
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,3072,1536,0.009197866916656494
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,3072,1024,0.004282666742801667
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,3072,1024,0.015434666474660238
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,3072,2048,0.010021332899729412
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,3072,768,0.0038250667353471124
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,3072,768,0.014503467082977294
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,3072,1024,0.007991466422875721
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,3072,512,0.003454933315515518
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,3072,768,0.007484800120194752
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,3072,512,0.014626133441925048
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,3072,256,0.0029706666866938275
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,3072,256,0.014447999993960061
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,3072,256,0.0070709332823753355
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,3072,128,0.0021407999098300933
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,3072,512,0.0075328002373377485
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,3072,128,0.014436266819636025
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,3072,64,0.0022154666483402254
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,3072,64,0.014418133099873862
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,3072,32,0.0022154666483402254
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,3072,32,0.014485333363215128
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,3072,128,0.006659199794133504
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,2560,65536,0.056124798456827794
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,2560,65536,0.04606186548868815
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,2560,16384,0.018143999576568603
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,2560,16384,0.02415999968846639
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,2560,16384,0.035869868596394856
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,2560,12288,0.013703466455141703
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,2560,12288,0.022629332542419434
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,3072,65536,0.1268725315729777
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,2560,10240,0.012475732962290447
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,2560,10240,0.020692267020543418
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,2560,12288,0.028462932507197065
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,2560,8192,0.01071573297182719
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,2560,8192,0.01909439961115519
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,2560,10240,0.02480106751124064
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,2560,7168,0.010263466835021972
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,2560,7168,0.021026132504145305
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,2560,8192,0.020719999074935914
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,2560,6144,0.009172266721725464
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,2560,6144,0.018604799111684164
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,2560,7168,0.01907093326250712
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,2560,6144,0.017017600933710735
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,2560,5120,0.008378666639328004
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,2560,5120,0.01791999936103821
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,2560,4096,0.00718506673971812
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,2560,4096,0.016682666540145875
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,2560,4096,0.01374399960041046
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,2560,3584,0.006323199967543285
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,2560,3584,0.016522666811943053
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,2560,5120,0.015794133146603904
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,2560,3072,0.007526400188604991
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,2560,3072,0.016714666287104288
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,2560,3584,0.012874666849772134
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,2560,2560,0.007528533538182576
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,2560,3072,0.01165013313293457
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,2560,2560,0.010918399691581726
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,2560,2048,0.006685866912206013
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,2560,2048,0.016722132762273155
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,2560,2048,0.01009386678536733
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,2560,1536,0.006498133142789205
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,2560,1536,0.016489600141843162
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,2560,1536,0.009238400061925252
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,2560,1024,0.005083733300367991
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,2560,1024,0.015017599860827128
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,2560,1024,0.0083146666487058
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,2560,768,0.004314666489760081
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,2560,768,0.014577066898345948
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,2560,768,0.007948799928029378
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,2560,512,0.0035157332817713416
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,2560,512,0.01451520025730133
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,2560,512,0.007573333382606506
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,2560,256,0.0029909332593282064
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,2560,256,0.014467199643452963
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,2560,256,0.0071050668756167095
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,2560,128,0.002217599997917811
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,2560,128,0.014503467082977294
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,2560,128,0.0067114666104316715
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,2560,64,0.002218666672706604
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,2560,64,0.014493866761525472
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,2560,32,0.002194133400917053
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,2560,32,0.01447466711203257
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,2048,65536,0.04640320142110189
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,2048,65536,0.041606398423512776
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,4096,32,0.0021930667261282604
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,2048,16384,0.015260799725850423
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,2048,16384,0.022668800751368203
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,2560,65536,0.12597013314565023
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,2048,12288,0.01169599990049998
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,2048,12288,0.020753065745035805
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,2048,16384,0.03552533388137817
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,2048,10240,0.010364799698193868
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,2048,10240,0.018732800086339315
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,2048,12288,0.028104533751805622
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,2048,8192,0.009703466296195984
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,2048,8192,0.01917440096537272
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,2048,10240,0.02445440093676249
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,2048,7168,0.00879039963086446
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,2048,7168,0.018629332383473717
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,2048,8192,0.0211082657178243
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,2048,6144,0.008066133161385854
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,2048,6144,0.0187285323937734
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,2048,6144,0.017518933614095053
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,2048,5120,0.006807466844717662
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,2048,5120,0.01716053287188212
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,2048,7168,0.01911999980608622
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,2048,4096,0.006449066599210103
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,2048,4096,0.017143466075261436
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,2048,4096,0.01378986636797587
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,2048,3584,0.005899733304977417
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,2048,3584,0.01759786605834961
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,2048,5120,0.01581333378950755
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,2048,3072,0.005890133480230967
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,2048,3072,0.017525333166122436
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,2048,3072,0.012122666835784912
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,2048,2560,0.005486933390299479
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,2048,3584,0.012843733032544455
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,2048,2560,0.016752000649770102
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,2048,2048,0.006867200136184692
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,2048,2048,0.01618880033493042
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,2048,2048,0.010014933347702027
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,2048,1536,0.005122133096059163
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,2048,1536,0.01511360009511312
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,2048,2560,0.010844799876213073
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,2048,1024,0.004204800228277842
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,2048,1024,0.014518400033315023
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,2048,1024,0.008340266346931458
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,2048,768,0.0037994667887687682
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,2048,768,0.014447999993960061
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,2048,1536,0.009096533060073853
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,2048,512,0.003479466587305069
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,2048,512,0.01458560029665629
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,2048,768,0.0075552001595497135
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,2048,512,0.007493333518505096
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,2048,256,0.002631466587384542
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,2048,256,0.014479999740918478
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,2048,256,0.007066666583220164
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,2048,128,0.002570666621128718
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,2048,128,0.013299199938774108
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,2048,128,0.006795733173688252
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,2048,64,0.0021514666577180227
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,2048,64,0.013722667098045349
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,2048,32,0.0022101332743962605
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,2048,32,0.014433067043622336
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,1536,65536,0.03544106483459473
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,1536,65536,0.03405119975407918
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,1536,16384,0.012220799922943115
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,1536,16384,0.02473599910736084
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,1536,16384,0.03551359971364339
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,1536,12288,0.015754666924476624
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,1536,12288,0.022733867168426514
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,2048,65536,0.12512319882710773
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,1536,10240,0.012852266430854797
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,1536,10240,0.0186954657236735
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,1536,12288,0.0281877338886261
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,1536,8192,0.011285332838694255
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,1536,8192,0.01906879941622416
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,1536,10240,0.02444373369216919
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,1536,7168,0.01036906639734904
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,1536,7168,0.01917653282483419
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,1536,7168,0.019169066349665323
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,1536,6144,0.009551999966303508
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,1536,6144,0.017854932943979898
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,1536,8192,0.020687999327977498
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,1536,5120,0.00631466656923294
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,1536,5120,0.017824000120162962
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,1536,5120,0.01541973352432251
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,1536,4096,0.006299733122189839
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,1536,4096,0.016687999169031777
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,1536,6144,0.0174453337987264
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,1536,3584,0.005926399926344554
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,1536,3584,0.016634666919708253
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,1536,3584,0.012918399771054587
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,1536,3072,0.005502933263778686
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,1536,3072,0.016539733608563742
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,1536,4096,0.013591466347376504
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,1536,2560,0.0051125332713127134
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,1536,2560,0.017307732502619425
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,1536,2560,0.010872532924016316
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,1536,2048,0.005421866476535797
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,1536,3072,0.012053333719571431
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,1536,2048,0.015085867047309876
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,1536,1536,0.004680533210436503
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,1536,1536,0.014569600423177084
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,1536,1536,0.0091839998960495
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,1536,1024,0.004385066529115042
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,1536,1024,0.014645333091417948
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,1536,2048,0.00993280013402303
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,1536,768,0.0035829332967599234
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,1536,768,0.01453013320763906
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,1536,1024,0.008362666765848795
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,1536,512,0.002962133288383484
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,1536,768,0.007877333462238312
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,1536,512,0.014503467082977294
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,1536,256,0.0026389333109060925
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,1536,256,0.014494933684666953
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,1536,256,0.0070602665344874065
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,1536,512,0.007108266651630402
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,1536,128,0.0026122666895389557
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,1536,128,0.014375467101732889
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,1536,64,0.0022463999688625337
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,1536,64,0.01448853313922882
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,1536,128,0.0064533332983652755
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,1536,32,0.002160000056028366
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,1536,32,0.014410666624704995
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,1024,65536,0.025357866287231447
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,1024,65536,0.029198932647705077
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,1024,16384,0.01302079955736796
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,1024,16384,0.020985599358876547
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,1024,16384,0.03522666692733765
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,1024,12288,0.011018666625022887
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,1024,12288,0.019426133235295615
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,1536,65536,0.12293226718902588
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,1024,10240,0.009325866897900898
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,1024,10240,0.018039466937383015
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,1024,12288,0.027999999125798543
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,1024,8192,0.00865600009759267
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,1024,8192,0.018592000007629395
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,1024,10240,0.024345600605010988
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,1024,7168,0.007665066421031952
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,1024,7168,0.01691733400026957
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,1024,8192,0.02067199945449829
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,1024,6144,0.006277333199977875
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,1024,6144,0.01850559910138448
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,1024,7168,0.018678400913874307
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,1024,5120,0.005545599758625031
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,1024,5120,0.017684266964594523
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,1024,6144,0.0173418660958608
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,1024,4096,0.005885866781075796
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,1024,4096,0.017052799463272095
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,1024,5120,0.015450666348139444
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,1024,3584,0.0054517333706219995
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,1024,4096,0.013683199882507324
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,1024,3584,0.01664959987004598
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,1024,3072,0.005456000069777171
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,1024,3072,0.016530133287111917
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,1024,3584,0.012865066528320312
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,1024,2560,0.005226666728655497
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,1024,2560,0.01665066679318746
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,1024,3072,0.011611732840538024
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,1024,2048,0.00624533345301946
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,1024,2048,0.015737600127855935
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,1024,2560,0.010830932855606079
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,1024,1536,0.005063466727733612
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,1024,1536,0.014589866995811463
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,1024,2048,0.010028800368309021
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,1024,1024,0.0046858668327331545
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,1024,1024,0.015969066818555196
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,1024,1536,0.009134933352470398
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,1024,768,0.0034645333886146545
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,1024,768,0.014472533265749613
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,1024,1024,0.007900799810886382
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,1024,512,0.003369600077470144
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,1024,512,0.01447466711203257
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,1024,768,0.007523199915885926
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,1024,256,0.0025642665723959604
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,1024,256,0.014177067081133523
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,1024,512,0.007118933399518331
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,1024,128,0.0024661332368850707
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,1024,256,0.006683733562628429
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,1024,128,0.01442026694615682
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,1024,64,0.0021920000513394673
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,1024,64,0.014477866888046264
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,1024,128,0.006698666512966156
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,1024,32,0.00223786657055219
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,1024,32,0.013268267114957174
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,768,65536,0.021079466740290324
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,768,65536,0.0322762668132782
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,768,16384,0.011322666207949321
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,768,16384,0.019997866948445638
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,1024,65536,0.12205866972605388
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,768,12288,0.009891200065612792
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,768,12288,0.01922559936841329
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,768,16384,0.03503359953562419
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,768,10240,0.00844373305638631
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,768,10240,0.018388267358144125
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,768,10240,0.024462932348251344
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,768,8192,0.007180800040562947
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,768,8192,0.018742400407791137
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,768,12288,0.028455466032028198
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,768,7168,0.005538133283456167
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,768,7168,0.01926400065422058
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,768,7168,0.019088000059127808
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,768,6144,0.0059232001503308615
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,768,8192,0.020680532852808634
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,768,6144,0.018538665771484376
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,768,5120,0.005876266459623972
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,768,5120,0.01849173307418823
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,768,5120,0.015729066729545594
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,768,4096,0.005932799975077311
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,768,4096,0.017195733388264973
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,768,6144,0.016978132724761962
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,768,3584,0.0058378666639328
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,768,3584,0.016530133287111917
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,768,3584,0.012857600053151449
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,768,3072,0.005468800167242686
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,768,4096,0.013645866513252258
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,768,3072,0.015732266505559287
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,768,2560,0.005457066496213277
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,768,2560,0.016459733247756958
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,768,3072,0.011677866180737812
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,768,2560,0.010761599739392598
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,768,2048,0.004276266694068909
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,768,2048,0.015827199816703795
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,768,1536,0.0046293333172798155
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,768,1536,0.01455893317858378
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,768,1536,0.009141332904497783
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,768,2048,0.009945600231488546
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,768,1024,0.0038165333370367683
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,768,1024,0.014747732877731323
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,768,768,0.003655466685692469
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,768,768,0.014528000354766845
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,768,1024,0.0079434668024381
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,768,512,0.0030421334008375804
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,768,768,0.007546666761239369
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,768,512,0.014457600315411887
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,768,256,0.0025802666942278544
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,768,256,0.013714133699735006
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,768,512,0.007420800129572551
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,768,128,0.0023562667270501454
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,768,256,0.00674773355325063
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,768,128,0.0144405335187912
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,768,64,0.0021920000513394673
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,768,128,0.006695466736952464
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,768,64,0.014428800344467163
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,768,32,0.0021514666577180227
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,768,32,0.014523733655611673
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,512,65536,0.0280458668867747
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,512,65536,0.02900693416595459
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,512,16384,0.008041599889596303
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,512,16384,0.018703999121983846
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,512,16384,0.03546239932378133
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,512,12288,0.00788800021012624
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,512,12288,0.018680532773335777
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,768,65536,0.12107093334197998
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,512,10240,0.007163733243942261
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,512,12288,0.028040534257888793
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,512,10240,0.018629332383473717
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,512,8192,0.0056874667604764305
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,512,8192,0.01856213410695394
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,512,10240,0.024353067080179848
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,512,7168,0.00550186683734258
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,512,7168,0.019400533040364584
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,512,8192,0.020670932531356812
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,512,6144,0.005841066439946493
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,512,6144,0.01847040057182312
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,512,7168,0.019006933768590292
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,512,5120,0.005485866467158
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,512,5120,0.01781760056813558
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,512,6144,0.017384533087412515
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,512,5120,0.015345066785812378
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,512,4096,0.018642133474349974
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,512,4096,0.01366933286190033
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,512,3584,0.005505066613356272
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,512,3584,0.01646080017089844
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,512,3584,0.012479999661445617
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,512,3072,0.005083733300367991
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,512,3072,0.016380799810091655
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,512,3072,0.011566932996114094
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,512,2560,0.005032533407211303
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,512,2560,0.016557866334915163
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,512,2560,0.010790399710337321
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,512,2048,0.004704000055789947
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,512,2048,0.014631467064221701
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,512,2048,0.009967999656995137
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,512,65536,0.1207466681798299
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,512,1536,0.00420906643072764
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,512,1536,0.01444906691710154
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,512,1024,0.0030303999781608583
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,512,1024,0.014506666858990987
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,512,1536,0.009230933586756388
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,512,768,0.0029909332593282064
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,512,1024,0.008310399949550629
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,512,768,0.01451520025730133
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,512,512,0.0026133333643277483
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,512,512,0.01444906691710154
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,512,768,0.007483733197053273
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,512,256,0.002570666621128718
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,512,512,0.007459199925263722
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,512,256,0.01383039951324463
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,512,128,0.0021503999829292296
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,512,128,0.012888532876968384
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,512,256,0.006728533407052357
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,512,64,0.0021269333859284716
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,512,64,0.014498133460680643
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,512,128,0.006660266717274983
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,512,32,0.0022005334496498106
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,256,65536,0.016852267583211265
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,512,32,0.013310933113098144
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,256,65536,0.022784000635147093
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,256,16384,0.005509333312511444
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,256,16384,0.018424532810846963
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,256,16384,0.035394132137298584
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,256,12288,0.005066666503747304
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,256,12288,0.018227199713389076
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,256,12288,0.028006400664647418
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,256,10240,0.005826133489608765
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,256,10240,0.01681706706682841
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,256,10240,0.02441493272781372
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,256,8192,0.005083733300367991
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,256,8192,0.01918506622314453
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,256,8192,0.02105813423792521
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,256,7168,0.00510506679614385
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,256,7168,0.018125865856806436
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,256,7168,0.01904639999071757
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,256,6144,0.005025066435337067
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,256,6144,0.016594133774439492
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,256,65536,0.12023039658864339
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,256,5120,0.005066666503747304
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,256,5120,0.016965333620707193
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,256,6144,0.016963199774424235
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,256,4096,0.0051125332713127134
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,256,4096,0.01667520006497701
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,256,5120,0.015310933192571005
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,256,3584,0.00468800018231074
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,256,3584,0.015052800377209982
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,256,4096,0.013649066289265951
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,256,3072,0.004665599763393402
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,256,3072,0.016458666324615477
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,256,3584,0.012443733215332032
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,256,2560,0.005015466610590617
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,256,2560,0.01513706644376119
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,256,3072,0.011683199803034465
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,256,2048,0.004605866471926371
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,256,2048,0.015672533710797628
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,256,2560,0.010749866565068562
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,256,1536,0.003828266759713491
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,256,1536,0.014509866635004679
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,256,2048,0.009610666831334432
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,256,1024,0.0030613332986831666
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,256,1024,0.014456533392270408
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,256,1536,0.008701866865158081
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,256,768,0.0029663999875386557
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,256,768,0.014444800217946372
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,256,1024,0.00832426647345225
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,256,512,0.002616533388694127
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,256,768,0.007546666761239369
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,256,512,0.01418560047944387
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,256,512,0.007162666817506154
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,256,256,0.0021568000316619872
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,256,256,0.01381226678689321
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,256,256,0.0066997334361076355
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,256,128,0.002182399978240331
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,256,128,0.01404906709988912
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,256,128,0.0062730665008227035
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,256,64,0.002145066608985265
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,256,64,0.013318399588267008
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,256,32,0.0021354667842388153
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,256,32,0.012928000092506409
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,128,65536,0.009549867113431294
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,128,65536,0.019054933389027914
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,128,16384,0.005976533393065134
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,128,16384,0.01898346741994222
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,128,16384,0.034677334626515705
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,128,12288,0.005029333134492239
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,128,12288,0.01802560091018677
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,128,12288,0.027980800469716387
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,128,10240,0.005432533224423727
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,128,10240,0.016646400094032288
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,128,10240,0.024734934171040855
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,128,8192,0.005044266581535339
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,128,8192,0.01689173380533854
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,128,65536,0.12037546634674072
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,128,7168,0.005103999873002371
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,128,7168,0.01688533425331116
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,128,8192,0.02067199945449829
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,128,6144,0.004676266511281332
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,128,6144,0.01664426624774933
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,128,7168,0.018569600582122803
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,128,5120,0.004628266890843709
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,128,5120,0.016565333803494772
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,128,6144,0.017272533973058064
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,128,4096,0.0042698666453361515
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,128,4096,0.016796799500783284
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,128,5120,0.015317333738009134
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,128,3584,0.004699733356634776
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,128,3584,0.015637333194414772
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,128,4096,0.013292800386746725
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,128,3072,0.004265599946180979
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,128,3584,0.012418133020401002
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,128,3072,0.015366400281588236
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,128,2560,0.0046528001626332605
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,128,2560,0.015862400333086647
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,128,3072,0.011588266491889954
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,128,2048,0.003828266759713491
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,128,2560,0.010743467013041179
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,128,2048,0.014706133802731832
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,128,1536,0.003804799914360046
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,128,1536,0.014519466956456503
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,128,2048,0.009925333658854167
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,128,1024,0.003035733352104823
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,128,1024,0.014453333616256715
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,128,1536,0.008725333213806152
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,128,768,0.0029845332105954488
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,128,768,0.014470400412877402
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,128,1024,0.008392533659934998
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,128,512,0.002583466718594233
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,128,768,0.007481599847475688
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,128,512,0.014461867014567056
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,128,256,0.0022079999248186748
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,128,256,0.014467199643452963
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,128,512,0.007123200098673503
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,128,128,0.002182399978240331
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,128,256,0.0071285332242647815
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,128,128,0.01411626636981964
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,128,64,0.0021429332594076794
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,4,128,128,0.0067104001839955645
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,128,64,0.012917332847913108
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,128,32,0.0018613333503405254
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,64,65536,0.006723199784755707
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,128,32,0.014139733711878457
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,64,16384,0.005438933273156484
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,64,65536,0.01923840045928955
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,64,12288,0.005043200155099233
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,64,16384,0.017308799425760905
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,64,10240,0.004810666541258494
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,64,12288,0.018054399887720743
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,64,8192,0.005426133175690969
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,64,10240,0.016571733355522155
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,64,7168,0.004681600133577982
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,64,8192,0.016531200210253397
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,64,6144,0.005076266825199127
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,64,7168,0.017340799172719322
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,64,5120,0.004198400179545084
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,64,6144,0.016613333423932394
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,64,4096,0.004266666869322458
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,64,5120,0.0164874662955602
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,64,3584,0.0042698666453361515
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,64,4096,0.01665173371632894
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,64,3072,0.004274133344491323
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,64,3584,0.014667733510335287
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,64,2560,0.0042133331298828125
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,64,3072,0.015406933426856995
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,64,2048,0.004284800092379252
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,64,2560,0.014457600315411887
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,64,1536,0.0034101332227389016
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,64,2048,0.01483626663684845
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,64,1536,0.014465066790580749
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,64,1024,0.003033600002527237
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,64,1024,0.014455466469128927
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,64,768,0.002665599932273229
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,64,768,0.014362667004267374
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,64,512,0.0025450666745503742
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,64,512,0.014429866274197897
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,64,256,0.002164266755183538
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,64,256,0.014346667130788169
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,64,128,0.002186666677395503
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,64,128,0.014491732915242514
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,64,64,0.002111999938885371
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,64,64,0.014459733168284097
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,64,32,0.0019050666441520054
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,64,32,0.014340266585350037
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,32,65536,0.00664213349421819
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,32,65536,0.02114880084991455
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,32,16384,0.005046399931112925
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,32,16384,0.01887893279393514
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,32,12288,0.00505920002857844
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,32,12288,0.016641066471735636
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,32,10240,0.004682666560014089
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,32,10240,0.016582399606704712
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,32,8192,0.004264533519744873
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,32,8192,0.01688213348388672
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,32,7168,0.0046965335806210835
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,32,7168,0.01725546717643738
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,32,6144,0.004278400043646494
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,32,6144,0.016614400347073875
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,32,5120,0.004226133227348328
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,32,5120,0.01763413349787394
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,32,4096,0.004204800228277842
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,32,3584,0.004430933296680451
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,32,4096,0.01665493349234263
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,32,3072,0.004244266450405121
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,32,3072,0.015358933806419372
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,32,3584,0.014811733365058899
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,32,2560,0.004615466793378194
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,32,2048,0.004249600072701773
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,32,2560,0.014588800072669984
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,32,2048,0.014526933431625366
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,32,1536,0.0037664001186688742
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,32,1024,0.0030986666679382324
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,32,1536,0.014528000354766845
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,32,768,0.0026079999903837843
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,32,1024,0.014605866869290671
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,32,768,0.014536533753077188
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,32,512,0.002477866659561793
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,32,256,0.0021877333521842955
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,32,512,0.014396799604098
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,32,128,0.002147199958562851
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,32,256,0.012659200032552085
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,32,128,0.014437333742777506
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,32,64,0.0018592000007629395
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,32,32,0.0018432000031073886
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,32,64,0.014549332857131957
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,32,32,0.013411200046539307
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,65536,16384,0.32373867034912107
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,65536,16384,0.17348373730977376
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,65536,12288,0.23325759569803872
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,65536,12288,0.1364565372467041
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,4,2560,2560,0.016532267133394875
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,65536,10240,0.1986826737721761
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,65536,10240,0.11799253622690838
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,4,512,4096,0.005940266450246175
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,65536,8192,0.16074879964192706
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,65536,8192,0.10266239643096924
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,65536,10240,0.07942186991373698
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,65536,7168,0.14060160319010417
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,65536,7168,0.08301333586374918
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,65536,12288,0.09205546379089355
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,65536,6144,0.12315200169881184
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,65536,6144,0.07209280331929525
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,65536,8192,0.06842346986134848
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,65536,5120,0.10406186580657958
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,65536,5120,0.06385066509246826
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,65536,7168,0.06289600133895874
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,65536,4096,0.0852128028869629
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,65536,4096,0.053556267420450845
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,65536,16384,0.12169600327809652
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,65536,3584,0.06981226603190104
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,65536,3584,0.049754667282104495
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,65536,6144,0.04875946839650472
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,65536,3072,0.06022506554921468
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,65536,3072,0.043619199593861895
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,65536,5120,0.04132800102233887
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,65536,2560,0.05008320013682047
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,65536,2560,0.040294400850931805
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,65536,3584,0.035897600650787356
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,65536,2048,0.0412778655687968
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,65536,2048,0.03621546824773152
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,65536,3072,0.030279467503229778
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,65536,1536,0.0321237325668335
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,65536,1536,0.02927253246307373
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,65536,2560,0.027013333638509114
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,65536,1024,0.02229866584142049
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,65536,1024,0.024848000208536784
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,65536,1024,0.015773866573969522
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,65536,768,0.017423999309539796
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,65536,768,0.022836265961329143
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,65536,4096,0.039700265725453696
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,65536,512,0.012519466876983642
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,65536,512,0.01933120091756185
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,65536,1536,0.019591466585795084
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,65536,256,0.007921066880226136
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,65536,256,0.016495999693870545
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,65536,512,0.011598933736483257
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,65536,128,0.005054933329423269
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,65536,128,0.015844266613324484
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,65536,256,0.009633066256841023
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,65536,64,0.004281599819660187
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,65536,64,0.01570453345775604
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,65536,32,0.0042367999752362575
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,65536,32,0.015085867047309876
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,16384,65536,0.31904961268107096
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,16384,65536,0.17770133018493653
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,65536,128,0.008400000135103862
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,16384,16384,0.08842133680979411
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,16384,16384,0.06110826730728149
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,65536,768,0.013722667098045349
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,16384,12288,0.06489493449529013
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,16384,12288,0.04465920130411784
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,65536,2048,0.023948800563812257
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,16384,10240,0.05633706649144491
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,16384,10240,0.04028799931208293
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,16384,10240,0.029656533400217695
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,16384,8192,0.04691733519236247
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,16384,8192,0.0350602666536967
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,16384,12288,0.037989334265391035
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,16384,7168,0.03618026574452718
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,16384,7168,0.032014934221903484
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,16384,16384,0.05057599941889444
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,16384,6144,0.03104426662127177
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,16384,6144,0.02937493324279785
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,16384,7168,0.024124799172083537
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,16384,5120,0.026922667026519777
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,16384,5120,0.026790400346120197
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,16384,8192,0.025187200307846068
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,16384,4096,0.022273067633310953
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,16384,4096,0.024782933791478477
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,16384,6144,0.021947733561197915
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,16384,3584,0.019407999515533448
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,16384,3584,0.022685867547988892
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,16384,4096,0.01623040040334066
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,16384,3072,0.01704746683438619
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,16384,3072,0.022563199202219643
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,16384,5120,0.019067732493082683
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,16384,2560,0.014545067151387533
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,16384,2560,0.021057067314783733
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,16384,2560,0.013700266679128012
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,16384,2048,0.012403200070063274
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,16384,2048,0.020381865898768108
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,16384,3072,0.014149333039919535
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,16384,1536,0.010026666522026061
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,16384,1536,0.018678400913874307
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,16384,3584,0.016408532857894897
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,16384,1024,0.007483733197053273
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,16384,1024,0.014654933412869772
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,16384,1024,0.00920853316783905
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,16384,768,0.005853866537412008
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,16384,768,0.014499200383822122
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,16384,1536,0.010045866171518963
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,16384,512,0.003516799956560135
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,16384,512,0.014572800199190775
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,16384,2048,0.012134400010108948
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,16384,256,0.0030410667260487873
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,16384,256,0.014452266693115234
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,16384,768,0.008387200037638346
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,16384,128,0.0025941332181294756
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,16384,128,0.013266133268674216
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,16384,512,0.007949866851170858
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,16384,64,0.002243199944496155
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,16384,128,0.0066655998428662615
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,16384,64,0.014425599575042724
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,16384,256,0.007485866546630859
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,16384,32,0.0021685334543387097
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,16384,32,0.01453439990679423
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,12288,65536,0.13909227053324383
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,12288,65536,0.2341610590616862
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,12288,16384,0.06781760056813559
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,12288,16384,0.050254933039347326
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,12288,12288,0.04899413188298543
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,12288,12288,0.040565331776936844
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,12288,12288,0.030558933814366657
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,12288,10240,0.04169066747029622
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,12288,10240,0.037773867448170975
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,12288,16384,0.03847893476486206
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,12288,8192,0.03516906499862671
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,12288,8192,0.03383040030797323
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,12288,10240,0.02682773272196452
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,12288,7168,0.030947200457255047
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,12288,7168,0.031600000460942586
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,12288,8192,0.022734934091567995
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,12288,6144,0.02722346584002177
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,12288,6144,0.028915200630823774
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,12288,7168,0.02113706668217977
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,12288,5120,0.023591466744740806
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,12288,5120,0.026846933364868163
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,12288,6144,0.01913493275642395
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,12288,4096,0.02037013371785482
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,12288,4096,0.021949867407480873
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,12288,5120,0.01776319940884908
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,12288,3584,0.017786665757497152
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,12288,3584,0.02074880003929138
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,12288,4096,0.015356799960136414
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,12288,3072,0.015969066818555196
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,12288,3072,0.020626133680343627
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,12288,3584,0.014383999506632486
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,12288,2560,0.013986133535703025
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,12288,2560,0.01904746691385905
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,12288,2560,0.011909332871437073
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,12288,2048,0.010291199882825215
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,12288,2048,0.017796266078948974
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,12288,3072,0.013105066617329917
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,12288,1536,0.008303999900817871
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,12288,1536,0.016235733032226564
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,12288,1536,0.009196799993515015
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,12288,1024,0.006277333199977875
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,12288,1024,0.01452906628449758
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,12288,2048,0.010466133554776508
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,12288,768,0.00432640016078949
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,12288,768,0.01454080045223236
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,12288,768,0.008378666639328004
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,12288,512,0.0035103999078273775
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,12288,1024,0.008330666522185007
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,12288,512,0.014571733276049294
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,12288,256,0.002833066632350286
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,12288,256,0.014553599556287131
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,12288,256,0.00713919997215271
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,12288,128,0.0025770666698614756
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,12288,128,0.014433067043622336
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,12288,128,0.006738133231798808
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,12288,64,0.0022240000466505687
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,12288,64,0.014428800344467163
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,12288,32,0.0022101332743962605
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,12288,32,0.014454399545987448
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,12288,512,0.007518933216730754
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,10240,65536,0.19548800786336262
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,10240,65536,0.11973439852396647
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,10240,16384,0.058405331770579015
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,10240,16384,0.043395201365153
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,10240,16384,0.03807466824849447
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,10240,12288,0.04118613402048747
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,10240,12288,0.037109335263570145
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,10240,12288,0.03011626601219177
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,10240,10240,0.035792001088460285
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,10240,10240,0.033148799339930216
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,16384,65536,0.15422933896382648
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,10240,8192,0.029659734169642134
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,10240,8192,0.029156267642974854
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,10240,10240,0.026716800530751546
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,10240,7168,0.026734934250513716
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,10240,7168,0.029303467273712157
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,10240,8192,0.022352000077565513
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,10240,6144,0.02403200070063273
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,10240,6144,0.027037866910298664
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,12288,65536,0.13320319652557372
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,10240,5120,0.020801067352294922
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,10240,5120,0.026105600595474242
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,10240,7168,0.020610133806864418
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,10240,4096,0.017058134078979492
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,10240,4096,0.022681599855422972
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,10240,6144,0.018636800845464072
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,10240,3584,0.0164192001024882
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,10240,3584,0.020665599902470907
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,10240,4096,0.015089066823323569
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,10240,3072,0.014451199769973755
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,10240,3072,0.02064746618270874
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,10240,5120,0.017214934031168617
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,10240,2560,0.01317759950955709
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,10240,2560,0.020512000719706217
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,10240,3584,0.013343999783198038
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,10240,2048,0.009585066636403402
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,10240,2048,0.018605866034825645
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,10240,2560,0.01130560040473938
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,10240,1536,0.008737066388130188
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,10240,1536,0.01667413314183553
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,10240,3072,0.012770133217175803
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,10240,1024,0.006728533407052357
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,10240,1024,0.015597866972287497
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,10240,2048,0.010427733262379963
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,10240,768,0.005103999873002371
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,10240,768,0.0145087997118632
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,10240,1024,0.008358400066693623
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,10240,512,0.004315733412901561
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,10240,512,0.014479999740918478
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,10240,1536,0.009240532914797466
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,10240,256,0.002717866748571396
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,10240,256,0.014485333363215128
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,10240,512,0.0075882668296496075
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,10240,128,0.002239999920129776
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,10240,128,0.013780267039934794
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,10240,256,0.007117866476376851
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,10240,64,0.002164266755183538
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,10240,64,0.014505599935849508
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,10240,32,0.0022453332940737406
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,10240,32,0.014503467082977294
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,10240,768,0.007942399879296621
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,10240,128,0.006763733426729838
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,8192,65536,0.15634454091389974
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,8192,65536,0.0998250643412272
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,8192,16384,0.04655146598815918
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,8192,16384,0.03697386582692464
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,8192,12288,0.03492906490961711
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,8192,12288,0.030155734221140547
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,8192,12288,0.029286400477091475
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,8192,10240,0.02977919975916545
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,8192,10240,0.028638933102289838
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,8192,16384,0.03708479801813762
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,8192,8192,0.022826667626698813
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,8192,8192,0.02524906595547994
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,8192,10240,0.026200532913208008
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,8192,7168,0.01988053321838379
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,8192,7168,0.022842667500178018
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,8192,8192,0.02203306754430135
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,8192,6144,0.017895466089248656
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,8192,6144,0.02266133427619934
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,10240,65536,0.12908693154652912
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,8192,5120,0.014957867066065469
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,8192,5120,0.021715199947357176
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,8192,7168,0.020725333690643312
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,8192,4096,0.012723199526468911
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,8192,4096,0.020159999529520668
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,8192,6144,0.01875413258870443
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,8192,3584,0.01157866617043813
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,8192,3584,0.01893226703008016
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,8192,5120,0.016924800475438435
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,8192,3072,0.0102293332417806
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,8192,3072,0.018832000096638997
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,8192,4096,0.014195199807484946
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,8192,2560,0.008858666817347208
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,8192,2560,0.01656000018119812
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,8192,3584,0.01297599971294403
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,8192,2048,0.008222933113574981
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,8192,2048,0.01653333306312561
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,8192,2560,0.011307733257611592
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,8192,1536,0.006364800035953522
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,8192,1536,0.014545067151387533
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,8192,3072,0.012072533369064331
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,8192,1024,0.004082133372624716
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,8192,1024,0.014582399527231851
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,8192,2048,0.010030933221181234
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,8192,768,0.0034495999415715536
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,8192,768,0.014458666245142618
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,8192,1024,0.00844266712665558
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,8192,512,0.00306986669699351
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,8192,512,0.014417066176732381
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,8192,1536,0.009237333138783773
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,8192,256,0.0026154667139053345
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,8192,256,0.013873066504796347
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,8192,512,0.007559466858704884
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,8192,128,0.002229333420594533
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,8192,768,0.007540266712506611
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,8192,128,0.014376533031463624
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,8192,64,0.002232533444960912
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,8192,64,0.014437333742777506
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,8192,32,0.002216533323129018
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,8192,32,0.014422399799029031
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,8192,128,0.006651733318964641
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,7168,65536,0.13739412625630695
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,8192,256,0.0070933332045873005
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,7168,16384,0.03938026825586955
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,7168,65536,0.08607467015584311
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,7168,16384,0.03672106663386027
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,7168,12288,0.030213334163029987
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,7168,12288,0.027322665850321455
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,7168,12288,0.029335467020670573
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,7168,10240,0.026172800858815508
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,7168,10240,0.02656000057856242
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,7168,16384,0.03664106527964274
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,7168,8192,0.021862399578094483
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,7168,8192,0.02297066648801168
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,7168,10240,0.02592746615409851
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,7168,7168,0.019396267334620156
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,7168,7168,0.02273919979731242
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,7168,8192,0.022051199277242025
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,7168,6144,0.015936000148455302
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,7168,6144,0.022381865978240968
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,7168,7168,0.020278400182723998
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,7168,5120,0.013912533720334372
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,7168,5120,0.02071146567662557
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,7168,6144,0.018649599949518838
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,7168,4096,0.01143999993801117
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,7168,4096,0.018633600076039633
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,7168,5120,0.01660160024960836
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,7168,3584,0.010643200079600016
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,7168,3584,0.018663465976715088
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,7168,4096,0.014146133263905843
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,7168,3072,0.009417600433031718
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,7168,3072,0.017774933576583864
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,7168,3584,0.012920533617337545
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,7168,2560,0.008322133123874665
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,7168,2560,0.01657386620839437
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,7168,3072,0.012066133817036946
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,7168,2048,0.007112533350785573
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,7168,2048,0.014461867014567056
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,7168,2560,0.010826667149861652
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,7168,1536,0.00558186670144399
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,7168,1536,0.014484266440073649
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,7168,2048,0.009987200299898785
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,7168,1024,0.004095999896526337
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,7168,1024,0.01454080045223236
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,7168,1536,0.00920853316783905
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,7168,768,0.0034634667138258614
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,7168,768,0.014467199643452963
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,7168,1024,0.008323200047016144
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,7168,512,0.002979200085004171
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,7168,512,0.01448319951693217
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,7168,768,0.007937066753705342
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,7168,256,0.002587733417749405
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,7168,256,0.014099199573198953
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,7168,512,0.0075093333919843035
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,7168,128,0.002152533332506816
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,7168,256,0.0069567998250325514
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,7168,128,0.013906133174896241
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,7168,64,0.002160000056028366
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,7168,64,0.013412266969680786
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,7168,32,0.002161066730817159
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,7168,32,0.013288533687591553
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,7168,128,0.006742399930953979
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,6144,65536,0.11843626499176026
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,6144,65536,0.07878186702728271
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,6144,16384,0.036397866408030194
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,6144,16384,0.03131413261095683
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,8192,65536,0.12751999696095784
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,6144,12288,0.02823786735534668
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,6144,12288,0.029554132620493574
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,6144,16384,0.03667733271916707
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,6144,10240,0.023191465934117635
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,6144,10240,0.026771199703216553
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,6144,12288,0.02928000092506409
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,6144,8192,0.020641066630681357
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,6144,8192,0.02345386743545532
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,6144,8192,0.022363734245300294
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,6144,7168,0.018345600366592406
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,6144,7168,0.023042132457097374
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,6144,10240,0.025676800807317095
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,6144,6144,0.015343999862670899
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,6144,6144,0.02066133419672648
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,6144,7168,0.020309333006540933
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,6144,5120,0.014544000228246054
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,6144,5120,0.020451200008392335
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,6144,6144,0.018380800882975258
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,6144,4096,0.010595200459162395
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,6144,4096,0.018755199511845906
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,6144,4096,0.013784533739089966
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,6144,3584,0.009699199597040813
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,6144,3584,0.017334399620691936
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,6144,5120,0.015848533312479655
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,6144,3072,0.008708266417185466
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,6144,3072,0.016694400707880655
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,6144,3584,0.012880000472068786
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,6144,3072,0.012026666601498922
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,6144,2560,0.007740800082683563
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,6144,2560,0.01648319959640503
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,6144,2048,0.006568533182144165
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,6144,2048,0.014681599537531533
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,7168,65536,0.1270069360733032
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,6144,1536,0.004982399940490723
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,6144,1536,0.014662399888038635
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,6144,1536,0.009293867150942485
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,6144,1024,0.004075733323891958
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,6144,1024,0.014658133188883463
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,6144,2048,0.010043733318646749
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,6144,768,0.0034944000343481696
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,6144,768,0.01455893317858378
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,6144,2560,0.010874666770299276
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,6144,512,0.0030421334008375804
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,6144,512,0.014443733294804893
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,6144,1024,0.008337066570917765
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,6144,256,0.002631466587384542
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,6144,768,0.0083146666487058
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,6144,256,0.01444586714108785
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,6144,128,0.0022463999688625337
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,6144,128,0.01446293294429779
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,6144,128,0.006742399930953979
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,6144,64,0.002165333429972331
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,6144,64,0.014403200149536133
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,6144,256,0.006702933212121327
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,6144,32,0.0021546666820844015
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,6144,32,0.014346667130788169
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,5120,65536,0.10083200136820476
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,5120,65536,0.07005120118459066
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,5120,16384,0.03184426625569661
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,5120,16384,0.030905600388844805
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,5120,16384,0.03622399965922038
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,5120,12288,0.02409600019454956
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,5120,12288,0.027054933706919353
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,5120,12288,0.02906559904416402
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,5120,10240,0.02143359978993734
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,5120,10240,0.026204800605773924
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,5120,10240,0.02604159911473592
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,5120,8192,0.01849493384361267
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,5120,8192,0.022337067127227783
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,6144,65536,0.1270357370376587
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,5120,7168,0.01602453291416168
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,5120,7168,0.021195733547210695
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,5120,8192,0.022290132443110146
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,5120,6144,0.014494933684666953
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,5120,7168,0.019882667064666747
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,5120,6144,0.019400533040364584
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,5120,5120,0.012530133128166199
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,5120,5120,0.020618667205174766
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,5120,5120,0.01622719963391622
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,5120,4096,0.009619200229644775
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,5120,4096,0.0187008003393809
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,5120,6144,0.017832533518473307
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,5120,3584,0.00872320036093394
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,5120,3584,0.016562133034070333
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,5120,3584,0.012971733013788858
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,5120,3072,0.007926400005817413
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,5120,3072,0.0162581334511439
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,5120,4096,0.013684266805648803
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,5120,2560,0.007150933146476746
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,5120,2560,0.016261333227157594
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,5120,2560,0.011165866255760193
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,5120,2048,0.00559146652619044
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,5120,2048,0.014636799693107605
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,5120,3072,0.011745066444079081
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,5120,1536,0.004508799811204275
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,5120,1536,0.014499200383822122
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,5120,1536,0.009204266468683879
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,5120,1024,0.0037429332733154297
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,5120,1024,0.014491732915242514
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,5120,2048,0.010001066327095031
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,5120,768,0.0032821332414944967
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,5120,768,0.014536533753077188
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,5120,768,0.007893333335717519
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,5120,512,0.002643200010061264
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,5120,1024,0.007926400005817413
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,5120,512,0.014593066771825156
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,5120,256,0.002609066665172577
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,5120,256,0.014459733168284097
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,5120,256,0.007236266632874806
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,5120,128,0.0022442666192849475
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,5120,128,0.014513066411018372
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,5120,512,0.007515733440717061
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,5120,64,0.002149333308140437
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,5120,128,0.006733866532643636
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,5120,64,0.013301333785057068
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,5120,32,0.002232533444960912
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,5120,32,0.014166399836540222
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,4096,65536,0.08203413486480712
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,4096,65536,0.059979732831319174
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,4096,16384,0.025569067398707075
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,4096,16384,0.02786453366279602
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,5120,65536,0.12649813493092854
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,4096,12288,0.02021333376566569
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,4096,12288,0.024833067258199056
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,4096,16384,0.03637760082880656
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,4096,10240,0.016785067319869996
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,4096,10240,0.022794665892918904
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,4096,10240,0.02520853281021118
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,4096,8192,0.014069333672523499
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,4096,8192,0.021988266706466676
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,4096,12288,0.02895359992980957
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,4096,7168,0.01337493360042572
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,4096,7168,0.023492266734441124
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,4096,7168,0.019773866732915243
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,4096,6144,0.012085333466529846
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,4096,6144,0.022651733954747517
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,4096,8192,0.021575466791788737
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,4096,5120,0.010776533683141073
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,4096,5120,0.021528534094492593
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,4096,5120,0.01581546664237976
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,4096,4096,0.009489066402117411
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,4096,4096,0.018582399686177573
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,4096,6144,0.01767146587371826
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,4096,3584,0.010812800129254658
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,4096,3584,0.01872746745745341
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,4096,3584,0.01295360028743744
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,4096,3072,0.00941439966360728
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,4096,3072,0.018029866615931193
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,4096,4096,0.013352533181508383
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,4096,2560,0.008313600222269695
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,4096,2560,0.017359999815622966
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,4096,3072,0.011768533786137899
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,4096,2048,0.0071487997968991595
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,4096,2048,0.016581333676973977
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,4096,2560,0.010798933108647664
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,4096,1536,0.005871999760468801
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,4096,1536,0.015753600001335143
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,6144,512,0.007493333518505096
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,4096,1024,0.005118933320045471
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,4096,1024,0.014470400412877402
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,4096,1536,0.00920960009098053
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,4096,768,0.004206933577855428
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,4096,768,0.014461867014567056
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,4096,1024,0.007942399879296621
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,4096,512,0.00340693344672521
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,4096,512,0.0144896000623703
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,4096,2048,0.010035199920336406
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,4096,256,0.0026730666557947796
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,4096,768,0.007897600034872691
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,4096,256,0.014403200149536133
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,4096,128,0.002221866697072983
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,4096,128,0.014424533645311991
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,4096,512,0.007482666770617168
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,4096,64,0.002169599880774816
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,4096,128,0.006704000135262807
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,4096,256,0.006723199784755707
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,4096,32,0.0021759999295075734
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,4096,64,0.01444586714108785
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,4096,32,0.013659733533859252
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,3584,65536,0.07351253032684327
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,3584,65536,0.0536629319190979
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,3584,16384,0.022241065899531044
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,3584,16384,0.026859732468922932
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,3584,12288,0.017496534188588462
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,3584,12288,0.022798933585484824
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,3584,12288,0.02882986664772034
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,3584,10240,0.01503679951032003
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,3584,10240,0.022709333896636964
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,3584,16384,0.03632853428522746
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,3584,8192,0.012920533617337545
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,3584,8192,0.020820266008377074
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,3584,10240,0.025460267066955568
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,3584,7168,0.012516267100969949
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,3584,7168,0.02109653353691101
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,3584,8192,0.021102933088938396
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,3584,6144,0.0108842670917511
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,3584,6144,0.02090453306833903
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,4096,65536,0.12753600279490154
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,3584,5120,0.010602666934331258
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,3584,5120,0.020687999327977498
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,3584,7168,0.01941439906756083
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,3584,4096,0.008823466300964356
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,3584,4096,0.017656532923380534
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,3584,6144,0.01742080052693685
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,3584,3584,0.010418132940928141
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,3584,3584,0.017534933487574258
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,3584,4096,0.013313066959381104
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,3584,3072,0.00915839970111847
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,3584,3072,0.01752106746037801
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,3584,3584,0.012942933042844138
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,3584,2560,0.007891199986139933
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,3584,2560,0.01649386684099833
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,3584,2560,0.010899200042088827
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,3584,2048,0.006691200037797292
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,3584,2048,0.016125866770744325
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,3584,3072,0.012034133076667786
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,3584,1536,0.00631039987007777
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,3584,1536,0.014491732915242514
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,3584,1536,0.009176533420880635
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,3584,1024,0.005016533533732096
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,3584,1024,0.01652479966481527
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,3584,2048,0.009731200337409974
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,3584,768,0.004205866654713949
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,3584,768,0.014465066790580749
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,3584,1024,0.008391466736793519
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,3584,512,0.0050357331832249965
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,3584,512,0.014451199769973755
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,3584,768,0.007536000013351441
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,3584,256,0.0025941332181294756
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,3584,256,0.01395093301932017
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,3584,256,0.00705813318490982
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,3584,128,0.002160000056028366
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,3584,128,0.013708800077438354
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,3584,512,0.0074879998962084455
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,3584,64,0.002194133400917053
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,3584,64,0.014219733079274497
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,3584,32,0.002178133279085159
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,3584,128,0.006656000018119812
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,3584,32,0.013380266229311624
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,3072,65536,0.06346880197525025
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,3072,65536,0.04982826709747314
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,3072,16384,0.02026559909184774
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,3072,16384,0.02451200087865194
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,3072,16384,0.03617599805196126
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,3072,12288,0.016150400042533875
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,3072,12288,0.022668800751368203
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,3072,12288,0.028854399919509888
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,3072,10240,0.014422399799029031
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,3072,10240,0.021592533588409422
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,3072,10240,0.025225599606831867
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,3072,8192,0.012405332922935487
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,3072,8192,0.020410666863123574
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,3584,65536,0.12613226572672526
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,3072,7168,0.01142080028851827
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,3072,7168,0.02135466734568278
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,3072,7168,0.019399466117223103
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,3072,6144,0.010542933146158855
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,3072,6144,0.020272000630696615
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,3072,8192,0.021141332387924195
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,3072,5120,0.009621333082516987
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,3072,5120,0.018568533658981323
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,3072,5120,0.015809067090352378
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,3072,4096,0.007924266656239827
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,3072,4096,0.01755519906679789
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,3072,6144,0.017848533391952515
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,3072,3584,0.009569066762924194
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,3072,3584,0.01817493240038554
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,3072,3584,0.012875733772913614
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,3072,3072,0.008346666892369587
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,3072,3072,0.01845759948094686
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,3072,4096,0.013660800457000733
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,3072,2560,0.007912533481915791
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,3072,2560,0.01790720025698344
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,3072,2560,0.010838400324185688
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,3072,3072,0.011902933319409687
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,3072,2048,0.0060138667623202005
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,3072,2048,0.01654293338457743
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,3072,1536,0.00644053320089976
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,3072,1536,0.01585813363393148
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,3072,2048,0.010135466853777569
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,3072,1024,0.005099733173847198
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,3072,1024,0.014665599664052328
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,3072,1536,0.009196799993515015
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,3072,768,0.004223999877770742
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,3072,768,0.01452906628449758
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,3072,1024,0.008361599842707316
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,3072,512,0.0034901333351929987
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,3072,512,0.014602667093276978
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,3072,768,0.007505066692829132
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,3072,256,0.002996266633272171
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,3072,512,0.007540266712506611
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,3072,256,0.01418239971001943
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,3072,128,0.0021802666286627453
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,3072,128,0.013361066579818726
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,3072,256,0.007067733506361644
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,3072,64,0.002143999934196472
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,3072,128,0.006682666639486949
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,3072,32,0.002143999934196472
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,3072,64,0.014471466342608133
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,3072,32,0.0137472003698349
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,2560,65536,0.05585600137710571
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,2560,65536,0.044846932093302413
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,2560,16384,0.018411733706792197
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,2560,16384,0.023705599705378215
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,3072,65536,0.1263818661371867
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,2560,12288,0.014004266262054444
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,2560,12288,0.021942400932312013
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,2560,16384,0.0359775980313619
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,2560,10240,0.012588799993197123
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,2560,10240,0.020745599269866945
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,2560,12288,0.02808319926261902
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,2560,8192,0.01112000048160553
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,2560,8192,0.018931200106938682
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,2560,10240,0.02485760052998861
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,2560,7168,0.009983999530474345
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,2560,7168,0.018681599696477254
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,2560,7168,0.019128533204396565
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,2560,6144,0.00921066701412201
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,2560,6144,0.018615466356277467
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,2560,8192,0.020932267109553017
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,2560,5120,0.008041599889596303
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,2560,5120,0.018576000134150186
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,2560,6144,0.01748159925142924
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,2560,4096,0.007156266768773396
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,2560,4096,0.01866453289985657
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,2560,5120,0.015709867080052696
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,2560,3584,0.006320000191529592
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,2560,3584,0.018547199169794717
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,2560,3584,0.012604799866676331
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,2560,3072,0.007189333438873291
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,2560,3072,0.016617600123087564
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,2560,4096,0.013706666231155396
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,2560,2560,0.006966400146484375
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,2560,2560,0.016528000434239708
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,2560,2560,0.010831999778747558
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,2560,3072,0.01202453374862671
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,2560,2048,0.0075328002373377485
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,2560,2048,0.015359999736150107
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,2560,1536,0.005724800129731497
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,2560,1536,0.015305599570274353
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,2560,1536,0.008845866719881693
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,2560,1024,0.005039999882380167
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,2560,1024,0.014468266566594442
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,2560,2048,0.010001066327095031
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,2560,768,0.0042463997999827065
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,2560,768,0.014484266440073649
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,2560,1024,0.00802346666653951
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,2560,512,0.003402666747570038
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,2560,512,0.014456533392270408
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,2560,768,0.00790293316046397
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,2560,256,0.0029685333371162414
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,2560,256,0.012939733266830445
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,2560,256,0.007086933155854543
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,2560,128,0.0021429332594076794
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,2560,128,0.013395200173060099
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,2560,128,0.007063466807206471
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,2560,64,0.0021770666042963665
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,2560,64,0.013056000073750814
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,2560,32,0.002178133279085159
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,2560,32,0.013532800475756326
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,2048,65536,0.04602133433024089
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,2048,65536,0.039374931653340654
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,2560,65536,0.12423893610636394
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,2048,16384,0.015275733669598899
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,2048,16384,0.02259413401285807
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,2048,16384,0.03584106763203938
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,2048,12288,0.012026666601498922
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,2048,12288,0.02132800022761027
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,2048,12288,0.028163200616836546
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,2048,10240,0.010664533575375874
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,2048,10240,0.02002453406651815
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,2048,10240,0.024836266040802003
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,2048,8192,0.009609599908192951
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,2048,8192,0.01876586675643921
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,2048,65536,0.12437866528828938
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,2048,7168,0.008833066622416178
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,2048,7168,0.01807253360748291
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,2048,8192,0.02074986696243286
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,2048,6144,0.0078015998005867004
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,2048,6144,0.018552533785502114
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,2048,7168,0.01906879941622416
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,2048,5120,0.006749866902828217
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,2048,5120,0.016619732975959776
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,2048,5120,0.01577279965082804
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,2048,6144,0.017378133535385133
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,2048,4096,0.006344533463319142
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,2048,4096,0.018733867009480796
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,2048,3584,0.005838933090368906
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,2048,3584,0.016752000649770102
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,2048,3584,0.012889599800109864
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,2048,3072,0.005905066430568695
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,2048,3072,0.017350399494171144
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,2048,4096,0.013714133699735006
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,2048,2560,0.005913599828879039
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,2048,2560,0.01653333306312561
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,2048,3072,0.011995733777681986
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,2048,2048,0.006222933530807495
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,2048,2048,0.016636799772580466
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,2048,2560,0.011154133081436157
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,2048,1536,0.006141866743564606
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,2048,1536,0.014578133821487427
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,2048,1536,0.009150933225949604
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,2048,1024,0.004638933142026265
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,2048,1024,0.014465066790580749
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,2048,2048,0.009592533111572266
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,2048,768,0.00413226659099261
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,2048,768,0.014604799946149192
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,2048,1024,0.008302933474381765
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,2048,512,0.003421866645415624
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,2048,768,0.007542400062084198
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,2048,512,0.014528000354766845
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,2048,256,0.002583466718594233
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,2048,256,0.014154666662216186
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,2048,256,0.007132799923419952
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,2048,128,0.0022485333184401194
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,2048,512,0.007458133498827617
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,2048,128,0.014760532975196838
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,2048,64,0.0021749332547187803
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,2048,64,0.01443839967250824
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,2048,128,0.006963199873765309
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,2048,32,0.0021856000026067098
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,2048,32,0.014512000481287637
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,1536,65536,0.03580693403879802
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,1536,65536,0.0336458683013916
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,1536,16384,0.01241386632124583
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,1536,16384,0.02493333419164022
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,3584,5120,0.01572266618410746
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,1536,12288,0.014602667093276978
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,1536,12288,0.02270080049832662
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,1536,16384,0.035255467891693114
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,1536,10240,0.012875733772913614
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,1536,10240,0.01997119983037313
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,1536,12288,0.028523733218510944
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,1536,8192,0.011272533734639486
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,1536,8192,0.018769067525863648
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,1536,10240,0.024423466126124064
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,1536,7168,0.01055999994277954
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,1536,7168,0.019222400585810342
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,1536,8192,0.021142399311065672
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,1536,6144,0.00941973328590393
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,1536,6144,0.018580265839894614
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,1536,7168,0.019042134284973145
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,1536,5120,0.006355200211207073
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,1536,5120,0.0165994664033254
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,1536,6144,0.017490132649739584
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,1536,4096,0.005963733295599619
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,1536,4096,0.01867093245188395
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,1536,5120,0.015801599621772765
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,1536,3584,0.005905066430568695
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,1536,3584,0.016571733355522155
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,1536,4096,0.013648000359535218
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,1536,3072,0.00595306654771169
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,1536,3072,0.017078399658203125
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,1536,3584,0.012862933675448099
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,1536,2560,0.005118933320045471
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,1536,2560,0.016591999928156534
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,1536,3072,0.011939199765523274
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,1536,2048,0.006814933319886525
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,1536,2048,0.016698666413625083
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,1536,2560,0.010870400071144103
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,1536,1536,0.005235200126965841
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,1536,1536,0.015126400192578635
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,1536,2048,0.009973333279291789
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,1536,1024,0.004198400179545084
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,1536,1024,0.014646400014559427
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,1536,1536,0.008772266904513042
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,1536,768,0.0037920000652472176
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,1536,768,0.014587733149528503
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,1536,1024,0.008303999900817871
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,1536,512,0.003115733216206233
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,1536,512,0.014472533265749613
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,1536,768,0.0075434664885203045
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,1536,256,0.002665599932273229
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,1536,512,0.0071370666225751235
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,1536,256,0.0144896000623703
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,1536,256,0.007113599777221679
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,1536,128,0.0022261333962281544
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,1536,128,0.014447999993960061
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,1536,128,0.006673066814740498
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,1536,64,0.002205866575241089
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,1536,64,0.014506666858990987
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,1536,32,0.0022485333184401194
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,1536,32,0.014459733168284097
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,1024,65536,0.02515946626663208
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,1024,65536,0.02797653277715047
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,1024,16384,0.013245866696039835
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,1024,16384,0.02108479936917623
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,1024,16384,0.03511786858240763
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,1024,12288,0.010652800401051838
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,1024,12288,0.018632533152898152
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,1024,12288,0.028200532992680865
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,1024,10240,0.009425066908200582
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,1024,10240,0.018557866414388023
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,1536,65536,0.1228554646174113
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,1024,8192,0.00853013296922048
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,1024,8192,0.019129600127538046
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,1024,10240,0.024793599049250284
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,1024,7168,0.007540266712506611
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,1024,7168,0.018634666999181114
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,1024,8192,0.020707199970881142
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,1024,6144,0.005985066791375478
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,1024,6144,0.016563199957211814
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,1024,7168,0.019056000312169395
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,1024,5120,0.005539200206597646
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,1024,5120,0.017755732933680216
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,1024,6144,0.017364267508188883
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,1024,4096,0.005876266459623972
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,1024,4096,0.01878933310508728
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,1024,5120,0.015347199638684592
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,1024,3584,0.0054624001185099285
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,1024,3584,0.017876267433166504
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,1024,4096,0.013724799950917563
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,1024,3072,0.005468800167242686
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,1024,3072,0.016537599762280784
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,1024,3584,0.012503467003504434
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,1024,2560,0.005072000126043955
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,1024,2560,0.017198934157689413
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,1024,3072,0.011640533804893494
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,1024,2048,0.007492266595363617
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,1024,2560,0.010804266730944315
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,1024,2048,0.016521599888801575
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,1024,1536,0.005846400062243144
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,1024,1536,0.015198933084805808
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,1024,1536,0.00879039963086446
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,1024,1024,0.004660266637802124
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,1024,1024,0.014460800091425577
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,1024,2048,0.00993280013402303
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,1024,768,0.0039488000174363455
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,1024,768,0.014586666226387024
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,1024,1024,0.007959466675917308
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,1024,512,0.00340693344672521
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,1024,768,0.007493333518505096
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,1024,512,0.0144405335187912
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,1024,256,0.0025898667673269907
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,1024,256,0.014455466469128927
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,1024,512,0.007142400244871776
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,1024,128,0.0025472000241279604
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,1024,256,0.0066655998428662615
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,1024,128,0.014421332875887552
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,1024,64,0.00222080002228419
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,1024,64,0.014452266693115234
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,1024,128,0.006323199967543285
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,1024,32,0.002197333425283432
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,1024,32,0.01441493332386017
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,768,65536,0.021388800938924153
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,768,65536,0.03383786678314209
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,768,16384,0.011225600043932598
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,768,16384,0.020296533902486168
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,1024,65536,0.12171306610107421
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,768,12288,0.00927786628405253
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,768,12288,0.019024000565210978
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,768,16384,0.03502933184305827
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,768,10240,0.007956266899903615
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,768,10240,0.017307732502619425
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,768,12288,0.027704532941182452
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,768,8192,0.007258666555086772
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,768,8192,0.01956053376197815
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,768,10240,0.024842667579650878
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,768,7168,0.005523199836413065
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,768,7168,0.017595734198888144
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,768,8192,0.020713599522908528
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,768,6144,0.005916800101598104
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,768,6144,0.018531199296315512
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,768,7168,0.019045333067576088
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,768,5120,0.0055189331372578945
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,768,5120,0.01651306649049123
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,768,6144,0.01743146578470866
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,768,4096,0.00591893345117569
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,768,4096,0.016568533579508462
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,768,5120,0.01569386621316274
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,768,3584,0.005443199972311656
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,768,3584,0.016724266608556113
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,768,4096,0.013642666737238565
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,768,3072,0.005082666873931885
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,768,3584,0.012925866246223449
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,768,3072,0.016562133034070333
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,768,3072,0.011702400445938111
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,768,2560,0.005063466727733612
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,768,2560,0.01652906636397044
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,768,2048,0.004507733384768168
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,768,2048,0.01681813398996989
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,768,2560,0.010802132884661357
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,768,1536,0.004730666677157084
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,768,2048,0.010021332899729412
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,768,1536,0.014471466342608133
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,768,1024,0.00415040006240209
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,768,1024,0.014444800217946372
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,768,1024,0.007916800181070964
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,768,1536,0.008778666456540424
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,768,768,0.0034688000877698263
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,768,768,0.014603733023007711
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,768,512,0.0030730667213598887
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,768,512,0.014475733041763306
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,768,768,0.007904000083605449
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,768,256,0.0025909334421157838
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,768,512,0.007178666690985362
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,768,256,0.013381333152453104
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,768,128,0.0022250667214393617
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,768,128,0.014441600441932679
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,768,256,0.006702933212121327
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,768,64,0.0021898667017618816
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,768,128,0.006385066608587901
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,768,64,0.014416000247001648
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,768,32,0.0021759999295075734
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,512,65536,0.029289599259694415
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,768,32,0.01444586714108785
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,512,65536,0.02688960035641988
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,512,16384,0.008318933347860973
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,512,16384,0.01760639945665995
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,512,16384,0.03506453434626262
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,512,12288,0.007149866720040639
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,512,12288,0.0184661328792572
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,768,65536,0.12178666591644287
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,512,10240,0.006786133348941803
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,512,10240,0.018781866629918417
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,512,12288,0.02770133415857951
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,512,8192,0.005481599768002828
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,512,8192,0.018823466698328652
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,512,10240,0.024791467189788818
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,512,7168,0.005504000186920166
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,512,8192,0.020617600282033285
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,512,7168,0.01883093317349752
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,512,6144,0.005852800110975901
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,512,6144,0.01835093299547831
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,512,6144,0.017433599630991618
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,512,7168,0.019045333067576088
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,512,5120,0.00550186683734258
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,512,5120,0.0165173331896464
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,512,4096,0.005539200206597646
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,512,4096,0.016544000307718913
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,512,5120,0.015362133582433065
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,512,3584,0.005475200215975444
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,512,4096,0.013286399841308593
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,512,3584,0.016500266393025716
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,512,3072,0.005083733300367991
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,512,3072,0.015773866573969522
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,512,3584,0.012428800264994305
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,512,2560,0.005037866532802582
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,512,3072,0.011601066589355469
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,512,2560,0.014454399545987448
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,512,2048,0.004621866842110952
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,512,2048,0.015084800124168397
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,512,2560,0.010789333780606588
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,512,1536,0.003870933254559835
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,512,1536,0.014569600423177084
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,512,2048,0.010051199793815612
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,512,1024,0.003387733300526937
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,512,1536,0.008806399504343669
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,512,1024,0.014452266693115234
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,512,768,0.003020799905061722
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,512,768,0.01446293294429779
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,512,1024,0.008285866677761078
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,512,512,0.0026335999369621276
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,512,768,0.00812799980243047
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,512,512,0.01446613371372223
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,512,256,0.00257493332028389
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,512,256,0.014375467101732889
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,512,512,0.0071381335457166035
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,512,128,0.0021738665799299877
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,512,256,0.007045333087444305
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,512,128,0.013438933094342551
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,512,64,0.0021759999295075734
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,512,64,0.0137855996688207
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,512,128,0.0067007998625437425
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,512,32,0.0021845333278179167
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,256,65536,0.01695786714553833
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,512,32,0.014432000120480857
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,256,65536,0.021143466234207153
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,256,16384,0.0054400001962979635
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,256,16384,0.01881600022315979
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,512,65536,0.12072319984436035
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,256,12288,0.005220266679922739
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,256,12288,0.01729493339856466
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,256,16384,0.03495146830876668
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,256,12288,0.02766079902648926
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,256,10240,0.005486933390299479
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,256,10240,0.016666666666666666
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,256,8192,0.005527466535568237
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,256,8192,0.016838399569193523
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,256,10240,0.024297600984573363
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,256,7168,0.005098666747411092
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,256,7168,0.018691200017929076
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,256,8192,0.020662399133046468
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,256,6144,0.005086933573087057
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,256,6144,0.01656426688035329
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,256,7168,0.018952532609303793
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,256,5120,0.00469760000705719
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,256,5120,0.016556800405184428
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,256,6144,0.017322667439778647
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,256,4096,0.005031466484069824
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,256,4096,0.016546133160591125
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,256,5120,0.015339733163515726
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,256,3584,0.005083733300367991
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,256,3584,0.015293866395950317
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,256,4096,0.013649066289265951
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,256,3072,0.004646400113900503
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,256,3584,0.012469333410263062
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,256,3072,0.014482133587201438
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,256,2560,0.005016533533732096
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,256,2560,0.01446293294429779
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,256,3072,0.011222400267918905
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,256,2048,0.004748799900213877
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,256,2560,0.01123520036538442
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,256,2048,0.014578133821487427
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,256,1536,0.0038101332883040107
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,256,1536,0.014596266547838846
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,256,1536,0.008779733379681905
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,256,1024,0.0030805334448814393
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,256,65536,0.12071893215179444
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,256,1024,0.014550399780273438
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,256,768,0.0030037333567937215
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,256,768,0.014548266927401224
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,256,1024,0.008363733688990276
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,256,512,0.0026538667579491934
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,256,768,0.007534933090209961
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,256,512,0.014457600315411887
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,256,256,0.0022858666876951856
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,256,512,0.007098666826883952
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,256,256,0.014419200023015341
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,256,128,0.002243199944496155
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,256,256,0.006720000008742015
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,256,128,0.014507733782132468
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,256,64,0.0021322667598724365
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,256,128,0.0063498665889104204
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,256,64,0.0145087997118632
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,256,32,0.002152533332506816
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,128,65536,0.009527466694513957
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,256,32,0.014528000354766845
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,128,65536,0.019163733720779418
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,128,16384,0.005402666827042898
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,128,16384,0.018663465976715088
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,128,16384,0.034976001580556235
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,128,12288,0.005058133105436961
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,128,12288,0.018184532721837364
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,128,12288,0.027589333057403565
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,128,10240,0.005444266895453135
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,128,10240,0.017072000106175742
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,128,10240,0.024330667654673257
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,128,8192,0.005083733300367991
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,128,8192,0.016798933347066246
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,128,8192,0.0206496000289917
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,128,7168,0.005095466474692027
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,128,65536,0.12089280287424724
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,128,7168,0.01653439998626709
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,128,6144,0.004691199958324432
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,128,6144,0.016547200083732606
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,128,7168,0.019000534216562906
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,128,5120,0.004282666742801667
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,128,6144,0.016961065928141277
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,128,5120,0.016612266500790916
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,128,4096,0.004295466840267182
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,128,4096,0.01653333306312561
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,128,5120,0.015686399737993875
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,128,3584,0.0042463997999827065
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,128,4096,0.013640532890955607
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,128,3584,0.015842133760452272
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,128,3072,0.004171733558177948
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,128,3072,0.014642133315404256
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,128,3584,0.01283519963423411
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,128,2560,0.004635733366012573
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,128,2560,0.014862933754920959
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,128,3072,0.011646933356920878
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,128,2048,0.004220800101757049
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,128,2048,0.014568533500035605
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,128,2560,0.010854400197664897
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,128,1536,0.003487999985615412
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,128,2048,0.009616000453631084
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,128,1536,0.014538666605949402
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,128,1024,0.003036800026893616
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,128,1024,0.014408533771832785
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,128,1536,0.009133866429328919
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,128,768,0.002677333354949951
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,128,1024,0.007997866471608479
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,128,768,0.014145066340764364
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,128,768,0.007896533111731212
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,128,512,0.002570666621128718
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,128,512,0.014233600099881491
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,128,512,0.007483733197053273
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,128,256,0.002170666555563609
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,128,256,0.01446613371372223
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,128,256,0.006885333359241486
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,128,128,0.0021888000269730886
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,128,128,0.014075733224550881
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,128,128,0.006714666883150737
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,128,64,0.0018346666047970455
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,128,64,0.014432000120480857
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,128,32,0.0018634666999181111
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,128,32,0.0144896000623703
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,64,65536,0.006735999882221222
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,64,65536,0.01949013272921244
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,64,16384,0.005916800101598104
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,64,16384,0.017846399545669557
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,64,12288,0.005098666747411092
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,64,12288,0.017842133839925133
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,64,10240,0.004702933132648468
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,64,10240,0.016563199957211814
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,64,8192,0.004670933385690053
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,64,8192,0.016698666413625083
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,64,7168,0.004620799918969473
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,64,7168,0.01665493349234263
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,64,6144,0.004266666869322458
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,64,5120,0.004242133100827535
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,64,5120,0.01667520006497701
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,64,4096,0.004310399790604909
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,64,4096,0.016641066471735636
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,64,3584,0.004626133541266123
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,64,3584,0.016459733247756958
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,64,3072,0.004273066421349844
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,64,3072,0.015734400351842245
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,64,2560,0.004680533210436503
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,64,2560,0.015009066462516785
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,64,2048,0.004294399917125702
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,64,2048,0.014692266782124838
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,64,1536,0.0034442665676275887
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,64,1536,0.01453013320763906
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,64,1024,0.003005866706371307
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,64,1024,0.014512000481287637
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,64,768,0.0026208000878492994
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,64,768,0.014484266440073649
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,64,512,0.0025973332424958544
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,64,512,0.014391466975212097
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,64,256,0.002164266755183538
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,64,256,0.014473600188891092
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,64,128,0.0021930667261282604
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,64,128,0.014379733800888061
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,64,64,0.001911466692884763
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,64,64,0.014102400342623393
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,64,32,0.0019765333582957584
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,64,32,0.014417066176732381
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,32,65536,0.006739200154940288
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,32,65536,0.020939733584721884
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,32,16384,0.0054282665252685545
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,32,16384,0.01732906699180603
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,32,12288,0.005436799923578898
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,32,12288,0.016530133287111917
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,32,10240,0.005047466854254405
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,32,10240,0.016901334126790367
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,32,8192,0.004276266694068909
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,32,8192,0.016551466782887776
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,32,7168,0.004695466657479604
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,32,7168,0.017097600301106772
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,32,6144,0.004300799965858459
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,32,6144,0.016545066237449647
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,32,5120,0.004261333247025808
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,32,5120,0.016614400347073875
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,32,4096,0.004198400179545084
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,32,4096,0.01685439944267273
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,32,3584,0.004321066538492838
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,32,3584,0.016545066237449647
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,32,3072,0.004222933451334635
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,32,3072,0.014997333288192749
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,32,2560,0.0045621335506439206
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,32,2560,0.014781866470972696
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,32,2048,0.004265599946180979
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,32,2048,0.014737066626548768
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,32,1536,0.0035285333792368567
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,32,1536,0.014430933197339377
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,32,1024,0.0029653333127498626
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,32,1024,0.014409599701563516
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,32,768,0.002621866762638092
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,32,768,0.014456533392270408
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,32,512,0.0022357332209746042
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,32,512,0.014472533265749613
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,32,256,0.0021888000269730886
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,32,256,0.014452266693115234
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,32,128,0.002198400100072225
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,32,128,0.014452266693115234
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,32,64,0.0018496000518401462
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,32,64,0.014365866780281067
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,2,32,32,0.0020853333175182343
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,32,32,0.014094932874043783
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,65536,16384,0.322602653503418
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,65536,16384,0.17335786819458007
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,2560,512,0.007523199915885926
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,65536,12288,0.2328554630279541
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,65536,12288,0.1355413277943929
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,2,256,2048,0.009954133629798889
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,65536,10240,0.19659199714660644
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,65536,10240,0.11726826826731365
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,2,64,6144,0.016705065965652466
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,65536,8192,0.15955840746561686
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,65536,8192,0.10101760228474935
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,65536,12288,0.08858239650726318
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,65536,7168,0.13891305923461914
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,65536,7168,0.08281173706054687
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,65536,8192,0.06834133466084799
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,65536,6144,0.12235519886016846
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,65536,6144,0.07377920150756836
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,65536,10240,0.08212800025939941
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,65536,5120,0.10263786315917969
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,65536,5120,0.06377280155817668
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,65536,16384,0.1207477331161499
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,65536,4096,0.08431466420491537
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,65536,4096,0.055070932706197104
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,65536,5120,0.04769173463185628
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,65536,3584,0.07008852958679199
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,65536,3584,0.04931626717249553
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,65536,6144,0.05267093181610107
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,65536,3072,0.06045120159784952
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,65536,3072,0.04427946805953979
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,65536,7168,0.06074453194936117
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,65536,2560,0.051284265518188474
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,65536,2560,0.03952000141143799
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,65536,4096,0.039613866806030275
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,65536,2048,0.042046932379404704
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,65536,2048,0.03537600040435791
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,65536,3584,0.03591146469116211
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,65536,1536,0.032110933462778726
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,65536,1536,0.029206399122873945
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,65536,2560,0.02640213370323181
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,65536,1024,0.022669865687688192
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,65536,1024,0.024752000967661537
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,65536,3072,0.03059626619021098
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,65536,768,0.017314134041468303
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,65536,768,0.02272426684697469
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,65536,2048,0.023669334252675374
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,65536,512,0.012578133742014566
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,65536,512,0.020555732647577922
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,65536,768,0.013763200243314108
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,65536,256,0.007546666761239369
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,65536,256,0.016598400473594666
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,65536,1024,0.016158933440844216
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,65536,128,0.0050453335046768185
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,65536,128,0.01556373337904612
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,65536,128,0.00848640004793803
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,65536,64,0.004257066547870636
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,65536,64,0.015379200379053751
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,65536,32,0.004194133480389913
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,65536,32,0.016116266449292503
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,16384,65536,0.3188405354817708
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,16384,65536,0.17578986485799153
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,65536,256,0.009571199615796406
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,16384,16384,0.0892736037572225
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,16384,16384,0.059380265076955166
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,65536,512,0.011684266726175944
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,16384,12288,0.06553706725438437
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,16384,12288,0.045422931512196854
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,65536,1536,0.020002132654190062
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,16384,10240,0.05599466562271118
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,16384,10240,0.03973013162612915
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,16384,10240,0.03242666721343994
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,16384,8192,0.04655253489812215
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,16384,8192,0.03445333242416382
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,16384,12288,0.03551573355992635
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,16384,7168,0.035898665587107345
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,16384,7168,0.0329258660475413
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,16384,16384,0.04807573159535726
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,16384,6144,0.03131200075149536
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,16384,6144,0.029576534032821657
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,16384,8192,0.02810346682866414
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,16384,5120,0.026368000109990436
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,16384,5120,0.026774400472640993
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,16384,7168,0.022685867547988892
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,16384,4096,0.02179946700731913
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,16384,4096,0.024791467189788818
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,16384,6144,0.021934932470321654
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,16384,3584,0.01944213310877482
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,16384,3584,0.023900800943374635
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,16384,5120,0.02028800050417582
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,16384,3072,0.016923733552296958
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,16384,3072,0.02249600092569987
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,16384,4096,0.016237866878509522
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,16384,2560,0.014923733472824097
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,16384,2560,0.020951465765635172
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,16384,2560,0.012963199615478515
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,16384,2048,0.012362666924794515
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,16384,2048,0.020483199755350748
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,16384,3584,0.01624853312969208
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,16384,1536,0.010041600465774536
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,16384,1536,0.018428800503412883
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,16384,3072,0.014987732966740927
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,16384,1024,0.0075434664885203045
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,16384,1024,0.014724266529083253
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,16384,1024,0.009201066692670186
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,16384,1536,0.010387200117111205
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,16384,768,0.0059456000725428265
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,16384,768,0.0146506667137146
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,16384,512,0.0041184000670909885
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,16384,512,0.014473600188891092
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,16384,2048,0.011686399579048157
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,16384,256,0.0029866665601730345
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,16384,256,0.014444800217946372
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,16384,768,0.008393599589665731
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,16384,128,0.00258240004380544
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,16384,128,0.014388266205787658
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,16384,512,0.007892266909281413
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,16384,64,0.002197333425283432
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,16384,64,0.013173333803812661
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,16384,32,0.002234666546185811
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,16384,32,0.014197333653767904
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,12288,65536,0.2304543972015381
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,16384,128,0.006775466601053874
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,12288,65536,0.13789226214090983
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,16384,256,0.007233066856861115
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,12288,16384,0.06887786388397217
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,12288,16384,0.04907413323720296
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,12288,12288,0.04915306568145752
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,12288,12288,0.04166933298110962
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,12288,12288,0.03029973308245341
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,12288,10240,0.042404266198476155
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,12288,10240,0.03920213381449382
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,12288,16384,0.0384768009185791
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,12288,8192,0.035086933771769205
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,12288,8192,0.03391893307367961
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,12288,10240,0.026422399282455444
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,12288,7168,0.030696533123652142
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,12288,7168,0.03219626744588216
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,12288,8192,0.022641066710154215
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,12288,6144,0.027896533409754436
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,12288,6144,0.028698666890462236
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,12288,7168,0.020785067478815714
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,12288,5120,0.024318933486938477
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,12288,5120,0.027858134110768634
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,12288,6144,0.01909546653429667
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,12288,4096,0.02034453352292379
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,12288,4096,0.022627200682957968
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,12288,5120,0.017439999183019004
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,12288,3584,0.018207999070485432
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,12288,3584,0.020785067478815714
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,12288,4096,0.015296000242233276
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,12288,3072,0.016084266702334087
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,12288,3072,0.02065599958101908
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,12288,3584,0.01430293321609497
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,12288,2560,0.014248533050219217
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,12288,2560,0.020150399208068846
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,12288,3072,0.012973866860071816
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,12288,2048,0.010261332988739014
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,12288,2048,0.018196266889572144
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,12288,2560,0.012150399883588155
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,12288,1536,0.008382933338483174
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,12288,1536,0.016445866227149962
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,12288,1536,0.009633066256841023
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,12288,1024,0.006691200037797292
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,12288,1024,0.01453439990679423
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,12288,2048,0.010058666268984478
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,12288,768,0.004305066665013632
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,12288,768,0.014468266566594442
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,12288,1024,0.008411733309427898
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,12288,512,0.0036906667053699495
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,12288,512,0.014568533500035605
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,12288,768,0.007980800171693166
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,12288,256,0.0029472000896930695
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,12288,256,0.014451199769973755
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,12288,512,0.007990399996439617
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,12288,128,0.002583466718594233
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,12288,128,0.014497066537539164
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,12288,256,0.007125333448251088
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,12288,64,0.0021685334543387097
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,12288,64,0.013714133699735006
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,12288,32,0.0021578667064507803
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,12288,32,0.012984533111254373
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,12288,128,0.006716800232728322
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,10240,65536,0.19265599250793458
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,10240,65536,0.11723840236663818
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,10240,16384,0.05819626649220785
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,10240,16384,0.04351360003153483
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,10240,16384,0.03810880184173584
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,10240,12288,0.04099520047505696
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,10240,12288,0.037767465909322104
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,10240,12288,0.02967039942741394
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,10240,10240,0.03570666710535685
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,10240,10240,0.03459839820861817
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,16384,65536,0.15434133211771647
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,10240,8192,0.029333333174387615
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,10240,8192,0.030067199468612672
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,10240,10240,0.02648746569951375
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,10240,7168,0.02649173339207967
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,10240,7168,0.029942399263381957
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,10240,8192,0.0223146657148997
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,10240,6144,0.024178133408228556
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,10240,6144,0.026601600646972656
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,12288,65536,0.13274239699045817
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,10240,5120,0.020846933126449585
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,10240,5120,0.025526400407155352
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,10240,7168,0.021384533246358237
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,10240,4096,0.016872533162434897
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,10240,4096,0.022657066583633423
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,10240,6144,0.019019732872645058
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,10240,3584,0.016331733266512553
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,10240,3584,0.020916267236073812
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,10240,5120,0.01702079971631368
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,10240,3072,0.014664533734321594
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,10240,3072,0.02073813279469808
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,10240,4096,0.015128533045450846
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,10240,2560,0.013121066490809121
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,10240,2560,0.02060906688372294
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,10240,3584,0.013794133067131042
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,10240,2048,0.010032000144322713
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,10240,2048,0.018565332889556883
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,10240,3072,0.012526933352152506
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,10240,1536,0.008737066388130188
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,10240,1536,0.016680532693862916
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,10240,2560,0.011708799997965496
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,10240,1024,0.0067007998625437425
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,10240,1024,0.016613333423932394
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,10240,1024,0.008483200271924337
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,10240,768,0.005487999816735586
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,10240,768,0.014646400014559427
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,10240,1536,0.00925546685854594
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,10240,512,0.004247466723124186
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,10240,512,0.014619732896486918
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,10240,2048,0.010040533542633057
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,10240,256,0.002569599946339925
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,10240,256,0.013863466183344521
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,10240,768,0.008368000388145447
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,10240,512,0.007558399935563405
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,10240,128,0.0022410665949185687
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,10240,128,0.01416853368282318
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,10240,64,0.0022154666483402254
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,10240,64,0.014153599739074707
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,10240,256,0.007097599903742473
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,10240,32,0.002238933245340983
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,10240,32,0.012982400258382163
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,10240,128,0.006647466619809468
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,8192,65536,0.1564789295196533
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,8192,16384,0.047680000464121505
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,8192,65536,0.0958400011062622
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,8192,16384,0.035563735167185466
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,8192,12288,0.0352725346883138
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,8192,12288,0.030624000231424968
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,8192,12288,0.029315199454625445
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,8192,10240,0.029949865738550824
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,8192,10240,0.028360533714294433
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,8192,16384,0.036645332972208663
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,8192,8192,0.022951465845108033
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,8192,8192,0.025041067600250246
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,8192,10240,0.02602880001068115
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,8192,7168,0.019871999820073448
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,8192,7168,0.023939200242360435
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,8192,8192,0.021964800357818604
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,8192,6144,0.017849600315093993
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,8192,6144,0.022705066204071044
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,8192,7168,0.020281600952148437
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,8192,5120,0.015145599842071533
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,8192,5120,0.02135146657625834
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,10240,65536,0.12894506454467775
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,8192,4096,0.012619733810424805
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,8192,4096,0.02020053267478943
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,8192,6144,0.01861013372739156
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,8192,3584,0.01162986656030019
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,8192,3584,0.01867199937502543
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,8192,5120,0.016923733552296958
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,8192,3072,0.010241066416104633
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,8192,3072,0.018544000387191773
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,8192,3584,0.013330133756001792
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,8192,2560,0.00913279950618744
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,8192,2560,0.016582399606704712
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,8192,4096,0.014151466886202493
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,8192,2048,0.00782400021950404
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,8192,2048,0.01653866668542226
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,8192,3072,0.01209493378798167
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,8192,1536,0.006007466713587443
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,8192,1536,0.015541332960128783
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,8192,2560,0.011242666840553283
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,8192,1024,0.004491733511288961
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,8192,1024,0.014487466216087342
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,8192,2048,0.010028800368309021
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,8192,768,0.003803733239571253
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,8192,768,0.0144896000623703
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,8192,1536,0.009300266702969868
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,8192,512,0.0033781332274278007
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,8192,512,0.014479999740918478
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,8192,1024,0.008372267087300617
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,8192,256,0.0025781333446502686
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,8192,256,0.013740799824396768
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,8192,768,0.007572266459465027
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,8192,128,0.002162133405605952
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,8192,512,0.0075093333919843035
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,8192,128,0.013944533467292786
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,8192,64,0.002178133279085159
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,8192,64,0.0137472003698349
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,8192,32,0.002181333303451538
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,8192,256,0.007095466554164887
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,8192,32,0.014133333166440328
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,7168,65536,0.13629973729451497
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,8192,128,0.0067562664548556015
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,7168,65536,0.08599039713541666
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,7168,16384,0.03993813196818034
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,7168,16384,0.03536213239034017
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,7168,12288,0.03113493323326111
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,7168,12288,0.027794132630030315
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,7168,12288,0.02926186720530192
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,7168,10240,0.026713599761327107
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,7168,10240,0.026280534267425538
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,7168,16384,0.03651626507441203
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,7168,8192,0.02206719915072123
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,7168,8192,0.02296853264172872
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,7168,10240,0.025704532861709595
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,7168,7168,0.019640533129374187
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,7168,7168,0.02272319992383321
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,7168,8192,0.022056533892949422
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,7168,6144,0.01593066652615865
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,7168,6144,0.02155733307202657
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,7168,7168,0.020665599902470907
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,7168,5120,0.013852799932161966
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,7168,5120,0.020694400866826376
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,7168,6144,0.01869973341623942
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,7168,4096,0.011602133512496948
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,7168,4096,0.018598399559656777
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,7168,5120,0.016616533199946083
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,7168,3584,0.010493866602579753
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,7168,3584,0.01849386692047119
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,7168,4096,0.014137599865595499
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,7168,3072,0.009543466567993163
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,7168,3072,0.016989866892496742
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,7168,3584,0.012900267044703165
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,7168,2560,0.008345599969228108
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,7168,2560,0.01654293338457743
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,7168,3072,0.012456533312797547
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,7168,2048,0.007239466905593872
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,7168,2048,0.014506666858990987
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,7168,2560,0.011313066879908244
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,7168,1536,0.005534933507442474
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,7168,1536,0.014566399653752646
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,7168,2048,0.010044800241788228
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,7168,1024,0.004356266558170318
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,7168,1024,0.014481066664059957
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,7168,1536,0.00920960009098053
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,7168,768,0.0037930667400360107
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,7168,768,0.01444586714108785
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,7168,1024,0.008338133494059246
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,7168,512,0.002996266633272171
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,7168,512,0.014454399545987448
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,7168,512,0.007186133166154225
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,7168,256,0.002551466723283132
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,7168,256,0.013795199990272521
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,7168,768,0.007904000083605449
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,7168,128,0.002229333420594533
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,7168,128,0.01362773378690084
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,7168,128,0.006683733562628429
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,7168,64,0.002169599880774816
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,7168,64,0.0132832000652949
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,7168,32,0.0021717332303524016
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,7168,32,0.012967466314633688
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,6144,65536,0.11847360134124756
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,6144,65536,0.07833920319875082
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,8192,65536,0.12757546901702882
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,6144,16384,0.036637866497039796
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,6144,16384,0.03285653392473857
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,6144,16384,0.03713066577911377
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,6144,12288,0.027965867519378663
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,6144,12288,0.028988800446192425
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,6144,12288,0.029286400477091475
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,6144,10240,0.023825067281723022
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,6144,10240,0.027884799242019653
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,7168,65536,0.12656959692637126
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,6144,8192,0.01936533252398173
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,6144,8192,0.022712532679239908
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,6144,10240,0.025809067487716674
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,6144,7168,0.01773759921391805
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,6144,7168,0.022247467438379923
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,6144,8192,0.02200746734937032
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,6144,6144,0.015412267049153647
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,6144,6144,0.02071146567662557
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,6144,7168,0.020257065693537392
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,6144,5120,0.01405333379904429
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,6144,5120,0.020539732774098714
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,6144,6144,0.0182261327902476
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,6144,4096,0.01071999967098236
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,6144,4096,0.01875200072924296
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,6144,5120,0.016567466656366985
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,6144,3584,0.009669333696365356
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,6144,3584,0.018613332509994508
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,6144,3584,0.013327999909718832
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,6144,3072,0.008869333068529765
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,6144,3072,0.01686720053354899
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,6144,4096,0.013716266552607218
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,6144,2560,0.007973333199818928
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,6144,2560,0.016484266519546507
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,6144,3072,0.01202880044778188
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,6144,2048,0.006925866504510244
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,6144,2048,0.015709867080052696
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,6144,2560,0.011193600296974183
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,6144,1536,0.00524479995171229
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,6144,1536,0.015191466609636942
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,6144,2048,0.010453333457310993
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,6144,1024,0.00425493319829305
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,6144,1024,0.014598400394121806
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,6144,1536,0.009180800120035807
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,6144,768,0.0038143999874591826
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,6144,768,0.014587733149528503
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,6144,1024,0.008377599716186523
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,6144,512,0.0032992000381151833
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,6144,512,0.014409599701563516
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,6144,768,0.007659733295440674
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,6144,256,0.0026186667382717133
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,6144,256,0.014401066303253173
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,6144,256,0.006807466844717662
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,6144,512,0.007533866663773854
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,6144,128,0.002209066599607468
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,6144,128,0.014493866761525472
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,6144,64,0.00211519996325175
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,6144,64,0.014412800470987955
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,6144,32,0.002125866711139679
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,6144,32,0.014446933070818582
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,6144,128,0.006682666639486949
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,5120,65536,0.10055359999338787
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,5120,65536,0.06912213166554769
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,5120,16384,0.03227413296699524
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,5120,16384,0.030728532870610552
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,6144,65536,0.1270090659459432
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,5120,12288,0.024000000953674317
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,5120,12288,0.026743467648824053
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,5120,16384,0.03623466491699219
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,5120,10240,0.021076265970865884
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,5120,10240,0.02532266577084859
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,7168,256,0.0070933332045873005
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,5120,8192,0.01821440060933431
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,5120,8192,0.022474666436513267
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,5120,10240,0.025645865996678667
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,5120,7168,0.016165332992871602
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,5120,7168,0.020845866203308104
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,5120,12288,0.029260800282160444
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,5120,6144,0.014363732933998109
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,5120,6144,0.019568000237147012
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,5120,8192,0.021734400590260824
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,5120,5120,0.012467199563980102
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,5120,5120,0.01898026665051778
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,5120,7168,0.019860267639160156
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,5120,4096,0.010016000270843506
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,5120,4096,0.018244266510009766
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,5120,6144,0.017924267053604125
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,5120,3584,0.008814932902654012
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,5120,3584,0.016637866695721946
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,5120,5120,0.015475199619928996
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,5120,3072,0.007924266656239827
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,5120,3072,0.01653333306312561
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,5120,3584,0.012957866986592612
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,5120,2560,0.007170133292675018
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,5120,2560,0.01454080045223236
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,5120,4096,0.013384532928466798
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,5120,2048,0.005966933568318685
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,5120,2048,0.014651733636856078
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,5120,2560,0.01125333309173584
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,5120,1536,0.005054933329423269
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,5120,2048,0.009941333532333374
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,5120,1536,0.014453333616256715
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,5120,1024,0.004228266576925913
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,5120,1024,0.014476799964904785
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,5120,3072,0.011865599950154623
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,5120,768,0.0034634667138258614
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,5120,768,0.014494933684666953
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,5120,1024,0.007977599898974102
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,5120,512,0.0030048000315825146
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,5120,512,0.014486400286356607
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,5120,768,0.007560533285140991
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,5120,256,0.002570666621128718
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,5120,256,0.013719466328620911
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,5120,1536,0.009197866916656494
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,5120,128,0.0022272000710169475
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,5120,128,0.014495999614397685
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,5120,256,0.007123200098673503
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,5120,64,0.002145066608985265
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,5120,64,0.01362986663977305
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,5120,32,0.002160000056028366
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,5120,32,0.013718400398890176
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,5120,128,0.006720000008742015
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,5120,512,0.007467733323574066
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,4096,65536,0.0820362647374471
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,4096,16384,0.02579946716626485
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,4096,65536,0.06027733484903971
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,4096,16384,0.026732800404230754
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,4096,12288,0.020321067174275717
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,4096,12288,0.024706133206685386
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,4096,12288,0.02861119906107585
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,4096,10240,0.016901334126790367
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,4096,10240,0.022666666905085245
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,4096,16384,0.03624106645584106
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,4096,8192,0.01462399959564209
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,4096,8192,0.02296533385912577
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,4096,10240,0.025204267104466754
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,4096,7168,0.012943999965985617
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,4096,7168,0.0249674657980601
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,4096,8192,0.021667200326919555
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,4096,6144,0.01218986709912618
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,4096,6144,0.02289066712061564
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,4096,7168,0.01953386664390564
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,4096,5120,0.010846933722496033
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,4096,5120,0.02064639925956726
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,4096,6144,0.017778132359186807
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,4096,4096,0.009545600414276123
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,4096,4096,0.02056106726328532
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,4096,5120,0.015737600127855935
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,4096,3584,0.012407466769218445
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,4096,3584,0.018564265966415406
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,4096,4096,0.013788800438245138
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,4096,3072,0.010775466759999592
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,4096,3072,0.01856106718381246
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,4096,3584,0.012915199995040894
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,4096,2560,0.009546666344006857
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,4096,2560,0.017480534315109254
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,4096,3072,0.011749333143234253
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,4096,2048,0.008258133133252462
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,4096,2048,0.01662826637427012
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,4096,2560,0.010830932855606079
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,4096,1536,0.006552533308664958
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,4096,1536,0.015824000040690102
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,4096,1536,0.00920746624469757
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,4096,1024,0.005063466727733612
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,4096,1024,0.014473600188891092
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,4096,2048,0.009993599851926167
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,4096,768,0.004298666616280874
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,4096,768,0.014588800072669984
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,4096,1024,0.00839573343594869
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,4096,512,0.0034634667138258614
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,4096,768,0.007921066880226136
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,4096,512,0.014107732971509298
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,4096,256,0.0025813333690166474
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,4096,256,0.014222932855288186
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,4096,256,0.00676800012588501
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,4096,128,0.002161066730817159
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,4096,128,0.013193600376447043
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,4096,512,0.007520000139872233
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,4096,64,0.00223786657055219
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,4096,64,0.01442346672217051
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,4096,32,0.0022282667458057405
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,4096,128,0.006725333134333293
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,4096,32,0.01448319951693217
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,3584,65536,0.07250239849090576
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,3584,65536,0.05417173306147257
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,3584,16384,0.022210133075714112
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,3584,16384,0.02564479907353719
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,5120,65536,0.12761493523915607
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,3584,12288,0.017528533935546875
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,3584,12288,0.024925865729649863
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,3584,12288,0.029337600866953535
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,3584,10240,0.015398400028546652
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,3584,10240,0.022961066166559855
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,3584,16384,0.03628160158793132
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,3584,8192,0.012872533003489176
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,3584,8192,0.021615999937057494
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,3584,10240,0.025272534290949507
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,3584,7168,0.012771200140317282
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,3584,7168,0.02359573245048523
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,3584,8192,0.021082667509714763
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,3584,6144,0.010863999525705974
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,3584,6144,0.02100800077120463
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,3584,7168,0.01934400002161662
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,3584,5120,0.010098133484522502
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,3584,5120,0.020651733875274657
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,3584,6144,0.017387733856836955
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,3584,4096,0.008757332960764568
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,3584,4096,0.018684800465901694
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,3584,4096,0.013351466258366904
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,3584,3584,0.011319466431935628
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,3584,3584,0.01854506731033325
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,3584,5120,0.015738667050997416
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,3584,3072,0.010886399944623312
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,3584,3072,0.017842133839925133
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,4096,65536,0.12653226852416993
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,3584,2560,0.00956053336461385
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,3584,2560,0.016563199957211814
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,3584,3072,0.012053333719571431
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,3584,2048,0.00811413327852885
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,3584,2048,0.016481066743532814
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,3584,3584,0.012840533256530761
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,3584,1536,0.006678399940331777
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,3584,1536,0.015734400351842245
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,3584,1536,0.00881706674893697
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,3584,1024,0.005123200019200643
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,3584,1024,0.01657600005467733
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,3584,2048,0.01009386678536733
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,3584,768,0.004279466470082601
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,3584,768,0.014498133460680643
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,3584,2560,0.011194666226704914
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,3584,512,0.005017599960168203
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,3584,512,0.014432000120480857
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,3584,1024,0.008378666639328004
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,3584,256,0.002587733417749405
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,3584,256,0.014344533284505209
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,3584,512,0.00745066652695338
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,3584,128,0.002195200075705846
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,3584,128,0.014123732844988504
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,3584,768,0.007886933286984761
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,3584,256,0.007043200234572093
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,3584,64,0.0021930667261282604
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,3584,64,0.014442666371663412
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,3584,128,0.006668800115585327
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,3584,32,0.0022101332743962605
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,3584,32,0.012959999839464822
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,3072,65536,0.06379520098368327
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,3072,65536,0.048189866542816165
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,3072,16384,0.02020906607309977
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,3072,16384,0.024410667022069295
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,3072,12288,0.015940266847610473
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,3072,12288,0.023694932460784912
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,3072,12288,0.028869332869847615
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,3072,10240,0.013754666845003764
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,3072,10240,0.021849600474039714
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,3072,16384,0.036246399084726974
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,3072,8192,0.012205866972605388
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,3072,8192,0.020623999834060668
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,3072,10240,0.025259733200073242
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,3072,7168,0.011015466849009196
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,3072,7168,0.022692267100016275
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,3072,8192,0.021102933088938396
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,3072,6144,0.010411733388900756
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,3072,6144,0.020645334323247274
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,3072,7168,0.019471999009450278
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,3072,5120,0.009149866302808125
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,3072,5120,0.018633600076039633
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,3072,6144,0.017460266749064125
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,3072,4096,0.007916800181070964
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,3072,4096,0.018573866287867228
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,3072,5120,0.015800533692042033
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,3072,3584,0.010104533036549885
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,3072,3584,0.018680532773335777
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,3072,4096,0.013710932930310568
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,3072,3072,0.008954667051633199
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,3072,3072,0.017061332861582436
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,3072,3072,0.01204266647497813
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,3072,2560,0.008473599950472515
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,3072,2560,0.016546133160591125
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,3584,65536,0.1262943983078003
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,3072,2048,0.007405866682529449
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,3072,2048,0.016522666811943053
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,3072,2560,0.011291733384132386
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,3072,1536,0.006382933259010315
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,3072,1536,0.01502293348312378
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,3072,2048,0.009662933150927226
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,3072,1024,0.005029333134492239
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,3072,1024,0.014459733168284097
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,3072,1536,0.00928106705347697
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,3072,768,0.0042250668009122215
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,3072,1024,0.008331733445326488
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,3072,768,0.014458666245142618
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,3072,512,0.0034527999659379324
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,3072,512,0.014455466469128927
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,3072,512,0.007562666634718577
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,3072,256,0.0025941332181294756
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,3072,256,0.014148267110188803
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,3072,768,0.0075082664688428235
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,3072,128,0.002218666672706604
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,3072,128,0.013171199957529703
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,3072,256,0.007177599767843883
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,3072,128,0.006698666512966156
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,3072,64,0.0021717332303524016
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,3072,32,0.0021898667017618816
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,3072,64,0.013097600142161051
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,3072,32,0.0141184002161026
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,2560,65536,0.05498239994049072
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,2560,65536,0.044898132483164474
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,2560,16384,0.018266665935516357
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,2560,16384,0.0243285338083903
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,2560,16384,0.035930665334065755
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,2560,12288,0.014519466956456503
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,2560,12288,0.022744532426198324
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,2560,12288,0.028531199693679808
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,2560,10240,0.012105600039164225
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,2560,10240,0.020878932873407998
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,3072,65536,0.12604479789733886
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,2560,8192,0.011026133100191753
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,2560,8192,0.02057173252105713
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,2560,10240,0.02477226654688517
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,2560,7168,0.009755733609199523
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,2560,7168,0.018602667252222697
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,2560,8192,0.02062293291091919
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,2560,6144,0.00876800020535787
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,2560,6144,0.019610667228698732
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,2560,7168,0.019128533204396565
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,2560,5120,0.008039466540018718
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,2560,5120,0.019016534090042114
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,2560,5120,0.015785599748293556
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,2560,4096,0.006750933329264323
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,2560,4096,0.017086933056513466
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,2560,6144,0.017388800779978432
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,2560,3584,0.006297599772612255
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,2560,3584,0.018620800971984864
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,2560,3584,0.01253546675046285
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,2560,3072,0.007948799928029378
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,2560,3072,0.018279467026392618
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,2560,4096,0.013757866621017457
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,2560,2560,0.007133866846561432
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,2560,2560,0.016489600141843162
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,2560,3072,0.012134400010108948
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,2560,2048,0.007513600091139476
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,2560,2048,0.016503467162450155
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,2560,2560,0.011210667093594869
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,2560,1536,0.0063178668419520065
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,2560,1536,0.014486400286356607
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,2560,2048,0.010021332899729412
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,2560,1024,0.005141333242257436
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,2560,1024,0.014681599537531533
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,2560,1536,0.009565866986910502
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,2560,768,0.004257066547870636
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,2560,1024,0.008335999647776286
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,2560,768,0.014460800091425577
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,2560,512,0.003419733295838038
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,2560,512,0.014484266440073649
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,2560,512,0.007603199779987335
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,2560,256,0.002603733291228612
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,2560,768,0.00796693315108617
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,2560,256,0.01288426617781321
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,2560,128,0.0022154666483402254
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,2560,128,0.013286399841308593
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,2560,256,0.006734933455785115
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,2560,64,0.002123733361562093
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,2560,128,0.007067733506361644
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,2560,32,0.0021888000269730886
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,2560,64,0.014499200383822122
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,2560,32,0.01404800017674764
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,2048,65536,0.04538346529006958
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,2048,65536,0.038832000891367596
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,2048,16384,0.015427199999491372
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,2048,16384,0.022974934180577597
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,2560,65536,0.1256170670191447
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,2048,12288,0.01195093293984731
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,2048,12288,0.020629332462946574
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,2048,16384,0.035496532917022705
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,2048,10240,0.010332799951235453
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,2048,10240,0.020143999656041463
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,2048,10240,0.024485333760579427
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,2048,8192,0.009206400314966837
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,2048,8192,0.018636800845464072
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,2048,8192,0.020729599396387736
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,2048,7168,0.00876586635907491
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,2048,7168,0.01834986607233683
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,2048,7168,0.019077332814534505
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,2048,6144,0.007971199850241344
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,2048,6144,0.018611200650533042
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,2048,6144,0.01768959959348043
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,2048,5120,0.006737066805362702
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,2048,5120,0.016586666305859886
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,2048,5120,0.015726932883262636
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,2048,4096,0.005875200033187866
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,2048,4096,0.01857173244158427
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,2048,4096,0.013700266679128012
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,2048,3584,0.0054954667886098225
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,2048,3584,0.018130133549372353
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,2048,3584,0.012889599800109864
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,2048,3072,0.006692266464233399
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,2048,3072,0.017043199141820273
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,2048,3072,0.01172693371772766
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,2048,2560,0.0058559998869895935
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,2048,2560,0.016522666811943053
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,2048,2560,0.010865066448847454
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,2048,2048,0.007590400179227193
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,2048,2048,0.016526933511098227
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,2048,2048,0.010010666648546855
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,2048,1536,0.006284800171852112
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,2048,1536,0.014660267035166421
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,2048,1536,0.008716799815495809
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,2048,1024,0.004611200094223023
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,2048,1024,0.014501333236694336
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,2048,1024,0.0083146666487058
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,2048,768,0.0038133333126703895
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,2048,768,0.014521599809328715
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,2048,768,0.0075914666056633
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,2048,512,0.0033781332274278007
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,2048,512,0.014505599935849508
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,2048,512,0.007176533341407776
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,2048,256,0.0026047999660174055
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,2048,256,0.01448319951693217
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,2048,256,0.0071381335457166035
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,2048,128,0.0022111999491850535
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,2048,128,0.013776000340779623
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,2048,128,0.006630399823188781
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,2048,64,0.002145066608985265
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,2048,64,0.014505599935849508
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,2048,32,0.002149333308140437
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,2048,32,0.014424533645311991
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,1536,65536,0.035658665498097736
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,1536,65536,0.03367040157318115
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,2048,65536,0.12417173385620117
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,1536,16384,0.012513066331545511
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,1536,16384,0.02476159930229187
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,1536,16384,0.035641598701477054
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,1536,12288,0.015432533621788026
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,1536,12288,0.023030400276184082
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,1536,12288,0.028064000606536865
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,1536,10240,0.013321600357691445
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,1536,10240,0.018465065956115724
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,1536,10240,0.024860799312591553
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,1536,8192,0.011668266852696736
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,1536,8192,0.018739199638366698
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,1536,8192,0.02071573336919149
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,1536,7168,0.010787199934323628
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,1536,7168,0.01863573392232259
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,1536,7168,0.01914986570676168
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,1536,6144,0.009607467055320739
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,1536,6144,0.018588799238204955
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,1536,6144,0.017443199952443443
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,1536,5120,0.006265600025653839
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,1536,5120,0.017528533935546875
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,1536,5120,0.015389866630236306
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,1536,4096,0.006107733150323232
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,1536,4096,0.018629332383473717
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,1536,65536,0.12355519930521648
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,1536,4096,0.013333333532015481
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,1536,3584,0.005509333312511444
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,1536,3584,0.016796799500783284
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,1536,3072,0.006105599800745646
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,1536,3072,0.01655466655890147
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,1536,3072,0.011665067076683045
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,1536,2560,0.00547733356555303
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,1536,3584,0.012828800082206725
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,1536,2560,0.01649173299471537
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,1536,2048,0.00639466643333435
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,1536,2048,0.01648319959640503
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,1536,2560,0.010405332843462626
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,1536,1536,0.005373866856098175
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,1536,1536,0.014506666858990987
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,1536,2048,0.009577600161234538
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,1536,1024,0.004228266576925913
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,1536,1024,0.014593066771825156
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,1536,1024,0.00795839975277583
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,1536,1536,0.009131733576456707
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,1536,768,0.0035818666219711304
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,1536,768,0.01444586714108785
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,1536,512,0.00297173336148262
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,1536,512,0.014469333489735923
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,1536,768,0.007538133362929027
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,1536,512,0.0071274667978286745
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,1536,256,0.002616533388694127
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,1536,256,0.014465066790580749
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,1536,128,0.002218666672706604
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,1536,128,0.014013866583506266
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,1536,256,0.007055999835332234
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,1536,64,0.0021536000072956084
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,1536,128,0.006295466423034668
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,1536,64,0.014494933684666953
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,1536,32,0.0022272000710169475
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,1536,32,0.014439466595649719
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,1024,65536,0.024892799059549966
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,1024,65536,0.028616533676783247
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,1024,16384,0.013091199596722922
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,1024,16384,0.020801067352294922
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,1024,16384,0.03545600175857544
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,1024,12288,0.011699199676513672
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,1024,12288,0.018742400407791137
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,1024,12288,0.02807253400484721
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,1024,10240,0.00958079993724823
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,1024,10240,0.01856213410695394
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,1024,10240,0.02445333401362101
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,1024,8192,0.009150933225949604
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,1024,8192,0.018675200144449868
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,1024,8192,0.020665599902470907
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,1024,7168,0.008113066852092742
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,1024,7168,0.01845653255780538
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,1024,7168,0.019054933389027914
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,1024,6144,0.0062496001521746315
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,1024,6144,0.01662613352139791
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,1024,6144,0.017399466037750243
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,1024,5120,0.005497600138187409
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,1024,5120,0.016858667135238647
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,1024,5120,0.015740799903869628
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,1024,4096,0.005635199944178263
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,1024,4096,0.018567466735839845
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,1024,4096,0.013345066706339517
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,1024,3584,0.005459199845790863
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,1024,3584,0.018971733252207437
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,1024,65536,0.1219551960627238
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,1024,3072,0.0055285334587097164
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,1024,3072,0.0166101336479187
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,1024,3584,0.0124917338291804
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,1024,2560,0.005077333251635233
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,1024,2560,0.01667520006497701
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,1024,3072,0.011942399541536967
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,1024,2048,0.0070709332823753355
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,1024,2048,0.016492799917856852
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,1024,2560,0.010413866241772969
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,1024,1536,0.005922133227189382
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,1024,1536,0.015442132949829102
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,1024,1536,0.009148800373077392
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,1024,1024,0.004661333560943603
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,1024,2048,0.009929600358009338
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,1024,1024,0.014603733023007711
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,1024,768,0.003857066730658213
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,1024,768,0.014545067151387533
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,1024,1024,0.008292266726493835
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,1024,768,0.007868800063927968
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,1024,512,0.0030602666238943735
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,1024,512,0.014510933558146158
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,1024,256,0.0026133333643277483
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,1024,512,0.007101866602897644
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,1024,256,0.014456533392270408
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,1024,256,0.006759466727574666
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,1024,128,0.0022240000466505687
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,1024,128,0.014456533392270408
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,1024,128,0.00628053347269694
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,1024,64,0.0022079999248186748
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,1024,64,0.014170666535695394
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,1024,32,0.002163200080394745
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,1024,32,0.014424533645311991
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,768,65536,0.02081813414891561
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,768,65536,0.031850665807724
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,768,16384,0.011776000261306763
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,768,16384,0.01919680039087931
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,768,16384,0.03512746493021647
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,768,12288,0.009469866752624512
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,768,12288,0.018641066551208497
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,768,12288,0.028092799584070842
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,768,10240,0.00806933343410492
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,768,10240,0.01898026665051778
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,3072,3584,0.012936533490816752
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,768,8192,0.007155199845631917
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,768,8192,0.01858773430188497
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,768,8192,0.020706133047739664
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,768,7168,0.005533866584300995
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,768,10240,0.02439893285433451
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,768,7168,0.018629332383473717
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,768,6144,0.005817600091298421
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,768,6144,0.017937066157658894
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,768,6144,0.01742186745007833
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,768,5120,0.005510400235652924
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,768,7168,0.018660267194112144
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,768,5120,0.017988266547520955
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,768,4096,0.006363733112812043
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,768,4096,0.01655893325805664
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,768,5120,0.015784533818562825
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,768,4096,0.013718400398890176
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,768,3584,0.005862399935722351
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,768,3584,0.016622933745384216
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,768,3072,0.005648000041643778
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,768,3072,0.01662720044453939
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,768,3584,0.012410666545232136
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,768,2560,0.00508480022350947
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,768,3072,0.012011733651161195
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,768,2560,0.016509866714477538
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,768,2048,0.004315733412901561
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,768,2048,0.01576640009880066
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,768,2560,0.010890666643778484
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,768,1536,0.0054293334484100345
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,768,2048,0.009573333462079366
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,768,1536,0.014705066879590353
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,768,1024,0.004250666499137879
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,768,1024,0.014519466956456503
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,768,1536,0.00916373332341512
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,768,1024,0.007942399879296621
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,768,768,0.014532267053922018
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,768,768,0.007627733548482259
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,768,512,0.0030080000559488933
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,768,512,0.014553599556287131
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,768,512,0.007269333302974701
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,768,256,0.002552533398071925
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,768,256,0.014537599682807923
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,768,256,0.007052800059318543
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,768,128,0.002238933245340983
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,768,128,0.014446933070818582
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,768,128,0.006807466844717662
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,768,64,0.0022154666483402254
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,768,64,0.014389333128929139
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,768,32,0.0021183999876181287
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,768,32,0.014364799857139588
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,512,65536,0.02884053389231364
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,512,65536,0.027189334233601887
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,768,65536,0.12143359978993733
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,512,16384,0.008360532919565837
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,512,16384,0.01685546636581421
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,512,16384,0.03466773430506388
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,512,12288,0.0075573335091273
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,512,12288,0.018389334281285606
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,512,12288,0.02768106659253438
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,512,10240,0.007067733506361644
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,512,10240,0.018492799997329713
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,512,10240,0.024375466505686443
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,512,8192,0.005115733544031779
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,512,8192,0.018557866414388023
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,512,8192,0.02066133419672648
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,512,7168,0.005086933573087057
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,512,7168,0.018283732732137046
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,512,7168,0.018921599785486857
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,512,6144,0.00586346685886383
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,512,6144,0.017412267128626504
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,512,65536,0.12115840117136638
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,512,5120,0.005973333120346069
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,512,6144,0.017375999689102174
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,512,5120,0.018644267320632936
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,512,4096,0.005527466535568237
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,512,4096,0.01735146641731262
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,512,5120,0.015280000368754067
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,512,3584,0.005432533224423727
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,512,4096,0.013289599617322286
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,512,3584,0.016548267006874083
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,512,3584,0.012522666652997335
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,512,3072,0.005093333125114441
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,512,3072,0.016569599509239197
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,512,3072,0.012126933534940083
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,512,2560,0.004700799783070883
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,512,2560,0.015421866377194723
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,512,2048,0.0048437332113583885
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,512,2048,0.014638933539390563
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,512,2560,0.011221333344777425
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,512,1536,0.0042250668009122215
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,512,1536,0.014685866236686707
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,512,2048,0.009648000200589497
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,512,1024,0.0033802665770053864
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,512,1024,0.014465066790580749
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,512,1536,0.009126399954160053
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,512,768,0.0029696000119050344
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,512,768,0.014428800344467163
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,512,1024,0.008297599852085114
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,512,512,0.0025898667673269907
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,512,512,0.014115200440088908
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,512,768,0.007955199976762136
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,512,256,0.0025770666698614756
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,512,512,0.007118933399518331
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,512,256,0.014401066303253173
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,512,128,0.002242133269707362
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,512,128,0.0144896000623703
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,512,256,0.0066453332702318835
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,512,64,0.0021407999098300933
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,512,64,0.014465066790580749
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,512,128,0.006308266520500183
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,512,32,0.002145066608985265
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,512,32,0.014493866761525472
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,256,65536,0.01657919983069102
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,256,65536,0.02285333275794983
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,256,16384,0.005533866584300995
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,256,16384,0.018604799111684164
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,256,16384,0.035045333703358966
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,256,12288,0.005037866532802582
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,256,12288,0.01662506659825643
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,256,12288,0.02797653277715047
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,256,10240,0.0058677335580190025
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,256,10240,0.017682133118311565
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,256,10240,0.02467306653658549
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,256,8192,0.005097599824269613
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,256,8192,0.018567466735839845
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,256,8192,0.020644267400105797
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,256,7168,0.005459199845790863
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,256,7168,0.017704532543818156
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,256,7168,0.01904746691385905
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,256,6144,0.005096533397833506
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,256,6144,0.01717653274536133
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,256,65536,0.12032639980316162
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,256,5120,0.004631466666857402
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,256,5120,0.01653333306312561
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,256,6144,0.017314134041468303
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,256,4096,0.00508480022350947
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,256,4096,0.01660586694876353
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,256,5120,0.01528320014476776
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,256,3584,0.005063466727733612
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,256,3584,0.015773866573969522
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,256,4096,0.01325440009435018
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,256,3072,0.004691199958324432
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,256,3072,0.01595626672108968
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,256,3584,0.01242026686668396
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,256,2560,0.005028266708056132
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,256,2560,0.014587733149528503
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,256,3072,0.011622400085131327
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,256,2048,0.004311466713746389
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,256,2048,0.014542933305104574
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,256,2560,0.01074026624361674
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,256,1536,0.0038015998899936674
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,256,1536,0.014484266440073649
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,256,2048,0.01032533347606659
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,256,1024,0.0030303999781608583
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,256,1024,0.014458666245142618
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,256,1536,0.009110400080680847
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,256,768,0.002796799937884013
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,256,768,0.014507733782132468
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,256,1024,0.008313600222269695
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,256,512,0.002629333237806956
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,256,512,0.014467199643452963
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,256,768,0.007477333148320516
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,256,256,0.0022357332209746042
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,256,256,0.013814399639765421
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,256,512,0.007529599964618683
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,256,128,0.002253866692384084
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,256,256,0.006678399940331777
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,256,128,0.014458666245142618
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,256,64,0.0021514666577180227
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,256,64,0.014152533809343972
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,256,128,0.006668800115585327
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,256,32,0.0021920000513394673
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,128,65536,0.009478400150934856
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,256,32,0.013591466347376504
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,128,65536,0.018592000007629395
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,128,16384,0.005067733426888784
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,128,16384,0.018639999628067016
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,128,16384,0.03538879950841268
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,128,12288,0.0050570666790008545
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,128,12288,0.016707199811935424
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,128,65536,0.12065386772155762
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,128,10240,0.005062399804592133
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,128,10240,0.017079466581344606
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,128,12288,0.02802773316701253
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,128,8192,0.0052255998055140175
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,128,8192,0.01641706625620524
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,128,10240,0.0247381329536438
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,128,7168,0.004653866589069367
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,128,7168,0.016525866587956746
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,128,8192,0.020637865861256918
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,128,6144,0.005038933455944061
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,128,6144,0.016664533813794454
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,128,7168,0.019038933515548705
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,128,5120,0.0043029333154360454
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,128,5120,0.016504533092180886
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,128,6144,0.017578667402267455
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,128,4096,0.004221866528193155
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,128,4096,0.016531200210253397
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,128,5120,0.015701333681742348
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,128,3584,0.004198400179545084
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,128,3584,0.014565333724021912
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,128,4096,0.013505066434542337
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,128,3072,0.004284800092379252
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,128,3072,0.015464533368746439
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,128,3584,0.01279253363609314
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,128,2560,0.004286933441956838
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,128,2560,0.014495999614397685
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,128,3072,0.012023466825485229
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,128,2048,0.004273066421349844
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,128,2560,0.011168000102043153
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,128,2048,0.01463573376337687
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,128,1536,0.0035103999078273775
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,128,1536,0.014599466323852539
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,128,2048,0.009939199686050415
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,128,1024,0.002985599885384242
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,128,1024,0.013697066903114319
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,128,1536,0.009125333031018574
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,128,768,0.0026346666117509207
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,128,768,0.014402133226394654
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,128,1024,0.008297599852085114
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,128,512,0.002218666672706604
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,128,768,0.007838933169841767
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,128,512,0.013316266735394797
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,128,256,0.0021568000316619872
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,128,256,0.013698133826255798
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,128,512,0.007524266839027405
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,128,128,0.0021546666820844015
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,128,256,0.006743466854095459
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,128,128,0.006656000018119812
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,128,64,0.00176959993938605
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,128,64,0.01260586678981781
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,128,32,0.0018154667069514591
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,128,32,0.01304533382256826
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,64,65536,0.0066549330949783325
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,64,65536,0.01842666665712992
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,64,16384,0.005885866781075796
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,64,16384,0.018590933084487914
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,64,12288,0.005022933085759481
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,64,12288,0.016531200210253397
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,64,10240,0.005017599960168203
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,64,10240,0.016526933511098227
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,64,8192,0.005468800167242686
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,64,8192,0.016721065839131674
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,64,7168,0.005086933573087057
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,64,7168,0.016503467162450155
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,64,6144,0.005063466727733612
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,64,6144,0.016590933005015053
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,64,5120,0.003832533210515976
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,64,5120,0.016506666938463845
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,64,4096,0.003822933385769526
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,64,4096,0.016495999693870545
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,64,3584,0.0043029333154360454
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,64,3584,0.01550933321317037
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,64,3072,0.004188799858093261
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,64,3072,0.014478933811187745
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,64,2560,0.004229333500067393
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,64,2560,0.014572800199190775
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,64,2048,0.004220800101757049
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,64,2048,0.014441600441932679
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,64,1536,0.0034229333202044168
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,64,1536,0.014439466595649719
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,64,1024,0.002998399982849757
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,64,1024,0.013778133193651834
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,64,768,0.002609066665172577
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,64,768,0.013100799918174744
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,64,512,0.0022304000953833262
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,64,512,0.013697066903114319
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,64,256,0.002152533332506816
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,64,256,0.01374399960041046
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,64,128,0.0021429332594076794
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,64,128,0.013411200046539307
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,64,64,0.0018314667046070098
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,64,64,0.013315199812253316
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,64,32,0.0018250666558742522
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,64,32,0.01409173309803009
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,32,65536,0.006696533163388569
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,32,65536,0.01653333306312561
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,32,16384,0.004634666442871094
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,32,16384,0.015915733575820924
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,32,12288,0.004693333307902018
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,32,12288,0.017116800944010416
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,32,10240,0.004397866626580557
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,32,10240,0.015988266468048094
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,32,8192,0.004345599810282389
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,32,8192,0.016683733463287352
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,32,7168,0.004724266628424326
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,32,7168,0.016587733229001363
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,32,6144,0.004355200131734212
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,32,6144,0.016562133034070333
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,32,5120,0.0038602667550245917
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,32,5120,0.016583466529846193
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,32,4096,0.0038474666575590765
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,32,4096,0.015200000007947287
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,32,3584,0.004240000247955322
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,32,3584,0.014769066373507181
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,32,3072,0.003881600002447764
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,32,3072,0.01458026667435964
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,32,2560,0.004226133227348328
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,32,2560,0.01453439990679423
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,32,2048,0.003884800026814143
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,32,2048,0.014481066664059957
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,32,1536,0.003425066669782003
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,32,1536,0.01448853313922882
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,32,1024,0.003014400104681651
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,32,1024,0.014265599846839904
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,32,768,0.0025770666698614756
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,32,768,0.013757866621017457
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,32,512,0.0025621332228183747
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,32,512,0.013796266913414002
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,32,256,0.002194133400917053
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,32,256,0.013383466998736063
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,32,128,0.0021802666286627453
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,32,128,0.013157332936922709
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,32,64,0.0018133333573738734
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,32,64,0.013668266932169595
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,32,32,0.0017621333400408429
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,32,32,0.013796266913414002
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,nvfp4,1,2048,12288,0.02808106740315755
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,float16,1,768,768,0.003403733422358831
TRTLLM,1.2.0rc5,NVIDIA GB200,gemm,torch_flow,fp8,1,128,128,0.014402133226394654
