framework,version,device,op_name,kernel_source,gemm_dtype,m,n,k,latency
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,65536,51200,72.269482421875
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,65536,16384,22.127288818359375
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,65536,16384,42.24506429036458
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,65536,12288,31.419598388671876
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,65536,12288,16.750584920247398
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,65536,51200,135.38187662760419
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,65536,10240,26.50440673828125
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,65536,10240,14.346412150065103
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,65536,10240,15.411936442057291
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,65536,8192,21.0932861328125
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,65536,8192,11.450221761067708
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,65536,12288,18.340362548828125
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,65536,7168,18.664990234375
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,65536,7168,9.638510131835938
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,65536,16384,24.11119181315104
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,65536,6144,15.865666707356771
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,65536,6144,8.196072387695313
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,65536,8192,12.329905192057293
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,65536,5120,13.123355102539062
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,65536,5120,6.902112833658855
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,65536,7168,10.675140380859375
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,65536,4096,10.785209147135417
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,65536,4096,5.549104817708334
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,65536,6144,9.272698974609375
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,65536,3584,8.90253397623698
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,65536,3584,4.752525838216146
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,65536,4096,6.165106201171875
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,65536,3072,7.8384852091471355
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,65536,3072,4.237933858235677
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,65536,5120,7.8708852132161455
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,65536,2560,6.537586975097656
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,65536,2560,3.4921236673990883
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,65536,3072,4.69122568766276
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,65536,3584,5.50725351969401
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,65536,2048,5.013115946451823
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,65536,2048,2.8164713541666666
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,65536,1536,4.075549825032552
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,65536,1536,2.091000493367513
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,65536,2560,3.5162378946940103
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,65536,1024,2.561370595296224
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,65536,1024,1.4507755279541015
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,65536,1024,1.8838175455729167
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,65536,768,1.874173863728841
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,65536,768,1.159442138671875
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,65536,1536,1.921508280436198
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,65536,512,1.356936518351237
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,65536,512,0.954473622639974
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,65536,768,1.5756895701090494
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,65536,256,0.8644202550252279
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,65536,256,0.76953919728597
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,65536,512,1.5583829243977865
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,65536,128,0.749132792154948
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,65536,128,0.658677355448405
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,65536,2048,2.506111907958984
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,65536,64,0.6366720199584961
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,65536,64,0.6772672017415364
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,65536,32,0.656394640604655
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,65536,32,0.6760074615478515
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,65536,128,1.6348799387613933
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,65536,256,1.5543626149495444
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,51200,65536,73.10724283854167
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,51200,65536,136.3540771484375
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,51200,51200,105.1383544921875
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,51200,51200,57.71318766276041
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,51200,16384,33.24975382486979
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,51200,16384,17.791107177734375
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,51200,16384,18.924418131510414
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,51200,12288,24.537117513020835
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,51200,12288,13.274437459309897
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,65536,51200,75.2330322265625
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,51200,10240,20.78642781575521
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,51200,10240,11.404163614908855
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,51200,10240,12.119546508789062
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,51200,8192,16.564497884114584
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,51200,8192,9.008866373697916
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,51200,12288,14.42080078125
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,51200,7168,14.485306803385416
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,51200,7168,7.653421020507812
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,51200,8192,9.717598470052083
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,51200,6144,12.362830607096354
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,51200,6144,6.572984313964843
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,51200,7168,8.398058573404949
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,51200,5120,10.700152587890624
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,51200,5120,5.491931660970052
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,51200,6144,7.275198872884114
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,51200,4096,8.496842447916666
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,51200,4096,4.288756306966146
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,51200,4096,4.759707641601563
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,51200,5120,6.127637227376303
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,51200,3584,3.831287384033203
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,51200,3584,7.098959859212239
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,51200,3072,5.944002278645834
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,51200,3072,3.3247370402018226
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,51200,3072,2.684558868408203
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,51200,2560,4.913230895996094
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,51200,2560,2.8496607462565104
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,51200,3584,3.8309076944986975
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,51200,2048,3.959178670247396
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,51200,2048,2.192230478922526
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,51200,2560,2.239000447591146
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,51200,1536,2.9415903727213544
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,51200,1536,1.6110997517903645
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,51200,2048,1.8735936482747395
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,51200,1024,1.9880320231119792
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,51200,1024,1.1719263712565104
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,51200,1024,1.6002367655436198
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,51200,768,1.4393397013346354
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,51200,768,0.9625919977823892
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,51200,1536,1.6147380828857423
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,51200,512,1.0101151784261069
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,51200,512,0.7587167739868164
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,51200,512,1.4053418477376303
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,51200,256,0.6818645477294922
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,51200,256,0.5918207804361979
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,51200,768,1.4163285573323567
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,51200,128,0.5875573476155599
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,51200,128,0.5148245175679524
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,51200,128,1.4312352498372396
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,51200,51200,59.126973470052086
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,51200,256,1.4011338551839194
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,51200,64,0.49304745992024734
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,51200,64,0.5350293477376302
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,51200,32,0.511959457397461
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,51200,32,0.5312095959981282
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,16384,65536,24.946317545572917
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,16384,65536,41.987255859375
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,16384,51200,33.41603393554688
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,16384,51200,20.144486490885416
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,16384,16384,10.313267008463542
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,16384,16384,6.168557739257812
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,16384,16384,4.03295669555664
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,16384,12288,7.719313049316407
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,16384,12288,4.625362141927083
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,16384,12288,3.019042205810547
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,16384,10240,6.39972178141276
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,16384,10240,3.9753781636555994
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,16384,10240,2.560862986246745
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,16384,8192,5.108044942220053
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,16384,8192,3.2225547790527345
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,51200,65536,78.36982421875
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,16384,7168,4.64873046875
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,16384,7168,2.7347946166992188
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,16384,8192,1.828883234659831
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,16384,6144,3.800501251220703
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,16384,6144,2.247900899251302
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,16384,7168,1.5372639973958333
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,16384,5120,3.169627634684245
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,16384,5120,1.8561524709065753
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,16384,51200,12.187413533528645
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,16384,4096,2.633373769124349
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,16384,4096,1.5246261596679687
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,16384,5120,1.0904458363850913
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,16384,6144,1.2447509765625
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,16384,3584,1.337176513671875
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,16384,3584,2.237834676106771
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,16384,3072,1.8288555145263672
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,16384,3072,1.1548394521077474
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,16384,4096,0.9210666656494141
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,16384,2560,1.4822410583496093
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,16384,2560,0.9883562723795573
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,16384,3072,0.6993024190266927
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,16384,2048,1.1917781829833984
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,16384,2048,0.857914670308431
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,16384,3584,0.8092063903808594
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,16384,2560,0.6809845606486002
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,16384,1536,0.832806396484375
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,16384,1536,0.5943861643473307
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,16384,1024,0.5876842498779297
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,16384,1024,0.4426421483357747
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,16384,1024,0.4992095947265625
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,16384,768,0.4457695960998535
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,16384,768,0.33371734619140625
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,16384,1536,0.513972282409668
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,16384,512,0.3249504089355469
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,16384,512,0.2678208033243815
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,16384,768,0.4142250696818034
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,16384,256,0.21989760398864747
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,16384,512,0.40509761174519854
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,16384,256,0.21354026794433595
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,16384,128,0.19257814089457195
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,16384,128,0.18776960372924806
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,16384,2048,0.5976352055867513
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,16384,64,0.165937074025472
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,16384,64,0.18742186228434246
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,16384,32,0.16763200759887695
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,16384,128,0.42797333399454757
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,16384,32,0.1853376070658366
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,16384,256,0.4027061462402344
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,12288,65536,19.994873046875
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,12288,65536,31.89493408203125
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,12288,51200,24.111236572265625
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,12288,51200,15.536329142252603
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,12288,16384,7.601880391438802
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,12288,16384,4.876237996419271
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,12288,16384,2.8758613586425783
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,12288,12288,5.588953653971354
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,12288,12288,3.6624682108561197
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,16384,65536,18.740519205729164
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,12288,10240,4.90052490234375
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,12288,10240,3.0681610107421875
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,12288,12288,2.0934122721354167
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,12288,8192,3.7995061238606773
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,12288,8192,2.432106781005859
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,12288,10240,1.7642720540364583
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,12288,7168,3.2785247802734374
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,12288,7168,2.1504170735677084
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,12288,8192,1.2319477081298829
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,12288,6144,2.7678741455078124
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,12288,6144,1.822652816772461
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,12288,6144,0.9484074910481771
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,12288,5120,2.2859947204589846
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,12288,5120,1.5297589619954428
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,12288,7168,1.1362922668457032
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,12288,4096,1.8136789957682293
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,12288,4096,1.2623701731363932
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,12288,5120,0.8424885431925455
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,12288,3584,1.5352031707763671
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,12288,3584,1.0738538106282554
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,12288,4096,0.7174805323282878
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,12288,3072,1.2908170064290365
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,12288,3072,0.9378901163736979
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,12288,3584,0.6367359797159831
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,12288,2560,1.10600217183431
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,12288,51200,8.648518880208332
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,12288,2560,0.7695733388264974
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,12288,2048,0.8454485575358073
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,12288,2048,0.6348960240681966
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,12288,3072,0.5661216100056966
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,12288,1536,0.6310986836751302
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,12288,1536,0.5267946561177571
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,12288,2048,0.5003733317057292
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,12288,1024,0.43167254130045574
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,12288,1024,0.33225812911987307
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,12288,2560,0.5493802388509115
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,12288,768,0.33134400049845375
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,12288,768,0.2771807988484701
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,12288,1536,0.4563925425211589
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,12288,512,0.24749013582865395
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,12288,512,0.214026673634847
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,12288,768,0.40287040074666336
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,12288,256,0.16396160125732423
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,12288,256,0.16735893885294598
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,12288,512,0.39384425481160484
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,12288,128,0.14560747146606445
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,12288,1024,0.44092480341593426
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,12288,128,0.1391146659851074
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,12288,64,0.11794453461964924
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,12288,64,0.14358399709065756
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,12288,32,0.12388799985249836
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,12288,32,0.14199360211690265
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,12288,256,0.3909994761149088
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,12288,128,0.3986069361368815
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,10240,65536,16.733673095703125
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,10240,65536,25.938924153645832
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,10240,51200,19.927305094401042
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,10240,51200,13.470808919270834
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,10240,16384,6.412890625
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,10240,16384,4.286567687988281
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,12288,65536,11.073642985026042
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,10240,12288,4.739961751302084
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,10240,12288,3.1777854919433595
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,10240,16384,2.3786496480305988
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,10240,10240,4.026302846272786
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,10240,12288,1.6321866353352863
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,10240,10240,2.6678761800130206
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,10240,8192,3.384754180908203
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,10240,8192,2.1508148193359373
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,10240,10240,1.3434538523356119
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,10240,7168,2.68677978515625
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,10240,7168,1.8616575876871746
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,10240,8192,1.0343744277954101
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,10240,6144,2.3235850016276043
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,10240,6144,1.615138117472331
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,10240,7168,0.9506378809611002
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,10240,5120,1.8720458984375
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,10240,5120,1.3353162129720053
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,10240,6144,0.8082688013712565
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,10240,4096,1.486236826578776
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,10240,4096,1.0844192504882812
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,10240,5120,0.7230496088663737
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,10240,3584,1.215996805826823
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,10240,3584,0.9543008168538412
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,10240,4096,0.6187466939290365
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,10240,3072,1.0777802785237631
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,10240,3072,0.8270250956217448
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,10240,3072,0.5007530530293782
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,10240,3584,0.5550090789794921
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,10240,2560,0.6774143854777018
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,10240,2560,0.8650954564412435
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,10240,2048,0.7663455963134765
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,10240,2048,0.5702047983805338
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,10240,2560,0.4861781438191731
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,10240,2048,0.4428373336791992
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,10240,1536,0.5203861236572266
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,10240,1536,0.42277971903483075
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,10240,1024,0.3624544143676758
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,10240,1024,0.31688747406005857
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,10240,1024,0.3878047943115234
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,10240,768,0.2830592155456543
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,10240,768,0.23288853963216147
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,10240,1536,0.4013728141784668
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,10240,512,0.20671680768330894
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,10240,512,0.1862346649169922
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,10240,768,0.3523680051167806
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,10240,256,0.13817493120829266
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,10240,256,0.1440170605977376
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,10240,512,0.34418986638387045
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,10240,128,0.12145919799804687
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,10240,256,0.34147841135660806
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,10240,128,0.12482240200042724
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,10240,64,0.10300479729970295
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,10240,64,0.12922879854838054
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,10240,32,0.10884693463643391
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,10240,32,0.12533653577168782
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,10240,128,0.3498378753662109
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,10240,51200,7.094987487792968
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,8192,65536,14.687776692708335
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,8192,65536,20.66299845377604
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,8192,51200,16.020287068684897
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,8192,51200,11.528438313802083
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,8192,16384,5.041236368815104
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,8192,16384,3.6824801127115885
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,10240,65536,9.132260131835938
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,8192,12288,3.919042205810547
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,8192,12288,2.7359263102213545
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,8192,16384,1.8546197255452472
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,8192,10240,3.19955317179362
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,8192,10240,2.3653045654296876
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,8192,12288,1.2642101287841796
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,8192,8192,2.5722379048665363
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,8192,8192,1.8626431783040363
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,8192,10240,1.084218724568685
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,8192,7168,2.240884145100911
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,8192,7168,1.6421984354654948
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,8192,8192,0.8574645360310873
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,8192,6144,1.8862954457600911
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,8192,6144,1.387444305419922
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,8192,7168,0.7931466420491536
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,8192,5120,1.5371988932291667
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,8192,5120,1.1560714721679688
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,8192,6144,0.6789471944173177
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,8192,4096,1.216324234008789
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,8192,4096,0.940492820739746
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,8192,5120,0.613432502746582
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,8192,3584,1.0333728154500326
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,8192,3584,0.8333333333333334
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,8192,4096,0.517409070332845
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,8192,3072,0.8390325546264649
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,8192,3072,0.7327381134033203
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,8192,3584,0.45692160924275715
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,8192,2560,0.686236826578776
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,8192,2560,0.651144536336263
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,8192,3072,0.3976085344950358
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,8192,2048,0.5524106979370117
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,8192,2048,0.4786325454711914
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,8192,2560,0.3823605219523112
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,8192,1536,0.4238368034362793
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,8192,1536,0.36599467595418295
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,8192,2048,0.3318666776021322
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,8192,51200,5.75333251953125
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,8192,1024,0.2887157440185547
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,8192,1024,0.2722400029500326
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,8192,768,0.2249589284261068
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,8192,768,0.2026048024495443
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,8192,1536,0.2861994743347168
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,8192,512,0.1647605260213216
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,8192,512,0.15788267453511556
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,8192,768,0.2262869358062744
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,8192,256,0.11099413235982258
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,8192,256,0.1215392033259074
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,8192,1024,0.27043838500976564
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,8192,128,0.09993279774983724
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,8192,128,0.1057696024576823
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,8192,256,0.21511573791503907
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,8192,64,0.08284266789754233
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,8192,64,0.10621333122253418
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,8192,32,0.08554133574167887
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,8192,128,0.23247040112813316
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,8192,32,0.10266773700714112
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,8192,512,0.21773120562235512
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,7168,65536,13.297081502278647
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,7168,65536,18.51008097330729
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,7168,51200,13.995429484049478
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,7168,51200,10.47017822265625
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,7168,16384,4.314403279622396
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,7168,16384,3.362929026285807
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,8192,65536,7.195700073242188
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,7168,12288,3.211921183268229
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,7168,12288,2.52315190633138
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,7168,16384,1.5801226298014321
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,7168,10240,2.820994059244792
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,7168,10240,2.138263448079427
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,7168,12288,1.1154229482014975
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,7168,8192,2.2302731831868488
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,7168,8192,1.6924959818522136
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,7168,10240,0.9677141189575196
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,7168,7168,1.8805834452311199
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,7168,7168,1.5222816467285156
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,7168,8192,0.767686398824056
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,7168,6144,1.6409472147623698
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,7168,6144,1.269979731241862
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,7168,7168,0.7122293472290039
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,7168,5120,1.3542570749918619
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,7168,5120,1.0841451009114584
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,7168,5120,0.5554261525472005
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,7168,4096,0.9798954645792642
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,7168,4096,0.8855093638102213
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,7168,6144,0.6093141555786132
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,7168,3584,0.8901034673055014
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,7168,3584,0.814969571431478
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,7168,3584,0.41436052322387695
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,7168,3072,0.7400351842244466
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,7168,4096,0.4706005414326985
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,7168,3072,0.7016821543375651
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,7168,2560,0.6052607854207357
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,7168,2560,0.6144416173299153
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,7168,2560,0.35273920694986977
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,7168,2048,0.49154132207234696
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,7168,2048,0.4747477213541667
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,7168,3072,0.36666345596313477
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,7168,1536,0.385865592956543
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,7168,1536,0.3559349377950033
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,7168,1536,0.2765856107076009
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,7168,1024,0.2653205394744873
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,7168,1024,0.22841493288675943
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,7168,2048,0.31401920318603516
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,7168,768,0.20419413248697915
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,7168,768,0.17910827000935872
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,7168,1024,0.26226560274759925
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,7168,512,0.14468480745951334
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,7168,512,0.1428426742553711
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,7168,768,0.2312170664469401
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,7168,256,0.09720426400502523
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,7168,256,0.11659306685129803
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,7168,256,0.22059413592020669
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,7168,128,0.08646612962086996
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,7168,128,0.09399466514587403
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,7168,512,0.22212160428365074
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,7168,64,0.07184960047403971
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,7168,64,0.09464213053385416
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,7168,32,0.07510826587677003
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,7168,128,0.23076906204223632
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,7168,32,0.09215466976165772
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,6144,65536,12.246176147460938
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,6144,65536,15.500936889648438
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,7168,51200,5.031520080566406
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,6144,51200,11.908756510416667
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,6144,51200,9.5102783203125
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,6144,16384,3.634680430094401
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,6144,16384,3.0598154703776044
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,6144,16384,1.3345226287841796
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,6144,12288,2.7758453369140623
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,6144,12288,2.3022750854492187
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,6144,12288,0.9788576126098633
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,7168,65536,6.470999654134114
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,6144,10240,2.3777450561523437
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,6144,10240,1.9522474924723308
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,6144,8192,1.8854591369628906
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,6144,8192,1.56790402730306
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,6144,10240,0.8570826848347982
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,6144,8192,0.6865909576416016
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,6144,7168,1.6628074645996094
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,6144,7168,1.376958974202474
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,6144,6144,1.3463264465332032
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,6144,6144,1.1952810923258463
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,6144,7168,0.6305077234903972
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,6144,5120,1.0908660888671875
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,6144,5120,0.9950997034708658
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,6144,6144,0.5493408203125
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,6144,4096,0.8681568145751953
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,6144,4096,0.805571174621582
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,6144,5120,0.49963518778483074
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,6144,51200,4.371965026855468
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,6144,4096,0.42147305806477864
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,6144,3584,0.737612787882487
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,6144,3584,0.6924906412760417
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,6144,3072,0.6858495712280274
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,6144,3072,0.5862762451171875
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,6144,2560,0.5221653302510579
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,6144,2560,0.5004885355631511
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,6144,2560,0.3137322743733724
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,6144,2048,0.42677866617838545
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,6144,2048,0.4071040153503418
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,6144,3072,0.32663787206014
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,6144,1536,0.3155850728352865
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,6144,1536,0.30576534271240235
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,6144,3584,0.37213331858317056
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,6144,1024,0.21821333567301432
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,6144,1024,0.2201024055480957
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,6144,1024,0.23309760093688964
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,6144,768,0.17051092783610028
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,6144,768,0.16462400754292805
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,6144,1536,0.24732054074605309
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,6144,512,0.12299093405405681
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,6144,512,0.12982719739278156
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,6144,2048,0.2792682647705078
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,6144,256,0.08377599716186523
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,6144,256,0.09950186411539713
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,6144,768,0.20722667376200357
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,6144,512,0.1985439936319987
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,6144,128,0.06632533470789591
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,6144,128,0.08049066861470541
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,6144,256,0.1965023994445801
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,6144,64,0.059305600325266515
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,6144,128,0.20601812998453775
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,6144,32,0.06275519927342733
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,6144,64,0.0832640012105306
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,6144,32,0.0798911968866984
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,5120,65536,11.08148905436198
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,5120,65536,12.607862345377605
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,5120,51200,8.548104858398437
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,5120,51200,9.881897989908854
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,5120,16384,3.162719980875651
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,5120,16384,2.778004201253255
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,6144,65536,5.6226252237955725
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,5120,12288,2.288665517171224
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,5120,12288,2.0881407419840494
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,5120,12288,0.8503573099772135
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,5120,10240,1.9962399800618489
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,5120,10240,1.7444491068522137
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,5120,16384,1.14017702738444
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,5120,8192,1.5363306681315103
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,5120,8192,1.4114229838053385
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,5120,10240,0.7537685394287109
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,5120,7168,1.4134997049967448
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,5120,8192,0.6054911931355794
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,5120,7168,1.2337738037109376
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,5120,6144,1.1235925038655599
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,5120,6144,1.0562431971232096
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,5120,6144,0.48395627339680986
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,5120,5120,0.8535381317138672
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,5120,5120,0.9055882771809897
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,5120,7168,0.5582752227783203
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,5120,4096,0.6829941431681316
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,5120,4096,0.7433557510375977
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,5120,5120,0.44077971776326497
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,5120,3584,0.5959882736206055
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,5120,3584,0.6999647776285808
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,5120,4096,0.36733547846476233
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,5120,3072,0.517250124613444
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,5120,3072,0.5424522399902344
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,5120,3584,0.33067626953125
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,5120,3072,0.2951221466064453
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,5120,2560,0.4384522755940755
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,5120,2560,0.4527146657307942
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,5120,2048,0.349397341410319
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,5120,2048,0.39618028004964195
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,5120,2560,0.28086932500203454
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,5120,1536,0.2666624069213867
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,5120,2048,0.2530752023061117
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,5120,1536,0.28093226750691735
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,5120,1024,0.1818336009979248
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,5120,1024,0.18595093091328938
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,5120,1024,0.21117866834004723
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,5120,768,0.1420629342397054
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,5120,768,0.1443178653717041
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,5120,1536,0.2258944034576416
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,5120,512,0.10459413528442382
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,5120,512,0.11653227011362713
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,5120,768,0.19100799560546874
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,5120,256,0.07067626317342122
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,5120,256,0.08812266985575358
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,5120,512,0.18291519482930502
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,5120,128,0.06350186665852865
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,5120,128,0.07513279914855957
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,5120,256,0.18025387128194173
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,5120,128,0.1884640057881673
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,5120,64,0.05382399956385294
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,5120,64,0.07633919715881347
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,5120,32,0.05675520102183024
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,5120,32,0.07373546759287516
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,4096,65536,10.030250040690104
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,4096,65536,9.988798014322917
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,5120,51200,3.667021942138672
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,4096,51200,7.861654154459636
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,4096,51200,7.688597106933594
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,4096,16384,2.474283854166667
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,4096,16384,2.4853759765625
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,4096,16384,0.9514666875203451
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,5120,65536,4.719666035970052
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,4096,12288,1.8678485870361328
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,4096,12288,1.9008586883544922
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,4096,10240,1.532909901936849
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,4096,10240,1.5773749033610025
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,4096,10240,0.6439306894938152
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,4096,8192,1.179692840576172
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,4096,8192,1.2562858581542968
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,4096,12288,0.7231701532999675
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,4096,7168,1.0630325317382812
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,4096,7168,1.0921834309895835
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,4096,7168,0.4764117240905762
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,4096,6144,0.835857073465983
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,4096,6144,0.9380320231119791
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,4096,8192,0.5160629272460937
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,4096,5120,0.6934719721476237
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,4096,5120,0.8379093170166015
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,4096,5120,0.3823552131652832
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,4096,4096,0.5420874913533529
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,4096,4096,0.6874453226725261
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,4096,6144,0.4179530779520671
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,4096,3584,0.4704074541727702
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,4096,3584,0.556164296468099
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,4096,4096,0.3161738713582357
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,4096,3072,0.41079146067301436
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,4096,3072,0.48190399805704753
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,4096,3584,0.2797504107157389
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,4096,2560,0.3371765454610189
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,4096,2560,0.40359894434611004
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,4096,3072,0.2422869364420573
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,4096,2048,0.2784330685933431
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,4096,2560,0.22929813067118326
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,4096,2048,0.32708158493041994
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,4096,1536,0.21009920438130697
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,4096,1536,0.24810986518859862
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,4096,1536,0.166318941116333
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,4096,1024,0.14597867329915365
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,4096,1024,0.16377706527709962
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,4096,2048,0.19687573115030926
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,4096,768,0.1119808038075765
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,4096,768,0.1276682694753011
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,4096,1024,0.15225812594095867
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,4096,512,0.08282132943471274
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,4096,768,0.12536746660868328
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,4096,512,0.10299200216929119
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,4096,256,0.05743039846420288
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,4096,256,0.07733866373697916
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,4096,256,0.11456000010172526
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,4096,128,0.04422933260599772
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,4096,512,0.11690560181935628
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,4096,128,0.06300693353017171
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,4096,64,0.04069120089213053
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,4096,64,0.0636031985282898
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,4096,32,0.04291520118713379
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,4096,128,0.12645546595255536
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,4096,32,0.060591999689737955
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,3584,65536,8.652154541015625
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,3584,65536,9.322122192382812
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,3584,51200,6.750306193033855
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,4096,51200,3.022866058349609
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,3584,51200,7.2881922403971355
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,3584,16384,2.1363934834798175
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,3584,16384,2.3663103739420572
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,4096,65536,3.85202891031901
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,3584,12288,1.6008597056070966
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,3584,12288,1.7873205820719402
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,3584,16384,0.8757504145304361
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,3584,10240,1.3195531209309896
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,3584,10240,1.4945290883382163
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,3584,12288,0.671128527323405
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,3584,8192,0.9944831848144531
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,3584,8192,1.200983428955078
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,3584,10240,0.596021334330241
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,3584,7168,0.8735904057820638
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,3584,7168,1.0521194458007812
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,3584,8192,0.4784277280171712
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,3584,6144,0.7555327733357747
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,3584,6144,0.8989568074544272
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,3584,7168,0.44240427017211914
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,3584,5120,0.615396245320638
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,3584,5120,0.8107722600301107
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,3584,6144,0.38878933588663733
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,3584,4096,0.49554878870646163
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,3584,4096,0.6309055964152018
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,3584,4096,0.2929621378580729
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,3584,3584,0.4347850799560547
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,3584,5120,0.35583359400431314
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,3584,3584,0.5349311828613281
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,3584,3072,0.3704000155131022
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,3584,3072,0.4682976086934407
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,3584,3584,0.2621760050455729
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,3584,3072,0.23181546529134117
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,3584,2560,0.3121845245361328
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,3584,2560,0.3892362594604492
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,3584,2048,0.25315839449564614
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,3584,2048,0.3154890696207682
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,3584,2560,0.21909653345743813
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,3584,2048,0.19325440724690754
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,3584,1536,0.19363412857055665
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,3584,1536,0.23592000007629393
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,3584,1024,0.13392747243245443
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,3584,1024,0.15278933842976888
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,3584,1024,0.15454187393188476
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,3584,768,0.1006325324376424
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,3584,1536,0.1686858654022217
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,3584,768,0.11961599985758464
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,3584,512,0.07332479953765869
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,3584,512,0.09634453455607096
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,3584,512,0.12798293431599933
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,3584,256,0.05054933230082194
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,3584,768,0.13635199864705402
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,3584,256,0.07157866954803467
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,3584,128,0.045665065447489425
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,3584,128,0.059579734007517496
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,3584,256,0.12596480051676434
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,3584,64,0.03854933182398478
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,3584,128,0.13507946332295734
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,3584,32,0.039510401089986165
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,3584,64,0.05970880190531412
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,3584,32,0.055912534395853676
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,3072,65536,7.355924479166667
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,3072,65536,8.722300211588543
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,3072,51200,5.702638753255209
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,3072,51200,6.783266194661458
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,3584,51200,2.726331837972005
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,3072,16384,1.8314805348714192
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,3072,16384,2.2190175374348957
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,3584,65536,3.45934092203776
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,3072,12288,1.3447157541910806
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,3072,12288,1.6565760294596354
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,3072,16384,0.8041056315104166
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,3072,10240,1.1672991434733073
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,3072,10240,1.4134602864583334
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,3072,12288,0.6088874816894532
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,3072,8192,0.9144800186157227
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,3072,8192,1.1288052876790364
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,3072,10240,0.5488063812255859
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,3072,7168,0.7566538492838542
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,3072,7168,0.9920202891031901
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,3072,8192,0.43820587793986004
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,3072,6144,0.6326698939005534
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,3072,6144,0.8681887944539388
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,3072,7168,0.4078559875488281
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,3072,6144,0.35886828104654944
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,3072,5120,0.531715202331543
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,3072,5120,0.7076149622599284
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,3072,4096,0.41640214920043944
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,3072,4096,0.5702378590901692
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,3072,4096,0.27085758845011393
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,3072,3584,0.36519254048665367
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,3072,5120,0.3313354810078939
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,3072,3584,0.5138549486796061
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,3072,3072,0.31318400700887045
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,3072,3072,0.4280394554138184
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,3072,3584,0.24181547164916992
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,3072,3072,0.21225706736246744
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,3072,2560,0.27017386754353845
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,3072,2560,0.36259838740030925
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,3072,2048,0.21659733454386393
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,3072,2048,0.2917994817097982
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,3072,2560,0.20019307136535644
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,3072,1536,0.16362454096476237
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,3072,2048,0.17503892580668132
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,3072,1536,0.2239317258199056
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,3072,1024,0.11233279705047608
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,3072,1024,0.14569600423177081
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,3072,1024,0.1380949338277181
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,3072,768,0.08795093695322673
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,3072,1536,0.1525152047475179
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,3072,768,0.11361707051595052
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,3072,512,0.06425386667251587
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,3072,512,0.09062080383300782
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,3072,51200,2.492535400390625
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,3072,256,0.04878400166829427
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,3072,256,0.0679744005203247
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,3072,512,0.11374506950378419
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,3072,768,0.12172053654988606
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,3072,256,0.11125013033548992
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,3072,128,0.03604160149892171
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,3072,64,0.03176213304201762
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,3072,128,0.053409067789713535
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,3072,32,0.03373759984970093
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,3072,64,0.05345600048700968
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,3072,32,0.05128959814707438
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,3072,128,0.12032106717427571
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,2560,65536,6.090776570638021
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,2560,65536,8.08892822265625
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,2560,51200,5.052151489257812
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,2560,51200,6.347094217936197
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,2560,16384,1.3945941925048828
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,2560,16384,2.051467768351237
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,3072,65536,3.1265279134114583
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,2560,12288,1.0506879806518554
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,2560,12288,1.5573482513427734
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,2560,16384,0.7125333150227864
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,2560,10240,0.9480661392211914
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,2560,10240,1.299065653483073
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,2560,12288,0.548908805847168
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,2560,8192,0.722170639038086
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,2560,8192,1.0454346974690756
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,2560,10240,0.48993705113728836
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,2560,7168,0.6107285181681316
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,2560,7168,0.9302015940348307
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,2560,8192,0.39374186197916666
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,2560,6144,0.5431754430135091
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,2560,6144,0.7914815902709961
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,2560,7168,0.3633546511332194
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,2560,5120,0.43088639577229815
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,2560,5120,0.6958581288655599
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,2560,6144,0.32451839447021485
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,2560,4096,0.35135679244995116
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,2560,4096,0.5302293459574382
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,2560,5120,0.2979999860127767
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,2560,3584,0.2988650639851888
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,2560,3584,0.4668405214945476
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,2560,4096,0.242851193745931
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,2560,3072,0.26295359929402673
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,2560,3072,0.40303999582926436
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,2560,3584,0.21440320014953612
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,2560,2560,0.21456425984700522
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,2560,2560,0.3413066546122233
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,2560,3072,0.187281068166097
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,2560,2048,0.1749567985534668
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,2560,2048,0.2731349309285482
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,2560,2560,0.17476479212443036
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,2560,1536,0.13614293734232585
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,2560,1536,0.2056981404622396
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,2560,2048,0.15072426795959473
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,2560,1024,0.0923637310663859
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,2560,1024,0.13549332618713378
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,2560,1536,0.12891519864400228
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,2560,768,0.07330986658732096
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,2560,768,0.10460800329844158
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,2560,1024,0.11449066797892253
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,2560,512,0.05255786577860514
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,2560,512,0.08234453201293945
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,2560,768,0.09860053062438964
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,2560,256,0.036954665184021
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,2560,256,0.060393599669138584
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,2560,512,0.09054186344146728
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,2560,256,0.08822826544443765
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,2560,128,0.032953600088755294
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,2560,128,0.0499295989672343
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,2560,64,0.02807040015856425
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,2560,64,0.0492415984471639
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,2560,128,0.09758826891581217
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,2560,32,0.02927359938621521
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,2560,32,0.04741119941075643
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,2048,65536,4.91866455078125
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,2048,65536,7.590392557779948
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,2048,51200,3.844379679361979
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,2560,51200,2.1400299072265625
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,2048,51200,5.925984191894531
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,2048,16384,1.1165855407714844
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,2048,16384,1.9252213795979816
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,2560,65536,2.703191375732422
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,2048,12288,0.8746474583943685
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,2048,12288,1.455678939819336
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,2048,16384,0.6351637522379557
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,2048,10240,0.7415722529093425
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,2048,12288,0.48946558634440107
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,2048,10240,1.216038386027018
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,2048,8192,0.5772138595581054
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,2048,8192,0.9723520278930664
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,2048,8192,0.35339733759562175
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,2048,10240,0.44116694132486983
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,2048,7168,0.5039573351542155
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,2048,7168,0.8557066599527994
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,2048,6144,0.4356789271036784
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,2048,6144,0.7571168263753255
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,2048,7168,0.3274794578552246
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,2048,5120,0.3530495961507162
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,2048,6144,0.29389546712239584
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,2048,5120,0.6152853647867839
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,2048,4096,0.2783434549967448
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,2048,4096,0.49669974644978837
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,2048,5120,0.2713397343953451
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,2048,3584,0.25058347384134927
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,2048,4096,0.21920639673868814
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,2048,3584,0.44068266550699875
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,2048,3072,0.21286826133728026
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,2048,3072,0.3782901446024577
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,2048,3584,0.19432214101155598
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,2048,2560,0.1763157367706299
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,2048,3072,0.16880000432332357
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,2048,2560,0.31693013509114587
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,2048,2048,0.14269013404846193
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,2048,2048,0.25467306772867837
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,2048,2560,0.1556341330210368
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,2048,2048,0.13244799772898358
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,2048,1536,0.11114880243937175
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,2048,1536,0.1933311939239502
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,2048,1024,0.07528746922810872
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,2048,1024,0.12611093521118164
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,2048,1024,0.09664639631907145
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,2048,768,0.05905919869740804
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,2048,768,0.09789226849873861
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,2048,1536,0.11013867060343425
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,2048,768,0.07871466477711996
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,2048,512,0.04379520018895467
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,2048,512,0.07703039646148682
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,2048,256,0.031222399075826007
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,2048,256,0.05554133256276449
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,2048,512,0.07145386536916097
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,2048,256,0.0686090628306071
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,2048,128,0.04440960089365641
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,2048,128,0.078438401222229
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,2048,64,0.02178879976272583
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,2048,64,0.04357973337173462
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,2048,32,0.022359466552734374
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,2048,32,0.0406272013982137
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,2048,128,0.023950932423273723
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,1536,65536,3.838388315836588
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,1536,65536,7.088246154785156
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,1536,51200,3.0290537516276044
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,2048,51200,1.8969579060872395
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,1536,51200,5.546843973795573
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,1536,16384,0.8514975865681966
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,1536,16384,1.7981557210286456
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,2048,65536,2.3947380065917967
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,1536,12288,0.6433258692423502
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,1536,12288,1.35240847269694
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,1536,16384,0.567518933614095
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,1536,10240,0.5487189610799154
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,1536,10240,1.13349978129069
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,1536,12288,0.4384885470072429
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,1536,8192,0.4432714780171712
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,1536,8192,0.9082880020141602
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,1536,10240,0.39598401387532556
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,1536,7168,0.38433065414428713
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,1536,7168,0.8141344070434571
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,1536,8192,0.3175061225891113
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,1536,6144,0.3359114646911621
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,1536,6144,0.6884255727132161
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,1536,7168,0.29484478632609046
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,1536,5120,0.27616214752197266
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,1536,5120,0.576800028483073
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,1536,6144,0.26528533299764
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,1536,4096,0.21960426966349283
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,1536,4096,0.4643424034118652
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,1536,5120,0.2470250606536865
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,1536,3584,0.1934325377146403
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,1536,3584,0.4075466791788737
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,1536,4096,0.1972607930501302
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,1536,3072,0.1681429386138916
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,1536,3072,0.354528013865153
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,1536,3584,0.17606080373128255
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,1536,2560,0.13934720357259114
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,1536,2560,0.2972522735595703
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,1536,3072,0.15444266001383464
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,1536,2048,0.11306133270263671
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,1536,2048,0.2396362622578939
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,1536,2560,0.14221226374308268
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,1536,1536,0.0862943967183431
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,1536,1536,0.18328426678975424
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,1536,2048,0.12242773373921711
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,1536,1024,0.06001706520716349
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,1536,1024,0.11853866577148438
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,1536,1536,0.10430186589558918
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,1536,768,0.04718079964319865
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,1536,768,0.091811203956604
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,1536,1024,0.09042879740397135
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,1536,512,0.03528639872868856
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,1536,512,0.06997653643290201
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,1536,512,0.0705087979634603
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,1536,256,0.025148799022038777
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,1536,256,0.05045653184254965
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,1536,768,0.07893439928690592
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,1536,128,0.019402666886647543
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,1536,256,0.06755733489990234
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,1536,128,0.039892268180847165
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,1536,64,0.0173418660958608
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,1536,64,0.037801599502563475
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,1536,128,0.07943999767303467
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,1536,32,0.017893334229787193
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,1536,51200,1.6752853393554688
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,1536,32,0.034983468055725095
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,1024,65536,2.578228251139323
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,1024,51200,1.9934399922688801
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,1024,65536,6.577966817220052
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,1024,51200,5.127692667643229
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,1024,16384,0.59966615041097
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,1024,16384,1.6645301818847655
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,1536,65536,2.096347681681315
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,1024,12288,0.4490837415059407
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,1024,12288,1.2547295888264975
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,1024,12288,0.3686645189921061
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,1024,10240,0.37365439732869465
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,1024,16384,0.47502934137980146
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,1024,10240,1.0667413075764975
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,1024,8192,0.2980298678080241
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,1024,8192,0.8344757080078125
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,1024,10240,0.33445440928141273
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,1024,7168,0.2684896151224772
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,1024,8192,0.26796798706054686
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,1024,7168,0.732040532430013
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,1024,6144,0.23009173075358072
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,1024,6144,0.630404281616211
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,1024,7168,0.24932692845662435
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,1024,5120,0.18852373758951824
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,1024,6144,0.22765013376871743
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,1024,5120,0.5266858736673992
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,1024,4096,0.15030825932820638
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,1024,4096,0.42470293045043944
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,1024,5120,0.2130293369293213
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,1024,3584,0.1340138594309489
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,1024,4096,0.16699946721394857
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,1024,3584,0.37320747375488283
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,1024,3072,0.11347306569417318
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,1024,3072,0.3227402687072754
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,1024,3584,0.14854399363199872
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,1024,2560,0.093339737256368
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,1024,3072,0.12952533562978108
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,1024,2560,0.2719008127848307
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,1024,2048,0.07535999615987142
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,1024,2048,0.22046507199605309
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,1024,2560,0.11699519952138264
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,1024,1536,0.058667735258738196
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,1024,2048,0.09822399616241455
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,1024,1536,0.16766293843587238
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,1024,1024,0.0411733349164327
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,1024,1536,0.08029867013295491
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,1024,1024,0.10800853570302327
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,1024,1024,0.06654400030771891
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,1024,768,0.03223680059115092
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,1024,51200,1.3827220916748046
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,1024,768,0.08200319608052573
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,1024,512,0.023857067028681435
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,1024,512,0.062174932161966956
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,1024,768,0.054131201903025304
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,1024,256,0.017092265685399375
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,1024,512,0.04442453384399414
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,1024,128,0.014054399728775025
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,1024,256,0.04325973192850749
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,1024,128,0.03297599951426188
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,1024,64,0.012409599622090657
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,1024,256,0.042822400728861496
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,1024,32,0.01333440045515696
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,1024,64,0.030853333075841267
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,1024,32,0.02880106568336487
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,1024,128,0.059180800120035806
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,768,65536,1.9644245147705077
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,768,51200,1.5100640614827474
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,1024,65536,1.7285451253255206
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,768,65536,6.296399943033854
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,768,16384,0.4656928062438965
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,768,51200,4.911868794759115
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,768,16384,1.5938837687174479
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,768,12288,0.35202452341715496
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,768,12288,1.2023743947347005
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,768,12288,0.35267839431762693
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,768,10240,0.29896532694498695
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,768,16384,0.45337813695271806
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,768,10240,1.0012863794962565
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,768,8192,0.23801280657450358
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,768,8192,0.8025056203206381
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,768,10240,0.31970561345418297
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,768,7168,0.2052138646443685
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,768,7168,0.7036725362141927
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,768,8192,0.2561471939086914
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,768,6144,0.1805727958679199
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,768,6144,0.605242665608724
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,768,7168,0.23861227035522461
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,768,5120,0.14694719314575194
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,768,5120,0.5051434516906739
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,768,6144,0.2185429255167643
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,768,4096,0.11984106699625652
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,768,4096,0.40862293243408204
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,768,5120,0.20502613385518392
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,768,3584,0.10182720025380451
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,768,4096,0.16049920717875163
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,768,3584,0.35912745793660483
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,768,3072,0.08729600111643473
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,768,3072,0.31013758977254235
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,768,3584,0.14391786257425945
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,768,2560,0.0757045348485311
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,768,3072,0.1253706693649292
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,768,2560,0.2618293285369873
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,768,2048,0.0614794651667277
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,768,2048,0.2120533307393392
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,768,2560,0.11307733058929444
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,768,1536,0.04587946732838948
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,768,1536,0.16219733556111654
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,768,2048,0.09555839697519938
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,768,1024,0.03356906572977702
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,768,1024,0.10234026908874512
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,768,1536,0.08013333479563395
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,768,768,0.026312534014383954
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,768,51200,1.3136778513590495
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,768,1024,0.06565866470336915
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,768,512,0.019012266397476198
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,768,768,0.07715626557668051
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,768,512,0.05744106769561767
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,768,256,0.012990933656692506
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,768,256,0.036857601006825766
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,768,512,0.04579413334528605
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,768,768,0.05522026618321737
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,768,128,0.011225600043932598
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,768,256,0.04618026812871297
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,768,128,0.028793599208196
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,768,64,0.010025599598884582
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,768,128,0.05549226601918539
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,768,32,0.011640533804893494
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,768,64,0.0288810670375824
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,768,32,0.02885119915008545
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,512,65536,1.2876778920491536
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,512,51200,0.9965813318888346
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,768,65536,1.6438826243082683
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,512,65536,5.967003885904948
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,512,16384,0.30613972345987955
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,512,51200,4.678048197428385
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,512,16384,1.5132651011149088
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,512,12288,0.231223471959432
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,512,12288,1.1396981557210286
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,512,16384,0.4104362805684407
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,512,12288,0.31882880528767904
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,512,10240,0.21161386171976723
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,512,10240,0.9610570907592774
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,512,8192,0.16860480308532716
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,512,8192,0.7713610967000325
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,512,10240,0.2915957450866699
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,512,8192,0.23244907061258951
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,512,7168,0.1479925314585368
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,512,7168,0.6707125345865885
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,512,6144,0.12906880378723146
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,512,6144,0.5806912104288737
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,512,7168,0.21671679814656578
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,512,5120,0.1065717299779256
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,512,5120,0.48603092829386396
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,512,6144,0.2001034736633301
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,512,4096,0.08782293001810709
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,512,5120,0.18806400299072265
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,512,4096,0.39428799947102866
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,512,3584,0.07407466570536295
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,512,3584,0.3476917266845703
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,512,51200,1.1726399739583333
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,512,3072,0.06263039906819662
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,512,4096,0.14454506238301595
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,512,3072,0.3004063924153646
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,512,2560,0.05301973422368368
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,512,3584,0.12972053686777751
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,512,2560,0.2535264015197754
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,512,2048,0.04315520127614339
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,512,65536,1.4669834136962892
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,512,2048,0.2055573304494222
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,512,3072,0.11275626818339031
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,512,1536,0.033690667152404784
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,512,1536,0.1547925313313802
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,512,1024,0.023494400580724082
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,512,2560,0.10012266635894776
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,512,1024,0.09736639658610026
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,512,768,0.017459199825922648
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,512,1536,0.0673087994257609
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,512,768,0.07193066279093424
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,512,2048,0.0831775983174642
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,512,512,0.012563199798266093
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,512,1024,0.05301760037740072
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,512,256,0.009531733393669129
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,512,512,0.049745066960652666
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,512,768,0.04064853191375732
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,512,256,0.03271786570549011
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,512,512,0.03288639982541402
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,512,256,0.03657173315684001
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,512,128,0.02584853370984395
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,512,64,0.0075445334116617845
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,512,128,0.04532586733500163
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,512,64,0.02486720085144043
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,512,32,0.009146666526794434
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,512,32,0.024589866399765015
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,256,65536,0.7298784255981445
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,512,128,0.008409600456555684
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,256,51200,0.5876725514729817
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,256,51200,4.466618855794271
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,256,65536,5.7187347412109375
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,256,16384,0.19829227129618326
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,256,16384,1.461890157063802
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,256,12288,0.1496778647104899
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,256,12288,1.1005994160970052
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,256,16384,0.37063252131144203
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,256,10240,0.11944746971130371
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,256,10240,0.9182037353515625
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,256,12288,0.28928000132242837
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,256,8192,0.09891520341237386
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,256,8192,0.7387093226114909
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,256,51200,1.0493216196695963
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,256,7168,0.08455359935760498
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,256,10240,0.265012264251709
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,256,7168,0.6487146377563476
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,256,6144,0.07496960163116455
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,256,6144,0.5589546839396159
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,256,8192,0.21122132937113441
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,256,7168,0.19780799547831218
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,256,5120,0.0627573331197103
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,256,5120,0.47095041275024413
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,256,4096,0.051351467768351235
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,256,65536,1.3091551462809243
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,256,4096,0.38100694020589193
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,256,5120,0.1747711976369222
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,256,3584,0.04582080046335856
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,256,6144,0.18366506894429524
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,256,3072,0.03972266515096028
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,256,3584,0.33525546391805017
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,256,4096,0.13132479985555012
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,256,3072,0.28868373235066735
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,256,2560,0.032839467128117875
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,256,2560,0.24313599268595376
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,256,3584,0.11871039867401123
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,256,2048,0.026680533091227216
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,256,3072,0.10266666412353516
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,256,2048,0.1961301326751709
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,256,1536,0.020338133970896403
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,256,2560,0.08994773228963217
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,256,1024,0.013209600249926248
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,256,1536,0.14796266555786133
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,256,2048,0.07362773418426513
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,256,768,0.010566400488217671
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,256,1536,0.05877866744995117
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,256,1024,0.0421834667523702
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,256,768,0.06355199813842774
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,256,512,0.007974400122960409
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,256,512,0.04147626558939616
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,256,768,0.03210986653963725
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,256,256,0.00631573349237442
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,256,512,0.026075732707977296
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,256,256,0.028983465830485028
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,256,128,0.0054730668663978575
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,256,256,0.028069333235422773
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,256,128,0.024612265825271606
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,256,64,0.005509333312511444
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,256,128,0.03707840045293172
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,256,32,0.006357333560784657
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,256,64,0.022855466604232787
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,256,32,0.022038400173187256
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,256,1024,0.09008533159891764
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,128,65536,0.6490624109903972
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,128,51200,0.514196268717448
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,128,51200,4.436378479003906
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,128,65536,5.673479715983072
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,128,16384,0.17013012568155925
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,128,16384,1.4494986216227213
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,128,12288,0.13167893091837565
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,128,65536,1.2976224263509115
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,128,16384,0.36396907170613607
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,128,10240,0.11000320116678874
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,128,51200,1.0340682983398437
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,128,12288,1.0911776224772134
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,128,8192,0.09469119707743326
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,128,10240,0.910802141825358
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,128,12288,0.2861568133036295
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,128,10240,0.2618229389190674
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,128,7168,0.07710506916046142
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,128,8192,0.7351551691691081
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,128,6144,0.06925546328226725
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,128,8192,0.20960960388183594
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,128,7168,0.64301331837972
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,128,5120,0.05831573406855265
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,128,7168,0.19495253562927245
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,128,6144,0.5545813242594402
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,128,4096,0.04686719973882039
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,128,5120,0.46517546971638996
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,128,6144,0.18189226786295573
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,128,3584,0.04221546649932861
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,128,4096,0.37728745142618814
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,128,5120,0.17388799985249836
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,128,3072,0.03663040002187093
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,128,3584,0.33053226470947267
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,128,4096,0.12981759707132975
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,128,2560,0.03110719919204712
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,128,3072,0.28484373092651366
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,128,3584,0.11687359809875489
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,128,2048,0.026264532407124834
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,128,2560,0.2384415944417318
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,128,3072,0.10086613496144611
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,128,2048,0.1921866734822591
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,128,1536,0.01801813244819641
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,128,2560,0.08855893611907958
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,128,1024,0.00885653297106425
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,128,1536,0.14464640617370605
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,128,2048,0.07196053663889566
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,128,768,0.007097599903742473
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,128,1024,0.08571306864420572
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,128,768,0.0566922664642334
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,128,1024,0.041595733165740965
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,128,512,0.005854933460553487
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,128,768,0.031310933828353885
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,128,256,0.004704000055789947
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,128,512,0.039293865362803146
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,128,512,0.024324266115824382
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,128,256,0.027093333005905152
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,128,128,0.00425600012143453
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,128,256,0.025198932488759356
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,128,128,0.022290132443110146
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,128,64,0.003874133278926214
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,128,32,0.006788266698519389
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,128,128,0.03457599878311157
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,128,64,0.023808000485102336
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,128,32,0.024421334266662598
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,64,51200,0.49475412368774413
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,64,65536,0.6369354883829753
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,32768,128,1536,0.05991679827372233
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,64,16384,0.17277119954427084
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,64,12288,0.12268906434377033
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,64,12288,1.0862805684407553
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,64,10240,0.10370666980743408
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,64,10240,0.9082730611165365
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,64,8192,0.08475200335184732
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,64,16384,1.4411776224772135
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,64,7168,0.07372480233510335
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,64,8192,0.7304693222045898
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,64,6144,0.06428693135579427
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,64,7168,0.6419274648030598
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,64,5120,0.054307198524475096
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,64,6144,0.5532021204630534
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,64,51200,4.43110097249349
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,64,5120,0.4645205179850261
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,64,3584,0.03868159850438436
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,64,4096,0.3763797442118327
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,64,65536,5.673670450846354
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,64,3072,0.03267306685447693
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,64,3584,0.3297151883443197
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,64,2560,0.026713599761327107
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,64,2560,0.23829545974731445
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,64,3072,0.28469759623209634
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,64,2048,0.021517866849899293
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,64,1536,0.01309653321901957
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,64,2048,0.19106987317403157
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,64,1536,0.14326613744099934
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,64,1024,0.007841066519419352
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,64,768,0.00626453310251236
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,64,1024,0.0844544013341268
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,64,4096,0.04342079957326253
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,64,512,0.005520000060399374
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,64,768,0.0540832002957662
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,64,256,0.004312533140182495
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,64,512,0.03734506766001384
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,64,128,0.0038304001092910765
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,64,256,0.027058132489522296
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,64,64,0.004726399978001913
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,64,32,0.004666666686534882
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,64,128,0.023389865954717003
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,64,64,0.020677334070205687
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,64,32,0.020733867088953653
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,32,51200,0.5416362762451172
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,32,16384,0.17673279444376627
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,32,65536,0.7141279856363932
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,32,12288,0.12147733370463054
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,32,16384,1.4391007741292319
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,32,10240,0.10432319641113282
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,32,12288,1.0858411153157552
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,32,8192,0.08363733291625977
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,32,10240,0.9067306518554688
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,32,7168,0.07275199890136719
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,32,8192,0.7292170842488607
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,32,6144,0.06299200057983398
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,32,7168,0.6408906936645508
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,32,6144,0.5507413228352864
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,32,5120,0.05342719952265421
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,32,4096,0.04479039907455444
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,32,5120,0.46189333597819016
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,32,4096,0.3736192067464193
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,32,3584,0.03906666835149129
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,32,3072,0.03371626536051432
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,32,3584,0.32882026036580403
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,32,2560,0.02932693362236023
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,32,3072,0.2817023913065592
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,32,2048,0.0236842672030131
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,32,2560,0.23664959271748862
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,32,2048,0.18917759259541828
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,32,65536,5.720435078938802
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,32,1024,0.0066538666685422255
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,32,1536,0.14211732546488445
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,32,768,0.0054517333706219995
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,32,1024,0.08237013022104898
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,32,768,0.05125866731007894
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,32,512,0.004612266520659129
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,32,256,0.003435733417669932
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,32,512,0.037037865320841475
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,32,51200,4.468647257486979
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,32,256,0.026716800530751546
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,32,128,0.003005866706371307
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,32,64,0.002648533384005229
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,32,32,0.004223999877770742
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,32,128,0.021908267339070638
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,32,64,0.02081386645634969
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,32768,32,32,0.019798400004704793
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,32768,32,1536,0.014239999651908874
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,65536,16384,20.619748942057292
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,65536,16384,11.00660909016927
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,65536,51200,36.26202392578125
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,65536,12288,15.699026489257813
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,65536,12288,8.357907104492188
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,65536,51200,66.58006184895834
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,65536,10240,13.028150431315103
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,65536,10240,7.091890970865886
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,65536,10240,4.761070760091146
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,65536,8192,10.774244181315105
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,65536,8192,5.579688517252604
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,65536,12288,5.714721171061198
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,65536,7168,9.567904663085937
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,65536,7168,4.8467859903971355
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,65536,16384,7.5971638997395825
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,65536,6144,7.965516662597656
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,65536,6144,4.114705149332682
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,65536,8192,3.291802724202474
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,65536,5120,6.3569890340169275
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,65536,5120,3.492075856526693
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,65536,6144,2.3624969482421876
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,65536,4096,5.029110209147135
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,65536,4096,2.78447265625
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,65536,5120,2.0115904490152996
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,65536,3584,4.441988118489584
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,65536,3584,2.3304415384928387
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,65536,7168,2.883678944905599
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,65536,3072,3.761742909749349
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,65536,3072,1.972433090209961
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,65536,4096,1.770904541015625
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,65536,2560,3.010176086425781
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,65536,2560,1.621838887532552
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,65536,3584,1.5801376342773437
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,65536,2048,2.3780886332194013
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,65536,2048,1.3251477559407552
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,65536,3072,1.3585674285888671
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,65536,1536,1.8335615793863933
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,65536,1536,1.001035753885905
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,65536,2560,1.325266138712565
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,65536,1024,1.21671142578125
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,65536,1024,0.7734912236531575
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,65536,2048,1.1801685333251952
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,65536,768,0.8856757481892904
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,65536,768,0.5909845352172851
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,65536,1536,1.0203797022501626
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,65536,512,0.6551392237345378
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,65536,512,0.5905226389567058
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,65536,768,0.9599509557088217
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,65536,256,0.43539838790893554
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,65536,256,0.38576745986938477
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,65536,1024,1.015609614054362
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,65536,128,0.38061758677164714
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,65536,128,0.3429312070210775
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,65536,256,0.9543584187825521
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,65536,64,0.3200415929158529
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,65536,64,0.34823786417643227
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,65536,128,0.9860383987426757
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,65536,32,0.33030293782552084
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,65536,32,0.3484970728556315
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,65536,512,0.9489973068237305
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,51200,65536,35.97529296875
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,51200,65536,66.50965983072916
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,51200,51200,51.80575358072916
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,51200,51200,28.065767415364583
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,51200,16384,16.254082234700522
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,51200,16384,8.826017252604167
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,51200,16384,5.926682535807291
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,51200,12288,12.149704996744791
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,51200,12288,6.483787536621094
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,51200,12288,4.444017028808593
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,51200,10240,10.614346313476563
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,51200,10240,5.524539693196615
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,65536,51200,22.65865681966146
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,51200,8192,8.266183471679687
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,51200,8192,4.53674570719401
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,51200,10240,3.7195648193359374
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,51200,7168,7.026104736328125
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,51200,7168,3.9868746439615883
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,51200,8192,2.6205429077148437
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,51200,6144,5.997967020670573
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,51200,6144,3.4957557678222657
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,51200,7168,2.180431874593099
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,51200,5120,4.940999348958334
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,51200,5120,2.6414655049641924
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,51200,6144,1.7848052978515625
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,51200,4096,3.956226094563802
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,51200,4096,2.107436752319336
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,51200,5120,1.5756788889567057
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,51200,3584,3.4527359008789062
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,51200,3584,1.8477653503417968
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,51200,3584,1.2192960103352866
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,51200,3072,2.950487518310547
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,51200,3072,1.5975104014078776
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,51200,4096,1.3708362579345703
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,51200,2560,2.3541033426920572
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,51200,2560,1.3225653330485025
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,51200,51200,17.78681437174479
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,51200,2048,1.8887893676757812
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,51200,2048,1.052505620320638
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,51200,2560,1.0330389022827149
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,51200,1536,1.375208536783854
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,51200,1536,0.8217514673868814
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,51200,3072,1.063270378112793
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,51200,1024,0.8952309290568033
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,51200,1024,0.6221322377522787
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,51200,2048,0.928875732421875
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,51200,768,0.688920529683431
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,51200,768,0.5324021339416504
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,51200,1536,0.8353034973144531
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,51200,512,0.5068021456400553
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,51200,512,0.4253493309020996
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,51200,1024,0.8262879689534506
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,51200,256,0.34096959431966145
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,51200,256,0.30644585291544596
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,51200,256,0.7706645329793294
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,51200,128,0.29698772430419923
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,51200,128,0.2690485318501791
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,51200,128,0.7902602513631185
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,51200,64,0.24872105916341147
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,51200,64,0.27559467951456706
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,51200,32,0.2578773339589437
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,51200,32,0.2751392046610514
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,51200,512,0.7678698857625326
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,51200,768,0.7714719772338867
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,16384,65536,12.49183858235677
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,16384,65536,20.690771484375
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,16384,51200,16.092953491210938
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,16384,51200,9.828181966145833
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,16384,16384,5.07684071858724
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,16384,16384,3.034449005126953
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,16384,16384,1.9585493723551433
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,16384,12288,3.673984018961588
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,16384,12288,2.2691019694010417
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,16384,12288,1.4856981913248697
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,16384,10240,3.204971822102865
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,16384,10240,1.9072682698567707
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,16384,10240,1.2592533111572266
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,16384,8192,2.568597412109375
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,16384,8192,1.5443604787190757
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,16384,8192,0.8932938893636069
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,16384,7168,2.2017494201660157
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,16384,7168,1.3625099182128906
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,16384,51200,5.9744105021158855
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,16384,6144,1.8395455678304036
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,16384,6144,1.1280298868815104
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,16384,7168,0.7668917338053386
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,16384,5120,1.5183733622233073
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,16384,5120,0.951091194152832
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,51200,65536,23.03223673502604
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,16384,4096,1.1948992411295571
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,16384,4096,0.7648725509643555
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,16384,6144,0.6262016296386719
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,16384,3584,1.0030314763387045
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,16384,3584,0.665443229675293
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,16384,5120,0.544985580444336
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,16384,3072,0.8399285634358724
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,16384,3072,0.588916269938151
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,16384,4096,0.4680021286010742
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,16384,2560,0.6864565531412761
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,16384,2560,0.5160426775614421
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,16384,2560,0.3482218742370605
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,16384,2048,0.5638751983642578
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,16384,2048,0.3964223861694336
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,16384,3584,0.41074771881103517
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,16384,1536,0.41921173731486006
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,16384,3072,0.3555264155069987
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,16384,1536,0.29406827290852866
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,16384,1024,0.30110079447428384
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,16384,1024,0.22195092837015787
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,16384,1024,0.2552618662516276
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,16384,768,0.2245002587636312
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,16384,768,0.17386239369710285
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,16384,1536,0.2636799971262614
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,16384,512,0.1645792007446289
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,16384,512,0.14194879531860352
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,16384,2048,0.3063935915629069
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,16384,256,0.11256106694539388
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,16384,256,0.11438079675038655
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,16384,768,0.2124938646952311
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,16384,128,0.09971840381622314
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,16384,128,0.10487893422444661
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,16384,512,0.20786666870117188
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,16384,64,0.08493546644846597
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,16384,64,0.10272106329600017
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,16384,32,0.08548266887664795
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,16384,32,0.10083626906077067
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,16384,256,0.21027626991271972
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,16384,128,0.22509652773539224
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,12288,65536,9.925628662109375
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,12288,65536,15.675872802734375
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,12288,51200,12.027098592122396
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,12288,51200,7.537689717610677
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,12288,16384,3.744120534261068
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,12288,16384,2.4267723083496096
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,16384,65536,8.531252034505208
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,12288,12288,2.8767733256022137
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,12288,12288,1.8350645701090496
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,12288,16384,1.3932596842447915
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,12288,10240,2.3687700907389324
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,12288,10240,1.5386709849039713
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,12288,12288,1.0361685434977213
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,12288,8192,1.9035658518473308
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,12288,8192,1.220535405476888
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,12288,10240,0.8510485331217448
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,12288,7168,1.5223648071289062
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,12288,7168,1.0876202901204428
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,12288,8192,0.6234528223673503
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,12288,6144,1.3157418568929038
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,12288,6144,0.9261290868123373
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,12288,7168,0.5632821400960286
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,12288,5120,1.0389471689860026
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,12288,5120,0.7769738515218099
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,12288,6144,0.4757418632507324
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,12288,4096,0.8803914388020834
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,12288,4096,0.6198229471842448
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,12288,5120,0.4287061373392741
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,12288,3584,0.7219082514444988
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,12288,3584,0.5757567723592122
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,12288,4096,0.3656181335449219
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,12288,3072,0.620900281270345
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,12288,3072,0.46041173934936525
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,12288,3584,0.32674878438313804
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,12288,2560,0.5270357449849447
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,12288,2560,0.3891530672709147
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,12288,3072,0.29004907608032227
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,12288,2048,0.4176608085632324
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,12288,2048,0.3082133293151855
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,12288,2560,0.284770139058431
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,12288,1536,0.3156832059224447
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,12288,1536,0.23293120066324868
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,12288,1536,0.23601600329081215
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,12288,1024,0.21893760363260903
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,12288,2048,0.26053226788838707
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,12288,1024,0.16778666178385418
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,12288,768,0.16994239489237467
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,12288,768,0.13806719779968263
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,12288,1024,0.22730666796366372
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,12288,512,0.12208426793416341
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,12288,512,0.11423253218332927
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,12288,768,0.20730346043904624
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,12288,256,0.08441279729207357
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,12288,256,0.09237226645151773
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,12288,256,0.20551253954569498
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,12288,128,0.06623040040334066
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,12288,128,0.08083413441975912
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,12288,512,0.20379306475321451
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,12288,64,0.060813868045806886
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,12288,64,0.08084586461385092
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,12288,32,0.06365333398183187
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,12288,32,0.07933759689331055
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,12288,128,0.21106239954630532
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,10240,65536,8.436115010579426
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,10240,65536,12.73218485514323
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,10240,51200,9.676755777994792
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,10240,51200,6.48838399251302
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,12288,51200,4.3336949666341145
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,10240,16384,3.0387914021809896
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,10240,16384,2.1540545145670573
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,10240,16384,1.1396991729736328
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,10240,12288,2.304991912841797
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,10240,12288,1.613058090209961
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,12288,65536,5.562497965494791
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,10240,10240,1.96386235555013
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,10240,10240,1.3378463745117188
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,10240,12288,0.8337514877319336
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,10240,8192,1.5663626352945963
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,10240,8192,1.0676128387451171
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,10240,10240,0.6801749547322591
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,10240,7168,1.3059818267822265
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,10240,7168,0.9433535893758138
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,10240,8192,0.5268661181131999
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,10240,6144,1.1227743784586588
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,10240,6144,0.8266464233398437
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,10240,7168,0.48388051986694336
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,10240,5120,0.9005738576253256
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,10240,5120,0.6909098943074544
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,10240,6144,0.41301441192626953
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,10240,4096,0.6867530822753907
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,10240,4096,0.5511391957600911
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,10240,5120,0.3698314666748047
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,10240,3584,0.5924053192138672
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,10240,3584,0.5207210540771484
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,10240,4096,0.31633173624674477
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,10240,3072,0.5157034556070964
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,10240,3072,0.41187413533528644
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,10240,3584,0.2851978619893392
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,10240,2560,0.4303882598876953
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,10240,2560,0.34178345998128257
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,10240,3072,0.2580992062886556
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,10240,2048,0.3503871917724609
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,10240,2048,0.27721068064371746
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,10240,2560,0.251309871673584
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,10240,1536,0.26473493576049806
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,10240,1536,0.20356693267822265
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,10240,51200,3.4937121073404946
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,10240,1024,0.18147625923156738
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,10240,1024,0.15529707272847493
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,10240,1536,0.20839893023173012
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,10240,768,0.14195733070373534
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,10240,768,0.11932586828867595
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,10240,2048,0.23012266159057618
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,10240,512,0.10453653335571289
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,10240,512,0.09988586902618408
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,10240,1024,0.19977706273396809
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,10240,256,0.07026453018188476
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,10240,768,0.18177493413289386
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,10240,256,0.0806272029876709
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,10240,128,0.06350613435109456
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,10240,128,0.07348266442616781
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,10240,128,0.18659092585245768
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,10240,64,0.05300480127334595
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,10240,64,0.07186240355173747
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,10240,32,0.055087999502817786
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,10240,256,0.18041067123413085
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,10240,512,0.1779381275177002
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,10240,32,0.07035306294759115
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,8192,65536,7.405608622233073
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,8192,65536,10.208358764648438
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,8192,51200,8.053995768229168
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,8192,51200,5.733400472005209
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,8192,16384,2.4273089090983073
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,8192,16384,1.8483562469482422
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,10240,65536,4.540972900390625
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,8192,12288,1.7862378438313802
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,8192,12288,1.3932533264160156
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,8192,16384,0.9045034408569336
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,8192,10240,1.5922239939371745
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,8192,10240,1.153113555908203
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,8192,12288,0.6343488057454427
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,8192,10240,0.5490421295166016
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,8192,8192,1.1568927764892578
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,8192,8192,0.9117706934611002
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,8192,7168,0.9820800145467123
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,8192,7168,0.8114965438842774
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,8192,7168,0.4046794573465983
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,8192,6144,0.8427263895670573
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,8192,6144,0.6906218846638997
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,8192,8192,0.4374058723449707
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,8192,5120,0.6691658655802409
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,8192,5120,0.6332821528116862
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,8192,5120,0.31280320485432944
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,8192,4096,0.5362656275431316
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,8192,4096,0.4869855880737305
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,8192,6144,0.34921067555745444
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,8192,3584,0.47612800598144533
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,8192,3584,0.40751040776570635
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,8192,3584,0.23478825887044272
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,8192,4096,0.26468159357706705
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,8192,3072,0.407317320505778
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,8192,3072,0.3492735862731934
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,8192,2560,0.338590939839681
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,8192,2560,0.2965514818827311
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,8192,2560,0.197596804300944
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,8192,2048,0.27571840286254884
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,8192,2048,0.22926932970682778
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,8192,3072,0.20453012784322105
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,8192,1536,0.21067412694295248
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,8192,1536,0.17358613014221191
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,8192,1536,0.14871145884195963
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,8192,1024,0.14667733510335285
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,8192,1024,0.1243018627166748
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,8192,2048,0.1726357301076253
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,8192,768,0.11357653141021729
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,8192,768,0.10370026429494221
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,8192,768,0.11826666990915935
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,8192,512,0.0826047976811727
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,8192,512,0.0866101344426473
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,8192,1024,0.13982826868693035
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,8192,256,0.056898132960001624
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,8192,256,0.06801706949869792
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,8192,256,0.11751893361409504
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,8192,128,0.05177919864654541
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,8192,128,0.06170239845911661
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,8192,512,0.11445973714192707
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,8192,64,0.04179840087890625
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,8192,64,0.06193066835403442
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,8192,128,0.12739626566569012
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,8192,32,0.04480746587117513
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,8192,32,0.06044160127639771
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,7168,65536,6.748305257161458
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,7168,65536,8.62750244140625
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,7168,51200,6.999451700846355
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,7168,51200,5.217257690429688
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,8192,51200,2.8196990966796873
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,7168,16384,2.137769571940104
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,7168,16384,1.705366388956706
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,7168,16384,0.7945536295572917
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,7168,12288,1.6303423563639323
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,7168,12288,1.3001195271809896
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,8192,65536,3.5570281982421874
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,7168,10240,1.3210112253824868
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,7168,10240,1.0610432306925455
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,7168,12288,0.5678911844889323
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,7168,8192,1.0616874694824219
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,7168,8192,0.8827893575032553
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,7168,10240,0.4935445467631022
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,7168,7168,0.8890517552693685
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,7168,7168,0.7513450622558594
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,7168,7168,0.36463359196980794
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,7168,6144,0.731009038289388
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,7168,6144,0.7125461578369141
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,7168,8192,0.39586772918701174
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,7168,5120,0.6201087951660156
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,7168,5120,0.5452981313069661
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,7168,5120,0.28556906382242836
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,7168,4096,0.4977717399597168
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,7168,4096,0.43619734446207686
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,7168,6144,0.31389760971069336
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,7168,3584,0.4216981252034505
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,7168,3584,0.3824650764465332
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,7168,3584,0.21361494064331055
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,7168,3072,0.36484587987264
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,7168,3072,0.3518421490987142
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,7168,4096,0.24172587394714357
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,7168,2560,0.3117568016052246
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,7168,2560,0.2751562754313151
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,7168,2560,0.18457919756571453
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,7168,2048,0.253276793162028
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,7168,2048,0.2175509293874105
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,7168,3072,0.1893781344095866
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,7168,1536,0.1918293317159017
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,7168,1536,0.15756799379984537
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,7168,1536,0.14496320088704426
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,7168,1024,0.12889280319213867
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,7168,1024,0.11451733112335205
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,7168,2048,0.16457494099934894
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,7168,768,0.10099199612935383
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,7168,768,0.09621333281199138
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,7168,768,0.12100266615549723
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,7168,512,0.07430506547292073
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,7168,512,0.08018666903177897
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,7168,1024,0.13653119405110675
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,7168,256,0.050424532095591224
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,7168,256,0.06378560066223145
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,7168,256,0.11987626552581787
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,7168,512,0.11741973559061687
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,7168,128,0.04592853387196859
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,7168,128,0.05760213136672974
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,7168,128,0.12661653359731037
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,7168,64,0.0385696013768514
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,7168,64,0.056848001480102536
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,7168,32,0.0395082672437032
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,7168,32,0.05346986850102743
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,6144,65536,7.380018107096355
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,6144,65536,6.068511962890625
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,7168,51200,2.4095700581868487
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,6144,51200,4.74072011311849
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,6144,51200,5.759200032552084
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,6144,16384,1.8529856363932293
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,6144,16384,1.5435039520263671
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,6144,16384,0.6840821584065755
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,7168,65536,3.1907625834147137
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,6144,12288,1.2669194539388022
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,6144,12288,1.166843795776367
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,6144,10240,1.126701863606771
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,6144,10240,0.9879498799641928
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,6144,10240,0.43969173431396485
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,6144,8192,0.9056959788004557
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,6144,8192,0.7873226801554363
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,6144,12288,0.49541012446085614
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,6144,7168,0.7431840260823568
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,6144,7168,0.7303871790568034
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,6144,8192,0.35443840026855467
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,6144,6144,0.6437546412150066
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,6144,6144,0.6071253458658854
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,6144,7168,0.3260277430216471
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,6144,5120,0.5169909477233887
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,6144,5120,0.509170118967692
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,6144,5120,0.2592202663421631
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,6144,4096,0.4139264106750488
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,6144,4096,0.39422613779703775
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,6144,6144,0.28412160873413084
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,6144,3584,0.3756671905517578
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,6144,3584,0.3628160158793131
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,6144,4096,0.21889066696166992
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,6144,3584,0.19492053985595703
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,6144,3072,0.3140170733133952
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,6144,3072,0.30011625289916993
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,6144,2560,0.2681013425191244
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,6144,2560,0.25090133349100746
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,6144,2560,0.16638399759928385
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,6144,2048,0.21834133466084799
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,6144,2048,0.1968448003133138
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,6144,3072,0.17206506729125975
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,6144,1536,0.15816747347513835
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,6144,1536,0.14233919779459636
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,6144,1536,0.1314527988433838
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,6144,1024,0.1094208002090454
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,6144,1024,0.10577812989552815
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,6144,2048,0.1481749375661214
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,6144,768,0.08662292957305909
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,6144,768,0.0875104029973348
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,6144,768,0.10931946436564129
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,6144,1024,0.12275199890136719
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,6144,512,0.06337279876073201
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,6144,512,0.07374186515808105
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,6144,256,0.04437546730041504
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,6144,256,0.06167893409729004
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,6144,256,0.10863253275553386
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,6144,128,0.03506879806518555
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,6144,128,0.054851198196411134
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,6144,512,0.1062165339787801
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,6144,64,0.034254932403564455
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,6144,64,0.05362879832585653
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,6144,32,0.0371338685353597
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,6144,128,0.11488320032755535
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,6144,32,0.052286934852600095
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,5120,65536,5.466230265299479
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,5120,65536,6.000379943847657
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,5120,51200,5.06104736328125
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,5120,51200,4.316214497884115
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,6144,51200,2.1395306905110676
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,5120,16384,1.4568714141845702
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,5120,16384,1.3957387288411458
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,5120,16384,0.5736511866251628
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,5120,12288,1.0777290344238282
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,5120,12288,1.0546890894571939
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,6144,65536,2.744435119628906
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,5120,10240,0.9090933481852213
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,5120,10240,0.8900917053222657
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,5120,10240,0.3856522560119629
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,5120,8192,0.713580830891927
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,5120,8192,0.740345573425293
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,5120,12288,0.4327232042948405
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,5120,7168,0.6428362528483074
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,5120,7168,0.6302154541015625
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,5120,7168,0.28511571884155273
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,5120,6144,0.5269749323527019
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,5120,6144,0.5476810455322265
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,5120,8192,0.3098698616027832
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,5120,5120,0.4349973360697429
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,5120,5120,0.44635305404663084
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,5120,5120,0.22797652880350747
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,5120,4096,0.34813334147135416
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,5120,4096,0.3633002599080404
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,5120,6144,0.2490623950958252
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,5120,3584,0.3045834541320801
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,5120,3584,0.3225269317626953
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,5120,3584,0.172161070505778
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,5120,4096,0.18923519452412924
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,5120,3072,0.2623487949371338
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,5120,3072,0.26995519002278645
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,5120,2560,0.21602346102396647
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,5120,2560,0.22288106282552084
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,5120,2560,0.14777812957763672
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,5120,2048,0.17474560737609862
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,5120,2048,0.17579414049784342
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,5120,3072,0.15384000142415363
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,5120,1536,0.13562347094217936
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,5120,1536,0.13137919902801515
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,5120,1536,0.11934933662414551
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,5120,1024,0.09390933513641357
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,5120,1024,0.0956063985824585
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,5120,2048,0.13338880538940429
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,5120,768,0.07216853300730387
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,5120,768,0.07981546719868979
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,5120,1024,0.11103999614715576
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,5120,512,0.05357760190963745
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,5120,768,0.1010474681854248
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,5120,512,0.06602240006128947
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,5120,256,0.0375765323638916
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,5120,256,0.0529749313990275
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,5120,256,0.09965439637502035
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,5120,128,0.030401066939036055
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,5120,128,0.047337599595387775
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,5120,512,0.09739306767781576
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,5120,64,0.02688213388125102
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,5120,64,0.04558826684951782
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,5120,32,0.028628265857696532
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,5120,32,0.04381226698557536
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,5120,128,0.10542933146158855
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,4096,65536,4.868332926432291
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,5120,51200,1.760136540730794
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,4096,65536,4.966179402669271
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,4096,51200,3.8092191060384115
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,4096,51200,3.8804896036783854
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,4096,16384,1.1694133758544922
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,4096,16384,1.2629535675048829
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,4096,16384,0.49164479573567704
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,4096,12288,0.8385599772135416
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,4096,12288,0.9577770868937174
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,5120,65536,2.2885409037272137
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,4096,10240,0.7312010447184245
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,4096,10240,0.8264191945393881
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,4096,12288,0.3749055862426758
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,4096,8192,0.5836640040079752
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,4096,8192,0.709665044148763
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,4096,10240,0.3328949292500814
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,4096,7168,0.5013258616129558
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,4096,7168,0.6117216110229492
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,4096,7168,0.24880426724751792
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,4096,6144,0.4220896085103353
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,4096,6144,0.5158474604288737
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,4096,8192,0.2688511848449707
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,4096,5120,0.35047785441080725
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,4096,5120,0.4015498797098796
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,4096,5120,0.2006293296813965
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,4096,4096,0.27992318471272787
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,4096,4096,0.3315978686014811
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,4096,6144,0.218832000096639
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,4096,3584,0.24245972633361818
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,4096,3584,0.28581120173136393
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,4096,3584,0.14827413558959962
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,4096,3072,0.208296537399292
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,4096,3072,0.2456831932067871
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,4096,4096,0.16720852851867676
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,4096,2560,0.17662612597147626
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,4096,2560,0.20459945996602377
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,4096,3072,0.12872213522593182
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,4096,2048,0.14401920636494953
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,4096,2560,0.12252266407012939
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,4096,2048,0.1554741382598877
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,4096,1536,0.10808426539103191
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,4096,1536,0.11529066562652587
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,4096,1536,0.0896170695622762
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,4096,1024,0.07548373540242513
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,4096,1024,0.08609066804250082
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,4096,2048,0.10635946591695149
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,4096,768,0.059431465466817227
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,4096,768,0.07267839908599853
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,4096,1024,0.08156373500823974
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,4096,512,0.04319253365198771
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,4096,512,0.06021973292032877
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,4096,768,0.06777066389719645
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,4096,256,0.030984532833099366
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,4096,256,0.04688426653544108
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,4096,256,0.06626133521397909
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,4096,128,0.027635200818379717
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,4096,512,0.06411413351694742
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,4096,128,0.04315626621246338
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,4096,64,0.023156267404556275
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,4096,64,0.03991786638895671
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,4096,32,0.023735467592875162
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,4096,128,0.07394986947377523
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,4096,32,0.03788373470306396
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,4096,51200,1.487884775797526
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,3584,65536,4.242230478922526
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,3584,65536,4.599818420410156
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,3584,51200,3.3345130920410155
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,3584,51200,3.6006507873535156
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,3584,16384,0.9640266418457031
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,3584,16384,1.17138671875
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,4096,65536,1.9112288157145183
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,3584,12288,0.7286133448282878
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,3584,12288,0.9088885625203451
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,3584,16384,0.44515412648518876
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,3584,10240,0.6118933359781902
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,3584,10240,0.775707753499349
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,3584,10240,0.30606613159179685
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,3584,8192,0.4903882662455241
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,3584,8192,0.597485860188802
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,3584,12288,0.34217812220255533
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,3584,7168,0.43583145141601565
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,3584,7168,0.5540138880411785
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,3584,8192,0.2453770637512207
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,3584,6144,0.36566934585571287
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,3584,7168,0.2288970629374186
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,3584,6144,0.4595456123352051
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,3584,5120,0.30084158579508463
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,3584,5120,0.3879167874654134
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,3584,5120,0.18515520095825194
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,3584,4096,0.2401301383972168
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,3584,4096,0.3048351923624674
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,3584,6144,0.20093866984049477
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,3584,3584,0.21288639704386392
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,3584,3584,0.27224747339884436
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,3584,3584,0.13709759712219238
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,3584,3072,0.1828213373819987
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,3584,4096,0.15347520510355633
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,3584,3072,0.2303541342417399
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,3584,2560,0.15221865971883136
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,3584,2560,0.1877845287322998
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,3584,2560,0.11650666395823162
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,3584,2048,0.12356159687042237
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,3584,2048,0.14545599619547528
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,3584,3072,0.12206400235493978
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,3584,1536,0.09355093638102213
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,3584,1536,0.11011093457539875
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,3584,1536,0.0904298702875773
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,3584,1024,0.0651690681775411
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,3584,1024,0.08219306468963623
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,3584,2048,0.10313493410746258
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,3584,768,0.051617066065470375
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,3584,768,0.06853866577148438
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,3584,768,0.07357439994812012
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,3584,1024,0.08266773223876953
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,3584,512,0.03862506548563639
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,3584,512,0.056651735305786134
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,3584,256,0.026870399713516235
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,3584,256,0.04415786663691203
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,3584,256,0.07204373677571616
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,3584,128,0.024780799945195518
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,3584,512,0.06989226341247559
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,3584,128,0.041382400194803874
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,3584,64,0.021112533410390218
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,3584,64,0.0390997330347697
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,3584,32,0.02142080068588257
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,3584,128,0.07830186684926352
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,3584,32,0.03569813172022502
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,3072,65536,3.8029919942220056
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,3072,65536,4.411480712890625
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,3584,51200,1.3441183725992838
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,3072,51200,2.950702921549479
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,3072,51200,3.4749013264973962
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,3072,16384,0.8554357528686524
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,3072,16384,1.1271530151367188
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,3584,65536,1.7029076894124349
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,3072,12288,0.6496298472086589
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,3072,16384,0.41761280695597336
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,3072,12288,0.8587466557820639
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,3072,10240,0.5442506790161132
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,3072,10240,0.7110474904378254
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,3072,10240,0.28760000864664714
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,3072,8192,0.43883520762125655
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,3072,8192,0.5897055943806966
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,3072,12288,0.3199978510538737
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,3072,7168,0.38081067403157554
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,3072,7168,0.5063178698221843
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,3072,8192,0.2324672063191732
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,3072,7168,0.21567254066467284
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,3072,6144,0.32713918685913085
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,3072,6144,0.4318389256795247
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,3072,5120,0.2697696050008138
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,3072,5120,0.3686837196350098
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,3072,5120,0.1747615973154704
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,3072,4096,0.21731840769449867
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,3072,4096,0.28991146087646485
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,3072,6144,0.18997227350870768
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,3072,3584,0.19162666002909343
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,3072,3584,0.2582752068837484
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,3072,3584,0.1296789328257243
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,3072,3072,0.16581652959187826
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,3072,3072,0.2251370588938395
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,3072,4096,0.14458346366882324
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,3072,2560,0.13683199882507324
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,3072,2560,0.18039147059122723
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,3072,3072,0.11501546700795491
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,3072,2048,0.11136000156402588
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,3072,2048,0.14008426666259766
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,3072,2560,0.10895679791768391
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,3072,1536,0.0840394655863444
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,3072,1536,0.10504106680552165
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,3072,2048,0.09630293051401774
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,3072,1024,0.05866346756617228
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,3072,1024,0.07840960025787354
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,3072,1536,0.0835541327794393
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,3072,768,0.045124268531799315
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,3072,768,0.06575466791788737
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,3072,1024,0.07564053535461426
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,3072,512,0.0352021336555481
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,3072,512,0.053514667352040614
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,3072,768,0.0667359987894694
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,3072,256,0.024779733022054037
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,3072,256,0.04141546487808227
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,3072,512,0.06303679943084717
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,3072,128,0.01946773330370585
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,3072,256,0.06531946659088135
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,3072,128,0.036849065621693926
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,3072,64,0.01716053287188212
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,3072,64,0.03462719917297363
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,3072,32,0.017754666010538735
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,3072,128,0.07182506720225015
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,3072,32,0.03169066707293193
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,3072,51200,1.2534805297851563
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,2560,65536,3.073869832356771
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,2560,65536,4.150658162434896
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,2560,51200,2.3688649495442706
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,2560,51200,3.2479326883951822
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,2560,16384,0.7127914428710938
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,2560,16384,1.0562549591064454
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,3072,65536,1.590294392903646
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,2560,12288,0.5360960006713867
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,2560,12288,0.7929962793986003
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,2560,16384,0.3725088119506836
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,2560,10240,0.4634688059488933
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,2560,10240,0.6659050623575846
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,2560,10240,0.25858559608459475
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,2560,8192,0.36934293111165367
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,2560,8192,0.5364063898722331
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,2560,12288,0.2860682805379232
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,2560,7168,0.3154815991719564
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,2560,7168,0.47775894800821944
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,2560,7168,0.19266986846923828
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,2560,6144,0.2652234713236491
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,2560,6144,0.41132265726725264
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,2560,8192,0.20816747347513834
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,2560,5120,0.2302560011545817
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,2560,5120,0.35571200052897134
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,2560,5120,0.15842132568359374
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,2560,4096,0.18466986020406087
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,2560,4096,0.272978146870931
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,2560,6144,0.17119253476460775
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,2560,3584,0.15945067405700683
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,2560,3584,0.23743467330932616
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,2560,3584,0.11549440224965413
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,2560,3072,0.13642452557881674
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,2560,3072,0.20371947288513184
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,2560,4096,0.1304106632868449
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,2560,2560,0.11633600393931072
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,2560,2560,0.16935359636942546
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,2560,2560,0.09601600170135498
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,2560,2048,0.09341440200805665
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,2560,2048,0.13124159971872967
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,2560,3072,0.10157333215077717
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,2560,1536,0.07222399711608887
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,2560,1536,0.09883306821187338
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,2560,1536,0.07157973448435465
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,2560,1024,0.04835093418757121
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,2560,1024,0.0720800002415975
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,2560,2048,0.08345920244852702
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,2560,768,0.038396799564361574
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,2560,768,0.059799468517303465
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,2560,1024,0.06283946832021078
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,2560,512,0.027473066250483198
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,2560,768,0.05386773347854614
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,2560,512,0.049876264731089276
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,2560,256,0.020088533560434975
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,2560,256,0.039262934525807695
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,2560,256,0.05307626724243164
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,2560,512,0.049865599473317465
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,2560,128,0.016716800133387246
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,2560,128,0.03506666819254557
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,2560,64,0.015980799992879234
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,2560,64,0.032789333661397295
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,2560,128,0.05999253193537394
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,2560,32,0.016522666811943053
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,2560,32,0.031048534313837688
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,2560,51200,1.1039349873860678
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,2048,65536,2.5152203877766928
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,2048,65536,3.873333231608073
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,2048,51200,1.966811752319336
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,2048,51200,3.0354891459147137
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,2048,16384,0.5889279683430989
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,2048,16384,0.9892810821533203
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,2048,16384,0.3250805219014486
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,2560,65536,1.3831914265950522
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,2048,12288,0.44317121505737306
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,2048,12288,0.7506741205851237
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,2048,10240,0.36849921544392905
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,2048,10240,0.6398176193237305
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,2048,10240,0.22619306246439613
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,2048,8192,0.2965973218282064
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,2048,8192,0.5254656155904134
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,2048,12288,0.25098133087158203
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,2048,7168,0.25419839223225915
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,2048,7168,0.4350282669067383
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,2048,7168,0.16971839269002278
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,2048,6144,0.21680000623067222
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,2048,6144,0.37508586247762044
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,2048,8192,0.18178666432698568
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,2048,5120,0.1844490687052409
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,2048,5120,0.31345173517862956
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,2048,5120,0.14056213696797687
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,2048,4096,0.14601066907246907
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,2048,4096,0.2510261376698812
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,2048,6144,0.1515552043914795
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,2048,3584,0.12464746634165447
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,2048,3584,0.22047999699910484
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,2048,3584,0.10205226739247639
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,2048,3072,0.10665813287099202
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,2048,3072,0.18976853688557943
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,2048,4096,0.11407360235850017
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,2048,2560,0.09038506348927816
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,2048,2560,0.1563434600830078
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,2048,2560,0.08378132979075113
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,2048,2048,0.07411946455637614
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,2048,3072,0.08975466887156168
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,2048,2048,0.12157546679178874
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,2048,1536,0.05634133418401083
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,2048,1536,0.09173226356506348
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,2048,51200,0.9534794489542643
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,2048,1024,0.039946667353312176
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,2048,1024,0.0676736036936442
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,2048,1536,0.059562667210896814
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,2048,768,0.031300266583760575
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,2048,768,0.05688639879226685
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,2048,2048,0.07107306321461995
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,2048,512,0.02226240038871765
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,2048,512,0.04695146481196086
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,2048,1024,0.050910933812459314
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,2048,256,0.01657919983069102
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,2048,256,0.033921066919962564
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,2048,768,0.04123520056406657
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,2048,128,0.013304533561070761
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,2048,256,0.041517865657806394
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,2048,128,0.030873600641886396
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,2048,512,0.03786773284276326
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,2048,64,0.01283093293507894
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,2048,32,0.013760000467300415
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,2048,64,0.029478400945663452
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,2048,32,0.028318933645884198
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,2048,128,0.04806720018386841
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,1536,65536,1.8835647583007813
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,1536,51200,1.4250347137451171
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,1536,65536,3.575536092122396
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,1536,51200,2.8008351643880207
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,1536,16384,0.4396053314208984
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,1536,16384,0.9133162816365561
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,2048,65536,1.1955178578694663
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,1536,12288,0.33287572860717773
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,1536,12288,0.6974890391031902
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,1536,16384,0.3010656038920084
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,1536,10240,0.2770304044087728
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,1536,10240,0.5855466842651367
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,1536,12288,0.233514674504598
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,1536,8192,0.23355627059936523
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,1536,8192,0.4608138720194499
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,1536,10240,0.21167999903361
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,1536,7168,0.19762773513793946
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,1536,7168,0.4052319844563802
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,1536,8192,0.17031466166178386
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,1536,7168,0.15878079732259115
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,1536,6144,0.1727776050567627
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,1536,6144,0.3479146639506022
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,1536,5120,0.13879040082295735
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,1536,5120,0.2910677274068197
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,1536,5120,0.13360427220662435
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,1536,4096,0.11066559950510661
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,1536,6144,0.14259840647379557
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,1536,4096,0.23524266878763833
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,1536,3584,0.09644800027211507
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,1536,3584,0.20624747276306152
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,1536,3584,0.0967146635055542
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,1536,3072,0.08313600222269693
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,1536,3072,0.17782506942749024
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,1536,4096,0.10729066530863444
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,1536,2560,0.071397336324056
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,1536,2560,0.1462389310201009
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,1536,2560,0.08049493630727132
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,1536,2048,0.05794346729914347
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,1536,2048,0.11381866931915283
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,1536,3072,0.08572800159454345
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,1536,1536,0.04407680034637451
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,1536,1536,0.08519786993662516
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,1536,2048,0.06938560009002685
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,1536,1536,0.058924798170725504
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,1536,1024,0.030586665868759154
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,1536,1024,0.06273386478424073
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,1536,1024,0.04939093192418416
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,1536,768,0.02314773400624593
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,1536,768,0.05138560136159261
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,1536,51200,0.8785631815592447
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,1536,512,0.017416532834370932
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,1536,512,0.03928639888763428
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,1536,768,0.043295999368031815
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,1536,256,0.01297706663608551
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,1536,256,0.030142933130264282
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,1536,128,0.011249066392580668
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,1536,512,0.040540798505147295
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,1536,256,0.04388800064722697
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,1536,128,0.02696320017178853
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,1536,64,0.010043733318646749
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,1536,32,0.011783466736475626
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,1536,64,0.026948267221450807
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,1536,128,0.052190931638081875
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,1024,65536,1.2041205088297526
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,1536,32,0.024987733364105223
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,1024,51200,0.9051509221394858
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,1024,65536,3.2801195780436196
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,1024,51200,2.567283121744792
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,1024,16384,0.2926261266072591
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,1024,16384,0.8402048110961914
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,1536,65536,1.1077247619628907
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,1024,12288,0.22534292538960776
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,1024,12288,0.6370741526285808
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,1024,16384,0.254964272181193
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,1024,10240,0.19432214101155598
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,1024,10240,0.5326613426208496
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,1024,12288,0.19906667073567708
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,1024,8192,0.15951040585835774
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,1024,8192,0.42780160903930664
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,1024,10240,0.1808095932006836
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,1024,7168,0.13458879788716632
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,1024,7168,0.37455574671427405
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,1024,7168,0.1358794689178467
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,1024,6144,0.11597973505655926
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,1024,8192,0.1456170717875163
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,1024,6144,0.3231242815653483
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,1024,5120,0.09540586471557617
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,1024,5120,0.2726218541463216
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,1024,5120,0.11614186763763427
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,1024,4096,0.07565013567606607
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,1024,6144,0.12351679801940918
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,1024,4096,0.2202400048573812
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,1024,3584,0.0667359987894694
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,1024,3584,0.19407679239908854
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,1024,3584,0.08261546293894449
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,1024,3072,0.0575925350189209
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,1024,4096,0.0923413356145223
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,1024,3072,0.16685546239217122
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,1024,2560,0.048853333791097006
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,1024,2560,0.13832319577534993
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,1024,3072,0.07234453360239665
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,1024,2048,0.040831998984018965
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,1024,2560,0.06638293266296387
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,1024,2048,0.10548693339029949
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,1024,1536,0.030723200241724653
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,1024,1536,0.07678933143615722
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,1024,1536,0.044438401858011886
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,1024,2048,0.0559061328570048
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,1024,1024,0.02138239940007528
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,1024,1024,0.05331946611404419
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,1024,768,0.016385066509246825
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,1024,768,0.04144959847132365
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,1024,51200,0.732151476542155
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,1024,512,0.012100266416867574
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,1024,512,0.03450133403142293
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,1024,1024,0.03618239959081014
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,1024,256,0.009465600053469341
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,1024,768,0.0300981342792511
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,1024,256,0.02718293269475301
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,1024,128,0.008318933347860973
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,1024,512,0.0272053341070811
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,1024,128,0.023562665780385336
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,1024,64,0.007501866420110066
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,1024,128,0.04149973392486572
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,1024,64,0.02404266595840454
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,1024,32,0.009172266721725464
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,1024,32,0.023025067647298177
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,768,65536,0.9186922709147135
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,1024,256,0.030906667311986286
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,768,51200,0.6905173619588216
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,768,65536,3.152252705891927
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,768,51200,2.460954793294271
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,768,16384,0.22564053535461426
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,768,16384,0.8015925089518229
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,1024,65536,0.915290641784668
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,768,12288,0.17311360041300455
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,768,12288,0.6048095703125
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,768,16384,0.23440319697062173
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,768,10240,0.14698987007141112
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,768,10240,0.5082549413045248
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,768,12288,0.1836693286895752
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,768,8192,0.11918613115946453
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,768,8192,0.40921812057495116
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,768,10240,0.16721386909484864
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,768,7168,0.1036959966023763
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,768,7168,0.36047252019246423
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,768,8192,0.1341503938039144
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,768,6144,0.08760106563568115
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,768,6144,0.3125301361083984
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,768,7168,0.12617386976877848
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,768,5120,0.07410666942596436
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,768,5120,0.26213226318359373
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,768,5120,0.10862826506296794
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,768,4096,0.059661865234375
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,768,4096,0.21133333841959634
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,768,6144,0.11508373419443767
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,768,3584,0.05235840082168579
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,768,3584,0.18765652974446614
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,768,4096,0.08601706822713216
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,768,3072,0.04594240188598633
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,768,3072,0.16077760060628254
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,768,3584,0.07801493008931479
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,768,2560,0.03869013388951619
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,768,2560,0.13394239743550618
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,768,51200,0.6646890640258789
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,768,3072,0.06859093507130941
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,768,2048,0.03213013410568237
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,768,2048,0.10078506469726563
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,768,1536,0.023919999599456787
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,768,1536,0.07358720302581787
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,768,2560,0.06260373195012411
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,768,1024,0.017056000232696534
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,768,1024,0.049211732546488446
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,768,2048,0.05347413221995036
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,768,1536,0.0419487992922465
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,768,768,0.013672533631324767
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,768,1024,0.034689064820607504
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,768,768,0.03864106734593709
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,768,512,0.010307199756304423
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,768,512,0.03484799861907959
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,768,256,0.009940266609191895
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,768,256,0.026971733570098876
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,768,768,0.03415679931640625
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,768,128,0.008746666709582011
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,768,128,0.024702932437260947
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,768,256,0.03425279855728149
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,768,64,0.007965866724650066
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,768,64,0.023197867472966514
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,768,128,0.040693334738413495
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,768,32,0.008334933718045553
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,768,32,0.020679465929667153
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,512,65536,0.5991477330525716
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,768,512,0.031122134129206343
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,512,51200,0.46207466125488283
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,512,65536,2.999254353841146
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,512,51200,2.3410634358723956
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,512,16384,0.15162986119588215
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,512,16384,0.7632256189982096
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,768,65536,0.8286933263142904
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,512,12288,0.11511466503143311
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,512,12288,0.5776671727498373
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,512,16384,0.2115605354309082
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,512,10240,0.1032981316248576
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,512,10240,0.4893749237060547
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,512,12288,0.16668267250061036
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,512,8192,0.08453013102213541
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,512,8192,0.39669440587361654
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,512,10240,0.1527370611826579
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,512,7168,0.07381013234456381
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,512,8192,0.1223637342453003
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,512,7168,0.3495701471964518
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,512,6144,0.06385173400243124
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,512,6144,0.30226561228434246
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,512,7168,0.11470613479614258
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,512,5120,0.05382506847381592
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,512,5120,0.25413440068562826
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,512,6144,0.10556480089823406
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,512,4096,0.04337919950485229
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,512,4096,0.20602666536966957
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,512,5120,0.09952106475830078
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,512,3584,0.03872106472651164
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,512,4096,0.07767786979675292
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,512,3584,0.18069653511047362
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,512,3072,0.03369813362757365
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,512,3072,0.1547327995300293
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,512,3584,0.0699722687403361
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,512,2560,0.027638399600982667
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,512,3072,0.06146666606267294
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,512,2560,0.1279584010442098
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,512,2048,0.023001599311828613
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,512,2048,0.0964458703994751
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,512,2560,0.0549888014793396
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,512,1536,0.01807039976119995
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,512,2048,0.04438826640446981
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,512,1536,0.06752959887186685
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,512,1024,0.012821333607037863
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,512,1024,0.043229866027832034
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,512,51200,0.5927743911743164
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,512,768,0.010526933272679647
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,512,1536,0.03450026512145996
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,512,768,0.036610134442647296
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,512,1024,0.027638399600982667
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,512,512,0.00812906672557195
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,512,512,0.030118399858474733
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,512,768,0.022708266973495483
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,512,256,0.005942399799823761
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,512,512,0.02026559909184774
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,512,256,0.026422399282455444
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,512,128,0.005423999826113383
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,512,128,0.02204479972521464
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,512,128,0.03334826628367106
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,512,64,0.005051733553409576
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,512,64,0.02091839909553528
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,512,32,0.006320000191529592
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,512,32,0.02064746618270874
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,512,256,0.024682666858037314
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,256,65536,0.3573376019795736
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,256,51200,0.286187744140625
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,512,65536,0.7396725336710612
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,256,65536,2.8700543721516927
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,256,16384,0.10195840199788411
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,256,16384,0.7364170710245769
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,256,51200,2.2477066040039064
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,256,12288,0.07286293506622314
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,256,12288,0.5582229614257812
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,256,16384,0.1908394654591878
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,256,10240,0.06234879891077677
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,256,12288,0.15138346354166668
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,256,10240,0.4679648081461589
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,256,8192,0.05160746574401855
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,256,8192,0.3780853271484375
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,256,10240,0.1387957255045573
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,256,7168,0.04449706474939982
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,256,8192,0.11009813149770101
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,256,7168,0.33268801371256507
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,256,6144,0.038680533568064376
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,256,6144,0.2866688092549642
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,256,7168,0.10288853645324707
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,256,5120,0.033061333497365314
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,256,5120,0.24157546361287435
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,256,6144,0.09649066925048828
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,256,4096,0.027876265843709308
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,256,4096,0.19448107083638508
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,256,51200,0.5320800145467122
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,256,3584,0.025006934007008867
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,256,5120,0.09141120115915934
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,256,3584,0.1706442674001058
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,256,3072,0.021643733978271483
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,256,4096,0.06980586846669515
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,256,3072,0.14629653294881184
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,256,2560,0.015642666816711427
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,256,3584,0.06314880053202311
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,256,2560,0.11942719618479411
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,256,2048,0.012834133704503379
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,256,3072,0.055544535319010414
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,256,65536,0.6701920191446941
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,256,2048,0.08815680344899496
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,256,1536,0.010702932874361675
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,256,1536,0.059443199634552
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,256,2560,0.049217065175374344
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,256,1024,0.007963733375072479
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,256,1024,0.03975253502527873
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,256,768,0.006734933455785115
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,256,2048,0.03838080167770386
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,256,1536,0.030112000306447347
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,256,768,0.03450666666030884
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,256,512,0.005843199789524078
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,256,1024,0.023526400327682495
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,256,256,0.004646400113900503
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,256,768,0.019734400510787963
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,256,256,0.02230506738026937
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,256,512,0.01742080052693685
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,256,256,0.021485867102940877
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,256,128,0.02071146567662557
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,256,64,0.003903999924659729
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,256,128,0.02756800055503845
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,256,32,0.004260266820589701
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,256,64,0.019434666633605956
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,256,32,0.017313067118326822
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,128,65536,0.32271785736083985
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,256,512,0.028539733091990156
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,256,128,0.003857066730658213
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,128,51200,0.25272000630696617
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,128,65536,2.8783414204915365
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,128,16384,0.09265173276265462
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,128,51200,2.23733647664388
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,128,16384,0.7314175923665365
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,128,12288,0.06459306478500366
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,128,12288,0.5525034586588542
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,128,16384,0.18897813161214191
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,128,10240,0.055148800214131675
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,128,10240,0.4634986559549968
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,128,12288,0.14911893208821614
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,128,8192,0.04541120131810506
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,128,51200,0.5241536140441895
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,128,8192,0.3742698669433594
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,128,7168,0.03948160012563069
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,128,10240,0.1363104025522868
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,128,65536,0.6596021016438802
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,128,6144,0.034840532143910724
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,128,7168,0.3286741256713867
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,128,8192,0.1093781312306722
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,128,5120,0.029556266466776532
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,128,6144,0.284665584564209
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,128,7168,0.10179306666056316
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,128,5120,0.23824000358581543
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,128,4096,0.02435093323389689
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,128,6144,0.09510933558146159
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,128,4096,0.19113173484802246
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,128,3584,0.021874133745829263
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,128,5120,0.0899072011311849
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,128,4096,0.06836053530375162
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,128,3584,0.16775360107421874
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,128,3072,0.016660267114639284
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,128,3072,0.14306453069051106
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,128,2560,0.010724266370137531
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,128,3584,0.06240959962209066
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,128,2560,0.11605439980824787
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,128,3072,0.054639999071757
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,128,2048,0.00834986666838328
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,128,2048,0.08474666277567545
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,128,1536,0.006818133095900218
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,128,2560,0.04828159809112549
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,128,1536,0.053814399242401126
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,128,2048,0.036924799283345536
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,128,1024,0.005444266895453135
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,128,1024,0.03826346794764201
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,128,768,0.005097599824269613
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,128,1024,0.023960532744725545
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,128,768,0.03318079908688863
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,128,512,0.004281599819660187
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,128,512,0.02717439929644267
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,128,768,0.01981333295504252
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,128,256,0.00344106654326121
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,128,512,0.01653866668542226
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,128,256,0.020574933290481566
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,128,128,0.0030421334008375804
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,128,128,0.020039467016855876
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,128,64,0.0029909332593282064
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,128,128,0.026868265867233277
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,128,32,0.0031690667072931922
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,128,32,0.01810773412386576
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,16384,128,1536,0.030551467339197797
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,64,65536,0.32344532012939453
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,64,51200,0.24865919748942056
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,128,256,0.0212501327196757
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,64,16384,0.09026239713033041
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,128,64,0.018373332420984902
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,64,12288,0.06396586497624715
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,64,16384,0.7306197484334309
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,64,10240,0.05401813189188639
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,64,12288,0.5529866536458333
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,64,8192,0.04468693335851033
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,64,10240,0.46346559524536135
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,64,7168,0.03848746617635091
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,64,8192,0.3743157386779785
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,64,6144,0.033905065059661864
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,64,7168,0.32865492502848304
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,64,5120,0.02838933269182841
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,64,51200,2.2557652791341143
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,64,6144,0.28318398793538413
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,64,65536,2.877452850341797
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,64,4096,0.023141332467397056
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,64,3584,0.020900267362594604
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,64,5120,0.23694933255513512
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,64,3072,0.01474240024884542
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,64,2560,0.00899733304977417
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,64,4096,0.19212907155354816
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,64,3584,0.16870187123616537
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,64,2048,0.009173333644866943
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,64,3072,0.14418667157491047
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,64,1536,0.007111466427644093
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,64,2560,0.11513280073801677
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,64,1024,0.005701333284378052
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,64,2048,0.08468266328175864
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,64,768,0.005039999882380167
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,64,1536,0.05313599904378256
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,64,512,0.004242133100827535
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,64,1024,0.0391103982925415
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,64,768,0.03431679805119832
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,64,256,0.003425066669782003
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,64,512,0.02786239981651306
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,64,128,0.002993066608905792
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,64,256,0.02068159977595011
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,64,64,0.00262719988822937
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,64,32,0.002812800059715907
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,64,64,0.01869973341623942
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,64,32,0.0160778671503067
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,32,51200,0.248087469736735
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,64,128,0.019341866175333657
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,32,16384,0.08718933264414469
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,32,65536,0.33501866658528645
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,32,12288,0.06053653160730997
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,32,16384,0.7293376286824544
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,32,10240,0.051412268479665124
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,32,12288,0.5495018641153971
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,32,51200,2.2429237365722656
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,32,8192,0.04217066764831543
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,32,7168,0.03677013317743937
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,32,65536,2.865827178955078
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,32,10240,0.4626367886861165
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,32,8192,0.3711541175842285
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,32,7168,0.3269941329956055
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,32,6144,0.28226025899251306
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,32,5120,0.23666133880615234
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,32,4096,0.022179200251897176
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,32,4096,0.18964373270670573
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,32,3584,0.01984106699625651
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,32,3584,0.16577280362447103
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,32,3072,0.012984533111254373
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,32,6144,0.03198826710383097
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,32,2560,0.008040533463160197
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,32,3072,0.14134186108907063
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,32,5120,0.0270687997341156
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,32,2560,0.11490560372670491
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,32,1536,0.0054666668176651
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,32,2048,0.08159786860148112
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,32,1024,0.0042453333735466
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,32,1536,0.05153813362121582
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,32,768,0.0038933334251244865
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,32,1024,0.03705919981002807
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,32,512,0.003420799970626831
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,32,768,0.03176533381144206
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,32,256,0.0026464000344276427
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,32,512,0.026371200879414875
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,32,128,0.002611200014750163
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,32,256,0.021259733041127524
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,32,64,0.002279466638962428
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,32,128,0.018641066551208497
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,32,32,0.0022069332500298817
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,32,64,0.017813332875569663
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,16384,32,32,0.015560533603032431
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,16384,32,2048,0.006638933221499126
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,65536,16384,9.9238037109375
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,65536,16384,5.549547831217447
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,65536,51200,17.95042521158854
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,65536,12288,7.733467610677083
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,65536,12288,4.196902465820313
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,65536,51200,32.526930745442705
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,65536,10240,6.545070902506511
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,65536,10240,3.550916290283203
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,65536,10240,1.8017247517903645
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,65536,8192,5.050836181640625
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,65536,8192,2.9024032592773437
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,65536,12288,2.194983418782552
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,65536,7168,4.45948486328125
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,65536,7168,2.3186314900716143
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,65536,16384,3.247101847330729
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,65536,6144,3.837553151448568
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,65536,6144,1.9468320210774739
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,65536,8192,1.4747605641682944
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,65536,5120,3.138025665283203
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,65536,5120,1.6014517466227214
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,65536,7168,1.3346591949462892
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,65536,4096,2.5170847574869795
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,65536,4096,1.2880362192789714
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,65536,6144,1.1526944478352865
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,65536,3584,2.176963297526042
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,65536,3584,1.0885269165039062
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,65536,5120,1.0380138397216796
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,65536,3072,1.6754379272460938
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,65536,3072,0.9411754608154297
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,65536,4096,0.9391328175862631
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,65536,2560,1.4620906829833984
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,65536,2560,0.8041823705037435
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,65536,3584,0.8631551742553711
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,65536,2048,1.1519306182861329
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,65536,2048,0.6551296234130859
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,65536,2560,0.7133450826009115
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,65536,1536,0.8472746531168619
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,65536,1536,0.5374720255533855
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,65536,3072,0.7432479858398438
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,65536,1024,0.5651498794555664
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,65536,1024,0.3977749188741048
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,65536,1024,0.5188032150268554
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,65536,768,0.439299201965332
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,65536,768,0.3129834810892741
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,65536,1536,0.5279274622599284
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,65536,512,0.33160107930501304
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,65536,512,0.25222400029500325
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,65536,512,0.48850453694661455
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,65536,256,0.22172907193501792
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,65536,256,0.20292159716288247
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,65536,2048,0.6435733159383138
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,65536,768,0.4969269434611003
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,65536,128,0.19314239819844564
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,65536,128,0.17954026858011882
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,65536,64,0.16291306813557943
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,65536,64,0.18277653058369953
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,65536,32,0.1681663990020752
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,65536,32,0.18287572860717774
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,65536,256,0.4884074529012044
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,65536,128,0.5063477198282877
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,51200,65536,18.20924275716146
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,51200,65536,32.460866292317704
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,51200,51200,25.760614013671876
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,51200,51200,13.596554565429688
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,51200,16384,7.973802693684895
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,51200,16384,4.450340270996094
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,51200,16384,2.513026173909505
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,51200,12288,5.757347106933594
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,51200,12288,3.444612375895182
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,51200,12288,1.6380128224690753
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,51200,10240,5.005894470214844
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,51200,10240,2.7025215148925783
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,51200,10240,1.3732138315836588
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,51200,8192,3.9373825073242186
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,51200,8192,2.1161802927652995
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,65536,51200,9.676608276367187
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,51200,7168,3.6330612182617186
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,51200,7168,1.8673355102539062
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,51200,8192,1.1163850148518881
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,51200,6144,2.907816569010417
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,51200,6144,1.5596895853678385
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,51200,7168,1.0374581019083657
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,51200,5120,2.4031967163085937
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,51200,5120,1.2543883005777994
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,51200,5120,0.8128437042236328
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,51200,4096,1.9174442291259766
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,51200,4096,1.089093271891276
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,51200,6144,0.8816458384195963
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,51200,3584,1.7020491282145183
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,51200,3584,0.8535253524780273
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,51200,3584,0.6510847727457683
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,51200,3072,1.4399935404459634
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,51200,3072,0.7487103780110677
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,51200,51200,7.658924865722656
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,51200,2560,1.0782527923583984
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,51200,2560,0.6554688135782878
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,51200,4096,0.7332021077473958
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,51200,2048,0.8658389409383138
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,51200,2048,0.5350261052449544
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,51200,2560,0.55797545115153
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,51200,1536,0.6579338709513347
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,51200,1536,0.40401919682820636
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,51200,2048,0.494052251180013
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,51200,1024,0.4497706731160481
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,51200,1024,0.2878570556640625
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,51200,3072,0.5751914978027344
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,51200,768,0.3427530606587728
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,51200,768,0.23733654022216796
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,51200,1024,0.4170271873474121
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,51200,512,0.2548362731933594
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,51200,512,0.1960063934326172
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,51200,768,0.391048526763916
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,51200,256,0.17115732828776042
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,51200,256,0.16300053596496583
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,51200,1536,0.4265898704528809
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,51200,128,0.13826133410135905
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,51200,128,0.14343466758728027
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,51200,256,0.3922527949015299
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,51200,64,0.12684053579966229
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,51200,64,0.1477471987406413
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,51200,32,0.13176426887512208
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,51200,32,0.14792213439941407
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,51200,128,0.3990090688069662
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,51200,512,0.3898559888203939
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,16384,65536,10.148816935221355
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,16384,65536,6.179010009765625
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,16384,51200,4.850374348958334
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,16384,51200,7.868553670247396
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,16384,16384,2.5402315775553386
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,16384,16384,1.505956268310547
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,16384,16384,0.96473388671875
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,16384,12288,1.8510122934977216
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,16384,12288,1.148043696085612
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,16384,12288,0.7436490376790365
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,16384,10240,1.542516326904297
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,16384,10240,0.9264511744181314
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,16384,10240,0.6344736099243165
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,16384,8192,1.2199211120605469
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,16384,8192,0.7979679743448893
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,16384,8192,0.45101226170857744
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,16384,7168,0.9639445622762045
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,16384,7168,0.6877749125162761
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,16384,51200,2.9604095458984374
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,16384,6144,0.8782133102416992
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,16384,6144,0.5987221399943035
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,16384,7168,0.3878986676534017
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,16384,5120,0.6707626978556316
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,16384,5120,0.47791039148966474
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,16384,5120,0.28095146814982097
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,16384,4096,0.5261589368184407
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,16384,4096,0.36618986129760744
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,16384,6144,0.32327359517415366
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,16384,3584,0.47931947708129885
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,16384,3584,0.3263733228047689
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,16384,3584,0.21353599230448403
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,16384,3072,0.41123412450154617
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,16384,3072,0.297216002146403
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,16384,4096,0.24060160319010415
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,16384,2560,0.336627197265625
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,16384,2560,0.23180267016092934
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,51200,65536,9.697804768880207
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,16384,2048,0.2893482526143392
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,16384,2048,0.1915551980336507
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,16384,2560,0.18130240440368653
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,16384,1536,0.20820159912109376
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,16384,1536,0.14704426129659015
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,16384,3072,0.18676907221476238
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,16384,1024,0.14848960240681966
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,16384,1024,0.11098453203837078
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,16384,1024,0.13215146859486898
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,16384,1536,0.13791252772013346
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,16384,768,0.11241707007090251
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,16384,2048,0.1589951992034912
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,16384,768,0.09423039754231771
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,16384,512,0.08245759805043539
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,16384,512,0.078765869140625
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,16384,256,0.05812906821568807
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,16384,256,0.06814080079396566
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,16384,256,0.11263999938964844
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,16384,128,0.05176639954249064
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,16384,128,0.06077653169631958
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,16384,128,0.11672639846801758
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,16384,512,0.11025386651357014
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,16384,64,0.045176533857981364
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,16384,32,0.04506880044937134
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,16384,64,0.05998506546020508
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,16384,32,0.05966933170954386
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,16384,768,0.1102677345275879
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,12288,65536,4.91231689453125
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,12288,65536,7.283648173014323
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,12288,51200,5.871755981445313
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,12288,51200,3.8223124186197914
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,12288,16384,1.8491701761881512
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,12288,16384,1.2129951477050782
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,16384,65536,3.9910219828287756
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,12288,12288,1.3362261454264321
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,12288,12288,0.9698517481486002
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,12288,16384,0.7071776072184245
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,12288,10240,1.0873920440673828
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,12288,10240,0.7770975748697917
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,12288,12288,0.5336650848388672
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,12288,8192,0.8918005625406901
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,12288,8192,0.7008565266927083
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,12288,8192,0.3255061467488607
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,12288,10240,0.4379712104797363
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,12288,7168,0.7411392211914063
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,12288,7168,0.5309802691141765
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,12288,6144,0.6146176020304362
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,12288,6144,0.4777482668558757
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,12288,6144,0.24914986292521157
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,12288,5120,0.5151487986246746
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,12288,5120,0.3818709373474121
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,12288,7168,0.29408534367879235
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,12288,4096,0.41502933502197265
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,12288,4096,0.30342079798380533
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,12288,4096,0.19418667157491049
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,12288,3584,0.365995724995931
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,12288,5120,0.22693653106689454
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,12288,3584,0.2629183928171793
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,12288,3072,0.3129834810892741
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,12288,3072,0.2238922595977783
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,12288,3072,0.15597333908081054
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,12288,2560,0.26202027002970374
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,12288,2560,0.1905226707458496
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,12288,3584,0.17365546226501466
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,12288,2048,0.20742719968159995
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,12288,2048,0.15452693303426107
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,12288,2048,0.13824853897094727
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,12288,1536,0.15797866185506185
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,12288,1536,0.12054293155670166
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,12288,2560,0.15235840479532878
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,12288,1024,0.1110357364018758
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,12288,1024,0.08967680136362711
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,12288,1024,0.12041066487630207
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,12288,768,0.08639787038167318
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,12288,768,0.07981013456980388
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,12288,1536,0.12523626486460368
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,12288,512,0.06279999812444051
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,12288,512,0.06603200038274129
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,12288,768,0.10935359795888264
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,12288,256,0.044461866219838456
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,12288,256,0.059701331456502284
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,12288,512,0.11001813411712646
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,12288,128,0.03541333278020223
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,12288,128,0.05168106555938721
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,12288,128,0.11207146644592285
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,12288,64,0.03416533470153808
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,12288,64,0.05262080033620199
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,12288,32,0.03715200026830037
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,12288,32,0.05142826636632284
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,12288,256,0.1111786683400472
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,10240,65536,4.18590087890625
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,10240,65536,6.04489491780599
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,10240,51200,5.035343933105469
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,10240,51200,3.262474568684896
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,12288,51200,2.125482686360677
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,10240,16384,1.5050837198893228
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,10240,16384,1.0741311391194661
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,10240,16384,0.5747711817423503
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,10240,12288,1.0579488118489584
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,10240,12288,0.8256448109944662
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,12288,65536,2.735620371500651
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,10240,10240,0.9069920221964518
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,10240,10240,0.6794069290161133
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,10240,12288,0.42259521484375
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,10240,8192,0.7178719838460286
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,10240,8192,0.5686464309692383
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,10240,10240,0.34728212356567384
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,10240,7168,0.6387744267781575
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,10240,7168,0.48975359598795576
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,10240,8192,0.27119038899739584
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,10240,6144,0.5340010960896809
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,10240,6144,0.417574405670166
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,10240,7168,0.2494709332784017
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,10240,5120,0.4242656071980794
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,10240,5120,0.331932799021403
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,10240,6144,0.21449813842773438
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,10240,4096,0.3414389292399088
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,10240,4096,0.276362673441569
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,10240,51200,1.6703210194905598
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,10240,3584,0.3155904134114583
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,10240,3584,0.23073280652364098
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,10240,5120,0.19299626350402832
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,10240,3072,0.25788586934407554
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,10240,3072,0.19619520505269367
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,10240,3584,0.15074987411499025
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,10240,2560,0.2124821345011393
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,10240,2560,0.1631818612416585
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,10240,4096,0.1661472002665202
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,10240,2048,0.17719359397888185
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,10240,2048,0.1369887987772624
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,10240,3072,0.13758079210917157
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,10240,1536,0.134227196375529
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,10240,1536,0.10535360177357991
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,10240,2048,0.12193493048350017
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,10240,1024,0.09353813330332438
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,10240,1024,0.08020266691843668
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,10240,2560,0.13416852951049804
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,10240,768,0.07105600039164225
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,10240,768,0.070360533396403
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,10240,1536,0.10979413191477458
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,10240,512,0.053201067447662356
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,10240,512,0.057650132973988855
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,10240,512,0.09650453726450602
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,10240,256,0.036831998825073244
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,10240,768,0.09612692991892496
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,10240,256,0.049617067972819014
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,10240,128,0.030381866296132404
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,10240,128,0.04400959809621175
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,10240,1024,0.10549866358439128
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,10240,64,0.02720106641451518
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,10240,64,0.04348160028457641
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,10240,32,0.028917332490285236
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,10240,32,0.04326826731363932
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,10240,128,0.09832106431325277
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,10240,256,0.09830719629923503
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,8192,65536,3.649862416585287
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,8192,65536,4.945493570963541
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,8192,51200,3.833759053548177
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,8192,51200,2.922019195556641
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,8192,16384,1.1367903391520182
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,8192,16384,0.9600309371948242
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,10240,65536,2.1904897054036456
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,8192,12288,0.9114784240722656
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,8192,12288,0.7105152130126953
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,8192,16384,0.46652266184488933
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,8192,10240,0.7346666971842448
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,8192,10240,0.6124490737915039
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,8192,10240,0.2863114674886068
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,8192,8192,0.5669269561767578
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,8192,8192,0.4687914530436198
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,8192,12288,0.3305919965108236
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,8192,7168,0.482366943359375
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,8192,7168,0.4174389203389485
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,8192,7168,0.2108405272165934
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,8192,6144,0.4096629460652669
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,8192,6144,0.3680224100748698
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,8192,8192,0.22891732851664223
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,8192,5120,0.34182399113972983
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,8192,5120,0.3094069480895996
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,8192,5120,0.16578133900960285
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,8192,4096,0.2733834584554037
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,8192,4096,0.22836373647054037
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,8192,6144,0.18175253868103028
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,8192,3584,0.2409290631612142
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,8192,3584,0.1988928000132243
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,8192,4096,0.14209920565287273
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,8192,3584,0.1256064017613729
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,8192,3072,0.20897706349690756
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,8192,3072,0.16952746709187824
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,8192,2560,0.1745525360107422
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,8192,2560,0.141978661219279
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,8192,2560,0.10711039702097576
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,8192,2048,0.14125119845072429
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,8192,2048,0.11684799989064534
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,8192,3072,0.11122666994730632
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,8192,1536,0.10885439713795979
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,8192,1536,0.09409386316935221
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,8192,1536,0.07996479670206705
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,8192,1024,0.07604053020477294
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,8192,1024,0.07049173514048258
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,8192,2048,0.09259413083394369
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,8192,768,0.05873706738154093
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,8192,768,0.06292693217595419
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,8192,768,0.06337493260701497
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,8192,1024,0.07544639905293783
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,8192,512,0.04358506600062052
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,8192,512,0.05298773447672526
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,8192,256,0.030916267633438112
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,8192,256,0.04530880053838094
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,8192,512,0.06349226633707682
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,8192,128,0.028149332602818804
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,8192,128,0.04016533295313517
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,8192,256,0.0653162678082784
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,8192,64,0.023126399517059325
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,8192,64,0.04010879993438721
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,8192,32,0.023912533124287923
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,8192,32,0.0372320016225179
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,8192,128,0.06762026945749919
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,7168,65536,3.3494847615559897
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,7168,65536,4.2554069519042965
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,7168,51200,3.359528605143229
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,7168,51200,2.624068196614583
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,8192,51200,1.3829877217610678
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,7168,16384,0.9659615834554037
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,7168,16384,0.8483402887980143
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,7168,16384,0.4019136110941569
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,7168,12288,0.7433855692545573
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,7168,12288,0.6760096232096354
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,8192,65536,1.7890997568766278
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,7168,10240,0.6152277628580729
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,7168,10240,0.5308469454447429
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,7168,12288,0.28997440338134767
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,7168,8192,0.4930495897928874
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,7168,8192,0.4722645441691081
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,7168,10240,0.25311466852823894
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,7168,7168,0.4167946815490723
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,7168,7168,0.371944522857666
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,7168,8192,0.2045205275217692
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,7168,6144,0.36512638727823893
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,7168,6144,0.31956799825032556
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,7168,7168,0.18913920720418292
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,7168,5120,0.2951562563578288
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,7168,5120,0.2633514722188314
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,7168,6144,0.16478506724039715
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,7168,4096,0.24189440409342447
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,7168,4096,0.2187338670094808
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,7168,5120,0.14900693893432618
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,7168,3584,0.2075434684753418
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,7168,3584,0.17834773063659667
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,7168,3584,0.11458240350087483
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,7168,3072,0.17892266909281412
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,7168,3072,0.152402130762736
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,7168,4096,0.1278431971867879
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,7168,2560,0.15337173144022626
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,7168,2560,0.12868159612019855
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,7168,3072,0.10238719781239827
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,7168,2048,0.121562663714091
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,7168,2048,0.10655679702758789
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,7168,2560,0.09868799845377604
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,7168,1536,0.09478933016459147
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,7168,1536,0.08419413566589355
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,7168,1536,0.07803520361582438
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,7168,1024,0.06577813227971395
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,7168,1024,0.06514453490575155
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,7168,51200,1.1662347157796225
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,7168,768,0.05356800158818563
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,7168,768,0.06602986653645834
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,7168,2048,0.08732586701711019
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,7168,512,0.0380181352297465
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,7168,512,0.049607467651367185
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,7168,768,0.06523733139038086
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,7168,1024,0.07352426846822104
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,7168,256,0.027012266715367633
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,7168,512,0.0649450659751892
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,7168,256,0.04293440183003743
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,7168,128,0.02481813430786133
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,7168,128,0.037221332391103104
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,7168,64,0.020760534207026164
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,7168,64,0.03518613179524739
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,7168,32,0.021127466360727945
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,7168,32,0.03654826482137044
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,7168,128,0.06748053232828775
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,7168,256,0.0670528014500936
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,6144,65536,3.050404357910156
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,6144,65536,3.771070861816406
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,6144,51200,2.9244148254394533
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,6144,51200,2.424521636962891
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,6144,16384,0.840334955851237
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,6144,16384,0.8107231775919596
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,6144,16384,0.3615551948547363
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,6144,12288,0.6297365188598633
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,7168,65536,1.5437365214029948
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,6144,12288,0.6646741231282551
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,6144,10240,0.5440928141276042
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,6144,10240,0.5065525372823079
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,6144,10240,0.23322772979736328
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,6144,8192,0.4398048082987468
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,6144,8192,0.40404586791992186
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,6144,12288,0.2638048013051351
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,6144,7168,0.37982826232910155
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,6144,7168,0.3605418523152669
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,6144,8192,0.1889791965484619
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,6144,6144,0.3242805480957031
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,6144,6144,0.2991626739501953
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,6144,7168,0.17511359850565594
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,6144,5120,0.26769065856933594
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,6144,5120,0.24896960258483886
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,6144,5120,0.13951999346415203
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,6144,4096,0.21440745989481608
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,6144,6144,0.15196372667948405
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,6144,4096,0.1939626693725586
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,6144,3584,0.18881600697835285
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,6144,3584,0.16813759803771972
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,6144,3584,0.10686399936676025
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,6144,3072,0.16529067357381183
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,6144,3072,0.1484074592590332
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,6144,4096,0.11940373579661052
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,6144,2560,0.1337994734446208
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,6144,2560,0.11999893188476562
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,6144,3072,0.09585279623667399
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,6144,2560,0.09194880326588949
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,6144,2048,0.11265493233998616
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,6144,2048,0.09842987060546875
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,6144,1536,0.08212693532307944
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,6144,1536,0.07951146761576335
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,6144,1536,0.0724234660466512
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,6144,1024,0.05676586627960205
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,6144,1024,0.05970026652018229
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,6144,2048,0.08118186791737875
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,6144,768,0.044955734411875406
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,6144,768,0.053597867488861084
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,6144,1024,0.06778346697489421
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,6144,512,0.032305065790812174
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,6144,512,0.05151679913202921
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,6144,768,0.05999893347422282
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,6144,256,0.03012053370475769
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,6144,256,0.04543466567993164
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,6144,512,0.06001919905344645
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,6144,128,0.019166932503382365
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,6144,256,0.06227733294169108
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,6144,128,0.0331007997194926
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,6144,64,0.017121066649754844
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,6144,64,0.03203093409538269
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,6144,32,0.01758079926172892
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,6144,32,0.03150186737378438
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,6144,128,0.06234026749928793
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,5120,65536,2.8484926859537762
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,5120,65536,3.195013427734375
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,5120,51200,2.467001597086589
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,5120,51200,2.203558349609375
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,6144,51200,1.070427703857422
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,5120,16384,0.7034965515136719
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,5120,16384,0.7309792200724284
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,5120,16384,0.3025909423828125
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,5120,12288,0.542578125
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,5120,12288,0.5677226384480794
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,6144,65536,1.3669994354248047
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,5120,10240,0.4449770609537761
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,5120,10240,0.47030188242594406
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,5120,12288,0.2288287957509359
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,5120,8192,0.3513514518737793
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,5120,8192,0.35791788101196287
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,5120,10240,0.2037013371785482
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,5120,7168,0.3238186518351237
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,5120,7168,0.32539733250935876
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,5120,8192,0.16516159375508627
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,5120,6144,0.27120854059855143
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,5120,6144,0.27093547185262046
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,5120,7168,0.1531295935312907
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,5120,5120,0.2218389352162679
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,5120,5120,0.22150932947794594
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,5120,6144,0.13403840065002443
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,5120,4096,0.1746112028757731
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,5120,4096,0.17034986813863118
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,5120,5120,0.12338879903157551
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,5120,3584,0.15350507100423177
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,5120,3584,0.14803733825683593
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,5120,3584,0.09603947003682455
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,5120,3072,0.13182186285654704
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,5120,3072,0.12804480393727619
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,5120,4096,0.10509119828542073
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,5120,2560,0.10925439993540446
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,5120,2560,0.10849599838256836
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,5120,2560,0.08367466926574707
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,5120,2048,0.08810133139292399
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,5120,2048,0.09064319928487143
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,5120,3072,0.0868234634399414
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,5120,1536,0.07140800158182779
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,5120,1536,0.07317759990692138
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,5120,1536,0.06635839939117431
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,5120,1024,0.04733653465906779
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,5120,1024,0.05550826787948608
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,5120,2048,0.07415786584218344
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,5120,768,0.036660265922546384
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,5120,768,0.05001920064290365
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,5120,51200,0.9122282663981119
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,5120,512,0.027088000377019243
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,5120,1024,0.062046933174133304
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,5120,512,0.04319039980570476
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,5120,256,0.020683733622233073
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,5120,256,0.035545599460601804
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,5120,768,0.05555946826934814
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,5120,128,0.016985599199930826
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,5120,128,0.03300586740175883
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,5120,256,0.05841173330942789
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,5120,64,0.01573013365268707
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,5120,64,0.030728532870610552
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,5120,32,0.016482133666674295
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,5120,128,0.05848639806111654
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,5120,512,0.056005334854125975
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,5120,32,0.030203733841578168
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,4096,65536,2.551660919189453
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,4096,65536,2.554162089029948
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,4096,51200,1.9155754089355468
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,4096,51200,2.0026004791259764
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,4096,16384,0.5943520228068034
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,4096,16384,0.6653610865275066
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,4096,16384,0.25218772888183594
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,4096,12288,0.44663467407226565
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,4096,12288,0.49906667073567706
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,5120,65536,1.1631317138671875
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,4096,10240,0.36161492665608724
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,4096,10240,0.42704960505167644
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,4096,10240,0.17323412895202636
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,4096,12288,0.19399147033691405
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,4096,8192,0.29432106018066406
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,4096,8192,0.32354987462361656
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,4096,7168,0.25228479703267415
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,4096,7168,0.2975071907043457
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,4096,7168,0.12934292952219645
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,4096,6144,0.21653760274251305
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,4096,6144,0.24329493840535482
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,4096,8192,0.13929279645284015
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,4096,5120,0.17940586407979328
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,4096,5120,0.2033397356669108
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,4096,6144,0.11431039969126384
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,4096,4096,0.14245866139729818
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,4096,4096,0.1522314707438151
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,4096,5120,0.10586667060852051
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,4096,3584,0.12510080337524415
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,4096,3584,0.13380160331726074
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,4096,3584,0.07947839895884196
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,4096,3072,0.10747733116149902
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,4096,3072,0.11488853295644123
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,4096,4096,0.08876480261484782
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,4096,2560,0.0897098700205485
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,4096,2560,0.09741013050079346
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,4096,3072,0.07077866395314535
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,4096,2560,0.06706453164418538
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,4096,2048,0.07389866511027018
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,4096,2048,0.08145919640858969
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,4096,1536,0.055661865075429286
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,4096,1536,0.06559786796569825
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,4096,1536,0.046855465571085615
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,4096,1024,0.038837333520253495
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,4096,1024,0.050589867432912196
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,4096,2048,0.056745600700378415
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,4096,768,0.030428800980250043
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,4096,768,0.046639998753865555
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,4096,1024,0.04286613464355469
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,4096,512,0.021899733940760294
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,4096,512,0.03739519913991292
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,4096,768,0.03586666584014893
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,4096,256,0.016359466314315795
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,4096,256,0.031231999397277832
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,4096,512,0.036690131823221846
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,4096,128,0.013327999909718832
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,4096,256,0.03909866809844971
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,4096,128,0.0293503999710083
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,4096,64,0.013296000162760415
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,4096,64,0.027154133717219038
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,4096,32,0.014105600118637086
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,4096,128,0.03986666599909465
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,4096,32,0.0269269327322642
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,3584,65536,2.0679370880126955
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,3584,65536,2.329986063639323
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,3584,51200,1.5643391927083334
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,3584,51200,1.8272618611653644
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,4096,51200,0.7443957646687825
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,3584,16384,0.46357227961222336
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,3584,16384,0.5969386418660482
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,3584,16384,0.2293834686279297
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,3584,12288,0.34856106440226237
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,3584,12288,0.44946667353312175
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,4096,65536,0.9413642883300781
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,3584,10240,0.31242453257242836
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,3584,10240,0.3791349411010742
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,3584,12288,0.17723946571350097
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,3584,8192,0.24849066734313965
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,3584,8192,0.3017888069152832
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,3584,10240,0.15830079714457196
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,3584,7168,0.22123519579569498
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,3584,7168,0.27696425120035806
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,3584,8192,0.12868586381276448
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,3584,6144,0.18636266390482586
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,3584,7168,0.12010026772816976
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,3584,6144,0.2263925393422445
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,3584,5120,0.15354347229003906
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,3584,5120,0.18555413881937663
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,3584,5120,0.09928533236185709
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,3584,4096,0.12112320264180501
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,3584,4096,0.14326507250467937
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,3584,6144,0.10684373378753662
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,3584,3584,0.10769920349121094
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,3584,3584,0.12444586753845215
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,3584,4096,0.08355093002319336
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,3584,3072,0.09141226609547934
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,3584,3584,0.0763754685719808
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,3584,3072,0.10772373676300048
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,3584,2560,0.0768992026646932
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,3584,2560,0.09201493263244628
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,3584,3072,0.06790613333384196
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,3584,2048,0.06316479841868082
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,3584,2048,0.0776917298634847
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,3584,2560,0.06405119895935059
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,3584,1536,0.04776959816614787
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,3584,1536,0.0620959997177124
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,3584,2048,0.0555295983950297
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,3584,1024,0.03335466782251994
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,3584,1024,0.047470935185750324
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,3584,1536,0.0491370677947998
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,3584,768,0.0252074658870697
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,3584,768,0.04294506708780925
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,3584,768,0.039854931831359866
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,3584,512,0.018965333700180054
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,3584,1024,0.04467093149820964
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,3584,512,0.0342741330464681
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,3584,256,0.014152533809343972
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,3584,256,0.029154133796691895
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,3584,256,0.04315946499506633
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,3584,128,0.01330560048421224
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,3584,128,0.027030400435129803
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,3584,512,0.04044479926427205
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,3584,64,0.011591466267903645
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,3584,64,0.024810665845870973
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,3584,32,0.011612799763679505
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,3584,128,0.043082666397094724
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,3584,32,0.024795732895533242
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,3072,65536,1.8053525288899739
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,3584,51200,0.6721941630045574
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,3072,51200,1.4281002044677735
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,3072,65536,2.2391029357910157
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,3072,51200,1.7644906361897785
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,3072,16384,0.4387893358866374
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,3072,16384,0.5803253173828125
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,3072,16384,0.22578667004903158
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,3072,12288,0.3313119888305664
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,3072,12288,0.4392618815104167
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,3584,65536,0.8466367721557617
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,3072,10240,0.28996159235636393
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,3072,10240,0.36582934061686195
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,3072,10240,0.15612053871154785
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,3072,8192,0.22939200401306153
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,3072,8192,0.29130239486694337
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,3072,12288,0.17374080022176105
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,3072,7168,0.19544960657755533
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,3072,7168,0.25355626742045084
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,3072,7168,0.11797119776407879
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,3072,6144,0.16707305908203124
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,3072,6144,0.21827306747436523
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,3072,8192,0.12705600261688232
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,3072,5120,0.14049919446309406
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,3072,5120,0.17936320304870607
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,3072,5120,0.09752106666564941
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,3072,4096,0.11259840329488117
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,3072,4096,0.13781332969665527
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,3072,6144,0.10482772986094158
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,3072,3584,0.0976085344950358
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,3072,3584,0.11855040391286213
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,3072,3584,0.07398293018341065
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,3072,4096,0.08190080324808756
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,3072,3072,0.08472959995269776
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,3072,3072,0.10274240175882976
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,3072,2560,0.07185386816660563
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,3072,2560,0.08818773428599039
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,3072,2560,0.06193919976552328
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,3072,2048,0.05708906650543213
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,3072,2048,0.0741866668065389
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,3072,3072,0.06597119967142741
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,3072,1536,0.043032534917195636
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,3072,1536,0.05778773228327433
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,3072,1536,0.04618560075759888
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,3072,1024,0.029308799902598066
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,3072,1024,0.043628799915313723
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,3072,2048,0.05334186553955078
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,3072,768,0.023162666956583658
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,3072,768,0.03906773328781128
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,3072,1024,0.042762664953867595
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,3072,768,0.03795093297958374
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,3072,512,0.01716266671816508
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,3072,512,0.03248320023218791
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,3072,256,0.012950399518013
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,3072,256,0.028563199440638225
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,3072,256,0.04103253285090129
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,3072,128,0.012030933300654094
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,3072,512,0.038306132952372236
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,3072,128,0.024779733022054037
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,3072,64,0.00998293360074361
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,3072,64,0.02303253412246704
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,3072,128,0.04129279851913452
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,3072,32,0.023059199253718056
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,3072,32,0.010790399710337321
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,2560,65536,1.5864128112792968
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,2560,65536,2.1050453186035156
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,3072,51200,0.6672768274943034
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,2560,51200,1.1842229207356771
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,2560,51200,1.6353173573811848
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,2560,16384,0.35663251876831054
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,2560,16384,0.5348874409993489
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,2560,16384,0.20249387423197426
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,2560,12288,0.2701183954874674
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,2560,12288,0.41863358815511065
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,3072,65536,0.8394271850585937
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,2560,10240,0.23771306673685708
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,2560,10240,0.34706026713053384
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,2560,10240,0.14146453539530437
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,2560,8192,0.19187733332316081
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,2560,8192,0.2792799949645996
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,2560,12288,0.15682346026102703
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,2560,7168,0.17046079635620118
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,2560,7168,0.24568533897399902
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,2560,7168,0.10714453061421711
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,2560,6144,0.14339626630147298
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,2560,6144,0.2110367933909098
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,2560,8192,0.11473493576049805
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,2560,5120,0.12206400235493978
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,2560,5120,0.17370880444844564
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,2560,5120,0.089628799756368
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,2560,4096,0.09819093545277914
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,2560,4096,0.13475947380065917
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,2560,6144,0.09602133433024088
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,2560,3584,0.08824426333109538
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,2560,3584,0.11683519681294759
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,2560,3584,0.06722346941630045
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,2560,3072,0.0764959971110026
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,2560,3072,0.10080640316009522
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,2560,4096,0.07470613320668539
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,2560,2560,0.06374080181121826
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,2560,2560,0.08634666601816812
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,2560,3072,0.05865493218104044
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,2560,2048,0.052162134647369386
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,2560,2048,0.07217919826507568
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,2560,2560,0.05411946773529053
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,2560,1536,0.03936320145924886
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,2560,1536,0.05670506556828817
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,2560,2048,0.04619199832280477
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,2560,1024,0.027948800722757978
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,2560,1536,0.03987413247426351
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,2560,1024,0.04176213343938191
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,2560,768,0.021988266706466676
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,2560,768,0.037212800979614255
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,2560,51200,0.5951391855875652
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,2560,512,0.016551466782887776
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,2560,512,0.03099520007769267
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,2560,768,0.031779199838638306
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,2560,256,0.011733333269755047
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,2560,1024,0.0363594651222229
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,2560,256,0.027296000719070436
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,2560,128,0.009239466985066731
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,2560,128,0.023090134064356484
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,2560,512,0.032321067651112874
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,2560,64,0.008306133250395458
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,2560,256,0.03544319868087768
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,2560,32,0.008335999647776286
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,2560,64,0.02257066567738851
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,2560,128,0.035417600472768145
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,2560,32,0.021156267325083414
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,2048,65536,1.1903456370035808
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,2048,65536,1.940775426228841
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,2048,51200,0.9568288167317709
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,2048,51200,1.5225802103678385
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,2048,16384,0.29046398798624673
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,2048,16384,0.5013802528381348
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,2048,16384,0.17862613995869953
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,2048,12288,0.22006932894388834
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,2048,12288,0.38124799728393555
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,2560,65536,0.7463584264119466
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,2048,10240,0.19367039998372396
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,2048,10240,0.32048425674438474
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,2048,10240,0.12515626748402914
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,2048,8192,0.15661013921101888
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,2048,8192,0.2525983969370524
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,2048,12288,0.13855466842651368
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,2048,7168,0.13548693656921387
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,2048,7168,0.22272747357686362
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,2048,7168,0.09518506526947021
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,2048,6144,0.11421973705291748
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,2048,6144,0.19109973907470704
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,2048,8192,0.10171626408894856
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,2048,5120,0.09438827037811279
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,2048,5120,0.15666666030883789
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,2048,5120,0.08009280363718668
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,2048,4096,0.07534293333689371
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,2048,4096,0.12169066270192463
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,2048,6144,0.08571093082427979
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,2048,3584,0.06557866732279459
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,2048,3584,0.10632960001627605
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,2048,3584,0.059154133001963295
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,2048,3072,0.05672533512115478
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,2048,3072,0.09064319928487143
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,2048,4096,0.06612053314844767
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,2048,2560,0.047571198145548506
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,2048,2560,0.07624853452046712
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,2048,2560,0.046528001626332596
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,2048,2048,0.038258135318756104
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,2048,2048,0.06210986773173014
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,2048,3072,0.05064746538798014
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,2048,1536,0.029346134265263873
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,2048,1536,0.04827306667963664
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,2048,2048,0.03921600182851155
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,2048,1024,0.02077546715736389
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,2048,1024,0.03714880148569743
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,2048,1536,0.03371520042419433
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,2048,1024,0.030117332935333252
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,2048,768,0.01592639982700348
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,2048,768,0.033162667353947954
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,2048,512,0.012152533729871113
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,2048,512,0.029714133342107135
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,2048,768,0.02521066665649414
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,2048,512,0.025938133398691814
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,2048,256,0.02491733431816101
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,2048,256,0.02849493424097697
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,2048,128,0.008319999774297078
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,2048,128,0.02282879948616028
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,2048,51200,0.5160031954447428
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,2048,64,0.0074986666440963745
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,2048,64,0.02176106572151184
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,2048,32,0.007889066636562348
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,2048,128,0.028849067290623982
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,2048,32,0.020972800254821778
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,2048,256,0.009572266538937887
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,1536,65536,0.8941077550252279
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,1536,51200,0.7501216252644857
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,1536,65536,1.8082037607828778
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,1536,51200,1.4119604746500651
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,1536,16384,0.22205866177876793
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,1536,16384,0.46801067988077805
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,1536,16384,0.15753173828125
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,1536,12288,0.16836585998535156
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,1536,12288,0.3528522809346517
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,2048,65536,0.6453962961832682
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,1536,10240,0.1435797373453776
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,1536,10240,0.2957749366760254
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,1536,12288,0.12344319820404052
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,1536,8192,0.11637653509775799
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,1536,8192,0.2378122647603353
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,1536,10240,0.11190506617228191
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,1536,7168,0.10137386322021484
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,1536,7168,0.20996479988098143
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,1536,8192,0.09158399899800619
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,1536,6144,0.08762346903483073
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,1536,6144,0.18193813959757488
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,1536,7168,0.08575466473897299
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,1536,5120,0.07157759666442871
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,1536,5120,0.14946133295694988
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,1536,6144,0.07867840131123861
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,1536,4096,0.057643731435139976
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,1536,4096,0.11491200129191081
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,1536,5120,0.0734602689743042
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,1536,3584,0.05060799916585287
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,1536,3584,0.09857707023620606
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,1536,3584,0.053806932767232266
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,1536,3072,0.045109331607818604
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,1536,4096,0.059845332304636636
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,1536,3072,0.08402453263600668
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,1536,2560,0.03691093524297078
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,1536,2560,0.07024853229522705
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,1536,2560,0.04319253365198771
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,1536,2048,0.030474666754404706
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,1536,2048,0.05655466715494791
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,1536,3072,0.04724800189336141
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,1536,1536,0.023669334252675374
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,1536,1536,0.04559040069580078
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,1536,51200,0.44409812291463213
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,1536,1536,0.0329312006632487
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,1536,1024,0.017372800906499227
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,1536,1024,0.036423468589782716
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,1536,768,0.013992533087730408
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,1536,768,0.03510506550470988
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,1536,2048,0.03819733460744222
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,1536,512,0.011660800377527872
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,1536,512,0.029346134265263873
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,1536,1024,0.029705599943796797
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,1536,768,0.029813333352406816
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,1536,256,0.009596799810727436
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,1536,512,0.030539733171463013
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,1536,256,0.02483946681022644
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,1536,128,0.02301973303159078
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,1536,256,0.033123199144999185
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,1536,64,0.007950933277606964
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,1536,128,0.032813866933186844
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,1536,32,0.0070826664566993715
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,1536,64,0.02067306637763977
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,1536,32,0.02053440014521281
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,1024,65536,0.579312006632487
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,1536,128,0.007994666695594788
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,1024,51200,0.44675092697143554
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,1024,65536,1.6497984568277997
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,1024,51200,1.2919349670410156
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,1024,16384,0.14571839968363445
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,1024,16384,0.4272128105163574
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,1024,16384,0.13372480074564616
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,1024,12288,0.11068906784057617
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,1024,12288,0.32493654886881507
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,1536,65536,0.5548778533935547
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,1024,10240,0.09624000390370688
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,1024,10240,0.27174294789632164
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,1024,12288,0.10561599731445312
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,1024,10240,0.09548799991607666
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,1024,8192,0.07720853487650553
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,1024,8192,0.22066453297932945
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,1024,7168,0.06846506595611572
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,1024,7168,0.19317973454793294
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,1024,8192,0.07744426727294922
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,1024,6144,0.05840426683425903
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,1024,6144,0.1665280024210612
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,1024,7168,0.07325546741485596
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,1024,5120,0.049180801709493
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,1024,5120,0.13707092603047688
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,1024,6144,0.06730559666951498
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,1024,4096,0.03965760072072347
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,1024,4096,0.1050282637278239
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,1024,5120,0.06362133423487346
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,1024,3584,0.03508266607920329
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,1024,3584,0.08836053212483724
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,1024,4096,0.050877865155537924
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,1024,3072,0.03025280038515727
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,1024,3072,0.07566293080647787
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,1024,3584,0.04423893292744954
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,1024,2560,0.02592960000038147
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,1024,2560,0.06121493180592855
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,1024,3072,0.037970133622487384
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,1024,2048,0.021610667308171592
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,1024,51200,0.3697728157043457
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,1024,2048,0.04929706652959188
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,1024,1536,0.016845866044362386
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,1024,2560,0.034783999125162765
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,1024,1536,0.04094933271408081
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,1024,1024,0.012241066495577494
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,1024,1024,0.0325055996576945
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,1024,1536,0.024734934171040855
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,1024,1024,0.02178666591644287
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,1024,768,0.010053333640098572
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,1024,768,0.02913706700007121
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,1024,512,0.007980800171693166
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,1024,2048,0.029224532842636108
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,1024,512,0.024821333090464272
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,1024,256,0.00594346672296524
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,1024,256,0.02255679965019226
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,1024,768,0.018178133169809978
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,1024,128,0.005842133363087972
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,1024,512,0.019028266270955406
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,1024,128,0.020922666788101195
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,1024,64,0.005010133484999338
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,1024,256,0.021437867482503255
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,1024,128,0.02413866718610128
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,1024,32,0.005123200019200643
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,1024,64,0.01851093371709188
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,1024,32,0.017443199952443443
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,768,65536,0.43784319559733076
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,768,51200,0.34961067835489906
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,768,65536,1.5813855489095052
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,768,51200,1.2403743743896485
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,768,16384,0.11248640219370525
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,768,16384,0.4089152018229167
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,1024,65536,0.46112534205118816
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,768,12288,0.08729386329650879
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,768,12288,0.3107114791870117
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,768,16384,0.1349269390106201
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,768,10240,0.07888320287068686
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,768,10240,0.2656266689300537
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,768,12288,0.1064639965693156
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,768,8192,0.06453333298365274
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,768,10240,0.09601919651031494
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,768,8192,0.21327679951985679
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,768,7168,0.05665813287099203
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,768,7168,0.18816852569580078
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,768,7168,0.07434453169504801
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,768,6144,0.048351999123891196
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,768,6144,0.16164053281148275
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,768,8192,0.07833066781361898
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,768,5120,0.040185598532358806
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,768,5120,0.13234879970550537
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,768,5120,0.06397546529769897
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,768,4096,0.0323360006014506
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,768,6144,0.06826666990915933
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,768,4096,0.09928426742553711
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,768,3584,0.02840320070584615
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,768,3584,0.08415679931640625
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,768,3584,0.044495999813079834
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,768,3072,0.024892799059549966
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,768,4096,0.050780800978342686
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,768,3072,0.07062186400095621
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,768,2560,0.021615999937057494
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,768,2560,0.05669866800308228
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,768,3072,0.03873173395792644
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,768,2560,0.036236798763275145
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,768,2048,0.01804373264312744
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,768,2048,0.04696213404337565
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,768,1536,0.014102400342623393
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,768,1536,0.03886826833089192
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,768,1536,0.027185066541035967
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,768,1024,0.01036906639734904
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,768,1024,0.03106879989306132
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,768,1024,0.023655466238657632
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,768,768,0.008467200398445129
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,768,51200,0.37175893783569336
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,768,768,0.02881386677424113
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,768,512,0.006791466474533081
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,768,2048,0.03094080090522766
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,768,512,0.024962133169174193
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,768,768,0.020615466435750327
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,768,256,0.0059445331494013464
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,768,256,0.02158613403638204
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,768,512,0.020873600244522096
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,768,128,0.005474133292833964
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,768,128,0.01999359925587972
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,768,256,0.024678399165471397
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,768,64,0.005082666873931885
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,768,64,0.018927999337514243
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,768,32,0.005082666873931885
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,768,32,0.01882773240407308
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,512,65536,0.3131861368815104
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,768,128,0.02516266703605652
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,512,65536,1.513698196411133
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,512,51200,0.24534506797790528
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,512,51200,1.187929662068685
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,512,16384,0.07958186467488607
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,512,16384,0.3931893348693848
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,768,65536,0.46721919377644855
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,512,12288,0.060899198055267334
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,512,12288,0.29847145080566406
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,512,16384,0.10998186270395915
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,512,10240,0.05128959814707438
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,512,10240,0.25132800738016764
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,512,12288,0.08697493076324463
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,512,8192,0.041917868455251056
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,512,8192,0.20304853121439614
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,512,10240,0.07970560391743978
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,512,7168,0.036823467413584395
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,512,7168,0.1788149356842041
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,512,8192,0.06466773351033529
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,512,6144,0.03274453282356262
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,512,6144,0.15351999600728353
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,512,7168,0.06184853315353393
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,512,5120,0.0271616001923879
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,512,5120,0.1270090659459432
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,512,6144,0.057538131872812905
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,512,4096,0.021792000532150267
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,512,4096,0.09476479689280191
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,512,5120,0.054218665758768714
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,512,3584,0.019690666596094766
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,512,3584,0.08013652960459391
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,512,4096,0.0408458669980367
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,512,3072,0.0172160009543101
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,512,3072,0.06402986844380697
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,512,51200,0.29987945556640627
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,512,2560,0.014853333433469137
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,512,3072,0.0318069338798523
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,512,3584,0.03662506739298503
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,512,2560,0.05036373138427734
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,512,2048,0.012530133128166199
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,512,2048,0.0428607980410258
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,512,1536,0.010084266463915508
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,512,1536,0.035239466031392414
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,512,2560,0.02892799973487854
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,512,2048,0.024358399709065757
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,512,1024,0.0079434668024381
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,512,1024,0.029428267478942872
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,512,1536,0.020770132541656494
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,512,768,0.0067552000284194945
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,512,768,0.02722346584002177
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,512,512,0.005469866593678792
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,512,512,0.022794665892918904
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,512,1024,0.017414400974909462
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,512,256,0.004634666442871094
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,512,768,0.014844800035158793
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,512,256,0.021223467588424683
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,512,128,0.004171733558177948
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,512,512,0.015702399611473083
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,512,256,0.019425066312154134
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,512,128,0.019102933009465535
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,512,64,0.003931733220815659
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,512,32,0.004242133100827535
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,512,64,0.016710400581359863
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,512,128,0.01945599913597107
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,512,32,0.01779200037320455
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,256,65536,0.20170987447102867
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,256,51200,0.15820693969726562
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,256,65536,1.4655072530110678
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,256,51200,1.1483285268147787
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,256,16384,0.0572160005569458
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,256,16384,0.3812970797220866
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,512,65536,0.3725909233093262
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,256,12288,0.041162665685017905
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,256,12288,0.2893141428629557
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,256,16384,0.10891413688659668
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,256,10240,0.035760001341501875
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,256,10240,0.24296533266703285
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,256,12288,0.0864128033320109
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,256,8192,0.030137600501378377
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,256,8192,0.19522239367167155
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,256,10240,0.07907093365987142
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,256,7168,0.026358399788538617
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,256,7168,0.17130346298217775
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,256,8192,0.06408960024515788
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,256,6144,0.022046933571497597
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,256,7168,0.06041813294092814
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,256,6144,0.14709547360738118
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,256,6144,0.056635733445485434
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,256,5120,0.01633280018965403
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,256,5120,0.11826559702555339
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,256,51200,0.2994122823079427
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,256,4096,0.012782933314641318
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,256,4096,0.08659733136494954
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,256,5120,0.05264000097910563
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,256,3584,0.011637333035469054
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,256,3584,0.07155626614888509
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,256,4096,0.03913493156433105
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,256,3072,0.010026666522026061
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,256,3584,0.035045333703358966
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,256,3072,0.05655786593755087
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,256,2560,0.008849066495895386
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,256,2560,0.04678080081939697
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,256,2048,0.0076223999261856076
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,256,3072,0.030852266152699787
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,256,2048,0.03936426639556885
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,256,1536,0.006718933085600535
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,256,2560,0.02802133361498515
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,256,1536,0.033191466331481935
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,256,65536,0.37073065439860026
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,256,2048,0.023521065711975098
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,256,1024,0.016914133230845133
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,256,768,0.004664533336957296
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,256,768,0.024870399634043375
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,256,768,0.014520532886187234
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,256,512,0.003835733234882355
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,256,512,0.020942932367324828
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,256,1024,0.005144533514976501
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,256,512,0.015285332997639975
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,256,256,0.003525333354870478
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,256,256,0.01975253423055013
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,256,1024,0.02765333255132039
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,256,128,0.0030261332790056865
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,256,128,0.018590933084487914
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,256,256,0.017847466468811034
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,256,64,0.0030239999294281008
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,256,32,0.0030570665995279947
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,256,64,0.016166399916013083
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,256,32,0.01516480048497518
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,128,65536,0.1749226729075114
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,256,1536,0.020180267095565797
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,128,51200,0.13496960004170735
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,256,128,0.017813332875569663
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,128,65536,1.4528362274169921
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,128,16384,0.04741973479588826
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,128,16384,0.37521705627441404
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,128,51200,1.140326436360677
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,128,12288,0.03428800106048584
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,128,12288,0.2845216115315755
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,128,16384,0.10702933470408123
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,128,10240,0.029214932521184282
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,128,10240,0.2382943948109945
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,128,12288,0.08460266590118408
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,128,8192,0.02435093323389689
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,128,8192,0.19156266848246256
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,128,10240,0.07794773578643799
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,128,7168,0.02177600065867106
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,128,8192,0.0630730668703715
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,128,7168,0.16833386421203614
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,128,6144,0.01609493295351664
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,128,6144,0.1438485304514567
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,128,51200,0.2945280075073242
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,128,5120,0.010664533575375874
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,128,7168,0.05931946833928427
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,128,5120,0.1151647965113322
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,128,65536,0.3661386807759603
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,128,4096,0.008748799562454224
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,128,6144,0.05518613258997599
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,128,4096,0.08308586279551188
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,128,3584,0.06801066398620606
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,128,5120,0.05178346633911133
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,128,3072,0.007124266525109608
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,128,4096,0.03819093306859334
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,128,3072,0.05379093488057455
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,128,2560,0.006513066589832306
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,128,3584,0.03412479956944783
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,128,2560,0.04484906593958537
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,128,2048,0.005779199798901876
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,128,3072,0.030053333441416426
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,128,2048,0.03850666681925456
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,128,1536,0.00505920002857844
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,128,2560,0.027221333980560303
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,128,1536,0.03283733328183492
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,128,2048,0.02297919988632202
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,128,1024,0.004193066557248434
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,128,1024,0.026745599508285523
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,128,768,0.003841066608826319
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,128,1536,0.019398399194081626
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,128,1024,0.01609493295351664
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,128,768,0.025073067347208662
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,128,3584,0.008388266960779826
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,128,512,0.0034186666210492453
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,128,256,0.002962133288383484
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,128,512,0.020758400360743202
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,128,768,0.01362773378690084
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,128,256,0.019359999895095827
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,128,128,0.0025898667673269907
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,128,512,0.014525866508483887
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,128,128,0.01765759984652201
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,128,256,0.016927999258041383
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,128,64,0.002570666621128718
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,8192,128,128,0.016965333620707193
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,128,32,0.0026186667382717133
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,128,64,0.01602133313814799
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,128,32,0.01469013293584188
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,64,65536,0.17290132840474445
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,64,51200,0.13090346654256185
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,64,16384,0.04811520179112752
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,64,16384,0.37377707163492835
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,64,12288,0.03367786804835002
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,64,12288,0.2831967989603678
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,64,51200,1.1421194712320963
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,64,10240,0.028407466411590577
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,64,8192,0.023685334126154582
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,64,65536,1.4538997650146483
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,64,7168,0.020778665939966835
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,64,10240,0.2378229300181071
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,64,8192,0.190449062983195
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,64,6144,0.014614400267601014
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,64,7168,0.16663573582967123
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,64,6144,0.14230720202128094
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,64,4096,0.007736533383528392
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,64,5120,0.1148085355758667
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,64,3584,0.00698773314555486
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,64,4096,0.08229333559672038
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,64,3072,0.006308266520500183
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,64,3584,0.06682666937510172
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,64,3072,0.05247999827067057
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,64,2560,0.0062943999965985615
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,64,2048,0.0054613331953684485
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,64,2560,0.04319039980570476
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,64,2048,0.037229867776234944
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,64,1536,0.004702933132648468
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,64,1024,0.003866666555404663
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,64,1536,0.03209813237190247
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,64,1024,0.02674773335456848
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,64,768,0.0035786665976047516
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,64,512,0.003033600002527237
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,64,768,0.02474986712137858
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,64,256,0.0026101333399613695
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,64,128,0.0025909334421157838
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,64,256,0.018607999881108603
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,64,512,0.020936532815297445
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,64,64,0.0022111999491850535
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,64,128,0.016807466745376587
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,64,32,0.002195200075705846
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,64,32,0.014672000209490457
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,64,64,0.01641813317934672
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,64,5120,0.009588266412417095
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,32,65536,0.16567360560099284
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,32,16384,0.05015573501586914
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,32,16384,0.37232745488484703
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,32,12288,0.032103466987609866
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,32,12288,0.2815199851989746
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,32,10240,0.027188267310460406
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,32,51200,1.133020782470703
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,32,8192,0.022447999318440756
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,32,65536,1.4443380991617838
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,32,10240,0.23659200668334962
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,32,7168,0.019802665710449217
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,32,8192,0.18971734046936034
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,32,6144,0.013266133268674216
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,32,5120,0.00876800020535787
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,32,51200,0.1301376024881999
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,32,7168,0.16594667434692384
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,32,4096,0.007287466526031494
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,32,6144,0.14215466181437175
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,32,3584,0.006696533163388569
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,32,5120,0.11436159610748291
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,32,3072,0.00586346685886383
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,32,4096,0.08220372994740804
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,32,2560,0.005422933399677277
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,32,3584,0.06713600158691406
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,32,2048,0.004678399860858917
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,32,3072,0.05139840046564738
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,32,1536,0.004661333560943603
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,32,1024,0.0038954667747020722
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,32,1536,0.032517333825429276
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,32,768,0.003409066547950109
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,32,1024,0.02693866689999898
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,32,512,0.0030400000512599947
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,32,512,0.021246933937072755
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,32,256,0.0027200000981489818
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,32,256,0.020360533396402994
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,32,128,0.002536533276240031
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,32,128,0.017621332406997682
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,32,64,0.002347733328739802
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,32,64,0.01606613298257192
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,8192,32,32,0.002259200066328049
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,32,2560,0.04325759808222453
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,32,32,0.014739200472831726
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,32,2048,0.037511467933654785
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,8192,32,768,0.025230934222539265
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,65536,16384,5.458574930826822
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,65536,16384,3.012792460123698
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,65536,51200,9.15856424967448
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,65536,12288,4.231659698486328
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,65536,12288,2.1924201965332033
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,65536,51200,17.974019368489586
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,65536,10240,3.491383361816406
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,65536,10240,1.7308714548746746
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,65536,10240,0.9463221232096354
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,65536,8192,2.772076924641927
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,65536,8192,1.3207135518391928
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,65536,12288,1.0777365366617837
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,65536,7168,2.4229054768880207
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,65536,7168,1.227667236328125
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,65536,16384,1.4592010498046875
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,65536,6144,1.955447514851888
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,65536,6144,1.0460949579874674
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,65536,8192,0.7822698593139649
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,65536,5120,1.6499659220377605
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,65536,5120,0.8272085189819336
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,65536,7168,0.727891222635905
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,65536,4096,1.2295188903808594
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,65536,4096,0.6959434509277344
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,65536,6144,0.6223818461100261
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,65536,3584,1.117485809326172
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,65536,3584,0.6450581232706706
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,65536,4096,0.508679453531901
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,65536,3072,0.9450133641560873
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,65536,3072,0.5450453440348307
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,65536,5120,0.5661781311035157
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,65536,2560,0.7472415924072265
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,65536,2560,0.43988587061564133
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,65536,3584,0.45442988077799484
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,65536,2048,0.6115562438964843
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,65536,2048,0.3643925348917643
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,65536,2560,0.3823925336201986
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,65536,1536,0.4554986635843913
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,65536,1536,0.28317012786865237
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,65536,3072,0.4037535985310872
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,65536,1024,0.3197183926900228
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,65536,1024,0.20566293398539223
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,65536,1024,0.281767463684082
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,65536,768,0.25477439562479653
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,65536,768,0.18154773712158204
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,65536,1536,0.29134934743245444
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,65536,512,0.18080746332804362
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,65536,512,0.14817919731140136
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,65536,768,0.2670655886332194
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,65536,256,0.12793280283610026
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,65536,256,0.12419093449910482
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,65536,2048,0.3425098737080892
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,65536,128,0.10836053689320882
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,65536,128,0.10893653233846028
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,65536,512,0.27234347661336267
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,65536,64,0.0935370683670044
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,65536,64,0.11300480365753174
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,65536,32,0.09927573204040527
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,65536,32,0.11365333398183186
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,65536,128,0.2736063957214355
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,65536,256,0.2642517407735189
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,51200,65536,9.367477416992188
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,51200,65536,17.90730183919271
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,51200,51200,13.776282755533853
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,51200,51200,7.319953918457031
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,51200,16384,4.285286458333333
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,51200,16384,2.3771125793457033
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,51200,16384,1.1409749348958333
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,51200,12288,3.1392308553059896
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,51200,12288,1.6315755208333331
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,51200,12288,0.8629365285237631
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,51200,10240,2.701873016357422
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,51200,10240,1.402564239501953
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,51200,10240,0.7513162612915039
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,51200,8192,2.1661707560221357
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,51200,8192,1.0666336059570312
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,65536,51200,4.916216532389322
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,51200,7168,1.724779764811198
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,51200,7168,0.9757098515828451
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,51200,8192,0.6128053029378255
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,51200,6144,1.5417354583740235
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,51200,6144,0.7895903905232747
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,51200,7168,0.5722965240478516
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,51200,5120,1.1934218088785806
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,51200,5120,0.7051839828491211
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,51200,6144,0.48765974044799804
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,51200,4096,0.9177312215169271
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,51200,4096,0.5764490763346355
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,51200,5120,0.44518613815307617
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,51200,3584,0.8211008071899414
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,51200,3584,0.4815093358357747
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,51200,51200,3.9402112325032554
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,51200,3072,0.7162069320678711
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,51200,3072,0.4160533269246419
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,51200,3584,0.3563381195068359
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,51200,2560,0.5831904093424479
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,51200,2560,0.3486250559488932
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,51200,4096,0.39203414916992185
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,51200,2048,0.46676266988118487
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,51200,2048,0.287225596110026
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,51200,2560,0.3004938761393229
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,51200,1536,0.36360960006713866
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,51200,1536,0.22497173945109047
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,51200,3072,0.31320425669352214
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,51200,1024,0.2521397272745768
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,51200,1024,0.17833174069722493
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,51200,2048,0.2666805267333984
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,51200,768,0.19934825897216796
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,51200,768,0.14726932843526203
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,51200,768,0.21464319229125978
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,51200,512,0.1506495952606201
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,51200,512,0.12114453315734863
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,51200,1536,0.2334122657775879
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,51200,256,0.10048000017801921
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,51200,256,0.10148906707763672
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,51200,1024,0.2286815961201986
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,51200,128,0.0883957306543986
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,51200,128,0.09233600298563639
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,51200,128,0.21656427383422852
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,51200,64,0.07550400098164876
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,51200,64,0.09355200131734212
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,51200,32,0.07884159882863363
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,51200,32,0.09380479653676352
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,51200,256,0.2127679983774821
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,51200,512,0.2150026639302572
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,16384,65536,3.210928090413412
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,16384,65536,5.335553995768229
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,16384,51200,2.55794677734375
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,16384,51200,4.193805948893229
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,16384,16384,1.2512587229410808
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,16384,16384,0.8186186472574869
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,16384,16384,0.40964800516764327
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,16384,12288,0.9279872258504233
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,16384,12288,0.6255626678466797
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,16384,12288,0.3085568110148112
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,16384,10240,0.798306147257487
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,16384,10240,0.558460807800293
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,16384,10240,0.273688538869222
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,16384,8192,0.6375647862752278
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,16384,8192,0.4474410692850749
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,16384,8192,0.22461226781209312
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,16384,7168,0.56146240234375
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,16384,7168,0.35501120885213217
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,16384,51200,1.2545951843261718
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,16384,6144,0.47207358678181965
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,16384,6144,0.32771307627360025
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,16384,7168,0.20926933288574218
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,16384,5120,0.38878294626871746
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,16384,5120,0.24959039688110352
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,16384,5120,0.16538346608479818
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,16384,4096,0.31032533645629884
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,16384,4096,0.20466987291971842
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,16384,6144,0.18085546493530275
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,16384,3584,0.25863253275553383
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,16384,3584,0.18138133684794108
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,16384,3584,0.12975680033365886
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,16384,3072,0.22380372683207192
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,16384,3072,0.1571242650349935
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,16384,4096,0.145144526163737
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,16384,2560,0.18929279645284017
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,16384,2560,0.1361024061838786
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,16384,3072,0.11557973225911458
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,16384,2048,0.15198079744974774
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,16384,2048,0.11119786898295085
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,16384,2560,0.1101194699605306
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,16384,1536,0.11782506306966145
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,16384,1536,0.09093120098114013
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,16384,1536,0.08417493502298991
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,16384,1024,0.08160320123036703
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,16384,1024,0.06795626481374105
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,16384,2048,0.09821120103200277
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,16384,768,0.06354879935582479
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,16384,768,0.0658890684445699
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,51200,65536,4.979712931315104
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,16384,512,0.046589867273966475
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,16384,512,0.05550293525060018
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,16384,1024,0.08080960114796956
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,16384,256,0.036345601081848145
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,16384,256,0.04576106468836467
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,16384,768,0.07854186693827311
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,16384,128,0.02972480058670044
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,16384,128,0.04019733270009358
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,16384,512,0.07483519713083903
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,16384,64,0.025564799706141155
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,16384,64,0.0409226655960083
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,16384,32,0.026707200209299724
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,16384,32,0.040329599380493165
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,16384,256,0.07056427001953125
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,16384,128,0.0722815990447998
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,12288,65536,2.579625701904297
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,12288,65536,4.027534993489583
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,12288,51200,2.047873051961263
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,12288,51200,3.0955764770507814
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,12288,16384,0.983946673075358
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,12288,16384,0.6619509379069011
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,16384,65536,1.633418655395508
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,12288,12288,0.6930655797322591
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,12288,12288,0.5570304234822591
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,12288,16384,0.3294047991434733
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,12288,10240,0.5822463989257812
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,12288,10240,0.4607978820800781
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,12288,12288,0.24707627296447754
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,12288,8192,0.4766101201375325
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,12288,8192,0.3476618766784668
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,12288,8192,0.18162026405334472
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,12288,7168,0.4142346700032552
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,12288,7168,0.28566080729166665
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,12288,10240,0.22094613711039224
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,12288,6144,0.34135573705037436
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,12288,6144,0.25010453859965004
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,12288,7168,0.16823147137959799
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,12288,5120,0.2865056037902832
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,12288,5120,0.2134474595387777
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,12288,6144,0.14627307256062824
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,12288,4096,0.2249354680379232
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,12288,4096,0.16320533752441407
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,12288,4096,0.11733333269755046
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,12288,3584,0.19676159222920736
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,12288,3584,0.14682772954305012
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,12288,5120,0.13324373563130695
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,12288,3072,0.17366506258646647
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,12288,3072,0.12914880116780597
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,12288,3072,0.09418666362762451
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,12288,3584,0.10590506394704183
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,12288,2560,0.11110933621724446
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,12288,2560,0.14575360616048177
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,12288,2048,0.11775786876678467
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,12288,2048,0.09217066764831543
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,12288,2048,0.0799562692642212
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,12288,1536,0.09034773508707682
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,12288,1536,0.0759061336517334
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,12288,2560,0.09021440347035727
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,12288,1024,0.06189333200454712
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,12288,1024,0.057728000481923426
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,12288,1024,0.06879999637603759
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,12288,768,0.04893440008163452
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,12288,768,0.05170559883117676
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,12288,1536,0.07157759666442871
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,12288,512,0.036212265491485596
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,12288,512,0.046214401721954346
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,12288,512,0.063646932442983
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,12288,256,0.025976532697677614
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,12288,256,0.03922986586888631
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,12288,768,0.0631274660428365
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,12288,128,0.022802132368087768
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,12288,128,0.03600746790568034
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,12288,256,0.06323839823404948
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,12288,64,0.01964906652768453
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,12288,64,0.03505920171737671
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,12288,32,0.020488532384236653
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,12288,32,0.03433813254038493
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,12288,128,0.06335039933522543
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,10240,65536,2.2662623087565104
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,10240,65536,3.32652587890625
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,10240,51200,2.62467524210612
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,10240,51200,1.8140608469645183
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,12288,51200,1.0266133626302083
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,10240,16384,0.7525813420613606
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,10240,16384,0.5766634623209635
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,10240,16384,0.28983786900838215
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,10240,12288,0.5670581181844075
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,10240,12288,0.4432970682779948
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,12288,65536,1.2970709482828775
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,10240,10240,0.47807998657226564
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,10240,10240,0.37357867558797203
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,10240,12288,0.21870187123616538
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,10240,8192,0.38928000132242835
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,10240,8192,0.28809601465861
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,10240,10240,0.19418986638387042
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,10240,7168,0.3337962786356608
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,10240,7168,0.24599787394205727
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,10240,8192,0.15931305885314942
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,10240,6144,0.28761491775512693
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,10240,6144,0.21339732805887857
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,10240,7168,0.14749546051025392
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,10240,5120,0.23381226857503257
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,10240,5120,0.17661973635355632
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,10240,6144,0.12897813320159912
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,10240,4096,0.18982079823811848
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,10240,4096,0.1439136028289795
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,10240,5120,0.11714133421579997
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,10240,3584,0.16352640787760417
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,10240,3584,0.13007786273956298
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,10240,51200,0.8858549118041992
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,10240,3072,0.14375146230061847
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,10240,3072,0.11828266779581706
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,10240,3584,0.09255466461181641
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,10240,4096,0.10219306945800781
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,10240,2560,0.11837013562520345
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,10240,2560,0.09863146940867105
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,10240,2048,0.09748906294504801
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,10240,2048,0.08083093166351318
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,10240,3072,0.08357439835866293
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,10240,1536,0.075709867477417
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,10240,1536,0.06879253387451172
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,10240,1536,0.06579519907633463
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,10240,1024,0.052641065915425625
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,10240,1024,0.05150933265686035
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,10240,2048,0.07227839628855387
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,10240,768,0.04125866492589315
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,10240,768,0.04729173183441162
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,10240,2560,0.07997439702351888
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,10240,512,0.0302239994208018
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,10240,512,0.04269333283106486
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,10240,1024,0.06304639975229899
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,10240,256,0.021915733814239502
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,10240,768,0.058609068393707275
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,10240,256,0.036508798599243164
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,10240,512,0.058913067976633704
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,10240,128,0.019462400674819948
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,10240,128,0.031956267356872556
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,10240,64,0.016888533035914102
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,10240,64,0.031309866905212404
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,10240,256,0.058363731702168783
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,10240,32,0.01745599905649821
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,10240,32,0.03067520062128703
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,10240,128,0.05871146519978841
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,8192,65536,1.9650688171386719
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,8192,65536,2.6747006734212238
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,8192,51200,2.0762666066487627
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,8192,51200,1.5489376068115235
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,8192,16384,0.6072693506876627
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,8192,16384,0.5050506591796875
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,10240,65536,1.1101834615071615
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,8192,12288,0.45059518814086913
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,8192,12288,0.3826837221781413
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,8192,16384,0.269157346089681
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,8192,10240,0.3964384078979492
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,8192,10240,0.3385216077168783
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,8192,10240,0.16605440775553387
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,8192,8192,0.31979519526163735
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,8192,8192,0.25987839698791504
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,8192,12288,0.19165439605712892
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,8192,7168,0.27087039947509767
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,8192,7168,0.21413866678873697
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,8192,7168,0.1232042630513509
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,8192,6144,0.23183253606160484
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,8192,6144,0.1827338695526123
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,8192,8192,0.1331786632537842
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,8192,5120,0.19354453086853027
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,8192,5120,0.1573525269826253
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,8192,6144,0.10806933244069417
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,8192,4096,0.15136853853861493
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,8192,4096,0.12573333581288654
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,8192,5120,0.09923306306203207
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,8192,3584,0.1350111961364746
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,8192,3584,0.11474773089090984
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,8192,3584,0.07621440092722574
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,8192,3072,0.11656426588694255
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,8192,3072,0.09855466683705648
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,8192,4096,0.08546026547749838
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,8192,2560,0.0960640033086141
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,8192,2560,0.08726399739583333
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,8192,2560,0.06373120148976644
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,8192,2048,0.07806080182393392
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,8192,3072,0.06672960122426351
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,8192,2048,0.07211200396219888
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,8192,1536,0.060337066650390625
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,8192,1536,0.061197865009307864
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,8192,1536,0.047532800833384195
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,8192,1024,0.0421834667523702
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,8192,1024,0.04527253309885661
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,8192,2048,0.05576426585515341
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,8192,768,0.03344426552454631
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,8192,768,0.042843735218048094
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,8192,1024,0.04482666651407878
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,8192,512,0.023856000105539957
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,8192,512,0.0372927983601888
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,8192,768,0.03919359842936198
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,8192,256,0.018037333091100057
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,8192,256,0.030903466542561847
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,8192,512,0.03950080076853434
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,8192,128,0.016167466839154564
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,8192,128,0.028417066733042402
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,8192,256,0.03935253222783407
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,8192,64,0.01366933286190033
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,8192,64,0.026779733101526898
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,8192,32,0.01406719982624054
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,8192,32,0.02695786754290263
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,8192,128,0.040428801377614336
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,7168,65536,1.805829366048177
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,7168,65536,2.3615071614583334
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,7168,51200,1.8361770629882812
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,7168,51200,1.4313140869140626
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,8192,51200,0.7664543787638347
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,7168,16384,0.5343231836954753
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,7168,16384,0.47725547154744463
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,7168,16384,0.2354464054107666
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,7168,12288,0.40085865656534836
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,7168,12288,0.3527071952819824
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,8192,65536,0.9674175898234049
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,7168,10240,0.3405194600423177
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,7168,10240,0.29104426701863606
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,7168,12288,0.17038079897562664
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,7168,8192,0.27586132685343423
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,7168,8192,0.22882134119669595
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,7168,10240,0.1490250587463379
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,7168,7168,0.2358954588572184
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,7168,7168,0.19577600161234537
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,7168,7168,0.1129962682723999
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,7168,6144,0.20041279792785643
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,7168,6144,0.1663317362467448
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,7168,8192,0.1209216038386027
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,7168,5120,0.1683242638905843
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,7168,5120,0.1401034673055013
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,7168,6144,0.09924586613972983
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,7168,5120,0.09169387022654216
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,7168,4096,0.13357760111490885
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,7168,4096,0.11507840156555176
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,7168,3584,0.11541439692179363
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,7168,3584,0.10513813495635986
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,7168,3584,0.07076160113016763
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,7168,3072,0.09938986301422119
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,7168,3072,0.09062186876932779
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,7168,4096,0.07876693407694499
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,7168,2560,0.08481386502583822
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,7168,2560,0.08082239627838135
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,7168,3072,0.06274133523305257
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,7168,2048,0.06800426642100016
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,7168,2048,0.06674559911092123
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,7168,2560,0.059818665186564125
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,7168,1536,0.05293866793314615
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,7168,1536,0.05647679964701334
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,7168,51200,0.6820085525512696
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,7168,1024,0.03850666681925456
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,7168,1024,0.04498986800511678
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,7168,1024,0.04460586706797282
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,7168,768,0.029293866952260335
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,7168,1536,0.04690239826838176
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,7168,768,0.039816534519195555
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,7168,512,0.021223467588424683
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,7168,512,0.035035733381907144
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,7168,2048,0.05303893486658732
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,7168,256,0.015851733088493348
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,7168,256,0.028935466210047407
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,7168,512,0.04147200187047322
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,7168,128,0.013666133085886637
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,7168,128,0.025538132588068647
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,7168,768,0.04072426557540894
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,7168,64,0.01162559986114502
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,7168,128,0.041484800974527995
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,7168,32,0.012434132893880208
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,7168,64,0.026070400079091387
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,7168,32,0.02473599910736084
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,7168,256,0.041069865226745605
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,6144,65536,1.6613535563151043
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,6144,65536,1.9890261332194012
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,6144,51200,1.4859444936116537
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,6144,51200,1.3000597635904947
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,6144,16384,0.4523306528727214
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,6144,16384,0.42612479527791336
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,6144,16384,0.20257813135782876
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,6144,12288,0.3416245460510254
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,6144,12288,0.33646825154622395
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,7168,65536,0.8615925470987955
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,6144,10240,0.29280106226603186
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,6144,10240,0.2797173182169596
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,6144,12288,0.14981973965962728
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,6144,8192,0.23840853373209633
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,6144,8192,0.20949759483337402
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,6144,10240,0.13347412745157877
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,6144,7168,0.2068064053853353
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,6144,7168,0.18396587371826173
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,6144,8192,0.1084714651107788
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,6144,6144,0.17696000734965006
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,6144,6144,0.15288000106811522
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,6144,7168,0.10164373715718586
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,6144,5120,0.1441226641337077
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,6144,5120,0.12866453329722088
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,6144,5120,0.0827999989191691
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,6144,4096,0.11432747046152751
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,6144,4096,0.10648000240325928
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,6144,6144,0.08963840007781983
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,6144,3584,0.100708270072937
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,6144,3584,0.0966474692026774
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,6144,4096,0.070741335550944
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,6144,3072,0.08644479910532633
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,6144,3072,0.08451519807179769
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,6144,3584,0.06290880044301352
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,6144,2560,0.07302186489105225
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,6144,2560,0.07461547056833903
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,6144,2560,0.05316160122553507
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,6144,2048,0.05941973527272543
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,6144,2048,0.061410133043924967
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,6144,3072,0.055961600939432775
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,6144,1536,0.04565973281860351
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,6144,1536,0.05144213438034058
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,6144,1536,0.04233173529307048
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,6144,1024,0.031497599681218465
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,6144,1024,0.039241600036621097
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,6144,2048,0.047186132272084555
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,6144,768,0.024678399165471397
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,6144,768,0.03502506812413533
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,6144,768,0.03674986759821574
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,6144,512,0.018157867590586345
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,6144,512,0.0325162669022878
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,6144,1024,0.04033386707305908
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,6144,256,0.014054399728775025
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,6144,256,0.02882560094197591
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,6144,256,0.03730453252792358
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,6144,128,0.014028799533843995
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,6144,128,0.024755199750264488
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,6144,512,0.037045331796010335
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,6144,64,0.012050132950146992
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,6144,64,0.02438933253288269
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,6144,32,0.012411733468373615
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,6144,128,0.03732159932454427
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,6144,32,0.023987199862798056
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,6144,51200,0.5924618403116862
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,5120,65536,1.6962411244710285
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,5120,65536,1.5262186686197916
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,5120,51200,1.2826826731363932
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,5120,51200,1.1907423655192058
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,5120,16384,0.3893194516499837
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,5120,16384,0.40381867090861007
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,5120,16384,0.17370667457580566
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,5120,12288,0.29385814666748045
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,5120,12288,0.2999370574951172
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,6144,65536,0.7439872105916341
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,5120,10240,0.2414367993672689
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,5120,10240,0.24087252616882324
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,5120,12288,0.1321887969970703
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,5120,8192,0.20008533795674643
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,5120,8192,0.18808746337890625
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,5120,10240,0.11842239697774251
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,5120,7168,0.17034986813863118
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,5120,7168,0.16132693290710448
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,5120,8192,0.09669866561889648
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,5120,6144,0.14880746205647785
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,5120,6144,0.13797866503397624
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,5120,7168,0.09136426448822021
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,5120,5120,0.11855786641438801
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,5120,5120,0.11743360360463459
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,5120,6144,0.08087039788564046
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,5120,4096,0.09620160261789958
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,5120,4096,0.09715946515401205
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,5120,5120,0.07484479745229086
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,5120,3584,0.08306240240732829
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,5120,3584,0.08884159723917642
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,5120,3584,0.056927998860677086
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,5120,3072,0.07186666329701742
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,5120,3072,0.07814506689707437
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,5120,4096,0.06290773153305054
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,5120,2560,0.06129599809646606
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,5120,2560,0.06752320130666098
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,5120,2560,0.04920426607131958
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,5120,2048,0.049608532587687174
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,5120,2048,0.054520531495412194
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,5120,3072,0.051151998837788905
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,5120,1536,0.037589331467946366
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,5120,1536,0.0460533340771993
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,5120,1536,0.03966720104217529
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,5120,1024,0.026051199436187743
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,5120,1024,0.03540266752243042
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,5120,2048,0.044120534261067705
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,5120,768,0.021010132630666097
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,5120,768,0.032553599278132124
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,5120,768,0.035930665334065755
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,5120,512,0.01530026694138845
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,5120,1024,0.03751786549886067
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,5120,512,0.028893866141637164
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,5120,256,0.011727999647458394
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,5120,256,0.02474666635195414
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,5120,256,0.036064000924428304
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,5120,128,0.009980799754460652
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,5120,128,0.024051199356714882
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,5120,512,0.03640213410059611
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,5120,64,0.009136000275611877
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,5120,64,0.023827199141184488
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,5120,32,0.009435733159383137
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,5120,128,0.035760001341501875
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,5120,32,0.02106346686681112
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,4096,65536,1.232792536417643
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,4096,65536,1.3289482116699218
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,5120,51200,0.5095861434936524
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,4096,51200,0.913312021891276
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,4096,51200,1.0588245391845703
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,4096,16384,0.30360425313313805
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,4096,16384,0.363155206044515
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,4096,16384,0.1461685339609782
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,4096,12288,0.23070720036824546
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,4096,12288,0.26713921229044596
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,5120,65536,0.6402069091796875
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,4096,10240,0.1978069305419922
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,4096,10240,0.21755414009094237
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,4096,12288,0.11319680213928222
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,4096,8192,0.16293120384216309
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,4096,8192,0.17114240328470867
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,4096,10240,0.10133866469065349
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,4096,7168,0.13917439778645832
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,4096,7168,0.14612693786621095
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,4096,8192,0.08309226830800374
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,4096,6144,0.12057387034098307
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,4096,6144,0.1253226677576701
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,4096,7168,0.07828373114267985
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,4096,5120,0.09764479796091716
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,4096,5120,0.1060640017191569
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,4096,6144,0.0704533338546753
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,4096,4096,0.07874240080515543
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,4096,4096,0.08754666646321616
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,4096,4096,0.05317226648330688
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,4096,3584,0.06904959678649902
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,4096,3584,0.08033386866251627
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,4096,5120,0.0633898655573527
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,4096,3072,0.05859520037968954
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,4096,3072,0.06881919701894125
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,4096,3584,0.0469589352607727
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,4096,2560,0.05006826718648275
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,4096,2560,0.05980799992879232
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,4096,3072,0.04107946554819743
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,4096,2048,0.04054506619771321
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,4096,2048,0.04965653419494629
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,4096,2560,0.03914239803949992
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,4096,1536,0.031650133927663165
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,4096,1536,0.041419732570648196
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,4096,2048,0.03370560010274251
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,4096,1024,0.021783467133839926
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,4096,1024,0.03295786579449971
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,4096,51200,0.4241663932800293
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,4096,768,0.01748159925142924
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,4096,768,0.030961066484451294
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,4096,1536,0.02922346591949463
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,4096,512,0.013133866588274637
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,4096,512,0.026778666178385417
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,4096,1024,0.026953599850336712
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,4096,256,0.009618133306503296
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,4096,256,0.024691200256347655
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,4096,512,0.025145600239435833
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,4096,768,0.024378667275110878
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,4096,128,0.008742400010426839
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,4096,128,0.022388267517089843
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,4096,64,0.007521066566308339
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,4096,256,0.024765866994857787
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,4096,64,0.020717867215474448
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,4096,32,0.008025600016117096
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,4096,32,0.021151999632517494
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,4096,128,0.024765866994857787
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,3584,65536,1.0593119939168294
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,3584,65536,1.2613739013671874
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,3584,51200,0.8934911727905274
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,3584,51200,0.9890858968098959
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,3584,16384,0.2603093306223551
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,3584,16384,0.3268191973368327
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,3584,16384,0.14600319862365724
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,3584,12288,0.19738772710164387
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,3584,12288,0.24984639485677085
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,4096,65536,0.5344010670979817
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,3584,10240,0.1676736036936442
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,3584,10240,0.2058186690012614
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,3584,10240,0.10160106817881268
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,3584,8192,0.13569386800130206
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,3584,12288,0.1128117322921753
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,3584,8192,0.16035520235697429
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,3584,7168,0.11885440349578857
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,3584,7168,0.13983252843221028
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,3584,7168,0.07910613218943277
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,3584,6144,0.10288106600443522
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,3584,6144,0.11994773546854656
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,3584,8192,0.0835658629735311
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,3584,5120,0.08525760173797607
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,3584,5120,0.10369493166605633
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,3584,6144,0.07034773031870524
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,3584,4096,0.06972160339355468
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,3584,4096,0.08551146984100341
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,3584,5120,0.06339733203252157
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,3584,3584,0.06119466622670492
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,3584,3584,0.07790186405181884
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,3584,3584,0.048877867062886556
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,3584,3072,0.05331093470255534
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,3584,3072,0.0648522655169169
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,3584,4096,0.05311786731084188
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,3584,2560,0.04549226760864258
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,3584,2560,0.057462398211161295
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,3584,3072,0.04377706845601399
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,3584,2048,0.03770879904429118
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,3584,2048,0.04803200165430705
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,3584,2560,0.04190400044123332
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,3584,1536,0.028154667218526202
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,3584,1536,0.040447998046875
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,3584,1536,0.032995200157165526
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,3584,1024,0.02102186679840088
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,3584,1024,0.03302293419837952
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,3584,2048,0.03715626796086629
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,3584,768,0.016785067319869996
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,3584,768,0.03161919911702474
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,3584,1024,0.030803199609120684
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,3584,512,0.014481066664059957
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,3584,512,0.028248532613118486
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,3584,768,0.029128533601760865
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,3584,256,0.010843732953071594
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,3584,256,0.02491413354873657
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,3584,512,0.02961919903755188
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,3584,128,0.007563733557860057
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,3584,256,0.029329067468643187
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,3584,128,0.02061226765314738
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,3584,64,0.006757333377997081
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,3584,64,0.018956800301869713
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,3584,32,0.007131733496983846
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,3584,32,0.02021119991938273
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,3584,128,0.029752532641092937
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,3072,65536,0.9051328023274741
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,3072,65536,1.1845450083414713
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,3072,51200,0.6952512105305989
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,3072,51200,0.934058698018392
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,3584,51200,0.42678292592366535
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,3072,16384,0.22897280057271324
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,3072,16384,0.30745598475138347
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,3584,65536,0.5380266825358073
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,3072,12288,0.17600639661153156
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,3072,12288,0.23691840171813966
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,3072,16384,0.12389226754506429
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,3072,10240,0.150764799118042
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,3072,10240,0.1937887986501058
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,3072,10240,0.08728533585866292
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,3072,8192,0.12128960291544597
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,3072,8192,0.15090559323628744
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,3072,12288,0.09665173689524333
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,3072,7168,0.10761706829071045
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,3072,7168,0.13053332964579265
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,3072,8192,0.07226133346557617
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,3072,6144,0.09021866321563721
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,3072,6144,0.11155200004577637
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,3072,7168,0.06834773222605386
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,3072,5120,0.07543253103892009
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,3072,5120,0.09432533582051596
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,3072,5120,0.05477013190587362
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,3072,4096,0.05960853497187296
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,3072,4096,0.07702399889628092
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,3072,6144,0.0603925347328186
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,3072,3584,0.051527468363444004
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,3072,3584,0.06939413547515869
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,3072,3584,0.041929598649342856
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,3072,3072,0.04460479815800984
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,3072,3072,0.0589738647143046
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,3072,4096,0.04586346546808879
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,3072,2560,0.03835306564966838
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,3072,2560,0.05214720169703165
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,3072,2560,0.03541546662648519
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,3072,2048,0.031523199876149495
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,3072,2048,0.04347626765569051
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,3072,3072,0.03709973494211833
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,3072,1536,0.02413546641667684
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,3072,1536,0.03798720041910807
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,3072,1536,0.028460800647735596
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,3072,1024,0.017206400632858276
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,3072,1024,0.030933332443237305
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,3072,2048,0.03179200092951457
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,3072,51200,0.34757865269978844
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,3072,768,0.013981866836547851
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,3072,768,0.02881386677424113
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,3072,512,0.01037013332049052
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,3072,512,0.025092266003290814
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,3072,1024,0.026348799467086792
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,3072,256,0.00839466651280721
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,3072,256,0.022731733322143555
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,3072,512,0.029263999064763385
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,3072,768,0.028871466716130573
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,3072,128,0.006762666503588359
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,3072,128,0.020881066719690956
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,3072,64,0.006236800054709116
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,3072,64,0.018823466698328652
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,3072,128,0.02744106650352478
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,3072,32,0.006345599889755249
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,3072,256,0.029250133037567138
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,3072,32,0.02081600030263265
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,2560,65536,0.8218581517537435
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,2560,65536,1.1404351552327474
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,2560,51200,0.6679840087890625
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,2560,51200,0.8955039978027344
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,2560,16384,0.20448320706685386
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,2560,16384,0.2954901377360026
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,2560,16384,0.12219733397165936
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,2560,12288,0.15486613909403485
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,2560,12288,0.22656000455220543
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,3072,65536,0.43617814381917314
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,2560,10240,0.14936960538228353
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,2560,10240,0.19193919499715167
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,2560,12288,0.0951370636622111
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,2560,8192,0.12211306889851888
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,2560,8192,0.14988800684611003
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,2560,10240,0.08630613485972086
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,2560,7168,0.1029098669687907
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,2560,7168,0.13137173652648926
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,2560,7168,0.06688746611277262
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,2560,6144,0.08741119702657064
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,2560,6144,0.11006506284077962
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,2560,8192,0.07113706270853679
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,2560,5120,0.0753941297531128
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,2560,5120,0.08805973529815674
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,2560,5120,0.05302506685256958
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,2560,6144,0.059689601262410484
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,2560,4096,0.05813866853713989
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,2560,4096,0.0710261344909668
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,2560,3584,0.05195519924163818
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,2560,3584,0.06250026623407999
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,2560,3584,0.0402890682220459
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,2560,3072,0.04664426644643148
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,2560,3072,0.05374720096588135
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,2560,4096,0.04456959962844849
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,2560,2560,0.032682667175928756
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,2560,2560,0.04819200038909912
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,2560,3072,0.03591573238372803
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,2560,2048,0.02664746642112732
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,2560,2560,0.03412160078684489
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,2560,2048,0.04100480079650879
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,2560,1536,0.02135146657625834
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,2560,1536,0.03517866532007853
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,2560,2048,0.02982826630274455
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,2560,1024,0.014435199896494546
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,2560,1024,0.02935466567675273
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,2560,1024,0.024519467353820802
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,2560,768,0.01167680025100708
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,2560,768,0.026989867289861042
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,2560,1536,0.026474666595458985
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,2560,512,0.009333333373069764
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,2560,512,0.024631466468175253
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,2560,512,0.02355946699778239
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,2560,768,0.023466666539510093
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,2560,256,0.0076341331005096436
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,2560,256,0.021626667181650797
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,2560,128,0.007100800176461537
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,2560,128,0.02008533279101054
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,2560,256,0.023407999674479166
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,2560,64,0.0061258668700853985
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,2560,128,0.023108265797297158
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,2560,32,0.00631466656923294
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,2560,64,0.018747733036677042
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,2560,32,0.01973653237024943
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,2048,65536,0.5824576059977213
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,2048,65536,1.029101880391439
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,2048,51200,0.45397332509358723
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,2560,51200,0.3478922526041667
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,2048,51200,0.8406965255737304
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,2048,16384,0.14866453806559246
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,2048,16384,0.26736745834350584
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,2048,16384,0.09936959743499756
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,2048,12288,0.11277546882629394
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,2048,12288,0.20660799344380698
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,2560,65536,0.43589547475179036
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,2048,10240,0.09889280001322429
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,2048,10240,0.17082667350769043
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,2048,12288,0.07772800127665201
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,2048,8192,0.08082026640574137
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,2048,8192,0.13266239960988363
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,2048,10240,0.07092053095499674
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,2048,7168,0.06957013607025146
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,2048,7168,0.11436053117116292
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,2048,8192,0.058671998977661136
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,2048,6144,0.060362664858500156
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,2048,6144,0.09658559958140055
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,2048,7168,0.05500160058339437
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,2048,5120,0.04910080035527547
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,2048,5120,0.0808778683344523
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,2048,5120,0.04382400115331014
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,2048,4096,0.039928531646728514
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,2048,6144,0.04786026477813721
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,2048,4096,0.06467839876810709
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,2048,3584,0.03509653409322103
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,2048,3584,0.0575978676478068
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,2048,3584,0.03382399876912435
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,2048,3072,0.03049280047416687
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,2048,3072,0.04939306577046712
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,2048,4096,0.036363732814788816
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,2048,2560,0.02609279950459798
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,2048,2560,0.04536426862080892
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,2048,3072,0.02959573268890381
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,2048,2560,0.027550933758417766
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,2048,2048,0.021631999810536703
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,2048,2048,0.039049601554870604
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,2048,1536,0.017237333456675212
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,2048,1536,0.03421760002772013
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,2048,1536,0.020677334070205687
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,2048,1024,0.012599466244379678
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,2048,1024,0.027854933341344194
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,2048,2048,0.023931733767191567
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,2048,768,0.010109866658846538
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,2048,768,0.02693333427111308
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,2048,1024,0.018682666619618735
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,2048,512,0.007926400005817413
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,2048,512,0.02384106715520223
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,2048,768,0.017427200078964235
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,2048,256,0.006297599772612255
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,2048,256,0.020756266514460244
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,2048,512,0.017851734161376955
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,2048,256,0.01775040030479431
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,2048,128,0.00553173323472341
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,2048,51200,0.27252267201741537
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,2048,64,0.005122133096059163
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,2048,128,0.01920213301976522
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,2048,32,0.005066666503747304
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,2048,64,0.018220800161361694
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,2048,32,0.017730132738749186
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,1536,65536,0.4621280034383138
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,1536,65536,0.9583648045857748
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,1536,51200,0.3471285184224447
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,2048,128,0.01780160069465637
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,1536,51200,0.7539648056030274
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,1536,16384,0.11374613444010417
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,1536,16384,0.24964480400085448
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,1536,16384,0.09990293184916178
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,1536,12288,0.08644586404164632
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,1536,12288,0.19264426231384277
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,2048,65536,0.33968639373779297
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,1536,10240,0.07852373123168946
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,1536,10240,0.15907200177510578
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,1536,12288,0.07858666578928629
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,1536,10240,0.07164693673451741
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,1536,8192,0.06362880071004232
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,1536,8192,0.12322986920674642
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,1536,7168,0.05449279944101969
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,1536,7168,0.1064575990041097
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,1536,8192,0.059741866588592527
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,1536,6144,0.04675519863764445
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,1536,6144,0.08903253078460693
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,1536,7168,0.055852798620859775
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,1536,5120,0.03797119855880737
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,1536,5120,0.07356586456298828
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,1536,5120,0.044100264708201095
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,1536,4096,0.030885332822799684
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,1536,4096,0.05818560123443604
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,1536,6144,0.047858134905497236
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,1536,3584,0.027430399258931475
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,1536,3584,0.05134079853693644
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,1536,3584,0.034594134489695234
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,1536,3072,0.023960532744725545
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,1536,3072,0.045347201824188235
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,1536,4096,0.03831786712010701
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,1536,2560,0.020616533358891805
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,1536,2560,0.041417598724365234
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,1536,2560,0.02984106739362081
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,1536,2048,0.01705066760381063
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,1536,3072,0.03130240043004354
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,1536,2048,0.035877335071563723
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,1536,1536,0.013262933492660523
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,1536,1536,0.03160426616668701
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,1536,1536,0.02350613276163737
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,1536,1024,0.010029866298039754
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,1536,1024,0.02688213388125102
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,1536,2048,0.026233599583307905
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,1536,768,0.008088533580303193
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,1536,768,0.024716800451278685
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,1536,1024,0.02186773419380188
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,1536,512,0.006698666512966156
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,1536,768,0.020617600282033285
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,1536,512,0.022760534286499025
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,1536,256,0.005525333185990652
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,1536,256,0.01930453379948934
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,1536,512,0.021062399943669638
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,1536,256,0.020914133389790854
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,1536,128,0.004673066735267639
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,1536,128,0.01805866758028666
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,1536,64,0.0042463997999827065
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,1536,64,0.01774186690648397
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,1536,128,0.022258132696151733
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,1536,32,0.004615466793378194
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,1536,32,0.016696532567342125
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,1024,65536,0.30618454615275065
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,1024,65536,0.8871285120646158
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,1536,51200,0.27312211990356444
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,1024,51200,0.24408532778422037
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,1024,51200,0.6966592152913411
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,1024,16384,0.07946133613586426
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,1024,16384,0.23102720578511557
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,1024,16384,0.07524053255716959
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,1024,12288,0.06019733349482218
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,1024,12288,0.17920533816019696
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,1536,65536,0.3435893376668294
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,1024,10240,0.05169386863708496
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,1024,10240,0.14869653383890788
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,1024,12288,0.060031998157501223
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,1024,8192,0.04105706612269084
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,1024,8192,0.1133344014485677
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,1024,10240,0.05553493499755859
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,1024,7168,0.03674346605936686
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,1024,7168,0.09608746369679769
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,1024,8192,0.04527039925257365
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,1024,6144,0.031117866436640423
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,1024,6144,0.07957119941711426
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,1024,7168,0.04203413327534993
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,1024,5120,0.026434133450190227
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,1024,5120,0.06349440018335978
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,1024,5120,0.03386559883753459
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,1024,4096,0.022009599208831786
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,1024,6144,0.03671679894129436
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,1024,4096,0.0509610652923584
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,1024,3584,0.019827200969060262
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,1024,3584,0.04631679852803548
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,1024,3584,0.02640746633211772
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,1024,3072,0.017292799552281697
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,1024,3072,0.041495466232299806
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,1024,4096,0.028449066480000812
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,1024,2560,0.014632532993952433
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,1024,2560,0.038355199495951335
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,1024,2560,0.021908267339070638
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,1024,2048,0.012313600381215413
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,1024,2048,0.0329802672068278
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,1024,3072,0.023976532618204753
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,1024,1536,0.00988266666730245
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,1024,1536,0.029095466931660968
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,1024,2048,0.01939199964205424
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,1024,1024,0.007594666878382365
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,1024,1024,0.025494400660196943
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,1024,51200,0.1998261292775472
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,1024,768,0.006772266825040181
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,1024,768,0.022861866156260173
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,1024,1536,0.0165802667538325
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,1024,1024,0.014909866452217101
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,1024,512,0.005515733361244201
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,1024,768,0.014521599809328715
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,1024,512,0.022375466426213582
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,1024,256,0.004600533346335093
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,1024,256,0.018663465976715088
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,1024,512,0.014550399780273438
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,1024,128,0.00382080003619194
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,1024,256,0.014879999558130899
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,1024,128,0.01703146696090698
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,1024,64,0.003421866645415624
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,1024,128,0.01613759994506836
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,1024,64,0.017303466796875
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,1024,32,0.0038101332883040107
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,1024,32,0.014512000481287637
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,768,65536,0.24242773056030273
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,768,65536,0.87728640238444
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,768,51200,0.19122133255004883
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,768,51200,0.6665482838948568
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,768,16384,0.062401068210601804
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,768,16384,0.2215797265370687
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,1024,65536,0.24710613886515298
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,768,12288,0.04576319853464762
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,768,12288,0.1708288033803304
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,768,16384,0.0756704012552897
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,768,10240,0.04003200133641561
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,768,10240,0.14203306833902996
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,768,12288,0.06115093231201172
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,768,8192,0.031196800867716472
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,768,8192,0.10857280095418294
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,768,10240,0.05646293163299561
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,768,7168,0.02765973409016927
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,768,7168,0.09192213217417398
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,768,7168,0.04207466840744019
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,768,8192,0.04573866526285807
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,768,6144,0.023413334290186563
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,768,6144,0.07524480024973551
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,768,5120,0.019320533672968546
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,768,5120,0.0596565326054891
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,768,5120,0.03415573438008626
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,768,4096,0.015983999768892924
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,768,4096,0.04762666622797648
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,768,6144,0.03713279962539673
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,768,3584,0.014680533607800802
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,768,3584,0.04412480195363362
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,768,3584,0.027592533826828004
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,768,3072,0.012505599856376648
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,768,4096,0.029676800966262816
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,768,3072,0.03911466598510742
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,768,2560,0.010798933108647664
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,768,2560,0.03613119920094808
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,768,3072,0.024791467189788818
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,768,2560,0.023108265797297158
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,768,2048,0.009245866537094116
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,768,2048,0.032758400837580366
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,768,1536,0.007946666578451793
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,768,1536,0.028674133618672687
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,768,2048,0.01990293264389038
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,768,1024,0.006292266647020976
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,768,1024,0.024678399165471397
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,768,1024,0.01618346671263377
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,768,768,0.005527466535568237
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,768,768,0.02277759909629822
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,768,1536,0.017755732933680216
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,768,512,0.004631466666857402
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,768,768,0.015345066785812378
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,768,512,0.020627200603485107
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,768,256,0.003836799909671148
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,768,256,0.018653867642084758
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,768,512,0.01578133304913839
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,768,128,0.0034847999612490333
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,768,256,0.016574933131535848
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,768,128,0.017629865805308023
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,768,64,0.0030623999734719592
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,768,128,0.016759467124938966
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,768,51200,0.19980799357096354
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,768,32,0.0034346667428811393
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,768,32,0.014877866705258688
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,512,65536,0.1814143975575765
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,512,65536,0.8248512268066406
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,512,51200,0.14043199221293132
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,768,64,0.017704532543818156
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,512,51200,0.6516789118448894
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,512,16384,0.047254399458567305
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,512,16384,0.2155061403910319
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,512,16384,0.07361066341400146
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,512,12288,0.036139734586079914
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,512,12288,0.16579093933105468
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,768,65536,0.24793386459350586
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,512,10240,0.030433066685994464
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,512,10240,0.1367743968963623
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,512,12288,0.058855466047922765
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,512,8192,0.023206400871276855
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,512,10240,0.05397440195083618
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,512,8192,0.10253973007202148
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,512,7168,0.019833600521087645
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,512,7168,0.08513279755910239
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,512,8192,0.04291520118713379
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,512,6144,0.017254400253295898
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,512,6144,0.06923200289408366
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,512,7168,0.0392031987508138
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,512,6144,0.03460799853006999
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,512,5120,0.014954666296641031
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,512,5120,0.05480959812800089
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,512,4096,0.012142933408419291
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,512,4096,0.04525013367335002
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,512,51200,0.19831573168436686
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,512,5120,0.03214613397916158
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,512,3584,0.011575466394424439
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,512,3584,0.04118186632792155
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,512,3072,0.009969066580136616
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,512,3072,0.03785173495610555
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,512,3584,0.025192532936731977
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,512,2560,0.008774399757385254
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,512,4096,0.02762346665064494
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,512,3072,0.02272640069325765
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,512,2560,0.0350655992825826
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,512,2048,0.007574399809042613
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,512,2048,0.030882134040196733
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,512,1536,0.006656000018119812
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,512,1536,0.027834665775299073
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,512,2560,0.02112213373184204
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,512,2048,0.018463999032974243
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,512,1024,0.0054624001185099285
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,512,1536,0.015773866573969522
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,512,768,0.0046847999095916745
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,512,768,0.02253440022468567
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,512,1024,0.014740266402562461
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,512,512,0.004268800218900045
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,512,512,0.020356265703837077
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,512,768,0.013660800457000733
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,512,256,0.0034293333689371743
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,512,256,0.018449066082636516
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,512,512,0.013730133573214212
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,512,128,0.003050666550795237
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,512,256,0.014523733655611673
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,512,128,0.016775466998418174
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,512,64,0.0029994666576385496
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,512,128,0.014470400412877402
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,512,64,0.014857600132624308
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,512,32,0.0030037333567937215
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,512,1024,0.023018666108449302
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,512,32,0.016189866264661155
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,256,65536,0.11867733001708984
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,256,51200,0.09317546685536703
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,256,65536,0.7924170811971029
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,256,51200,0.6231903711954753
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,256,16384,0.032969599962234496
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,256,16384,0.21129280726114907
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,512,65536,0.24621866544087728
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,256,12288,0.025309866666793822
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,256,12288,0.1618058681488037
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,256,12288,0.058822401364644374
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,256,10240,0.021677867571512858
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,256,16384,0.07345706621805827
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,256,10240,0.13325546582539877
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,256,8192,0.017913599809010826
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,256,8192,0.09880959987640381
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,256,10240,0.053819731871287024
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,256,7168,0.014984533190727234
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,256,8192,0.041885864734649655
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,256,7168,0.08176853656768798
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,256,6144,0.013291733463605246
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,256,6144,0.06523840030034384
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,256,7168,0.038230399290720626
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,256,5120,0.011870933572451274
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,256,6144,0.03425600131352742
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,256,5120,0.04952960014343262
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,256,4096,0.010291199882825215
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,256,4096,0.042084264755249026
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,256,5120,0.031701334317525226
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,256,3584,0.009321600198745728
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,256,4096,0.02685760060946147
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,256,3584,0.039979732036590575
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,256,3072,0.007291733225186665
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,256,3072,0.03619199991226196
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,256,3584,0.025090134143829344
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,256,2560,0.00784746656815211
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,256,2560,0.03305813272794088
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,256,3072,0.02264639933904012
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,256,2048,0.005937066674232483
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,256,2048,0.029028266668319702
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,256,2560,0.020742400487263998
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,256,1536,0.005530666808287303
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,256,2048,0.01816213329633077
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,256,1536,0.026808534065882367
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,256,1024,0.004226133227348328
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,256,1024,0.0225600004196167
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,256,1536,0.01578879952430725
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,256,768,0.0038634667793909705
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,256,1024,0.014120533068974813
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,256,768,0.0217141330242157
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,256,512,0.003454933315515518
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,256,512,0.021517866849899293
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,256,768,0.013223466277122498
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,256,256,0.0029845332105954488
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,256,512,0.013605333367983499
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,256,256,0.01767573356628418
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,256,128,0.002605866640806198
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,256,256,0.013765333096186319
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,256,128,0.01583999991416931
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,256,128,0.014064000050226847
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,256,64,0.0025888000925381976
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,256,51200,0.1981760025024414
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,256,64,0.016212266683578492
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,256,32,0.0026378666361172995
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,256,32,0.014497066537539164
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,128,65536,0.09548373222351074
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,256,65536,0.2458570639292399
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,128,51200,0.07492480278015137
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,128,65536,0.7836202621459961
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,128,51200,0.6173322677612305
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,128,16384,0.032491733630498246
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,128,16384,0.20524586041768395
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,128,12288,0.024765866994857787
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,128,12288,0.15779733657836914
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,128,16384,0.07213653723398844
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,128,10240,0.014565333724021912
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,128,10240,0.1288597345352173
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,128,12288,0.058259201049804685
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,128,8192,0.01260373294353485
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,128,10240,0.05235093434651693
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,128,51200,0.19432746569315593
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,128,8192,0.04027093251546224
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,128,7168,0.010868266224861145
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,128,7168,0.07858666578928629
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,128,6144,0.00965333382288615
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,128,6144,0.06073386669158935
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,128,8192,0.0959007978439331
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,128,5120,0.008731733759244282
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,128,5120,0.04835520188013713
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,128,7168,0.03688533306121826
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,128,4096,0.007155199845631917
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,128,4096,0.04106346766153972
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,128,6144,0.03330666621526082
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,128,3584,0.00699946681658427
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,128,3584,0.03790186643600464
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,128,5120,0.03090133269627889
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,128,3072,0.006276266773541768
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,128,4096,0.026370133956273394
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,128,3072,0.03509866793950399
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,128,2560,0.005459199845790863
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,128,3584,0.023945599794387817
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,128,2560,0.03289706707000732
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,128,2048,0.0050240000089009605
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,128,3072,0.021492266654968263
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,128,2048,0.02877226670583089
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,128,1536,0.004251733422279358
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,128,2560,0.020205867290496827
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,128,1536,0.027870933214823406
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,128,1024,0.0037610667447249093
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,128,2048,0.017051732540130614
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,128,1024,0.022645332415898643
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,128,1536,0.014877866705258688
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,128,768,0.0033781332274278007
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,128,768,0.020785067478815714
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,128,1024,0.013103999694188437
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,128,512,0.003036800026893616
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,128,768,0.012423466642697651
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,128,512,0.018675200144449868
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,128,256,0.0026261332134405774
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,128,512,0.012894933422406515
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,128,256,0.01665173371632894
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,128,128,0.002221866697072983
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,128,256,0.012920533617337545
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,128,65536,0.24158933957417808
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,128,64,0.0022367998957633973
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,128,32,0.002569599946339925
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4353,128,128,0.012829867005348206
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,128,32,0.014968533317248026
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,64,65536,0.09612586498260497
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,128,64,0.016116266449292503
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,64,51200,0.07302186489105225
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,64,16384,0.027565866708755493
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,64,16384,0.20489066441853843
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,64,12288,0.02094399929046631
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,64,65536,0.778598403930664
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,64,51200,0.6137343724568685
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,64,12288,0.15684587160746258
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,64,10240,0.012795733412106833
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,64,8192,0.011592533191045125
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,64,7168,0.010368000467618306
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,64,8192,0.09530346393585205
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,64,6144,0.009262933333714803
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,128,128,0.015711999932924905
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,64,7168,0.07742400169372558
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,64,10240,0.12864960034688314
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,64,5120,0.00792746643225352
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,64,6144,0.05976639986038208
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,64,5120,0.04726719856262207
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,64,3584,0.006308266520500183
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,64,4096,0.0073525334397951765
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,64,4096,0.04047360022862752
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,64,3072,0.005859200159708659
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,64,3584,0.03730666637420654
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,64,2560,0.0051242664456367494
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,64,2048,0.004690133531888326
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,64,2560,0.03305813272794088
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,64,1536,0.003930666546026865
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,64,2048,0.027828266223271687
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,64,1024,0.0034101332227389016
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,64,1536,0.026563199361165364
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,64,1024,0.021709867318471274
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,64,768,0.0030165334542592366
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,64,512,0.002937600016593933
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,64,768,0.020754132668177286
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,64,256,0.0025621332228183747
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,64,512,0.019504000743230186
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,64,128,0.0022240000466505687
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,64,256,0.016897066434224447
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,64,64,0.002219733347495397
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,64,128,0.01570026675860087
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,64,32,0.00225600004196167
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,64,64,0.015006933609644571
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,64,32,0.01458453337351481
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,32,65536,0.09332906405131022
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,32,51200,0.07339306672414145
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,64,3072,0.03508266607920329
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,32,16384,0.02637333273887634
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,32,65536,0.7788607915242513
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,32,51200,0.6131509145100911
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,32,12288,0.018603734175364175
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,32,10240,0.011225600043932598
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,32,12288,0.15746347109476727
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,32,8192,0.009955199559529622
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,32,16384,0.20469226837158203
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,32,10240,0.13174293041229249
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,32,7168,0.009336533149083455
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,32,8192,0.09622399806976319
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,32,6144,0.008573866883913676
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,32,6144,0.061510399977366126
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,32,5120,0.01188053290049235
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,32,4096,0.009924266735712688
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,32,5120,0.04930026531219482
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,32,3584,0.009086933732032777
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,32,3584,0.037170132001241044
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,32,4096,0.04070933262507121
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,32,3072,0.007940266529719036
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,32,2560,0.007097599903742473
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,32,2560,0.03154346744219462
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,32,3072,0.03489813407262166
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,32,2048,0.004728533327579498
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,32,2048,0.02773226698239644
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,32,1536,0.004103466620047887
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,32,1536,0.024969599644343057
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,32,1024,0.003402666747570038
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,32,768,0.0029706666866938275
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,32,512,0.0025962665677070618
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,32,768,0.020682666699091592
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,32,1024,0.02291946609814962
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,32,256,0.002593066543340683
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,32,512,0.019364267587661743
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,32,256,0.017941333850224814
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,32,7168,0.07897386550903321
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,32,64,0.002455466737349828
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,32,32,0.002311466634273529
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,32,64,0.01590079963207245
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,32,32,0.014849066734313965
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4353,32,128,0.0021898667017618816
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4353,32,128,0.015058133006095886
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,65536,16384,5.190756225585938
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,65536,51200,8.921270751953125
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,65536,16384,2.8656651814778646
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,65536,12288,4.134769185384115
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,65536,12288,2.064243189493815
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,65536,51200,16.987969970703126
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,65536,10240,3.6570281982421875
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,65536,10240,1.8537930806477863
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,65536,10240,0.9581951777140298
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,65536,8192,2.8593897501627605
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,65536,8192,1.4146347045898438
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,65536,12288,1.1008714040120444
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,65536,7168,2.5032981872558593
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,65536,7168,1.199844233194987
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,65536,16384,1.4396981557210287
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,65536,6144,1.9717674255371094
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,65536,6144,1.0056565602620444
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,65536,8192,0.7917792002360027
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,65536,5120,1.6771221160888672
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,65536,5120,0.8944917043050131
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,65536,7168,0.748693339029948
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,65536,4096,1.2653141021728516
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,65536,4096,0.6650079727172852
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,65536,5120,0.5823359807332357
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,65536,3584,1.0626922607421876
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,65536,3584,0.6225845336914062
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,65536,4096,0.5278954823811849
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,65536,3072,0.9468575795491537
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,65536,3072,0.5086965243021647
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,65536,6144,0.646281623840332
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,65536,2560,0.7686656316121419
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,65536,2560,0.4390442530314128
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,65536,2560,0.39848321278889975
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,65536,2048,0.5845674514770508
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,65536,2048,0.34714988072713215
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,65536,3072,0.41834131876627606
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,65536,1536,0.4530624071756999
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,65536,1536,0.2887658754984538
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,65536,3584,0.4641557375590007
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,65536,1024,0.3029397328694662
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,65536,1024,0.19682346979777018
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,65536,1024,0.2807701428731283
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,65536,768,0.23688853581746422
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,65536,768,0.16609813372294108
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,65536,1536,0.29514026641845703
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,65536,512,0.17824212710062665
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,65536,512,0.13999679883321126
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,65536,2048,0.3515029271443685
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,65536,512,0.27016534805297854
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,65536,256,0.1183125336964925
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,65536,256,0.11522133350372314
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,65536,128,0.1046026627222697
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,65536,128,0.10261120001475017
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,65536,128,0.27254292170206706
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,65536,64,0.0883850653966268
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,65536,64,0.10580693085988362
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,65536,32,0.09146880308787028
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,65536,32,0.10606079896291096
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,65536,768,0.26692905426025393
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,65536,256,0.27250773111979165
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,51200,65536,9.351999918619793
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,51200,65536,17.832722981770836
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,51200,51200,13.55972696940104
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,51200,51200,7.4532958984375
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,51200,16384,4.405240376790365
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,51200,16384,2.2846132914225263
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,51200,16384,1.1582346598307292
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,51200,12288,3.159654490152995
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,51200,12288,1.640546162923177
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,51200,12288,0.8502837498982748
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,51200,10240,2.6325055440266927
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,51200,10240,1.3298858642578124
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,51200,10240,0.753062375386556
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,51200,8192,2.0709482828776045
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,51200,8192,1.0589408238728841
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,65536,51200,4.797822062174479
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,51200,7168,1.7057696024576825
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,51200,7168,0.9560106913248697
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,51200,8192,0.618395741780599
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,51200,6144,1.4045876820882162
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,51200,6144,0.7787743886311849
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,51200,7168,0.5767679850260417
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,51200,5120,1.1843743642171225
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,51200,5120,0.6984640121459961
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,51200,6144,0.4966549237569173
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,51200,4096,0.9282346725463867
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,51200,4096,0.5643605550130208
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,51200,5120,0.4550325393676758
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,51200,3584,0.8081535975138345
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,51200,3584,0.47575680414835614
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,51200,3584,0.3626506805419922
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,51200,3072,0.680284818013509
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,51200,3072,0.4135423978169759
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,51200,51200,3.7526710510253904
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,51200,2560,0.5553290685017903
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,51200,2560,0.34548587799072267
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,51200,4096,0.40427840550740557
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,51200,2048,0.4622933387756348
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,51200,2048,0.27894293467203773
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,51200,2560,0.3095861434936523
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,51200,1536,0.34472745259602866
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,51200,1536,0.2156149387359619
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,51200,2048,0.27318506240844725
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,51200,1024,0.2443295955657959
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,51200,1024,0.1612650712331136
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,51200,3072,0.3240746815999349
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,51200,768,0.18818453152974446
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,51200,768,0.13403840065002443
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,51200,768,0.21302293141682943
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,51200,512,0.13833279609680177
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,51200,512,0.11300266583760579
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,51200,1024,0.22788693110148112
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,51200,256,0.09423253536224366
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,51200,256,0.09435199896494548
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,51200,256,0.21328320503234863
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,51200,128,0.07548159758249919
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,51200,128,0.08442026774088542
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,51200,128,0.2162048021952311
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,51200,64,0.07059306303660075
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,51200,64,0.08409386475880941
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,51200,32,0.07083840370178222
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,51200,32,0.08584960301717123
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,51200,1536,0.23234240214029947
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,51200,512,0.21311039924621583
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,16384,65536,3.2543604532877604
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,16384,65536,5.349885050455729
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,16384,51200,2.597996775309245
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,16384,51200,4.159553019205729
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,16384,16384,1.3460875193277995
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,16384,16384,0.8241504033406576
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,16384,16384,0.4119999885559082
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,16384,12288,0.9410677591959635
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,16384,12288,0.6363082885742187
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,16384,12288,0.31050240198771156
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,16384,10240,0.7503722508748372
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,16384,10240,0.5127306620279948
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,16384,10240,0.27707306543986
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,16384,8192,0.6113887786865234
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,16384,8192,0.40452054341634114
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,16384,8192,0.22885440190633138
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,16384,7168,0.5086698532104492
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,16384,7168,0.34168532689412434
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,16384,51200,1.2795178731282553
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,16384,6144,0.4453610738118489
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,16384,6144,0.3084021250406901
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,16384,7168,0.21225813229878746
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,16384,5120,0.35713494618733727
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,16384,5120,0.25072107315063474
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,16384,5120,0.16897600491841633
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,16384,4096,0.2896362622578939
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,16384,4096,0.2000864028930664
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,16384,6144,0.18629439671834308
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,16384,3584,0.25831359227498374
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,16384,3584,0.18299412727355957
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,16384,4096,0.14930027325948078
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,16384,3072,0.2189781347910563
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,16384,3584,0.13613866170247396
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,16384,3072,0.1540895938873291
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,16384,2560,0.1822762648264567
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,16384,2560,0.1328970670700073
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,16384,2560,0.1144704023996989
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,16384,2048,0.1485919952392578
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,16384,2048,0.10817493597666424
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,16384,3072,0.12065280278523763
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,16384,1536,0.11666026910146077
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,16384,1536,0.08991146882375081
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,16384,1536,0.08737706343332927
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,16384,1024,0.08114346663157145
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,16384,1024,0.06774826844533285
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,16384,2048,0.09875840346018473
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,16384,768,0.06244800090789795
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,16384,768,0.060845867792765296
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,16384,768,0.07573973337809245
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,16384,512,0.046288001537323
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,16384,512,0.053572265307108555
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,16384,1024,0.08394880294799804
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,16384,256,0.03305173317591349
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,16384,256,0.058841598033905027
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,16384,256,0.07039679686228434
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,16384,128,0.03023359974225362
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,16384,128,0.039792001247406006
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,16384,512,0.0732479969660441
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,16384,64,0.025574400027592974
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,16384,64,0.04166293144226074
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,16384,32,0.026415999730428057
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,16384,32,0.03944106499354045
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,16384,128,0.07180373668670655
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,12288,65536,2.6083338419596354
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,12288,65536,3.9683359781901038
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,12288,51200,3.1223691304524737
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,12288,51200,2.0244330088297526
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,16384,65536,1.6096682230631512
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,12288,16384,0.9171264012654623
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,12288,16384,0.6543872197469075
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,51200,65536,4.799673461914063
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,12288,12288,0.6742613474527995
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,12288,12288,0.5015690803527832
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,12288,12288,0.2418037255605062
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,12288,10240,0.5600298563639323
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,12288,10240,0.4006656010945638
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,12288,16384,0.3199680010477702
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,12288,8192,0.449292786916097
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,12288,8192,0.35896533330281577
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,12288,8192,0.17769066492716473
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,12288,7168,0.3910069465637207
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,12288,7168,0.2803743998209635
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,12288,10240,0.2155402660369873
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,12288,6144,0.32818771998087565
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,12288,6144,0.2337183952331543
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,12288,6144,0.1436309337615967
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,12288,5120,0.2737781206766764
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,12288,5120,0.1987786610921224
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,12288,7168,0.16435839335123698
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,12288,4096,0.21980586051940917
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,12288,4096,0.16240320205688477
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,12288,4096,0.11662720044453938
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,12288,3584,0.1935637315114339
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,12288,3584,0.14353706041971842
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,12288,5120,0.1317632039388021
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,12288,3072,0.165775998433431
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,12288,3072,0.12400106588999431
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,12288,3584,0.10480319658915202
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,12288,2560,0.14031786918640138
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,12288,2560,0.10685973167419434
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,12288,3072,0.09313066800435385
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,12288,2048,0.1129642645517985
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,12288,2048,0.08896959622701009
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,12288,2048,0.07784319718678792
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,12288,1536,0.08647466500600179
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,12288,1536,0.07388266722361246
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,12288,2560,0.08886933326721191
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,12288,1024,0.06050346692403158
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,12288,1024,0.0555178682009379
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,12288,1024,0.06741226514180501
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,12288,768,0.04737600088119507
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,12288,768,0.05112959941228231
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,12288,1536,0.06977386474609375
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,12288,512,0.03522133429845174
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,12288,512,0.045424000422159834
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,12288,768,0.06280320088068644
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,12288,512,0.06327040195465088
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,12288,256,0.02473919987678528
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,12288,256,0.03712639808654785
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,12288,128,0.02047146757443746
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,12288,128,0.03322239915529887
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,12288,256,0.06285760005315146
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,12288,64,0.01823893388112386
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,12288,64,0.03318399985631307
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,12288,32,0.019407999515533448
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,12288,32,0.033240532875061034
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,12288,128,0.06292373339335124
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,10240,65536,2.28102289835612
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,10240,65536,3.321158345540365
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,10240,51200,2.6206036885579427
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,10240,51200,1.7827733357747395
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,12288,51200,0.9602783838907877
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,10240,16384,0.7999359766642253
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,10240,16384,0.5751125335693359
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,10240,16384,0.2712565422058105
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,10240,12288,0.5562378565470378
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,10240,12288,0.45226240158081055
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,12288,65536,1.2237247467041015
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,10240,10240,0.469160525004069
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,10240,10240,0.37404054005940757
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,10240,12288,0.2056682745615641
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,10240,8192,0.36412054697672525
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,10240,8192,0.2733375867207845
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,10240,10240,0.18495680491129557
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,10240,7168,0.32988160451253257
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,10240,7168,0.2504693349202474
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,10240,8192,0.1504672050476074
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,10240,6144,0.27071361541748046
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,10240,6144,0.20306453704833985
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,10240,7168,0.14196799596150717
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,10240,5120,0.23112640380859376
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,10240,5120,0.17609705924987792
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,10240,6144,0.12307519912719726
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,10240,4096,0.17893120447794597
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,10240,4096,0.13771840731302898
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,10240,5120,0.11450880368550617
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,10240,3584,0.16274666786193848
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,10240,3584,0.1280128002166748
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,10240,51200,0.8071690877278647
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,10240,3072,0.136735995610555
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,10240,3072,0.10745813051859539
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,10240,4096,0.09675306479136149
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,10240,2560,0.11439680258433024
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,10240,2560,0.09487679799397787
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,10240,3584,0.08914773464202881
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,10240,2048,0.09577066898345947
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,10240,2048,0.07901866436004638
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,10240,2560,0.07559466361999512
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,10240,1536,0.07090240319569906
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,10240,1536,0.06618880033493042
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,10240,3072,0.07940906683603922
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,10240,1024,0.05052586793899536
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,10240,1024,0.05115626653035482
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,10240,2048,0.06879893143971762
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,10240,768,0.03993173440297444
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,10240,768,0.04718933502833049
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,10240,768,0.05601173241933187
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,10240,512,0.029896533489227294
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,10240,512,0.041162665685017905
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,10240,1024,0.059860265254974364
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,10240,256,0.021081600586573282
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,10240,1536,0.06218026479085287
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,10240,256,0.03486293156941732
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,10240,128,0.017385600010553996
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,10240,128,0.03137173255284627
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,10240,128,0.05591786702473959
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,10240,256,0.05576320091883341
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,10240,64,0.015357866883277893
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,10240,32,0.01607360045115153
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,10240,64,0.030646399656931562
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,10240,32,0.02983040014902751
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,10240,512,0.05648213227589925
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,8192,65536,1.9420096079508464
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,8192,65536,2.6739222208658857
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,8192,51200,1.5151967366536458
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,8192,51200,1.9895552317301433
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,8192,16384,0.6930773417154948
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,8192,16384,0.5302687962849935
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,10240,65536,1.0219008127848306
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,8192,12288,0.45679359436035155
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,8192,12288,0.3810485204060873
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,8192,16384,0.2651690642038981
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,8192,10240,0.40127147038777666
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,8192,10240,0.33919572830200195
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,8192,12288,0.19170986811319987
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,8192,8192,0.32432851791381834
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,8192,8192,0.2646997292836507
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,8192,10240,0.16518294016520182
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,8192,7168,0.2825557390848795
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,8192,7168,0.22079253196716309
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,8192,7168,0.12329813639322917
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,8192,6144,0.2408991972605387
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,8192,8192,0.1323199987411499
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,8192,6144,0.1876405398050944
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,8192,5120,0.19718720118204752
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,8192,5120,0.15609599749247233
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,8192,5120,0.09907413323720296
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,8192,4096,0.15683840115865072
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,8192,4096,0.12871999740600587
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,8192,6144,0.10798079967498779
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,8192,3584,0.13891414006551106
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,8192,3584,0.11509973208109539
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,8192,3584,0.07581226825714112
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,8192,3072,0.11913706461588543
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,8192,3072,0.09945279757181803
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,8192,4096,0.08503039677937826
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,8192,2560,0.09939093589782715
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,8192,2560,0.08930880228678385
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,8192,3072,0.06691839694976806
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,8192,2048,0.0809706687927246
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,8192,2048,0.0741312026977539
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,8192,2560,0.0637386679649353
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,8192,1536,0.06260159810384115
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,8192,1536,0.06153706709543864
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,8192,1536,0.047508267561594646
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,8192,1024,0.04430186748504639
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,8192,1024,0.04551253318786621
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,8192,2048,0.05511039892832438
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,8192,768,0.03308586676915486
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,8192,768,0.04236479997634888
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,8192,1024,0.04498026768366496
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,8192,512,0.02357333302497864
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,8192,512,0.043510401248931886
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,8192,768,0.039061331748962404
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,8192,256,0.0247871994972229
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,8192,256,0.037062398592631024
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,8192,256,0.039095465342203775
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,8192,128,0.0165173331896464
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,8192,128,0.027254400650660197
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,8192,512,0.03938133319218953
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,8192,64,0.01416213313738505
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,8192,128,0.039953064918518064
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,8192,64,0.029233066240946452
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,8192,32,0.01402453382809957
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,8192,32,0.02752000093460083
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,7168,65536,1.8827103932698566
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,7168,65536,2.51169916788737
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,7168,51200,1.9548021952311199
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,7168,51200,1.4859808603922526
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,8192,51200,0.761894416809082
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,7168,16384,0.5611061096191406
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,7168,16384,0.49422934850056965
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,7168,16384,0.23439359664916992
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,7168,12288,0.4210997263590495
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,7168,12288,0.36656106313069664
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,8192,65536,0.9828437169392904
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,7168,10240,0.3587669372558594
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,7168,10240,0.3100778579711914
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,7168,12288,0.1699285348256429
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,7168,8192,0.27969172795613606
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,7168,8192,0.23205119768778482
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,7168,10240,0.14966293970743816
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,7168,7168,0.24979947408040365
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,7168,7168,0.20402560234069825
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,7168,8192,0.12133440176645915
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,7168,6144,0.208518393834432
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,7168,6144,0.16998720169067383
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,7168,7168,0.11336320241292316
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,7168,5120,0.17569279670715332
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,7168,5120,0.14686187108357746
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,7168,5120,0.09186240037282309
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,7168,4096,0.13958080609639484
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,7168,4096,0.11876906553904217
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,7168,6144,0.0994592030843099
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,7168,3584,0.11967999935150146
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,7168,3584,0.10724266370137532
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,7168,3584,0.07074879805246989
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,7168,4096,0.07882346312204996
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,7168,3072,0.10288426876068116
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,7168,3072,0.09361386299133301
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,7168,2560,0.08868266741434733
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,7168,2560,0.08321706453959146
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,7168,2560,0.05963626702626547
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,7168,51200,0.6809823989868165
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,7168,2048,0.07257066567738851
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,7168,2048,0.06771413485209146
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,7168,1536,0.054871467749277744
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,7168,1536,0.05699839989344278
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,7168,3072,0.06277226607004802
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,7168,1024,0.03825920025507609
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,7168,1024,0.04393279949824015
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,7168,1536,0.04707093238830566
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,7168,768,0.030011733373006184
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,7168,1024,0.04483413298924764
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,7168,768,0.04047786792119344
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,7168,512,0.022124799092610677
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,7168,512,0.035004798571268717
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,7168,512,0.04124906857808431
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,7168,256,0.016515200336774193
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,7168,256,0.02958826621373494
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,7168,2048,0.05272320111592611
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,7168,128,0.012965333461761475
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,7168,128,0.03091199994087219
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,7168,768,0.040746665000915526
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,7168,64,0.016532267133394875
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,7168,64,0.02834986646970113
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,7168,128,0.04155946572621663
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,7168,32,0.015020799636840821
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,7168,256,0.0409333348274231
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,7168,32,0.027213867505391436
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,6144,65536,1.6605791727701824
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,6144,65536,1.947537104288737
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,6144,51200,1.5474933624267577
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,6144,51200,1.2760351816813151
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,6144,16384,0.44518400828043625
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,6144,16384,0.42344961166381834
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,6144,16384,0.2017002741495768
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,6144,12288,0.33648745218912757
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,6144,12288,0.32079998652140296
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,7168,65536,0.8577162424723307
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,6144,10240,0.30049705505371094
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,6144,10240,0.2712394714355469
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,6144,10240,0.13278079827626546
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,6144,8192,0.24159040451049804
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,6144,8192,0.21268800099690757
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,6144,12288,0.14951786994934083
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,6144,7168,0.209716272354126
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,6144,7168,0.18296319643656414
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,6144,7168,0.10151573022206624
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,6144,6144,0.1789888064066569
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,6144,6144,0.15533653895060223
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,6144,8192,0.10827946662902832
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,6144,5120,0.1479434649149577
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,6144,5120,0.13149867057800294
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,6144,6144,0.08998933633168539
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,6144,4096,0.11696853637695312
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,6144,4096,0.10841066837310791
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,6144,5120,0.08261866569519043
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,6144,3584,0.10365440050760906
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,6144,3584,0.09963626861572265
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,6144,3584,0.06270613272984824
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,6144,3072,0.08945173422495524
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,6144,3072,0.0864736000696818
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,6144,4096,0.07066559791564941
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,6144,2560,0.07455146312713623
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,6144,2560,0.07707839806874593
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,6144,2560,0.05393813451131185
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,6144,2048,0.060773332913716636
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,6144,2048,0.063318399588267
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,6144,3072,0.0555893341700236
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,6144,1536,0.04755093256632487
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,6144,1536,0.052849066257476804
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,6144,1536,0.04199359814325969
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,6144,1024,0.03301546573638916
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,6144,1024,0.040677332878112794
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,6144,2048,0.04734933376312256
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,6144,768,0.025831466913223265
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,6144,768,0.03915839989980062
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,6144,768,0.037001601854960126
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,6144,512,0.01909439961115519
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,6144,512,0.03492586612701416
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,6144,512,0.03745599985122681
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,6144,256,0.016962132851282754
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,6144,256,0.02991039951642354
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,6144,1024,0.039946667353312176
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,6144,128,0.011193600296974183
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,6144,128,0.02900159955024719
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,6144,256,0.037050668398539224
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,6144,128,0.037028264999389646
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,6144,64,0.013332266608874002
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,6144,32,0.014077867070833841
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,6144,64,0.0284768005212148
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,6144,32,0.026979200045267743
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,5120,65536,1.5686965942382813
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,5120,65536,1.7767968495686848
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,5120,51200,1.3312479654947915
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,5120,51200,1.2328107198079428
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,6144,51200,0.5903872172037761
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,5120,16384,0.4217535972595215
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,5120,16384,0.41114559173583987
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,5120,16384,0.17313812573750814
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,5120,12288,0.31532694498697916
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,5120,12288,0.3153781255086263
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,6144,65536,0.7511605580647787
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,5120,10240,0.25421226819356285
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,5120,10240,0.2629749298095703
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,5120,12288,0.13112746874491374
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,5120,8192,0.20425173441569008
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,5120,8192,0.1909898598988851
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,5120,10240,0.11828052997589111
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,5120,7168,0.1807594617207845
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,5120,7168,0.16749332745869955
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,5120,8192,0.09647253354390463
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,5120,6144,0.15264533360799154
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,5120,6144,0.14016106923421223
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,5120,7168,0.09100373586018881
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,5120,5120,0.1263797362645467
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,5120,5120,0.11997333367665608
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,5120,5120,0.07463359832763672
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,5120,4096,0.10015359719594319
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,5120,4096,0.09899413585662842
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,5120,6144,0.0807925303777059
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,5120,3584,0.08724799950917562
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,5120,3584,0.09053866863250733
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,5120,3584,0.057513598601023355
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,5120,3072,0.07508479754130046
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,5120,4096,0.06316800117492676
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,5120,3072,0.07938026587168376
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,5120,2560,0.06333653529485067
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,5120,2560,0.06919786930084229
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,5120,2560,0.04912426471710205
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,5120,2048,0.051658666133880614
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,5120,2048,0.05576320091883341
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,5120,3072,0.05102719863255819
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,5120,1536,0.040440531571706136
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,5120,1536,0.04777599970499675
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,5120,1536,0.03946773211161296
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,5120,1024,0.028333866596221925
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,5120,1024,0.03656853437423706
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,5120,2048,0.044078934192657473
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,5120,768,0.022114133834838866
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,5120,768,0.03317973415056864
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,5120,1024,0.03746666510899861
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,5120,512,0.016709333658218382
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,5120,512,0.032994133234024045
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,5120,768,0.03589119911193848
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,5120,256,0.01617280046145121
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,5120,256,0.024897066752115886
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,5120,512,0.03613119920094808
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,5120,256,0.035979731877644854
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,5120,128,0.009958400328954061
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,5120,128,0.023090134064356484
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,5120,64,0.009177600344022114
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,5120,64,0.023481599489847817
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,5120,32,0.00960640013217926
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,5120,32,0.021169066429138184
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,5120,128,0.03599040110905965
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,5120,51200,0.5087509473164876
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,4096,65536,1.2699979146321616
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,4096,65536,1.3514539082845052
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,4096,51200,0.9836106618245444
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,4096,51200,1.0588383992513022
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,4096,16384,0.30397332509358727
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,4096,16384,0.34931519826253254
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,4096,16384,0.1459093411763509
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,4096,12288,0.2303178628285726
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,4096,12288,0.26723734537760413
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,5120,65536,0.6384703954060872
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,4096,10240,0.19953173001607258
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,4096,10240,0.2203221321105957
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,4096,10240,0.10057919820149738
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,4096,8192,0.16074132919311523
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,4096,8192,0.17089600563049318
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,4096,12288,0.11219093004862468
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,4096,7168,0.1452074686686198
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,4096,7168,0.14768533706665038
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,4096,7168,0.07811093330383301
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,4096,6144,0.12363093694051105
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,4096,6144,0.12658987045288086
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,4096,8192,0.08258346716562906
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,4096,5120,0.09814613660176595
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,4096,5120,0.10737813313802083
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,4096,5120,0.06341546773910522
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,4096,4096,0.07878186702728271
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,4096,4096,0.08856746355692545
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,4096,6144,0.06952319939931234
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,4096,3584,0.07125013669331869
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,4096,3584,0.08175466855367025
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,4096,3584,0.04631253480911255
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,4096,3072,0.061860267321268717
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,4096,3072,0.06943786938985189
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,4096,4096,0.05253760019938151
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,4096,2560,0.05118186473846435
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,4096,2560,0.06076906522115072
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,4096,2560,0.03911360104878743
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,4096,2048,0.04158186515172323
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,4096,2048,0.049556267261505124
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,4096,3072,0.041554133097330734
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,4096,1536,0.032813866933186844
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,4096,1536,0.042294398943583174
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,4096,1536,0.029257599512736005
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,4096,1024,0.023117866118748984
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,4096,1024,0.034635734558105466
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,4096,2048,0.033403734366099044
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,4096,768,0.018312533696492515
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,4096,768,0.03095146616299947
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,4096,1024,0.027266132831573486
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,4096,512,0.01369599997997284
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,4096,512,0.02677759925524394
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,4096,768,0.0243776003519694
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,4096,256,0.009639466802279156
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,4096,256,0.02350506583849589
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,4096,256,0.024627200762430825
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,4096,128,0.010418132940928141
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,4096,512,0.024719999233881632
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,4096,128,0.023025067647298177
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,4096,64,0.008844799796740214
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,4096,64,0.02186453342437744
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,4096,32,0.009175466497739156
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,4096,128,0.02477546731630961
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,3584,65536,1.189135996500651
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,4096,32,0.021432532866795858
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,3584,65536,1.3098475138346353
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,3584,51200,0.8884181340535482
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,3584,51200,1.0325696309407553
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,4096,51200,0.424179204305013
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,3584,16384,0.28158613840738933
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,3584,16384,0.3357269287109375
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,3584,16384,0.1456778685251872
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,3584,12288,0.2129749298095703
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,3584,12288,0.25536212921142576
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,4096,65536,0.5336362838745117
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,3584,10240,0.17714346249898275
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,3584,10240,0.21294506390889487
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,3584,12288,0.11225706736246746
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,3584,8192,0.14383039474487305
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,3584,8192,0.16513387362162274
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,3584,10240,0.10174612998962403
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,3584,7168,0.12583146890004476
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,3584,7168,0.15075519879659016
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,3584,7168,0.07896426518758139
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,3584,6144,0.10886080265045166
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,3584,6144,0.12789653142293295
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,3584,8192,0.08385813236236572
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,3584,5120,0.09176426728566488
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,3584,5120,0.10891306400299072
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,3584,5120,0.06335573196411133
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,3584,4096,0.07436266740163168
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,3584,4096,0.09065279960632325
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,3584,6144,0.06996373335520426
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,3584,3584,0.07187519868214926
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,3584,3584,0.0818506638209025
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,3584,3584,0.04845013221104939
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,3584,4096,0.05281173388163248
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,3584,3072,0.06358933448791504
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,3584,3072,0.06090559959411621
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,3584,2560,0.05228266716003418
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,3584,2560,0.05808746814727783
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,3584,3072,0.04389653205871582
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,3584,2048,0.04310613473256429
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,3584,2048,0.04816853205362956
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,3584,2560,0.04211733341217041
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,3584,1536,0.02795093258221944
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,3584,1536,0.04344853162765503
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,3584,51200,0.42772480646769206
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,3584,1024,0.023034665981928507
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,3584,1024,0.036183468500773114
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,3584,1536,0.03285226623217265
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,3584,768,0.018819200992584228
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,3584,2048,0.03725546598434448
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,3584,1024,0.031207466125488283
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,3584,512,0.015846400459607442
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,3584,512,0.030295467376708983
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,3584,768,0.03299626708030701
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,3584,256,0.012532266974449157
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,3584,256,0.02650559941927592
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,3584,768,0.02929813265800476
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,3584,128,0.008378666639328004
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,3584,128,0.02318933407465617
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,3584,256,0.02921813329060872
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,3584,128,0.02954346736272176
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,3584,64,0.010006399949391682
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,3584,32,0.01037013332049052
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,3584,64,0.024728532632191977
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,3584,32,0.023321600755055745
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,3072,65536,0.9264458974202474
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,3584,512,0.029625600576400755
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,3072,65536,1.1902762095133463
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,3072,51200,0.7147029240926106
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,3072,51200,0.9369909286499023
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,3072,16384,0.228165340423584
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,3072,16384,0.30847787857055664
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,3072,16384,0.12262612978617352
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,3072,12288,0.17261013984680176
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,3072,12288,0.23479040463765463
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,3584,65536,0.5361824035644531
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,3072,10240,0.1461525281270345
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,3072,10240,0.19458133379618328
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,3072,10240,0.08705066839853923
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,3072,8192,0.11826986471811932
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,3072,8192,0.15081706047058105
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,3072,12288,0.09626239935557047
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,3072,7168,0.10396160284678142
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,3072,7168,0.13184320131937663
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,3072,8192,0.07199573516845703
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,3072,7168,0.06770880222320556
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,3072,6144,0.0884447971979777
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,3072,6144,0.1127509355545044
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,3072,5120,0.07440319856007895
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,3072,5120,0.0955135981241862
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,3072,5120,0.05482986768086752
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,3072,4096,0.06027093331019083
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,3072,4096,0.07765440146128336
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,3072,6144,0.06092906792958578
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,3072,3584,0.05291093190511068
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,3072,3584,0.06969280242919922
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,3072,4096,0.045719468593597413
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,3072,3072,0.04628266493479411
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,3072,3584,0.04148373206456502
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,3072,3072,0.05903253157933554
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,3072,2560,0.03886613448460897
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,3072,2560,0.05238613287607828
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,3072,2560,0.035493334134419754
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,3072,2048,0.03179946740468343
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,3072,2048,0.043852798144022626
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,3072,3072,0.037623465061187744
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,3072,1536,0.024601600567499795
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,3072,1536,0.03914026816685994
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,3072,2048,0.03162026604016622
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,3072,1024,0.018231467405954997
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,3072,1024,0.03235200047492981
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,3072,1536,0.028505599498748778
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,3072,768,0.014452266693115234
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,3072,768,0.028832000494003297
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,3072,1024,0.026365866263707478
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,3072,768,0.028514132897059126
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,3072,512,0.01092800001303355
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,3072,512,0.026476800441741943
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,3072,256,0.00844266712665558
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,3072,256,0.024155733982721965
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,3072,256,0.029074132442474365
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,3072,128,0.007163733243942261
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,3072,512,0.029268266757329305
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,3072,128,0.022592000166575112
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,3072,64,0.00800960014263789
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,3072,64,0.021901865800221763
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,3072,32,0.008351999521255492
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,3072,32,0.020666666825612388
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,3072,128,0.026834134260813398
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,2560,65536,0.772109858194987
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,2560,65536,1.126317850748698
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,2560,51200,0.6023146947224934
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,2560,51200,0.8855509440104166
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,3072,51200,0.34842348098754883
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,2560,16384,0.2002474625905355
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,2560,16384,0.2935914675394694
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,2560,16384,0.12246399720509846
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,2560,12288,0.1527242660522461
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,2560,12288,0.22475306193033853
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,3072,65536,0.43584321339925125
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,2560,10240,0.14675733248392742
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,2560,10240,0.19354987144470215
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,2560,12288,0.09512853622436523
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,2560,8192,0.1151466687520345
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,2560,8192,0.15105493863423664
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,2560,10240,0.08611413637797037
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,2560,7168,0.1015893300374349
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,2560,7168,0.1310538689295451
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,2560,8192,0.07094720204671225
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,2560,6144,0.09146986802419027
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,2560,6144,0.11166933377583821
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,2560,7168,0.06668586730957031
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,2560,5120,0.07320746580759684
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,2560,5120,0.08806933561960856
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,2560,6144,0.059483734766642246
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,2560,4096,0.05989760160446167
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,2560,4096,0.07111252943674723
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,2560,5120,0.05258133411407471
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,2560,3584,0.05622400045394897
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,2560,3584,0.06282026767730713
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,2560,3584,0.040140799681345624
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,2560,3072,0.045747200647989913
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,2560,3072,0.05384106636047363
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,2560,4096,0.044489598274230956
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,2560,2560,0.03321066697438558
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,2560,2560,0.04817599852879842
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,2560,3072,0.03586346705754598
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,2560,2048,0.02693013350168864
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,2560,2048,0.040533332029978435
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,2560,2560,0.03415786822636922
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,2560,1536,0.0215445339679718
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,2560,1536,0.0350816011428833
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,2560,2048,0.030060799916585286
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,2560,1024,0.014472533265749613
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,2560,1536,0.026371200879414875
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,2560,1024,0.030109866460164385
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,2560,768,0.011709866921106975
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,2560,768,0.026885332663853963
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,2560,768,0.022680532932281495
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,2560,512,0.00958720048268636
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,2560,512,0.02495786746342977
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,2560,1024,0.024701867500940958
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,2560,256,0.007521066566308339
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,2560,256,0.021654399236043294
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,2560,256,0.02321066657702128
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,2560,128,0.006759466727574666
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,2560,512,0.023101866245269775
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,2560,128,0.02041706641515096
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,2560,64,0.006668800115585327
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,2560,64,0.018631466229756675
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,2560,32,0.006307200094064077
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,2560,128,0.023119999965031942
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,2560,32,0.018577067057291667
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,2048,65536,0.6266709645589192
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,2048,65536,1.0359722773234048
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,2560,51200,0.34770027796427405
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,2048,51200,0.4619690577189128
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,2048,51200,0.8087637583414713
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,2048,16384,0.14829440116882325
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,2048,16384,0.2674709320068359
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,2048,16384,0.09921173254648843
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,2048,12288,0.11268266836802165
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,2048,12288,0.20647892951965333
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,2560,65536,0.43659413655598955
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,2048,10240,0.09877226352691651
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,2048,10240,0.171888001759847
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,2048,12288,0.07789440155029297
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,2048,8192,0.08038720289866129
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,2048,8192,0.1328437328338623
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,2048,10240,0.07047146956125895
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,2048,7168,0.06942826906840006
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,2048,7168,0.11450560092926025
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,2048,8192,0.058664532502492275
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,2048,6144,0.0603007992108663
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,2048,6144,0.09678080081939697
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,2048,7168,0.05431253512700399
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,2048,5120,0.05015679995218912
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,2048,5120,0.08142826557159424
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,2048,6144,0.048709332942962646
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,2048,4096,0.041015466054280594
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,2048,4096,0.06355626583099365
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,2048,5120,0.042786133289337155
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,2048,3584,0.035699200630187986
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,2048,3584,0.057518935203552245
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,2048,3584,0.03338026603062948
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,2048,3072,0.031177600224812824
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,2048,3072,0.05119786659876505
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,2048,4096,0.036702934900919595
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,2048,2560,0.026945066452026368
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,2048,2560,0.04536426862080892
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,2048,3072,0.029443200429280596
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,2048,2048,0.022392533222834268
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,2048,2560,0.027436800797780353
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,2048,2048,0.03927786747614543
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,2048,1536,0.017682133118311565
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,2048,1536,0.03507839838663737
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,2048,1536,0.020989867051442464
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,2048,1024,0.012551466623942057
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,2048,1024,0.02757546703020732
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,2048,2048,0.023913600047429404
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,2048,768,0.010387200117111205
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,2048,768,0.026539733012517292
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,2048,1024,0.019016534090042114
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,2048,512,0.008205866813659668
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,2048,512,0.023509333531061806
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,2048,51200,0.27192961374918617
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,2048,256,0.006321066617965698
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,2048,256,0.02110613385836283
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,2048,768,0.016952532529830932
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,2048,512,0.017825067043304443
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,2048,128,0.00594346672296524
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,2048,256,0.01738986571629842
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,2048,64,0.005038933455944061
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,2048,128,0.019080533583958944
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,2048,64,0.017805866400400796
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,2048,32,0.00517439991235733
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,2048,32,0.016535466909408568
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,2048,128,0.017879466215769448
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,1536,65536,0.44386879603068036
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,1536,51200,0.35527467727661133
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,1536,65536,0.9649173100789389
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,1536,51200,0.7541450500488281
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,1536,16384,0.11373546918233235
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,1536,16384,0.2486133257548014
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,1536,16384,0.09938027064005533
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,1536,12288,0.08639466762542725
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,1536,12288,0.19263680775960285
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,2048,65536,0.3383573214213053
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,1536,10240,0.07914453347524007
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,1536,10240,0.16011093457539877
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,1536,12288,0.07818666299184164
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,1536,8192,0.06222933530807495
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,1536,8192,0.12381546497344971
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,1536,10240,0.07177386283874512
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,1536,7168,0.05530773401260376
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,1536,7168,0.10594027042388916
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,1536,8192,0.059095466136932374
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,1536,6144,0.047202134132385255
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,1536,7168,0.05531520048777262
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,1536,6144,0.08993173440297444
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,1536,5120,0.03947093486785889
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,1536,5120,0.07396372954050699
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,1536,5120,0.044088534514109295
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,1536,4096,0.032076799869537355
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,1536,4096,0.05751786629358927
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,1536,6144,0.04819200038909912
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,1536,3584,0.028500266869862872
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,1536,3584,0.05197866757710775
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,1536,3584,0.034628268082936606
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,1536,4096,0.037904000282287596
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,1536,3072,0.02477653423945109
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,1536,3072,0.04657173156738281
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,1536,2560,0.02148373325665792
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,1536,2560,0.04209920167922974
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,1536,2560,0.029639466603597002
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,1536,2048,0.01812160015106201
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,1536,2048,0.03694399992624919
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,1536,3072,0.031227733691533404
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,1536,1536,0.013892267147699991
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,1536,1536,0.030961066484451294
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,1536,1536,0.023386667172114052
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,1536,1024,0.010061867038408915
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,1536,1024,0.02605973283449809
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,1536,2048,0.02641493280728658
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,1536,768,0.008397866288820903
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,1536,768,0.025303467114766436
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,1536,1024,0.021990400552749634
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,1536,51200,0.27383467356363933
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,1536,512,0.006656000018119812
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,1536,512,0.022401066621144612
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,1536,768,0.02013333241144816
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,1536,256,0.005433600147565206
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,1536,256,0.019361066818237304
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,1536,256,0.02095466653505961
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,1536,512,0.020835200945536293
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,1536,128,0.005037866532802582
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,1536,128,0.01819733381271362
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,1536,64,0.0046304002404212955
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,1536,128,0.022232532501220703
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,1536,64,0.01665173371632894
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,1536,32,0.004715733230113983
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,1536,32,0.0165802667538325
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,1024,65536,0.3170271873474121
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,1024,51200,0.2464896043141683
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,1024,65536,0.8913653055826822
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,1024,51200,0.6970357259114583
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,1024,16384,0.08061227003733316
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,1024,16384,0.23154773712158203
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,1024,16384,0.07462506294250489
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,1024,12288,0.061426134904225674
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,1024,12288,0.1789621353149414
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,1536,65536,0.3434432029724121
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,1024,10240,0.05183999935785929
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,1024,10240,0.14805226325988768
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,1024,12288,0.05936533212661743
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,1024,8192,0.04198293288548787
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,1024,8192,0.11321600278218587
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,1024,10240,0.054876800378163657
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,1024,7168,0.03628906806310018
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,1024,7168,0.09594240188598632
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,1024,8192,0.045653335253397626
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,1024,6144,0.03096853295962016
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,1024,6144,0.07852906386057536
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,1024,7168,0.04102079868316651
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,1024,5120,0.02643199960390727
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,1024,5120,0.06373013257980346
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,1024,6144,0.03628373146057129
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,1024,4096,0.021782400210698445
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,1024,4096,0.050876800219217935
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,1024,4096,0.028435200452804565
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,1024,3584,0.019474132855733236
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,1024,3584,0.047075200080871585
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,1024,5120,0.03396053314208984
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,1024,3072,0.01727786660194397
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,1024,3072,0.04199893474578857
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,1024,3584,0.026387200752894087
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,1024,2560,0.014612266421318054
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,1024,2560,0.03747626543045044
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,1024,3072,0.024235733350118003
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,1024,2048,0.012521599729855856
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,1024,2048,0.03344213167826335
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,1024,51200,0.19967786471048993
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,1024,1536,0.010412800312042236
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,1024,1536,0.029912533362706502
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,1024,2560,0.022241065899531044
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,1024,1024,0.008003200093905132
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,1024,1024,0.024714666604995727
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,1024,1024,0.014974932869275412
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,1024,768,0.007129600147406261
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,1024,768,0.022962133089701332
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,1024,2048,0.019057067235310872
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,1024,512,0.005895466605822245
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,1024,512,0.02153279980023702
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,1024,768,0.014453333616256715
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,1024,256,0.004695466657479604
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,1024,256,0.018823466698328652
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,1024,256,0.01453439990679423
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,1024,1536,0.0165503998597463
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,1024,512,0.01479573349157969
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,1024,128,0.01671253244082133
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,1024,64,0.0037994667887687682
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,1024,64,0.017064533631006875
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,1024,32,0.0038218667109807336
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,1024,32,0.01446399986743927
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,1024,128,0.015405866503715514
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,768,65536,0.2540448029836019
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,1024,128,0.004244266450405121
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,768,65536,0.8626442591349284
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,768,51200,0.19585067431132
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,768,51200,0.6813034693400065
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,768,16384,0.06452800035476684
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,768,16384,0.22444799741109214
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,768,16384,0.07562239964803061
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,768,12288,0.04886293411254883
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,1024,65536,0.24691519737243653
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,768,12288,0.17336319287618002
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,768,10240,0.04149653514226277
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,768,10240,0.14243839581807455
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,768,10240,0.05591466824213663
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,768,8192,0.03323199947675069
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,768,8192,0.10847040017445882
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,768,12288,0.06051199833552042
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,768,7168,0.028553599119186403
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,768,7168,0.0915712038675944
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,768,8192,0.045954132080078126
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,768,7168,0.041915734608968094
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,768,6144,0.024514132738113405
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,768,6144,0.07404586474100748
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,768,6144,0.037544532616933184
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,768,5120,0.021356799205144248
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,768,5120,0.059623467922210696
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,768,4096,0.01771519978841146
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,768,4096,0.04802986780802409
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,768,5120,0.034194131692250565
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,768,3584,0.015899733702341715
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,768,3584,0.04514026641845703
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,768,4096,0.029201066493988036
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,768,3072,0.013659733533859252
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,768,3072,0.0395359992980957
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,768,3584,0.027221333980560303
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,768,2560,0.012519466876983642
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,768,2560,0.037113598982493085
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,768,2560,0.0232149342695872
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,768,2048,0.01011306643486023
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,768,51200,0.19928959210713704
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,768,2048,0.03184213240941365
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,768,1536,0.008356266220410665
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,768,3072,0.024800000588099162
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,768,1536,0.02909333308537801
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,768,1024,0.006726400057474773
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,768,1024,0.02483733296394348
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,768,1024,0.016139733791351318
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,768,2048,0.020267732938130698
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,768,768,0.0058773333827654515
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,768,768,0.022706133127212525
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,768,512,0.0050005331635475155
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,768,512,0.021319466829299926
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,768,1536,0.01763199965159098
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,768,768,0.01534293293952942
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,768,256,0.0042698666453361515
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,768,256,0.018874667088190713
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,768,512,0.01569706698258718
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,768,128,0.003435733417669932
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,768,128,0.01699519952138265
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,768,64,0.003403733422358831
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,768,256,0.016617600123087564
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,768,128,0.016565333803494772
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,768,64,0.01567039986451467
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,768,32,0.015598932902018229
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,512,65536,0.17700160344441732
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,512,65536,0.8244053522745768
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,512,51200,0.13800427118937175
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,768,32,0.0034624000390370687
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,512,51200,0.650823465983073
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,512,16384,0.0464906652768453
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,512,16384,0.21565866470336914
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,768,65536,0.246834135055542
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,512,12288,0.03624213139216105
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,512,12288,0.16661760012308757
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,512,16384,0.07373546759287516
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,512,10240,0.030504532655080158
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,512,10240,0.13713173866271972
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,512,12288,0.05867413282394409
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,512,8192,0.023548799753189086
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,512,8192,0.10240426858266194
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,512,10240,0.053990399837493895
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,512,7168,0.019778132438659668
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,512,7168,0.08642986615498861
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,512,8192,0.04276906649271647
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,512,6144,0.01750613252321879
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,512,6144,0.0695360024770101
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,512,7168,0.038734932740529374
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,512,5120,0.014829867084821067
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,512,5120,0.05415893395741781
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,512,6144,0.034543999036153156
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,512,4096,0.012588799993197123
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,512,4096,0.04503893454869588
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,512,5120,0.03206933339436849
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,512,3584,0.01113706628481547
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,512,3584,0.04165759881337484
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,512,4096,0.027037866910298664
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,512,3072,0.01001706620057424
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,512,3072,0.038466131687164305
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,512,3584,0.02524159948031108
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,512,2560,0.008803199728329976
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,512,2560,0.03543893496195476
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,512,3072,0.022670932610829673
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,512,2048,0.007506133119265239
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,512,2048,0.0295360008875529
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,512,2560,0.021078399817148843
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,512,1536,0.006738133231798808
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,512,1536,0.028595199187596638
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,512,2048,0.018266665935516357
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,512,1536,0.015736533204714458
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,512,1024,0.005067733426888784
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,512,1024,0.023336533705393472
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,512,768,0.004683733483155568
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,512,768,0.022500266631444298
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,512,1024,0.014045866330464682
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,512,512,0.003870933254559835
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,512,512,0.020195200045903524
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,512,768,0.013356799880663553
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,512,256,0.0034058667719364167
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,512,256,0.018403200308481853
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,512,512,0.013683199882507324
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,512,128,0.0029845332105954488
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,512,256,0.014084266622861228
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,512,51200,0.19803412755330402
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,512,128,0.01660480002562205
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,512,64,0.0029898665845394133
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,512,64,0.016952532529830932
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,512,32,0.0030421334008375804
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,512,32,0.0151146670182546
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,512,128,0.014514133334159851
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,256,65536,0.13186773459116619
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,256,65536,0.7923231760660807
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,256,51200,0.10388267040252686
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,256,51200,0.6243690490722656
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,256,16384,0.03697173198064168
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,256,16384,0.21067093213399252
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,512,65536,0.24560640652974447
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,256,12288,0.027939200401306152
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,256,12288,0.1609440008799235
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,256,16384,0.0737237294514974
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,256,10240,0.021478400627772013
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,256,10240,0.13288640181223552
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,256,12288,0.05825599829355875
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,256,8192,0.01734613378842672
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,256,8192,0.0978986660639445
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,256,10240,0.05377920071283976
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,256,7168,0.014573867122332254
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,256,7168,0.08118826548258463
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,256,8192,0.04180373350779216
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,256,6144,0.012823466459910074
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,256,6144,0.06395839850107829
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,256,7168,0.03771200180053711
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,256,5120,0.010849066575368245
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,256,6144,0.034135464827219644
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,256,5120,0.050811731815338136
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,256,4096,0.00922986666361491
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,256,4096,0.041536001364390056
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,256,5120,0.031692800919214885
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,256,3584,0.008711466193199157
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,256,4096,0.027167999744415285
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,256,51200,0.19755306243896484
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,256,3072,0.007739733159542084
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,256,3072,0.03672533432642619
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,256,3584,0.024757333596547446
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,256,2560,0.006748799979686737
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,256,2560,0.0329802672068278
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,256,3072,0.022770132621129355
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,256,2048,0.006017066538333893
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,256,3584,0.03917866547902425
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,256,2048,0.02919253309567769
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,256,2560,0.021036799748738608
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,256,1536,0.005515733361244201
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,256,65536,0.24486613273620605
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,256,1536,0.026789333422978717
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,256,1024,0.004315733412901561
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,256,1024,0.02286826570828756
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,256,768,0.0038176000118255613
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,256,2048,0.018179200092951455
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,256,768,0.02204479972521464
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,256,512,0.0034154665966828666
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,256,1536,0.015413332978884378
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,256,512,0.018802134195963542
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,256,768,0.013176533579826354
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,256,256,0.0029685333371162414
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,256,256,0.01776533325513204
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,256,128,0.0025802666942278544
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,256,256,0.013660800457000733
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,256,512,0.013767466942469279
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,256,128,0.015125333269437154
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,256,64,0.0028277332584063213
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,256,128,0.013722667098045349
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,256,32,0.0030389333764712016
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,256,32,0.015066666404406228
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,128,65536,0.09453120231628417
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,256,1024,0.014043733477592468
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,128,65536,0.7830432256062825
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,128,51200,0.07511253356933593
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,256,64,0.01555519998073578
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,128,16384,0.02948586742083232
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,128,51200,0.6161365509033203
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,128,16384,0.20506347020467125
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,128,12288,0.02134400010108948
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,128,12288,0.1568394660949707
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,128,16384,0.07179413636525472
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,128,10240,0.0151146670182546
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,128,12288,0.05706773201624552
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,128,10240,0.129694930712382
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,128,8192,0.012442666292190551
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,128,8192,0.09505919615427652
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,128,10240,0.05245653390884399
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,128,7168,0.010769066214561463
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,128,7168,0.07818453311920166
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,128,8192,0.04030826489130656
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,128,6144,0.009732266267140705
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,128,6144,0.0628981351852417
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,128,7168,0.03720426559448242
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,128,5120,0.00844266712665558
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,128,6144,0.033351465066274004
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,128,51200,0.19360532760620117
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,128,5120,0.04807146787643433
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,128,4096,0.007180800040562947
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,128,4096,0.040666667620340984
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,128,5120,0.0308896005153656
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,128,4096,0.026337067286173504
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,128,3584,0.0067071999112765
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,128,3584,0.03904853264490764
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,128,3072,0.006281599899133046
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,128,3072,0.03514773448308309
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,128,3584,0.02430400053660075
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,128,2560,0.005543466905752818
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,128,2560,0.03265066742897034
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,128,3072,0.021485867102940877
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,128,2048,0.0046965335806210835
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,128,65536,0.24134292602539062
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,128,2048,0.028859732548395793
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,128,1536,0.004335999985535939
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,128,2560,0.01997013290723165
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,128,1536,0.026719999313354493
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,128,1024,0.003499733408292135
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,128,1024,0.023545600970586143
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,128,2048,0.0172106663386027
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,128,768,0.0034080001215140024
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,128,1024,0.013298133015632629
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,128,768,0.021236266692479452
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,128,768,0.012054399649302164
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,128,512,0.0030239999294281008
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,128,512,0.019019732872645058
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,128,256,0.0026346666117509207
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,128,256,0.01808213392893473
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,128,512,0.012889599800109864
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,128,128,0.0026677332818508146
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,128,256,0.012893866499265036
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,128,128,0.016420267025629678
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,128,128,0.01286186675230662
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,128,64,0.0023605334262053173
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,128,32,0.002549333373705546
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,128,64,0.017164800564448038
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4352,128,1536,0.01495573321978251
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,128,32,0.014840533336003622
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,64,65536,0.09854613145192465
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,64,51200,0.07297279834747314
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,64,16384,0.0276256004969279
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,64,12288,0.02062506675720215
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,64,16384,0.2047797362009684
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,64,12288,0.15746560096740722
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,64,10240,0.12957013448079427
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,64,8192,0.011664000153541566
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,64,51200,0.6181845347086589
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,64,7168,0.010488532980283101
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,64,65536,0.7856885274251302
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,64,8192,0.09504746596018473
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,64,6144,0.00927786628405253
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,64,7168,0.07707839806874593
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,64,6144,0.05988053480784098
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,64,5120,0.008051200211048127
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,64,4096,0.006725333134333293
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,64,3584,0.006308266520500183
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,64,5120,0.047577599684397384
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,64,3072,0.005896533528963724
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,64,3584,0.03851413329442342
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,64,4096,0.04122560024261475
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,64,2560,0.0051242664456367494
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,64,2048,0.004622933268547058
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,64,3072,0.03477866649627685
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,64,1536,0.004147200038035711
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,64,2048,0.029033599297205608
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,64,2560,0.03328106602032979
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,64,1024,0.0034517332911491393
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,64,1536,0.02515733242034912
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,64,768,0.0030026666820049284
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,64,512,0.0030261332790056865
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,64,768,0.020678400993347168
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,64,1024,0.02269759972890218
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,64,256,0.0025898667673269907
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,64,10240,0.01423466702302297
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,64,512,0.01999680002530416
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,64,128,0.0022122666239738466
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,64,64,0.0022122666239738466
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,64,32,0.002218666672706604
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,64,128,0.015170133113861084
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,64,256,0.01730560064315796
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,64,64,0.014867200454076131
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,64,32,0.015889066457748412
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,32,65536,0.09315626621246338
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,32,51200,0.07351893583933512
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,32,16384,0.025731199979782106
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,32,12288,0.018959999084472656
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,32,16384,0.20352959632873535
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,32,10240,0.011166933178901672
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,32,12288,0.15675840377807618
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,32,8192,0.009649067123730978
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,32,10240,0.13065813382466634
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,32,51200,0.6122869491577149
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,32,8192,0.09647146860758463
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,32,6144,0.008237866560618083
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,32,65536,0.7793333053588867
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,32,7168,0.07821333408355713
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,32,5120,0.011172266801198323
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,32,6144,0.06082773208618164
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,32,5120,0.04948800007502238
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,32,4096,0.009408000111579894
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,32,4096,0.03983466625213623
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,32,3584,0.008363733688990276
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,32,3072,0.007559466858704884
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,32,3584,0.038179198900858566
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,32,3072,0.03498560190200806
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,32,2560,0.006899199883143107
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,32,2048,0.004645333190759023
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,32,2560,0.032902399698893234
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,32,1536,0.0038965334494908653
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,32,1536,0.026437334219614667
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,32,2048,0.028626133998235066
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,32,1024,0.0034272000193595886
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,32,768,0.0030293333033720653
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,32,1024,0.02302293380101522
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,32,768,0.020971733331680297
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,32,512,0.003018666555484136
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,32,512,0.019381332397460937
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,32,256,0.0027615999182065325
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,32,7168,0.00905386706193288
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,32,128,0.0023210667073726653
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,32,256,0.0172106663386027
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,32,64,0.002216533323129018
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,32,128,0.015387733777364096
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4352,32,32,0.002234666546185811
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,32,32,0.014265599846839904
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4352,32,64,0.016284799575805663
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,65536,16384,5.145493570963541
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,65536,51200,8.842836507161458
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,65536,16384,2.816323089599609
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,65536,12288,4.012411753336588
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,65536,12288,2.0783060709635417
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,65536,51200,17.176407877604166
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,65536,10240,3.458406321207682
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,65536,10240,1.7248266855875651
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,65536,10240,0.8983434677124024
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,65536,8192,2.8176233927408854
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,65536,8192,1.3486347198486328
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,65536,12288,1.0362538655598958
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,65536,7168,2.432423400878906
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,65536,7168,1.174525833129883
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,65536,16384,1.4218858083089194
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,65536,6144,2.001089096069336
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,65536,6144,1.013372802734375
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,65536,7168,0.687552007039388
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,65536,5120,1.576592000325521
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,65536,5120,0.8838848114013672
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,65536,8192,0.7374303817749024
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,65536,4096,1.2538805643717448
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,65536,4096,0.6877258936564128
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,65536,6144,0.5895509084065755
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,65536,3584,1.0987840016682944
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,65536,3584,0.5990037282307943
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,65536,4096,0.47573334376017257
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,65536,3072,0.8959957122802734
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,65536,3072,0.5216074625651042
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,65536,5120,0.5362869262695312
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,65536,2560,0.7558624267578125
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,65536,2560,0.4665450731913249
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,65536,3584,0.42946348190307615
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,65536,2048,0.5752533594767253
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,65536,2560,0.36369279225667317
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,65536,2048,0.3430879910786947
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,65536,1536,0.4829610824584961
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,65536,1536,0.2865109443664551
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,65536,3072,0.3782069206237793
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,65536,1024,0.3024181365966797
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,65536,1024,0.19370452562967938
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,65536,1024,0.2676543871561686
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,65536,768,0.2418432076772054
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,65536,768,0.18276373545328778
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,65536,1536,0.27435413996378577
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,65536,512,0.17131627400716146
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,65536,512,0.1458954652150472
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,65536,512,0.25060799916585286
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,65536,256,0.12646719614664714
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,65536,256,0.12218773365020752
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,65536,768,0.2558133284250895
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,65536,128,0.10180586973826092
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,65536,128,0.1067402680714925
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,65536,256,0.24686826070149742
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,65536,128,0.25847360293070476
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,65536,64,0.0907146692276001
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,65536,32,0.09598293304443359
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,65536,32,0.1107968012491862
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,65536,2048,0.3221205393473307
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,65536,64,0.11088106632232667
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,51200,65536,8.984803263346354
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,51200,65536,17.214841715494792
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,51200,51200,12.962619018554687
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,51200,51200,6.867210896809896
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,51200,16384,4.135213979085287
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,51200,16384,2.268821207682292
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,51200,16384,1.0859444936116538
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,51200,12288,2.9502708435058596
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,51200,12288,1.618994140625
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,51200,12288,0.8119125366210938
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,51200,10240,2.5779563903808596
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,51200,10240,1.3047167460123696
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,51200,10240,0.7167989095052083
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,51200,8192,1.976909891764323
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,51200,8192,1.0251679738362631
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,65536,51200,4.664193216959635
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,51200,7168,1.5818986256917318
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,51200,7168,0.8954559961954752
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,51200,7168,0.539411226908366
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,51200,6144,1.4028394063313803
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,51200,6144,0.7781887690226237
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,51200,8192,0.5830538431803386
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,51200,5120,1.1792799631754556
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,51200,5120,0.6614762624104817
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,51200,5120,0.41822293599446614
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,51200,4096,0.8744757334391275
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,51200,4096,0.5181290626525878
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,51200,6144,0.46101334889729817
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,51200,3584,0.7521024068196615
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,51200,3584,0.47364266713460285
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,51200,3584,0.3369333267211914
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,51200,3072,0.6476021448771159
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,51200,3072,0.41649920145670577
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,51200,4096,0.3711455980936686
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,51200,2560,0.5568639755249023
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,51200,2560,0.3374538739522298
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,51200,51200,3.6726720174153646
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,51200,2048,0.4607594807942708
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,51200,2048,0.27728853225708006
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,51200,2560,0.2820202509562174
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,51200,1536,0.36087465286254883
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,51200,1536,0.23262826601664224
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,51200,2048,0.2529184023539225
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,51200,1024,0.24845760663350425
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,51200,1024,0.1580341339111328
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,51200,3072,0.2982794761657715
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,51200,768,0.18093120257059733
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,51200,768,0.13131519953409831
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,51200,768,0.2025226593017578
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,51200,512,0.1328384002049764
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,51200,512,0.11324053605397541
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,51200,1024,0.21625706354777016
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,51200,256,0.09463146527608236
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,51200,256,0.0964128017425537
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,51200,1536,0.22024854024251303
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,51200,512,0.20258879661560059
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,51200,128,0.07977813084920247
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,51200,256,0.201802666982015
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,51200,128,0.08762986660003662
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,51200,64,0.07092373371124268
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,51200,64,0.09040106932322184
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,51200,32,0.07526933352152507
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,51200,32,0.09223786989847818
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,16384,65536,3.0517247517903643
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,51200,128,0.2041536013285319
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,16384,65536,5.04806162516276
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,16384,51200,2.412505594889323
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,16384,51200,3.9524767557779947
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,16384,16384,1.1937216440836589
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,16384,16384,0.8062218983968099
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,16384,16384,0.3851253191630045
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,16384,12288,0.8854186375935873
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,16384,12288,0.5951168060302734
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,16384,12288,0.2919861475626627
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,16384,10240,0.7252757390340169
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,16384,10240,0.5100778579711914
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,16384,10240,0.259173329671224
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,16384,8192,0.5739840189615886
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,16384,8192,0.3849034627278646
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,16384,8192,0.21274986267089843
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,16384,7168,0.5086965243021647
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,16384,7168,0.3568458557128906
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,16384,51200,1.1812213897705077
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,16384,6144,0.44669332504272463
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,16384,6144,0.30755093892415364
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,16384,7168,0.1966826597849528
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,16384,5120,0.36283092498779296
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,16384,5120,0.2522496064503988
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,16384,5120,0.15554879506429037
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,16384,4096,0.29328212738037107
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,16384,4096,0.20429546038309732
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,16384,6144,0.17142613728841144
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,16384,3584,0.26080640157063806
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,16384,3584,0.18477439880371094
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,16384,3584,0.12371093432108562
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,16384,3072,0.2362880071004232
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,16384,3072,0.1596384048461914
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,16384,4096,0.13637545903523762
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,16384,2560,0.19826666514078778
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,16384,2560,0.13462613423665365
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,16384,2560,0.10402026971181233
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,16384,2048,0.15971840222676595
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,16384,2048,0.10877866744995117
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,16384,3072,0.10973119735717773
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,16384,1536,0.12244799931844075
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,16384,1536,0.09096213181813559
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,16384,1536,0.08206400076548258
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,16384,1024,0.08420266310373942
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,16384,1024,0.06829866568247477
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,16384,2048,0.09232107003529867
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,16384,768,0.0653877337773641
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,16384,768,0.06362133423487346
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,16384,1024,0.07749120394388834
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,16384,512,0.049415465195973715
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,16384,512,0.05357120037078857
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,16384,768,0.07046293417612712
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,16384,256,0.03373226722081502
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,16384,256,0.045049599806467694
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,16384,256,0.06283200184504191
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,16384,128,0.02807253400484721
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,16384,128,0.03939839998881022
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,16384,128,0.06348053216934205
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,16384,64,0.02442880074183146
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,16384,64,0.040278398990631105
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,16384,32,0.025818665822347004
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,16384,512,0.06803840001424154
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,16384,32,0.039046398798624676
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,12288,65536,2.4569300333658854
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,12288,65536,3.788760630289713
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,12288,51200,2.957898712158203
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,12288,51200,1.9328768412272137
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,16384,65536,1.5251764933268228
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,12288,16384,0.8333354949951172
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,12288,16384,0.6119829177856445
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,51200,65536,4.643304443359375
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,12288,12288,0.6378165562947591
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,12288,12288,0.46505387624104816
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,12288,12288,0.23624746004740396
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,12288,10240,0.5304341316223145
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,12288,10240,0.3931893348693848
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,12288,16384,0.3120661417643229
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,12288,8192,0.4231253306070964
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,12288,8192,0.30919040044148766
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,12288,8192,0.17224853833516438
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,12288,7168,0.37305707931518556
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,12288,7168,0.27570133209228515
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,12288,10240,0.2109440008799235
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,12288,6144,0.3258517265319824
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,12288,6144,0.22683520317077638
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,12288,7168,0.15954772631327313
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,12288,5120,0.26479786237080893
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,12288,5120,0.1935413360595703
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,12288,6144,0.13892693519592286
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,12288,4096,0.21470719973246255
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,12288,4096,0.155621337890625
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,12288,4096,0.11036372979482015
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,12288,3584,0.19705066680908204
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,12288,3584,0.13956906000773112
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,12288,5120,0.12571199735005695
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,12288,3072,0.1697450637817383
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,12288,3072,0.11989546616872152
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,12288,3072,0.08796906471252441
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,12288,2560,0.14179520606994628
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,12288,2560,0.10801920096079509
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,12288,3584,0.09849920272827148
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,12288,2048,0.1105130672454834
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,12288,2048,0.090338134765625
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,12288,2048,0.07621546586354574
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,12288,1536,0.08483839829762777
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,12288,1536,0.07225706577301025
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,12288,2560,0.0844223976135254
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,12288,1024,0.05825493335723877
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,12288,1024,0.055438931783040366
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,12288,1024,0.06667733192443848
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,12288,768,0.04571733474731445
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,12288,768,0.05201280117034912
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,12288,1536,0.06880213419596354
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,12288,512,0.03486080169677734
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,12288,512,0.04548799991607666
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,12288,512,0.06218346754709879
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,12288,256,0.0247925341129303
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,12288,256,0.03789013226826986
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,12288,768,0.06163520018259684
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,12288,128,0.02188053329785665
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,12288,128,0.03488959868748982
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,12288,128,0.06157653331756592
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,12288,64,0.01904639999071757
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,12288,64,0.03371520042419433
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,12288,32,0.0200000007947286
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,12288,32,0.034686934947967527
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,12288,256,0.06204373439153036
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,10240,65536,3.1394304911295574
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,10240,65536,2.1546709696451822
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,10240,51200,2.443677775065104
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,10240,51200,1.6703520456949872
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,12288,51200,0.9545397440592447
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,10240,16384,0.7134634653727214
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,10240,16384,0.5486015955607096
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,10240,16384,0.2664927959442139
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,10240,12288,0.5407050450642903
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,10240,12288,0.4138250668843587
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,12288,65536,1.2074986775716146
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,10240,10240,0.46136960983276365
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,10240,10240,0.35248851776123047
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,10240,12288,0.2028127988179525
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,10240,8192,0.3726997375488281
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,10240,8192,0.2814698537190755
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,10240,10240,0.18067092895507814
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,10240,7168,0.32114985783894856
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,10240,7168,0.2523530642191569
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,10240,7168,0.13713280359903973
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,10240,6144,0.27420161565144857
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,10240,6144,0.20140159924825035
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,10240,8192,0.14768959681193033
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,10240,5120,0.2268298625946045
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,10240,5120,0.16987733840942382
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,10240,5120,0.10880959828694661
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,10240,4096,0.1867967923482259
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,10240,4096,0.137007999420166
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,10240,6144,0.11932799816131592
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,10240,3584,0.1591978708902995
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,10240,3584,0.12137386798858643
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,10240,3584,0.0851039965947469
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,10240,3072,0.13597013155619303
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,10240,3072,0.10350933074951171
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,10240,4096,0.09448746840159097
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,10240,2560,0.11828906536102295
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,10240,2560,0.09293333689371744
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,10240,2560,0.07435839970906576
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,10240,2048,0.09346986611684163
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,10240,2048,0.07875520388285319
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,10240,3072,0.0764842669169108
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,10240,1536,0.07096213499704997
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,10240,1536,0.06635306676228842
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,10240,51200,0.8024170557657877
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,10240,1024,0.049821865558624265
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,10240,1536,0.061495467027028405
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,10240,1024,0.052741332848866784
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,10240,768,0.038355199495951335
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,10240,768,0.04727253516515096
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,10240,2048,0.06787839730580648
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,10240,512,0.029064534107844035
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,10240,512,0.04062933524449666
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,10240,512,0.05510400136311849
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,10240,256,0.020906666914621987
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,10240,256,0.03504000107447307
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,10240,768,0.05487253268559774
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,10240,128,0.01849386692047119
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,10240,128,0.031267199913660684
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,10240,256,0.054714667797088626
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,10240,64,0.01614400049050649
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,10240,64,0.030525867144266767
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,10240,32,0.016926934321721397
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,10240,32,0.030065067609151203
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,10240,1024,0.05869013468424479
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,10240,128,0.05497599840164184
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,8192,65536,1.8467540740966797
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,8192,65536,2.503509267171224
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,8192,51200,1.4636629740397136
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,8192,51200,1.9326773325602216
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,8192,16384,0.5808554967244466
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,8192,16384,0.47214291890462234
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,8192,16384,0.24028587341308594
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,8192,12288,0.42995732625325517
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,8192,12288,0.36044371922810875
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,10240,65536,1.0138421376546225
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,8192,10240,0.355018679300944
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,8192,10240,0.3030794779459635
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,8192,12288,0.17439573605855305
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,8192,8192,0.2885525385538737
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,8192,8192,0.2312544027964274
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,8192,10240,0.15062185923258464
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,8192,7168,0.2501599947611491
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,8192,7168,0.2007157325744629
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,8192,8192,0.12179839611053467
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,8192,6144,0.21942720413208008
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,8192,6144,0.17654293378194172
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,8192,7168,0.11367039680480957
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,8192,5120,0.18144319852193197
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,8192,5120,0.14480214118957518
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,8192,5120,0.09143359661102295
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,8192,4096,0.1466101328531901
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,8192,4096,0.11895679632822673
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,8192,6144,0.09953920046488443
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,8192,3584,0.13072213331858318
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,8192,3584,0.10748480161031086
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,8192,3584,0.06985387007395426
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,8192,4096,0.07851093610127767
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,8192,3072,0.11560959815979004
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,8192,3072,0.093995730082194
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,8192,2560,0.0947498639424642
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,8192,2560,0.08512106736501059
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,8192,2560,0.058482134342193605
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,8192,2048,0.07728319962819417
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,8192,2048,0.06980160077412924
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,8192,3072,0.06158506472905477
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,8192,1536,0.060432000954945886
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,8192,1536,0.06004800001780192
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,8192,1536,0.04352746804555257
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,8192,1024,0.04268800020217896
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,8192,1024,0.04563200076421102
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,8192,2048,0.05054080088933309
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,8192,768,0.03341013193130493
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,8192,768,0.04328639904658
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,8192,768,0.03635306755701701
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,8192,512,0.02490239938100179
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,8192,512,0.03688426812489827
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,8192,1024,0.0406826655069987
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,8192,256,0.017819732427597046
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,8192,256,0.030869332949320476
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,8192,256,0.036432000001271565
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,8192,128,0.015340800086657206
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,8192,128,0.027460267146428425
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,8192,512,0.03691093524297078
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,8192,64,0.01330880026022593
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,8192,128,0.03702720006306966
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,8192,64,0.026900267601013182
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,8192,32,0.01370133360226949
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,8192,32,0.026770132780075073
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,7168,65536,2.207906087239583
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,7168,65536,1.7168224334716797
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,7168,51200,1.732266616821289
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,7168,51200,1.3399232228597007
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,8192,51200,0.6893130620320638
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,7168,16384,0.5061408042907715
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,7168,16384,0.4415562629699707
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,7168,16384,0.22869866689046225
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,7168,12288,0.38101228078206384
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,7168,12288,0.33200213114420574
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,8192,65536,0.8740533192952474
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,7168,10240,0.3166709264119466
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,7168,10240,0.2762869199117025
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,7168,12288,0.16528746287027996
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,7168,8192,0.27279678980509436
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,7168,8192,0.22656426429748536
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,7168,10240,0.14401599566141765
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,7168,7168,0.23340266545613608
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,7168,7168,0.1933695952097575
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,7168,8192,0.1163061300913493
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,7168,6144,0.20038720766703286
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,7168,6144,0.16437759399414062
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,7168,7168,0.10812479654947918
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,7168,5120,0.16320746739705402
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,7168,5120,0.13545494079589843
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,7168,6144,0.0948746681213379
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,7168,4096,0.13249173164367675
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,7168,4096,0.11241280237833659
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,7168,5120,0.08643306891123453
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,7168,3584,0.11923306783040363
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,7168,3584,0.10193386872609456
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,7168,3584,0.06609599987665812
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,7168,3072,0.09825066725413004
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,7168,3072,0.0902026653289795
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,7168,4096,0.07401813666025797
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,7168,2560,0.08424853483835856
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,7168,2560,0.08087039788564046
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,7168,2560,0.05617813269297282
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,7168,2048,0.06829013029734293
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,7168,2048,0.06477013429005941
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,7168,3072,0.05885119835535685
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,7168,1536,0.05258133411407471
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,7168,1536,0.05439893404642741
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,7168,1536,0.043109333515167235
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,7168,1024,0.035877335071563723
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,7168,1024,0.043185067176818845
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,7168,2048,0.0488480011622111
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,7168,768,0.028621866305669146
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,7168,768,0.04100373188654582
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,7168,768,0.03764479955037435
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,7168,512,0.02205866575241089
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,7168,512,0.03509013255437215
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,7168,1024,0.04110613266626994
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,7168,256,0.016396799683570863
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,7168,256,0.02903146743774414
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,7168,256,0.037588266531626384
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,7168,128,0.013225600123405457
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,7168,128,0.024909865856170655
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,7168,512,0.03784000078837077
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,7168,64,0.011203199625015259
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,7168,51200,0.6675807952880859
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,7168,64,0.024936532974243163
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,7168,32,0.013331199685732523
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,7168,32,0.02482666571935018
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,7168,128,0.03771626551946004
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,6144,65536,1.5692213694254558
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,6144,65536,1.840133285522461
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,6144,51200,1.4449930826822917
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,6144,51200,1.221410115559896
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,6144,16384,0.4259498596191406
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,6144,16384,0.43233706156412766
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,6144,16384,0.19944747289021808
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,6144,12288,0.3220117251078288
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,6144,12288,0.30117546717325844
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,7168,65536,0.841428248087565
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,6144,10240,0.2721738815307617
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,6144,10240,0.2553781350453695
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,6144,12288,0.1473024050394694
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,6144,8192,0.21868693033854164
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,6144,8192,0.19383254051208496
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,6144,10240,0.13068799972534179
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,6144,7168,0.18868692715962726
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,6144,7168,0.16875306765238446
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,6144,8192,0.10632106463114421
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,6144,6144,0.16597867012023926
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,6144,6144,0.14618453979492188
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,6144,7168,0.09945173263549804
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,6144,5120,0.13703786532084147
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,6144,5120,0.11836160024007161
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,6144,5120,0.0801749308904012
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,6144,4096,0.11086400349934895
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,6144,4096,0.09895679950714112
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,6144,6144,0.08740266958872477
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,6144,3584,0.09881707032521567
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,6144,3584,0.09060693581899007
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,6144,4096,0.06799573103586833
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,6144,3584,0.06125119924545288
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,6144,3072,0.08540480136871338
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,6144,3072,0.07793707052866618
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,6144,2560,0.06889386971791586
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,6144,2560,0.07030826409657796
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,6144,3072,0.05395093361536661
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,6144,2048,0.0568618655204773
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,6144,2560,0.05189973513285319
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,6144,2048,0.05731840133666992
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,6144,1536,0.04337600072224935
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,6144,1536,0.04738560120264689
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,6144,1536,0.04110506772994995
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,6144,1024,0.029709867636362713
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,6144,1024,0.03698773384094238
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,6144,2048,0.04606933196385701
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,6144,768,0.023060266176859537
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,6144,768,0.0353653351465861
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,6144,1024,0.038713598251342775
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,6144,768,0.03582186698913574
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,6144,512,0.01739413340886434
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,6144,512,0.03099946578343709
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,6144,256,0.013226667046546936
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,6144,256,0.027432533105214436
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,6144,256,0.03659520149230957
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,6144,128,0.01320746640364329
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,6144,128,0.025092266003290814
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,6144,512,0.03662399848302205
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,6144,64,0.011947733163833619
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,6144,64,0.024897066752115886
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,6144,128,0.036593067646026614
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,6144,51200,0.5881685256958008
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,6144,32,0.02508373260498047
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,6144,32,0.01227946678797404
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,5120,65536,1.5282410939534505
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,5120,65536,1.421393076578776
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,5120,51200,1.2301770528157552
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,5120,51200,1.1207776387532553
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,5120,16384,0.361521053314209
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,5120,16384,0.3757802645365397
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,5120,16384,0.1720138708750407
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,5120,12288,0.27104107538859046
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,5120,12288,0.2780373255411784
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,6144,65536,0.7409791946411133
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,5120,10240,0.24669547080993653
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,5120,10240,0.23281173706054686
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,5120,12288,0.12957013448079427
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,5120,8192,0.198088534673055
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,5120,8192,0.18000106811523436
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,5120,10240,0.11687359809875489
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,5120,7168,0.17498879432678222
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,5120,7168,0.15562559763590494
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,5120,8192,0.09550399780273437
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,5120,6144,0.147598934173584
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,5120,6144,0.1339157263437907
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,5120,7168,0.08998719851175943
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,5120,5120,0.1220479965209961
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,5120,5120,0.11479039986928304
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,5120,6144,0.07979199886322022
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,5120,4096,0.09667306741078695
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,5120,4096,0.0957909345626831
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,5120,5120,0.07274666627248129
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,5120,3584,0.08565333684285482
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,5120,3584,0.08807573318481446
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,5120,4096,0.061035732428232826
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,5120,3072,0.07317972977956136
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,5120,3072,0.07596159776051839
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,5120,3072,0.050280535221099855
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,5120,2560,0.06187839905420939
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,5120,2560,0.06701013247172037
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,5120,3584,0.05616533358891805
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,5120,2048,0.050918400287628174
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,5120,2048,0.05287466843922933
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,5120,2048,0.04367680152257283
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,5120,1536,0.03968853155771891
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,5120,1536,0.043911465009053546
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,5120,2560,0.048604798316955564
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,5120,1024,0.025051732858022053
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,5120,1024,0.035181868076324466
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,5120,51200,0.506280517578125
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,5120,768,0.02030613422393799
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,5120,768,0.03304746747016907
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,5120,1024,0.03708373308181763
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,5120,512,0.015079466501871744
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,5120,512,0.029819732904434203
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,5120,768,0.035225598017374675
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,5120,512,0.03583679993947347
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,5120,256,0.011588266491889954
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,5120,256,0.02488320072491964
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,5120,128,0.009270399808883667
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,5120,128,0.021924267212549843
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,5120,128,0.03540586630503337
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,5120,64,0.00843519965807597
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,5120,1536,0.039027198155721025
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,5120,64,0.022734934091567995
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,5120,32,0.009675733248392741
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,5120,32,0.022753065824508666
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,4096,65536,1.1622506459554036
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,5120,256,0.0353493332862854
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,4096,65536,1.256218719482422
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,4096,51200,0.9114453633626303
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,4096,51200,0.9851733525594076
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,4096,16384,0.2889888127644857
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,4096,16384,0.3288021405537923
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,4096,16384,0.14342613220214845
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,4096,12288,0.21784960428873695
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,4096,12288,0.24804587364196778
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,5120,65536,0.6373013178507487
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,4096,10240,0.17959146499633788
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,4096,10240,0.20292800267537436
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,4096,12288,0.11057493686676026
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,4096,8192,0.14732054074605305
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,4096,8192,0.1586901346842448
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,4096,10240,0.09913600285847982
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,4096,7168,0.12760213216145833
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,4096,7168,0.1367861270904541
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,4096,8192,0.08131946722666422
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,4096,6144,0.11174186865488689
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,4096,6144,0.12003520329793294
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,4096,7168,0.0767146666844686
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,4096,5120,0.09260799884796142
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,4096,5120,0.10265279610951741
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,4096,6144,0.0687882661819458
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,4096,4096,0.07595840295155844
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,4096,4096,0.08198293050130209
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,4096,5120,0.06148906548817953
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,4096,3584,0.0660373330116272
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,4096,3584,0.07436479727427164
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,4096,3584,0.045765332380930585
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,4096,3072,0.05726933479309082
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,4096,3072,0.06357226769129434
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,4096,4096,0.05168746709823609
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,4096,2560,0.04938666820526123
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,4096,2560,0.05672213236490885
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,4096,2560,0.038152531782786055
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,4096,2048,0.03883093198140462
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,4096,2048,0.04695039987564087
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,4096,3072,0.04043519894282023
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,4096,1536,0.029516800244649248
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,4096,1536,0.039390933513641355
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,4096,1536,0.028486400842666626
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,4096,1024,0.020629332462946574
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,4096,1024,0.03198506633440654
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,4096,2048,0.03304746747016907
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,4096,768,0.016481066743532814
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,4096,768,0.031112533807754517
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,4096,51200,0.4213290532430013
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,4096,512,0.012050132950146992
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,4096,512,0.026973867416381837
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,4096,1024,0.026804266373316447
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,4096,256,0.009585066636403402
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,4096,256,0.02449173331260681
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,4096,512,0.02453546722730001
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,4096,128,0.008761599659919739
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,4096,128,0.021345067024230956
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,4096,768,0.024272000789642333
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,4096,64,0.007540266712506611
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,4096,64,0.020786132415135702
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,4096,32,0.007994666695594788
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,4096,256,0.024266666173934935
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,4096,128,0.02478933334350586
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,3584,65536,1.0178570429484048
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,3584,65536,1.1816715240478515
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,4096,32,0.021988266706466676
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,3584,51200,0.7495381037394206
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,3584,51200,0.9381077448527018
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,3584,16384,0.2493216037750244
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,3584,16384,0.3137184143066406
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,3584,16384,0.14068800608317059
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,3584,12288,0.1896949291229248
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,3584,12288,0.23362879753112792
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,4096,65536,0.5309354782104492
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,3584,10240,0.15605546633402506
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,3584,10240,0.18987199465433757
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,3584,12288,0.10757439931233723
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,3584,8192,0.1283509333928426
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,3584,8192,0.14880107243855795
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,3584,10240,0.09715946515401205
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,3584,7168,0.11056319872538249
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,3584,7168,0.12885119915008544
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,3584,8192,0.0790496031443278
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,3584,6144,0.09731733004252116
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,3584,6144,0.11239893436431884
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,3584,7168,0.07410986423492431
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,3584,5120,0.0801472028096517
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,3584,5120,0.09610986709594727
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,3584,6144,0.06456640164057413
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,3584,4096,0.06607573429743449
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,3584,4096,0.0810144027074178
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,3584,5120,0.05742079814275106
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,3584,3584,0.05778666734695435
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,3584,3584,0.07234026590983073
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,3584,4096,0.048681600888570147
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,3584,3072,0.05076800187428793
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,3584,3072,0.061393066247304284
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,3584,3584,0.044302932421366376
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,3584,2560,0.042827733357747394
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,3584,2560,0.0527018666267395
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,3584,2560,0.03770346641540527
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,3584,2048,0.035743999481201175
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,3584,2048,0.044012800852457686
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,3584,3072,0.03934613466262817
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,3584,1536,0.028410667181015016
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,3584,1536,0.03967039982477824
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,3584,2048,0.033190399408340454
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,3584,1024,0.018583466609319053
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,3584,1024,0.032363732655843094
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,3584,51200,0.4132053375244141
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,3584,768,0.014818132917086283
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,3584,1536,0.028806400299072266
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,3584,768,0.03072426716486613
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,3584,512,0.011731200416882833
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,3584,512,0.02667306661605835
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,3584,1024,0.02706986665725708
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,3584,256,0.009180800120035807
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,3584,256,0.022718934218088786
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,3584,512,0.029348266124725342
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,3584,128,0.007574399809042613
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,3584,128,0.02082879940668742
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,3584,256,0.029314132531483968
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,3584,64,0.007122133175532024
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,3584,128,0.028702932596206664
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,3584,64,0.02081813414891561
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,3584,32,0.007530666887760162
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,3584,32,0.019501866896947224
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,3584,768,0.02881493369738261
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,3072,65536,0.8661578496297201
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,3072,65536,1.1451124827067056
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,3072,51200,0.7205343882242838
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,3072,51200,0.8977855682373047
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,3072,16384,0.2248074690500895
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,3072,16384,0.2955039978027344
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,3072,16384,0.12090773582458496
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,3072,12288,0.17076373100280762
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,3072,12288,0.22257599830627442
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,3584,65536,0.5239455858866374
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,3072,10240,0.14752000172932941
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,3072,10240,0.1816757361094157
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,3072,12288,0.0940831979115804
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,3072,8192,0.12035733064015705
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,3072,8192,0.14489919344584148
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,3072,10240,0.08584960301717123
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,3072,7168,0.10420587062835693
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,3072,7168,0.1235584020614624
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,3072,8192,0.07102399667104085
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,3072,6144,0.08987413247426351
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,3072,6144,0.10595413049062093
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,3072,7168,0.06671146551767984
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,3072,5120,0.07274453639984131
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,3072,5120,0.08849279880523682
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,3072,5120,0.052952531973520914
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,3072,4096,0.05777279933293661
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,3072,4096,0.07290773391723633
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,3072,6144,0.05944213469823202
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,3072,3584,0.053692801793416345
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,3072,3584,0.06561599969863892
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,3072,3584,0.041221332550048825
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,3072,3072,0.04448426564534505
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,3072,3072,0.0555189331372579
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,3072,4096,0.04492053190867106
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,3072,2560,0.036994131406148274
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,3072,2560,0.049428268273671465
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,3072,2560,0.0346837321917216
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,3072,2048,0.0305461327234904
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,3072,3072,0.03667946656545003
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,3072,2048,0.04121599992116292
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,3072,1536,0.024795732895533242
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,3072,1536,0.03704426685969035
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,3072,1536,0.027832533915837603
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,3072,1024,0.01699840029080709
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,3072,1024,0.02930026650428772
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,3072,2048,0.031648000081380204
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,3072,768,0.013963733116785684
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,3072,768,0.02794346610705058
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,3072,1024,0.025883734226226807
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,3072,512,0.010397866368293762
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,3072,512,0.024961066246032716
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,3072,768,0.028406399488449096
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,3072,256,0.008360532919565837
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,3072,256,0.02373440066973368
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,3072,512,0.02876373330752055
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,3072,256,0.02880106568336487
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,3072,128,0.0067104001839955645
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,3072,128,0.019783467054367065
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,3072,64,0.0059114664793014525
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,3072,64,0.020261333386103312
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,3072,128,0.0262773334980011
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,3072,32,0.0074986666440963745
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,3072,32,0.01928000052769979
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,2560,65536,0.7673898696899414
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,2560,65536,1.0679220835367838
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,2560,51200,0.5783807754516601
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,2560,51200,0.837770652770996
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,3072,51200,0.3436095873514811
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,2560,16384,0.1860607941945394
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,2560,16384,0.2745269457499186
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,2560,16384,0.11974399884541828
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,2560,12288,0.14074880282084148
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,2560,12288,0.20768960316975912
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,3072,65536,0.43113279342651367
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,2560,10240,0.11977173487345379
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,2560,10240,0.16617493629455565
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,2560,12288,0.09366719722747803
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,2560,8192,0.0944543997446696
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,2560,8192,0.12779200077056885
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,2560,10240,0.08448533217112222
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,2560,7168,0.08331733544667562
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,2560,7168,0.11167466640472412
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,2560,8192,0.06935359636942545
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,2560,6144,0.06944426695505777
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,2560,6144,0.09618986447652181
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,2560,7168,0.06480426788330078
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,2560,5120,0.05869439840316772
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,2560,5120,0.08085119724273682
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,2560,6144,0.057733333110809325
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,2560,4096,0.046263468265533444
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,2560,4096,0.06470506588617961
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,2560,5120,0.051235198974609375
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,2560,3584,0.04087786674499512
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,2560,3584,0.058847999572753905
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,2560,4096,0.04420373439788818
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,2560,3072,0.03471039930979411
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,2560,3072,0.04953066507975261
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,2560,3584,0.039494399229685465
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,2560,2560,0.02962239980697632
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,2560,2560,0.04530666669209798
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,2560,3072,0.0353877345720927
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,2560,2048,0.02404266595840454
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,2560,2048,0.03900479873021444
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,2560,2560,0.03325973351796468
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,2560,1536,0.01918720006942749
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,2560,1536,0.03412160078684489
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,2560,2048,0.029922133684158324
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,2560,1024,0.013274666666984559
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,2560,1024,0.02749546567598979
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,2560,51200,0.344702943166097
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,2560,768,0.010898133118947346
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,2560,768,0.026795732975006103
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,2560,1536,0.02579093376795451
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,2560,512,0.00869653324286143
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,2560,512,0.025146667162577314
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,2560,1024,0.02434773246447245
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,2560,256,0.0066890666882197065
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,2560,256,0.02116586764653524
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,2560,256,0.02345386743545532
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,2560,768,0.022718934218088786
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,2560,128,0.005946666498978933
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,2560,128,0.018971733252207437
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,2560,64,0.005470933516820272
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,2560,512,0.02281173268953959
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,2560,64,0.0172160009543101
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,2560,128,0.02297919988632202
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,2560,32,0.0058773333827654515
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,2048,65536,0.5464960098266601
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,2048,65536,0.9705408096313477
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,2560,32,0.01773759921391805
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,2048,51200,0.4278143882751465
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,2048,51200,0.7624319712320964
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,2048,16384,0.14106879234313965
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,2048,16384,0.2513397375742594
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,2048,16384,0.09665706952412924
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,2048,12288,0.10714133580525716
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,2048,12288,0.19082345962524414
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,2560,65536,0.43344958623250324
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,2048,10240,0.09047253131866455
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,2048,10240,0.160698668162028
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,2048,10240,0.06930239995320639
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,2048,8192,0.07327253023783366
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,2048,8192,0.12267200152079265
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,2048,12288,0.07624320189158121
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,2048,7168,0.07301013469696045
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,2048,7168,0.10570666790008545
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,2048,7168,0.05225280125935873
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,2048,6144,0.061673601468404136
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,2048,6144,0.0898026704788208
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,2048,8192,0.057048531373341885
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,2048,5120,0.051125331719716394
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,2048,5120,0.07600639661153158
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,2048,5120,0.04232426484425862
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,2048,4096,0.03963306744893392
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,2048,4096,0.060266665617624914
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,2048,6144,0.04650026559829712
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,2048,3584,0.034281599521636966
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,2048,3584,0.053547732035319005
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,2048,3584,0.03217386603355408
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,2048,3072,0.029573333263397217
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,2048,3072,0.04690560102462769
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,2048,4096,0.03606826861699422
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,2048,2560,0.02537386616071065
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,2048,2560,0.04298346837361654
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,2048,3072,0.028483200073242187
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,2048,2048,0.020547199249267577
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,2048,2560,0.026868265867233277
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,2048,2048,0.03654293219248454
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,2048,1536,0.01584106683731079
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,2048,1536,0.032629332939783734
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,2048,1536,0.020275199413299562
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,2048,1024,0.011538133025169373
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,2048,1024,0.026807467142740887
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,2048,2048,0.023536000649134317
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,2048,768,0.009692800045013428
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,2048,1024,0.018217599391937254
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,2048,768,0.025726934274037678
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,2048,768,0.017340799172719322
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,2048,512,0.007667199770609538
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,2048,512,0.022689066330591836
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,2048,256,0.006434133152167003
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,2048,512,0.017318399747212727
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,2048,256,0.02069759964942932
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,2048,128,0.005585066477457682
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,2048,128,0.01829119920730591
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,2048,256,0.017338667313257852
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,2048,64,0.005020800232887268
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,2048,64,0.018572799364725747
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,2048,32,0.005048533280690512
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,2048,128,0.017767467101415
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,2048,32,0.016834133863449098
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,1536,65536,0.44271039962768555
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,1536,65536,0.9116778691609702
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,1536,51200,0.3427978515625
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,1536,51200,0.7167935689290365
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,2048,51200,0.2674336115519206
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,1536,16384,0.11189333597819011
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,1536,16384,0.23765546480814614
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,1536,16384,0.09721919695536295
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,1536,12288,0.08481280008951822
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,1536,12288,0.18025919596354167
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,2048,65536,0.33315200805664064
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,1536,10240,0.07463466326395671
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,1536,10240,0.14830400149027506
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,1536,12288,0.07712213198343912
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,1536,8192,0.06095679998397827
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,1536,8192,0.11416320006052653
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,1536,10240,0.0697983980178833
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,1536,7168,0.052035200595855716
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,1536,7168,0.09781866868336996
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,1536,8192,0.057760000228881836
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,1536,6144,0.045005865891774494
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,1536,6144,0.08302826881408691
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,1536,7168,0.052789334456125894
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,1536,5120,0.037197868029276535
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,1536,5120,0.0686784029006958
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,1536,6144,0.047134931882222494
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,1536,4096,0.03054080009460449
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,1536,4096,0.055379199981689456
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,1536,5120,0.04320853153864543
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,1536,3584,0.026822400093078614
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,1536,3584,0.0493834654490153
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,1536,3584,0.034219733874003094
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,1536,3072,0.023626667261123658
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,1536,3072,0.04326719840367635
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,1536,4096,0.03709973494211833
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,1536,2560,0.020637865861256918
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,1536,2560,0.039510401089986165
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,1536,3072,0.03097066680590312
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,1536,2560,0.02910826603571574
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,1536,2048,0.016945066054662068
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,1536,2048,0.035010135173797606
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,1536,1536,0.0132533331712087
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,1536,1536,0.03189333279927571
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,1536,2048,0.02566080093383789
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,1536,1024,0.009809066851933796
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,1536,1536,0.02300800085067749
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,1536,1024,0.025493333737055462
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,1536,768,0.008319999774297078
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,1536,768,0.024731733401616416
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,1536,768,0.02029013236363729
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,1536,512,0.006733866532643636
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,1536,512,0.022980266809463502
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,1536,51200,0.2689845403035482
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,1536,256,0.00556160012880961
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,1536,256,0.01991999944051107
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,1536,512,0.020512000719706217
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,1536,128,0.004690133531888326
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,1536,1024,0.021473066012064616
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,1536,128,0.01804479956626892
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,1536,64,0.004394666850566864
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,1536,256,0.02055466572443644
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,1536,64,0.017847466468811034
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,1536,32,0.0046517332394917805
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,1536,128,0.021511467297871907
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,1024,65536,0.3276672045389811
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,1024,65536,0.8538314819335937
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,1024,51200,0.25745813051859534
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,1536,32,0.017834667364756265
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,1024,51200,0.6727359771728516
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,1024,16384,0.08539626598358155
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,1024,16384,0.2255253314971924
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,1024,16384,0.07335039774576822
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,1024,12288,0.06561066706975302
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,1024,12288,0.1703338623046875
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,1536,65536,0.3382805188496908
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,1024,10240,0.055567999680837
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,1024,10240,0.13932159741719563
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,1024,12288,0.058466132481892905
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,1024,8192,0.04438079992930095
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,1024,8192,0.1046453317006429
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,1024,10240,0.05376000006993612
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,1024,7168,0.03842560052871704
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,1024,7168,0.08828266461690268
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,1024,8192,0.04250880082448323
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,1024,7168,0.039078398545583086
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,1024,6144,0.032739200194676715
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,1024,6144,0.07271467049916586
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,1024,5120,0.026869332790374754
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,1024,5120,0.061020799477895106
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,1024,6144,0.03509333531061808
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,1024,4096,0.0208512008190155
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,1024,4096,0.04782400131225586
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,1024,5120,0.032620799541473386
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,1024,3584,0.018402133385340372
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,1024,3584,0.045100800196329754
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,1024,4096,0.028516266743342084
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,1024,3072,0.016108799974123636
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,1024,3072,0.039383467038472494
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,1024,3584,0.025674666961034136
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,1024,2560,0.01358080009619395
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,1024,2560,0.03710079987843831
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,1024,3072,0.0230730672677358
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,1024,2048,0.011849600076675414
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,1024,2048,0.03134933312733968
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,1024,2560,0.02176426649093628
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,1024,1536,0.009220266342163086
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,1024,1536,0.029095466931660968
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,1024,51200,0.19550186793009441
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,1024,1024,0.007321600119272869
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,1024,1024,0.024658133586247764
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,1024,2048,0.018564265966415406
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,1024,768,0.006301866471767425
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,1024,768,0.022845866282780965
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,1024,1024,0.01487573285897573
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,1024,512,0.005534933507442474
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,1024,1536,0.016612266500790916
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,1024,512,0.0220960001150767
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,1024,256,0.0046079998215039575
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,1024,256,0.018707199891408285
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,1024,768,0.014083199699719749
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,1024,128,0.0038624001046021783
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,1024,128,0.016520532965660095
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,1024,256,0.014881066481272378
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,1024,64,0.003542399903138479
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,1024,512,0.014594133694966635
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,1024,64,0.01643946667512258
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,1024,128,0.015312000115712484
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,1024,32,0.003849600007136663
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,1024,32,0.017612799008687337
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,768,65536,0.23895680109659828
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,768,51200,0.18712320327758789
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,768,65536,0.8340031941731771
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,768,51200,0.6533482869466145
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,768,16384,0.060601600011189784
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,768,16384,0.21122986475626626
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,1024,65536,0.2419648011525472
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,768,12288,0.04618026812871297
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,768,12288,0.15918292999267578
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,768,16384,0.07368853092193603
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,768,10240,0.039325865109761556
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,768,10240,0.13035093148549398
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,768,10240,0.05437653462092081
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,768,8192,0.031905066967010495
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,768,12288,0.059157331784566246
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,768,8192,0.09841919740041097
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,768,7168,0.027639466524124145
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,768,7168,0.08363733291625977
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,768,7168,0.04113493363062541
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,768,6144,0.0240831991036733
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,768,6144,0.06789226531982422
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,768,8192,0.04323199987411499
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,768,5120,0.02023893396059672
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,768,5120,0.055166931947072353
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,768,5120,0.03344000180562337
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,768,4096,0.017025067408879598
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,768,6144,0.035979731877644854
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,768,4096,0.045201067129770914
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,768,3584,0.014975999792416891
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,768,3584,0.04282559951146443
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,768,3584,0.026800000667572023
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,768,4096,0.02927466630935669
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,768,3072,0.013195733229319254
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,768,3072,0.03756159941355387
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,768,2560,0.011480533083279927
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,768,2560,0.03672959804534912
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,768,3072,0.023998934030532836
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,768,2560,0.02227413256963094
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,768,2048,0.00960640013217926
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,768,2048,0.030931200583775836
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,768,1536,0.008162133395671844
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,768,1536,0.027108265956242876
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,768,1536,0.017321600516637167
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,768,2048,0.019504000743230186
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,768,1024,0.006694399813810985
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,768,1024,0.024660267432530723
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,768,768,0.005494399865468343
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,768,768,0.022778666019439696
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,768,1024,0.015347199638684592
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,768,512,0.005074133475621542
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,768,768,0.014938666423161825
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,768,512,0.020857600371042888
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,768,256,0.00421973317861557
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,768,256,0.01874986688296
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,768,512,0.015285332997639975
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,768,256,0.016152532895406087
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,768,128,0.0038111999630928038
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,768,128,0.016498133540153503
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,768,128,0.016321067015329996
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,768,64,0.0034304000437259674
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,768,64,0.016455466548601784
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,768,32,0.0034122665723164878
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,768,32,0.015564800302187601
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,512,65536,0.1744106610616048
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,768,51200,0.19574613571166993
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,512,65536,0.7808767954508464
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,512,51200,0.13718080520629883
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,512,51200,0.6165226618448894
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,512,16384,0.0460042675336202
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,512,16384,0.20175466537475586
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,512,16384,0.07188906669616699
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,512,12288,0.03498026529947917
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,512,12288,0.15167039235432941
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,768,65536,0.24145493507385254
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,512,10240,0.029550933837890626
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,512,10240,0.12448320388793946
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,512,12288,0.05749013423919678
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,512,8192,0.02101866602897644
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,512,8192,0.09122560024261475
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,512,10240,0.052205868562062586
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,512,7168,0.01755839983622233
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,512,7168,0.07730027039845785
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,512,8192,0.040855467319488525
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,512,6144,0.015453867117563882
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,512,7168,0.03712426821390788
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,512,6144,0.06112639904022217
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,512,5120,0.013358933726946512
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,512,5120,0.05041066805521647
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,512,5120,0.031318400303522745
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,512,4096,0.01097813347975413
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,512,6144,0.03337386846542358
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,512,4096,0.04132159948348999
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,512,3584,0.009593600034713745
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,512,3584,0.039043200016021726
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,512,4096,0.027106134096781413
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,512,3584,0.024733867247899374
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,512,3072,0.008810666203498841
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,512,3072,0.03500159978866577
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,512,2560,0.007843199868996937
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,512,2560,0.03291626572608948
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,512,51200,0.19346879323323568
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,512,2048,0.006920533378918965
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,512,2048,0.029021867116292316
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,512,3072,0.02219520012537638
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,512,1536,0.005942399799823761
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,512,2560,0.02066133419672648
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,512,1536,0.026858667532602947
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,512,1024,0.00470719983180364
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,512,1024,0.023618133862813313
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,512,1536,0.015702399611473083
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,512,768,0.004310399790604909
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,512,2048,0.018178133169809978
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,512,1024,0.014102400342623393
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,512,512,0.0038165333370367683
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,512,512,0.020487467447916664
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,512,768,0.013265066345532737
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,512,256,0.0034293333689371743
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,512,256,0.017633066574732462
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,512,256,0.01404159963130951
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,512,128,0.003053866575161616
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,512,128,0.016545066237449647
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,512,128,0.014033066232999167
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,512,64,0.0027594665686289472
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,512,64,0.014607999722162882
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,512,768,0.02211839954058329
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,512,32,0.0030069333811601003
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,512,32,0.015003732840220132
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,256,65536,0.11063679854075115
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,512,512,0.013704533378283182
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,256,51200,0.08532586892445883
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,256,65536,0.7478282928466797
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,256,51200,0.588375473022461
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,256,16384,0.03668266534805298
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,256,16384,0.1983136018117269
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,512,65536,0.24099626541137695
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,256,12288,0.026903466383616133
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,256,12288,0.14810986518859864
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,256,16384,0.07158506711324056
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,256,10240,0.019661867618560792
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,256,10240,0.1210261344909668
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,256,12288,0.057251199086507165
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,256,8192,0.014190933108329773
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,256,8192,0.09005119800567626
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,256,10240,0.05044053395589193
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,256,7168,0.011970133582750956
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,256,7168,0.07319146792093913
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,256,7168,0.036263465881347656
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,256,6144,0.010453333457310993
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,256,6144,0.05791039864222208
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,256,8192,0.03948479890823364
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,256,5120,0.009238400061925252
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,256,5120,0.04744960069656372
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,256,6144,0.03336639801661174
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,256,4096,0.007777066528797149
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,256,4096,0.03983786503473918
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,256,51200,0.1929759979248047
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,256,5120,0.030987733602523805
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,256,4096,0.026426666975021364
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,256,3584,0.007187200089295705
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,256,3584,0.03703039884567261
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,256,3072,0.006716800232728322
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,256,3072,0.03392746845881144
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,256,3584,0.02434879938761393
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,256,2560,0.006111999849478403
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,256,3072,0.02190293272336324
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,256,2560,0.032307199637095135
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,256,2048,0.005438933273156484
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,256,2048,0.029179733991622925
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,256,2560,0.020191999276479085
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,256,1536,0.004666666686534882
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,256,1536,0.025766400496164958
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,256,65536,0.24047999382019042
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,256,1024,0.003969066590070724
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,256,1536,0.01530880033969879
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,256,2048,0.01781439979871114
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,256,768,0.0034517332911491393
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,256,768,0.021594667434692384
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,256,512,0.0034154665966828666
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,256,768,0.012871467073758445
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,256,1024,0.014139733711878457
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,256,256,0.0031797334551811217
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,256,512,0.019471999009450278
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,256,256,0.01839253306388855
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,256,512,0.013247999548912048
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,256,128,0.0026954665780067446
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,256,128,0.016165332992871602
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,256,64,0.0026122666895389557
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,256,256,0.013671466708183288
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,256,64,0.014784000317255654
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,256,32,0.0026933332284291584
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,256,32,0.014696533481280008
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,128,65536,0.09430507024129232
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,256,1024,0.022036266326904298
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,128,51200,0.07371626694997152
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,128,65536,0.7389450709025065
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,128,51200,0.581218147277832
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,256,128,0.0137855996688207
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,128,16384,0.19497812589009603
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,128,16384,0.03156053423881531
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,128,12288,0.02093013326327006
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,128,12288,0.14545173645019532
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,128,16384,0.0703765312830607
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,128,10240,0.01439786652723948
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,128,10240,0.11833600203196208
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,128,12288,0.0556885321935018
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,128,8192,0.012100266416867574
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,128,8192,0.08478720188140869
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,128,10240,0.05135360161463419
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,128,7168,0.010857599973678588
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,128,7168,0.06855466365814208
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,128,51200,0.18951039314270018
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,128,6144,0.009590400258700053
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,128,8192,0.03827519814173381
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,128,6144,0.055581867694854736
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,128,7168,0.03550293445587158
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,128,5120,0.008354133367538452
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,128,5120,0.045371735095977785
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,128,4096,0.007216000060240428
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,128,4096,0.039137065410614014
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,128,6144,0.03253866632779439
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,128,3584,0.006682666639486949
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,128,5120,0.030015999078750612
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,128,3584,0.03697813351949056
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,128,3072,0.00625600020090739
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,128,4096,0.025898667176564534
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,128,3584,0.02351893385251363
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,128,2560,0.0055285334587097164
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,128,2560,0.03094826738039653
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,128,3072,0.021116799116134642
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,128,2048,0.005051733553409576
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,128,2048,0.027115732431411743
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,128,2560,0.019402666886647543
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,128,1536,0.004273066421349844
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,128,1536,0.024844799439112344
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,128,2048,0.01730666756629944
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,128,1024,0.003669333209594091
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,128,1024,0.023630932966868082
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,128,1536,0.014457600315411887
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,128,768,0.0034154665966828666
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,128,3072,0.03301759958267212
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,128,768,0.02095573345820109
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,128,1024,0.012888532876968384
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,128,512,0.0030005333324273427
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,128,512,0.019189333915710448
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,128,256,0.002696533252795537
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,128,768,0.0123690664768219
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,128,512,0.012823466459910074
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,128,128,0.0025813333690166474
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,128,256,0.017914666732152303
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,128,256,0.012475732962290447
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,128,128,0.01639253298441569
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,128,64,0.0025546667476495106
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,128,32,0.002584533393383026
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,128,64,0.015289599696795145
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,128,32,0.014531200130780539
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,64,65536,0.09580372969309489
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,128,65536,0.23657600084940592
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,64,65536,0.732419204711914
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,64,51200,0.06955413023630777
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,64,16384,0.027074132363001508
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4097,128,128,0.012489599982897441
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,64,12288,0.016810667514801026
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,64,16384,0.19454933802286783
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,64,51200,0.5769194920857748
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,64,10240,0.012139733632405598
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,64,12288,0.1449109395345052
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,64,8192,0.01095360020796458
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,64,10240,0.11643733183542888
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,64,7168,0.01048533320426941
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,64,8192,0.08472426732381186
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,64,6144,0.009265066186587016
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,64,5120,0.008088533580303193
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,64,6144,0.0556277314821879
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,64,4096,0.0067669332027435304
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,64,5120,0.045347201824188235
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,64,4096,0.03913386662801106
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,64,3584,0.006306133170922597
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,64,3072,0.005873066683610281
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,64,3584,0.037062398592631024
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,64,2560,0.005076266825199127
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,64,3072,0.033011200030644734
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,64,2048,0.0046634669105211895
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,64,2560,0.0310591995716095
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,64,1536,0.004181333382924398
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,64,2048,0.027345067262649535
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,64,1024,0.0034783999125162757
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,64,1536,0.02505706747372945
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,64,768,0.0030464000999927522
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,64,1024,0.021067732572555543
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,64,768,0.020906666914621987
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,64,512,0.003115733216206233
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,64,256,0.002567466596762339
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,64,128,0.0022474666436513263
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,64,512,0.01976319948832194
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,64,256,0.018074667453765868
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,64,128,0.015991466244061787
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,64,64,0.002266666789849599
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,64,32,0.0022122666239738466
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,64,64,0.014708266655604044
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,32,65536,0.08428373336791992
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,64,32,0.014729600151379904
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,32,51200,0.06758613586425781
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,32,16384,0.024913066625595094
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,32,16384,0.19341546694437664
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,32,12288,0.015687466661135355
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,32,51200,0.5782389322916667
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,32,12288,0.14506773948669432
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,32,65536,0.7327552159627279
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,32,10240,0.010903466741243999
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,32,8192,0.009665066997210186
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,32,7168,0.009005866448084513
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,32,10240,0.1200287977854411
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,32,6144,0.008332799871762593
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,32,8192,0.08653759956359863
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,32,7168,0.07026027043660482
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,32,6144,0.055941331386566165
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,32,5120,0.010809600353240967
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,32,4096,0.009296000003814697
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,64,7168,0.06788480281829834
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,32,3584,0.008358400066693623
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,32,5120,0.04531626701354981
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,32,3072,0.007540266712506611
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,32,3584,0.03633600076039632
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,32,2560,0.006656000018119812
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,32,3072,0.03290773431460063
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,32,2048,0.004648533463478088
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,32,2560,0.030855466922124226
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,32,1536,0.0038965334494908653
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,32,2048,0.026833067337671917
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,32,1024,0.003409066547950109
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,32,1536,0.025613866249720257
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,32,768,0.0030432000756263735
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,32,512,0.0026677332818508146
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,32,768,0.020662399133046468
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,32,256,0.002573866645495097
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,32,512,0.019862399498621622
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,32,128,0.0022314667701721193
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,32,256,0.016977065801620485
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,32,64,0.0022122666239738466
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,32,64,0.014453333616256715
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4097,32,32,0.0021503999829292296
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,32,32,0.014467199643452963
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,32,4096,0.03883519967397054
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,32,1024,0.020631466309229532
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4097,32,128,0.015828266739845276
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,65536,16384,4.915745035807292
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,65536,16384,2.6870208740234376
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,65536,51200,8.469432576497395
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,65536,12288,3.856180318196615
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,65536,12288,2.0594197591145833
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,65536,51200,16.473178100585937
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,65536,10240,3.175459289550781
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,65536,10240,1.6238325754801433
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,65536,10240,0.9130165100097656
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,65536,8192,2.5639328002929687
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,65536,8192,1.2728501637776692
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,65536,12288,1.0146538416544595
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,65536,7168,2.1827423095703127
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,65536,7168,1.1399487813313802
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,65536,16384,1.3734485626220703
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,65536,6144,1.755509312947591
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,65536,6144,0.9129215876261393
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,65536,8192,0.7563519795735677
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,65536,5120,1.4234976450602213
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,65536,5120,0.8210762659708658
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,65536,7168,0.7021525065104167
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,65536,4096,1.0839221954345704
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,65536,4096,0.6512586593627929
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,65536,6144,0.6159807840983074
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,65536,3584,0.9375370661417642
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,65536,3584,0.568887456258138
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,65536,5120,0.562004280090332
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,65536,3072,0.81527468363444
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,65536,3072,0.48575468063354493
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,65536,4096,0.49964052836100264
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,65536,2560,0.6783797581990559
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,65536,2560,0.43793493906656905
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,65536,3584,0.45197121302286786
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,65536,2048,0.5413471857706706
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,65536,2048,0.3483637491861979
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,65536,3072,0.3998048146565755
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,65536,1536,0.42383359273274734
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,65536,1536,0.27361812591552737
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,65536,2560,0.3724309285481771
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,65536,1024,0.2900917371114095
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,65536,1024,0.18712533315022786
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,65536,1024,0.2674271901448568
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,65536,768,0.2272501309712728
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,65536,768,0.15972906748453777
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,65536,1536,0.2785557428995768
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,65536,512,0.16634772618611654
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,65536,512,0.1314357360204061
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,65536,512,0.24944106737772623
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,65536,256,0.11189760367075603
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,65536,768,0.2558314641316732
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,65536,256,0.11022079785664876
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,65536,128,0.09989120165506998
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,65536,128,0.09788800080617269
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,65536,2048,0.33007787068684896
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,65536,64,0.0834719975789388
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,65536,64,0.10109333197275798
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,65536,32,0.08631892999013266
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,65536,32,0.1019477367401123
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,65536,128,0.25301760037740073
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,65536,256,0.253326940536499
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,51200,65536,8.637366739908854
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,51200,65536,16.281864420572916
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,51200,51200,12.606868489583333
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,51200,51200,6.695454915364583
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,51200,16384,3.9332331339518225
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,51200,16384,2.0407039642333986
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,51200,16384,1.0627594629923502
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,51200,12288,2.8669087727864584
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,51200,12288,1.599829355875651
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,51200,12288,0.8038314819335938
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,51200,10240,2.497289530436198
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,51200,10240,1.3063477834065755
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,51200,10240,0.7103541056315105
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,51200,8192,1.9344768524169922
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,51200,8192,0.9970090866088868
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,65536,51200,4.552979024251302
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,51200,7168,1.575002670288086
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,51200,7168,0.854092788696289
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,51200,8192,0.5862464269002279
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,51200,6144,1.3437503814697265
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,51200,6144,0.7211818695068359
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,51200,7168,0.5548437118530274
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,51200,5120,1.1335435231526694
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,51200,5120,0.6738911946614583
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,51200,5120,0.4354912122090657
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,51200,4096,0.861845334370931
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,51200,4096,0.5339455922444661
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,51200,6144,0.4779701232910156
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,51200,3584,0.7329696019490559
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,51200,3584,0.4344223976135254
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,51200,3584,0.3478133201599121
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,51200,3072,0.6339882532755534
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,51200,3072,0.377234141031901
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,51200,4096,0.3913013458251953
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,51200,2560,0.5409109115600585
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,51200,2560,0.3515221277872721
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,51200,51200,3.5742207845052087
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,51200,2048,0.43200639088948567
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,51200,2048,0.2606634616851807
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,51200,2560,0.2917279879252116
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,51200,1536,0.33599252700805665
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,51200,1536,0.20636266072591147
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,51200,2048,0.25774933497111
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,51200,1024,0.22718507448832193
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,51200,1024,0.14758507410685223
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,51200,3072,0.3046026547749837
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,51200,768,0.17360320091247558
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,51200,768,0.1287168025970459
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,51200,768,0.20185066858927408
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,51200,512,0.1247594674428304
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,51200,512,0.10830399990081788
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,51200,1024,0.2139392058054606
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,51200,256,0.0881119966506958
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,51200,256,0.11719573338826497
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,51200,1536,0.22134186426798502
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,51200,128,0.070742400487264
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,51200,128,0.08430720170338948
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,51200,128,0.20433707237243653
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,51200,64,0.06914347012837728
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,51200,256,0.20012052853902182
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,51200,64,0.08773866494496664
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,51200,32,0.08841919898986816
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,51200,32,0.07350827058156331
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,16384,65536,3.0305099487304688
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,16384,65536,4.890746561686198
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,51200,512,0.20250239372253417
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,16384,51200,3.8023231506347654
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,16384,51200,2.3343839009602863
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,16384,16384,1.129205322265625
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,16384,16384,0.7412000020345052
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,16384,16384,0.38917652765909827
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,16384,12288,0.8205706914265951
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,16384,12288,0.5690762837727864
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,16384,12288,0.29344107309977213
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,16384,10240,0.7000789642333984
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,16384,10240,0.4760138511657715
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,16384,10240,0.26200853983561195
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,16384,8192,0.5517077128092448
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,16384,8192,0.3859231948852539
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,16384,8192,0.21698667208353678
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,16384,7168,0.4787647883097331
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,16384,7168,0.3211722691853841
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,16384,51200,1.1909355163574218
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,16384,6144,0.4251872062683105
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,16384,6144,0.2867466608683268
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,16384,7168,0.20147199630737306
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,16384,5120,0.33615147272745766
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,16384,5120,0.22857173283894858
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,16384,5120,0.1606335957845052
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,16384,4096,0.26974827448527017
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,16384,4096,0.1925002733866374
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,16384,6144,0.17667519251505534
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,16384,3584,0.24226133028666177
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,16384,3584,0.17118933995564778
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,16384,4096,0.14204479853312174
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,16384,3072,0.2046816031138102
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,16384,3072,0.14322667121887206
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,16384,3584,0.12861226399739584
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,16384,2560,0.176199467976888
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,16384,2560,0.12774186929066975
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,16384,2560,0.10867520173390706
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,16384,2048,0.14296107292175292
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,16384,2048,0.10384960174560547
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,16384,3072,0.1129418690999349
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,16384,1536,0.10675520102183025
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,16384,1536,0.08440213203430176
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,16384,1536,0.08207360108693441
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,16384,1024,0.07414720058441163
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,16384,1024,0.06375786860783896
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,16384,2048,0.0961184024810791
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,16384,768,0.05973226626714071
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,16384,768,0.05921813249588013
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,16384,1024,0.07801813284556071
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,16384,512,0.04360426664352417
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,16384,512,0.051344001293182374
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,16384,768,0.07196586926778158
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,16384,256,0.030986666679382324
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,16384,256,0.04269013404846191
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,16384,256,0.0645525336265564
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,16384,128,0.027908267577489217
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,16384,128,0.03768959840138753
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,16384,512,0.06975680192311605
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,16384,64,0.023898667097091673
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,16384,64,0.038179198900858566
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,16384,32,0.024514132738113405
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,16384,128,0.06397013266881307
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,16384,32,0.03747946818669637
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,12288,65536,2.436726379394531
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,12288,65536,3.7478294372558594
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,12288,51200,2.9013748168945312
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,12288,51200,1.9317887624104817
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,16384,65536,1.534661356608073
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,12288,16384,0.8565216064453125
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,12288,16384,0.616325314839681
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,51200,65536,4.573001607259115
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,12288,12288,0.6155381520589193
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,12288,12288,0.46808640162150067
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,12288,12288,0.23790507316589354
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,12288,10240,0.5376341501871745
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,12288,10240,0.4045546531677246
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,12288,16384,0.3151189486185709
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,12288,8192,0.43385705947875974
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,12288,8192,0.31960748036702474
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,12288,10240,0.21173866589864096
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,12288,7168,0.3752245267232259
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,12288,7168,0.2590378602345785
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,12288,8192,0.17437440554300945
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,12288,6144,0.3195647875467936
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,12288,6144,0.233351469039917
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,12288,7168,0.1621503988901774
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,12288,5120,0.26319360733032227
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,12288,5120,0.18781973520914713
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,12288,6144,0.14161386489868164
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,12288,4096,0.21026879946390786
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,12288,4096,0.15202453931172688
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,12288,4096,0.11370346546173096
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,12288,3584,0.18452480634053547
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,12288,3584,0.13718719482421876
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,12288,5120,0.12960533301035565
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,12288,3072,0.1641866683959961
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,12288,3072,0.12228159904479981
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,12288,3072,0.0903498649597168
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,12288,2560,0.13188052972157796
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,12288,2560,0.102783997853597
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,12288,3584,0.1013983964920044
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,12288,2048,0.10760213534037273
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,12288,2048,0.08227840264638266
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,12288,2048,0.07643093268076578
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,12288,1536,0.08223360379536947
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,12288,1536,0.0694976011912028
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,12288,2560,0.08707733154296875
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,12288,1024,0.056738134225209555
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,12288,1024,0.05357546806335449
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,12288,1024,0.06634026765823364
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,12288,768,0.044758399327596024
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,12288,768,0.05040853420893351
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,12288,1536,0.06853973070780436
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,12288,512,0.03295573393503825
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,12288,512,0.05053973197937012
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,12288,512,0.06159893274307251
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,12288,256,0.030060799916585286
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,12288,256,0.042683732509613034
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,12288,768,0.0615658680597941
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,12288,128,0.019208532571792603
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,12288,128,0.03186666568120321
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,12288,256,0.06157653331756592
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,12288,64,0.017083734273910522
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,12288,64,0.03094826738039653
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,12288,32,0.01772586703300476
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,12288,32,0.03163733283678691
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,12288,128,0.06128426790237427
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,10240,65536,3.2218368530273436
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,10240,65536,2.15347417195638
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,10240,51200,2.4963946024576824
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,10240,51200,1.6974773406982422
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,12288,51200,0.9534751892089843
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,10240,16384,0.7253962834676106
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,10240,16384,0.5687957127888997
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,10240,16384,0.2689258575439453
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,10240,12288,0.5487071990966796
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,10240,12288,0.42956587473551433
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,12288,65536,1.2046112060546874
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,10240,10240,0.4405834515889485
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,10240,10240,0.34173758824666345
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,10240,12288,0.2039797306060791
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,10240,8192,0.3596874554951986
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,10240,8192,0.28071358998616536
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,10240,10240,0.1812405268351237
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,10240,7168,0.31932160059611003
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,10240,7168,0.2291093349456787
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,10240,8192,0.15017174084981283
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,10240,6144,0.2800970713297526
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,10240,6144,0.20244266192118326
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,10240,7168,0.13839680353800457
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,10240,5120,0.21505813598632811
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,10240,5120,0.1622762680053711
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,10240,5120,0.11072213649749756
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,10240,4096,0.1716383934020996
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,10240,4096,0.13202133178710937
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,10240,6144,0.12168853282928467
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,10240,3584,0.15503466924031575
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,10240,3584,0.12160212993621826
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,10240,4096,0.09632746378580728
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,10240,3584,0.08794346650441488
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,10240,3072,0.12917760213216145
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,10240,3072,0.10436480045318604
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,10240,2560,0.11181973616282145
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,10240,2560,0.09203733603159586
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,10240,2560,0.0757578690846761
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,10240,2048,0.09050986766815186
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,10240,2048,0.0744266668955485
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,10240,3072,0.07789119879404703
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,10240,1536,0.07076906363169352
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,10240,1536,0.06364479859670004
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,10240,1536,0.060729598999023436
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,10240,1024,0.04695253372192383
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,10240,1024,0.04866773287455241
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,10240,2048,0.0676362673441569
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,10240,768,0.03796159823735555
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,10240,768,0.045424000422159834
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,10240,768,0.055180799961090085
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,10240,512,0.027949867645899455
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,10240,512,0.03927359978357951
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,10240,1024,0.05886186758677164
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,10240,256,0.019849600394566853
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,10240,256,0.041127467155456544
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,10240,256,0.05472853183746338
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,10240,512,0.05529173215230306
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,10240,128,0.016663466890652977
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,10240,128,0.029377067089080812
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,10240,128,0.0548746665318807
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,10240,64,0.01532799998919169
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,10240,64,0.029452800750732422
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,10240,32,0.016552533706029257
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,10240,32,0.029975465933481854
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,8192,65536,2.5712554931640623
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,8192,65536,1.8939659118652343
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,10240,51200,0.798906644185384
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,8192,51200,1.96257807413737
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,8192,51200,1.4648053487141928
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,8192,16384,0.5773450851440429
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,8192,16384,0.47393172581990556
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,8192,16384,0.23895146052042643
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,8192,12288,0.43145173390706376
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,8192,12288,0.36010348002115883
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,10240,65536,1.0092479705810546
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,8192,10240,0.3559626579284668
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,8192,10240,0.3057098706563314
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,8192,12288,0.17327146530151366
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,8192,8192,0.2866847991943359
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,8192,8192,0.2288053353627523
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,8192,10240,0.1504405339558919
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,8192,7168,0.25178133646647133
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,8192,7168,0.19588799476623536
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,8192,8192,0.12139413356781006
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,8192,6144,0.21208640734354653
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,8192,6144,0.1657471974690755
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,8192,7168,0.11317546367645263
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,8192,5120,0.17682986259460448
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,8192,5120,0.1396106719970703
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,8192,5120,0.09115520318349203
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,8192,4096,0.14071787198384603
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,8192,4096,0.1134218692779541
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,8192,6144,0.09979306856791179
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,8192,3584,0.12263360023498535
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,8192,3584,0.10554986794789631
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,8192,3584,0.06996053059895833
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,8192,3072,0.10522133509318035
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,8192,3072,0.0905514637629191
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,8192,4096,0.07837546666463216
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,8192,2560,0.0893344004948934
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,8192,2560,0.08007146517435709
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,8192,2560,0.058238931496938075
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,8192,2048,0.0723743995030721
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,8192,2048,0.06575040022532144
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,8192,3072,0.06140586535135904
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,8192,1536,0.05549439986546835
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,8192,1536,0.056944000720977786
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,8192,1536,0.04299413363138835
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,8192,1024,0.03838506539662679
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,8192,1024,0.044121599197387694
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,8192,2048,0.05012266635894776
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,8192,768,0.029849600791931153
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,8192,768,0.03970133463541667
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,8192,1024,0.041110400358835855
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,8192,512,0.021943465868632
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,8192,512,0.035252265135447186
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,8192,768,0.03624213139216105
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,8192,256,0.016178133090337117
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,8192,256,0.02895680069923401
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,8192,256,0.0363317330678304
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,8192,128,0.015404799580574035
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,8192,128,0.031616000334421794
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,8192,128,0.037043201923370364
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,8192,64,0.016921599706014
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,8192,64,0.03171093265215556
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,8192,32,0.017734400431315103
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,8192,32,0.029949865738550824
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,8192,512,0.0369269331296285
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,7168,65536,2.09521598815918
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,7168,65536,1.6667456309000652
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,7168,51200,1.50513916015625
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,7168,51200,1.2937141418457032
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,8192,51200,0.6864437103271485
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,7168,16384,0.4785440127054851
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,7168,16384,0.47510827382405596
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,7168,16384,0.20900160471598306
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,7168,12288,0.3627039909362793
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,7168,12288,0.34018131891886394
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,8192,65536,0.8585472106933594
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,7168,10240,0.30360320409138997
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,7168,10240,0.2825418790181478
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,7168,12288,0.15378133455912274
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,7168,8192,0.2518496036529541
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,7168,8192,0.2059328079223633
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,7168,10240,0.1343850612640381
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,7168,7168,0.21415146191914877
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,7168,7168,0.17624853452046713
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,7168,8192,0.10999146302541096
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,7168,6144,0.18757440249125162
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,7168,6144,0.15341226259867352
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,7168,7168,0.1024821360905965
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,7168,5120,0.1507754643758138
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,7168,5120,0.12740800380706788
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,7168,6144,0.09093333085378011
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,7168,4096,0.12339306672414144
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,7168,4096,0.10598293145497639
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,7168,5120,0.08288959662119547
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,7168,3584,0.10510719617207845
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,7168,3584,0.09609706401824951
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,7168,4096,0.07093760172526041
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,7168,3072,0.09369386831919352
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,7168,3072,0.08338560263315836
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,7168,3584,0.06368533372879029
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,7168,2560,0.07625173727671306
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,7168,2560,0.07446293036142984
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,7168,2560,0.053384534517923986
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,7168,2048,0.06191466649373373
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,7168,2048,0.06134719848632812
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,7168,3072,0.05658133427302042
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,7168,1536,0.048681600888570147
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,7168,1536,0.05226879914601644
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,7168,1536,0.04278826713562012
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,7168,1024,0.03371306657791138
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,7168,1024,0.04071573416392009
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,7168,2048,0.04747519890467326
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,7168,768,0.02532479961713155
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,7168,768,0.03736533323923747
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,7168,1024,0.04024853308995564
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,7168,512,0.019630932807922365
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,7168,512,0.0330570658047994
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,7168,768,0.037324798107147214
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,7168,256,0.014114133516947427
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,7168,256,0.02722666660944621
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,7168,256,0.03709973494211833
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,7168,128,0.0132832000652949
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,7168,512,0.03746666510899861
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,7168,128,0.02521066665649414
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,7168,64,0.011583999792734782
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,7168,64,0.02477546731630961
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,7168,32,0.01159999966621399
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,7168,32,0.02472213308016459
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,7168,128,0.03740479946136475
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,6144,65536,1.8832511901855469
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,6144,65536,1.5888960520426432
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,6144,51200,1.4310197194417318
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,6144,51200,1.2453290303548177
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,7168,51200,0.5951338450113932
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,6144,16384,0.4493077278137207
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,6144,16384,0.42412694295247394
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,6144,16384,0.1987925370534261
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,6144,12288,0.33285013834635413
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,6144,12288,0.30935999552408855
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,7168,65536,0.7556159973144532
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,6144,10240,0.28508265813191735
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,6144,10240,0.24853973388671874
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,6144,10240,0.1311850627263387
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,6144,8192,0.2269045352935791
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,6144,8192,0.19330239295959473
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,6144,12288,0.14646399815877278
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,6144,7168,0.1948959986368815
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,6144,7168,0.16479253768920898
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,6144,7168,0.09972480138142904
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,6144,6144,0.16573972702026368
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,6144,6144,0.1391328016916911
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,6144,8192,0.10656106472015381
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,6144,5120,0.1380736033121745
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,6144,5120,0.11821333567301433
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,6144,5120,0.07954986890157065
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,6144,4096,0.11045440038045247
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,6144,4096,0.09741226832071939
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,6144,6144,0.08793280124664307
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,6144,3584,0.09602239926656088
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,6144,3584,0.09150933424631755
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,6144,3584,0.06130666732788086
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,6144,3072,0.08244160016377768
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,6144,3072,0.07869333426157633
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,6144,4096,0.06819626490275064
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,6144,2560,0.07005759874979654
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,6144,2560,0.06990826924641927
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,6144,3072,0.05477866729100546
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,6144,2048,0.056377601623535153
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,6144,2048,0.05789866844813028
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,6144,2560,0.0521781325340271
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,6144,1536,0.043584001064300534
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,6144,1536,0.04791146516799927
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,6144,1536,0.04113493363062541
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,6144,1024,0.02965013384819031
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,6144,1024,0.03773866494496663
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,6144,2048,0.04690133333206177
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,6144,768,0.02350826660792033
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,6144,768,0.035892268021901444
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,6144,1024,0.03911786476771037
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,6144,512,0.017542399962743125
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,6144,512,0.030477867523829145
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,6144,768,0.036278398831685384
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,6144,256,0.013241599996884665
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,6144,256,0.0272597332795461
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,6144,512,0.03621546824773152
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,6144,128,0.012035199999809265
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,6144,128,0.022792534033457438
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,6144,256,0.0363103985786438
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,6144,128,0.03659626642862956
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,6144,64,0.022802132368087768
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,6144,64,0.010121599833170573
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,6144,32,0.010568533341089885
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,6144,32,0.02295893430709839
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,5120,65536,1.526052220662435
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,5120,65536,1.4169012705485025
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,5120,51200,1.160202662150065
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,5120,51200,1.1154324849446615
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,6144,51200,0.5844490687052409
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,5120,16384,0.3597034772237142
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,5120,16384,0.37650667826334633
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,5120,16384,0.17267413139343263
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,5120,12288,0.2728533426920573
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,5120,12288,0.27770878473917643
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,6144,65536,0.7348384221394857
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,5120,10240,0.23703893025716147
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,5120,10240,0.23137920697530112
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,5120,12288,0.13017173608144122
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,5120,8192,0.1938474655151367
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,5120,8192,0.18121387163798014
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,5120,10240,0.11596266428629558
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,5120,7168,0.17355306943257648
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,5120,7168,0.15543999671936035
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,5120,8192,0.09585386912027995
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,5120,6144,0.14369813601175946
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,5120,6144,0.1332757314046224
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,5120,7168,0.08945706685384115
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,5120,5120,0.1155509312947591
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,5120,5120,0.11263573169708252
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,5120,6144,0.0797322670618693
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,5120,4096,0.09386666615804037
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,5120,4096,0.09349439938863119
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,5120,5120,0.07270293235778809
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,5120,3584,0.08116693496704101
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,5120,3584,0.08608746528625488
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,5120,4096,0.06159893274307251
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,5120,3072,0.07125439643859863
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,5120,3072,0.0740832010904948
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,5120,3584,0.055954134464263915
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,5120,2560,0.05886293252309164
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,5120,2560,0.06551466782887777
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,5120,3072,0.05034559965133667
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,5120,2048,0.0486624002456665
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,5120,2048,0.05191786686579386
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,5120,2560,0.04853333234786987
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,5120,1536,0.03700053294499715
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,5120,1536,0.04354453484217326
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,5120,2048,0.04369920094807943
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,5120,1024,0.0252895991007487
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,5120,1024,0.03502293427785237
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,5120,1536,0.03901653289794922
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,5120,768,0.019802665710449217
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,5120,768,0.032978133360544844
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,5120,1024,0.03703573147455851
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,5120,512,0.01532799998919169
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,5120,768,0.035148799419403076
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,5120,512,0.02874559958775838
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,5120,256,0.012027733524640401
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,5120,256,0.02477653423945109
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,5120,256,0.035412267843882246
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,5120,128,0.009156266848246258
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,5120,128,0.02063680092493693
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,5120,512,0.03573760191599528
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,5120,64,0.008365866541862488
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,5120,64,0.02458666761716207
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,5120,128,0.03537493149439494
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,5120,32,0.010403199990590414
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,5120,32,0.022686932484308878
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,4096,65536,1.1774954477945963
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,4096,65536,1.2860212961832682
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,4096,51200,0.9133525212605795
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,4096,51200,1.0090058644612632
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,5120,51200,0.5054922739664713
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,4096,16384,0.29075520833333335
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,4096,16384,0.33032000859578453
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,4096,16384,0.14352107048034668
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,4096,12288,0.22016746203104653
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,4096,12288,0.24902186393737794
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,5120,65536,0.638810666402181
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,4096,10240,0.19246293703715006
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,4096,10240,0.20078399976094566
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,4096,12288,0.1106005350748698
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,4096,8192,0.15353493690490722
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,4096,8192,0.1551424026489258
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,4096,10240,0.09920746485392253
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,4096,7168,0.1344970703125
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,4096,7168,0.13456320762634277
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,4096,7168,0.07626986503601074
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,4096,6144,0.11317546367645263
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,4096,6144,0.11610559622446696
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,4096,8192,0.0813045342763265
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,4096,5120,0.09218346277872722
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,4096,5120,0.09959359963734946
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,4096,5120,0.06056106487909953
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,4096,4096,0.07331519921620687
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,4096,6144,0.06829120318094889
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,4096,4096,0.08236052989959716
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,4096,3584,0.06442026694615682
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,4096,3584,0.07541866302490234
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,4096,3584,0.04570666551589966
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,4096,3072,0.05554879903793335
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,4096,3072,0.06403413216272989
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,4096,4096,0.05110933383305868
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,4096,2560,0.047126400470733645
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,4096,2560,0.05623466571172079
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,4096,3072,0.04046613375345866
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,4096,2048,0.038114134470621744
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,4096,2560,0.03844053347905477
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,4096,2048,0.045553068319956466
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,4096,1536,0.029989333947499593
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,4096,1536,0.04069013198216756
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,4096,2048,0.03300586740175883
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,4096,1024,0.020836265881856282
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,4096,1024,0.03090133269627889
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,4096,1536,0.028716800610224406
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,4096,768,0.01664426624774933
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,4096,768,0.029418667157491047
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,4096,1024,0.026824533939361572
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,4096,768,0.024001065889994302
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,4096,512,0.012035199999809265
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,4096,512,0.027040000756581622
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,4096,256,0.00944533348083496
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,4096,256,0.024091732501983643
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,4096,256,0.02434879938761393
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,4096,128,0.008362666765848795
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,4096,512,0.024341332912445068
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,4096,128,0.021081600586573282
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,4096,64,0.007560533285140991
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,4096,64,0.021075199047724404
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,4096,32,0.00790719985961914
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,4096,128,0.024707200129826863
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,4096,32,0.026601600646972656
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,3584,65536,1.0920575459798179
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,3584,65536,1.2391743977864584
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,3584,51200,0.8521237055460611
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,3584,51200,0.9736053466796875
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,4096,51200,0.4231370608011882
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,3584,16384,0.2757034619649251
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,3584,16384,0.32290239334106446
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,3584,16384,0.12292906443277996
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,3584,12288,0.2092970689137777
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,3584,12288,0.2445653279622396
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,4096,65536,0.5302144050598144
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,3584,10240,0.15385066668192546
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,3584,10240,0.18421972592671712
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,3584,12288,0.09610133171081543
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,3584,8192,0.12792960007985432
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,3584,8192,0.1432693322499593
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,3584,10240,0.08733119964599609
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,3584,7168,0.10800426801045734
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,3584,7168,0.12340799967447917
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,3584,8192,0.0722752014795939
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,3584,6144,0.09513813654581706
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,3584,6144,0.10704320271809895
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,3584,7168,0.06809493700663248
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,3584,5120,0.07619093259175619
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,3584,5120,0.0908128023147583
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,3584,5120,0.0540832002957662
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,3584,4096,0.0607317328453064
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,3584,4096,0.07539199988047282
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,3584,6144,0.06050026814142863
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,3584,3584,0.055085865656534824
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,3584,3584,0.06840213139851889
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,3584,3584,0.041870931784311935
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,3584,3072,0.04724053144454956
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,3584,3072,0.05783040126164755
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,3584,4096,0.045865599314371744
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,3584,2560,0.038574934005737305
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,3584,2560,0.050552535057067874
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,3584,2560,0.0353983998298645
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,3584,2048,0.03134186665217082
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,3584,3072,0.037646933396657305
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,3584,2048,0.04346880118052165
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,3584,1536,0.024420267343521117
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,3584,1536,0.03727039893468221
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,3584,1536,0.0286080002784729
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,3584,1024,0.017509333292643228
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,3584,1024,0.03049280047416687
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,3584,2048,0.03175786733627319
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,3584,768,0.014069333672523499
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,3584,768,0.02924586733182271
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,3584,1024,0.026708267132441205
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,3584,512,0.010493866602579753
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,3584,512,0.02537920077641805
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,3584,768,0.0286954661210378
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,3584,256,0.007975466549396515
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,3584,256,0.02266133427619934
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,3584,512,0.02884693344434102
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,3584,128,0.0070933332045873005
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,3584,256,0.028758400678634645
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,3584,128,0.020662399133046468
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,3584,64,0.006794666747252147
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,3584,64,0.020717867215474448
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,3584,32,0.007155199845631917
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,3584,128,0.028830933570861816
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,3584,32,0.01992853283882141
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,3072,65536,0.9142293294270833
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,3072,65536,1.131722640991211
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,3072,51200,0.6896981557210287
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,3072,51200,0.8924949645996094
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,3584,51200,0.34575039545694985
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,3072,16384,0.22305386861165366
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,3072,16384,0.29352105458577477
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,3072,16384,0.12113173007965088
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,3072,12288,0.16815679868062336
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,3072,12288,0.22115519841512046
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,3584,65536,0.43401492436726885
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,3072,10240,0.14470613797505696
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,3072,10240,0.17583146095275878
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,3072,12288,0.0945077339808146
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,3072,8192,0.11436053117116292
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,3072,8192,0.13731733957926434
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,3072,10240,0.0856223980585734
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,3072,7168,0.09932586352030436
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,3072,7168,0.11802773475646973
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,3072,7168,0.06641173362731934
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,3072,6144,0.08722026348114013
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,3072,6144,0.10220479965209961
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,3072,8192,0.07096532980600992
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,3072,5120,0.06946132977803549
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,3072,5120,0.08753813107808431
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,3072,5120,0.052985600630442296
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,3072,4096,0.05580906470616659
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,3072,4096,0.07166399955749511
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,3072,6144,0.05903146664301554
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,3072,3584,0.049293867746988934
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,3072,3584,0.06408106486002604
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,3072,4096,0.04473280111948649
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,3072,3584,0.0408245325088501
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,3072,3072,0.04238400061925252
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,3072,3072,0.053973333040873206
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,3072,2560,0.035717332363128663
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,3072,2560,0.04862399895985921
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,3072,2560,0.034731733798980716
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,3072,2048,0.029125332832336426
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,3072,2048,0.04117439985275269
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,3072,3072,0.03625919818878174
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,3072,1536,0.022778666019439696
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,3072,1536,0.03711680173873901
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,3072,2048,0.030898133913675945
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,3072,1024,0.016090666254361473
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,3072,1024,0.03025813301404317
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,3072,1536,0.027642667293548584
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,3072,768,0.013255467017491659
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,3072,768,0.029895466566085816
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,3072,1024,0.025755733251571655
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,3072,512,0.011210667093594869
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,3072,512,0.02680319945017497
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,3072,512,0.028833067417144774
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,3072,256,0.009202133615811665
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,3072,256,0.02477759917577108
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,3072,256,0.02847040096918742
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,3072,128,0.009114666779836019
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,3072,768,0.028434133529663085
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,3072,128,0.022120533386866252
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,3072,64,0.008538666367530822
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,3072,64,0.020933334032694498
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,3072,32,0.008729599912961324
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,3072,32,0.02065920035044352
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,3072,128,0.02637333273887634
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,2560,65536,0.7636128107706706
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,2560,65536,1.0758794148763022
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,2560,51200,0.592521603902181
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,2560,51200,0.8638741175333658
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,3072,51200,0.3439210573832194
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,2560,16384,0.19218026796976725
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,2560,16384,0.2788319905598958
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,2560,16384,0.11946453253428142
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,2560,12288,0.14610026677449545
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,2560,12288,0.21018239657084145
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,3072,65536,0.4318559964497884
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,2560,10240,0.12108053366343181
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,2560,10240,0.1662432034810384
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,2560,10240,0.08520426750183105
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,2560,8192,0.09658346970876058
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,2560,8192,0.12904640038808185
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,2560,12288,0.09348159631093343
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,2560,7168,0.08165226777394613
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,2560,7168,0.11145706971486409
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,2560,7168,0.06510826746622721
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,2560,6144,0.0695904016494751
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,2560,6144,0.09541760285695394
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,2560,8192,0.06992426713307699
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,2560,5120,0.05925973256429037
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,2560,5120,0.08149973551432291
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,2560,6144,0.05663359959920248
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,2560,4096,0.04613546530405681
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,2560,4096,0.06499413251876832
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,2560,5120,0.05101973215738932
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,2560,3584,0.04162666797637939
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,2560,3584,0.059070932865142825
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,2560,3584,0.03970453341801961
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,2560,3072,0.03545813163121541
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,2560,3072,0.05003306468327841
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,2560,4096,0.04400959809621175
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,2560,2560,0.02919573386510213
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,2560,2560,0.04553279876708984
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,2560,3072,0.03529706796010335
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,2560,2048,0.024703999360402427
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,2560,2048,0.038898134231567384
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,2560,2560,0.03341333468755086
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,2560,1536,0.01883413394292196
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,2560,1536,0.0340394655863444
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,2560,1536,0.025970133145650227
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,2560,1024,0.013346133629480996
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,2560,1024,0.028520532449086505
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,2560,2048,0.029239465792973835
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,2560,1024,0.024166399240493776
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,2560,768,0.010840533177057902
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,2560,768,0.02691199978192647
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,2560,768,0.022709333896636964
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,2560,512,0.008772266904513042
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,2560,512,0.024293333292007446
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,2560,256,0.006728533407052357
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,2560,256,0.021817600727081297
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,2560,256,0.022742400566736855
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,2560,128,0.0059114664793014525
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,2560,128,0.019117865959803262
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,2560,512,0.022641066710154215
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,2560,64,0.005460266768932342
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,2560,64,0.018235733111699425
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,2560,32,0.005532800157864889
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,2560,128,0.02285760045051575
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,2560,32,0.019086933135986327
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,2048,65536,0.5668778737386068
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,2048,65536,0.9806378682454427
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,2048,51200,0.44612585703531904
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,2048,51200,0.7700394948323568
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,2560,51200,0.3448341369628906
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,2048,16384,0.14453333218892414
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,2048,16384,0.25526933670043944
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,2048,16384,0.09710826873779296
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,2048,12288,0.11003200213114421
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,2048,12288,0.19277226130167643
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,2560,65536,0.4345631917317708
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,2048,10240,0.09927146434783936
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,2048,10240,0.16076800028483074
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,2048,10240,0.06971306800842285
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,2048,12288,0.07640853722890219
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,2048,8192,0.0812554677327474
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,2048,8192,0.12497920195261639
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,2048,7168,0.07056639989217123
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,2048,7168,0.10792000293731689
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,2048,7168,0.0525056004524231
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,2048,6144,0.059582932790120446
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,2048,6144,0.09119359652201334
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,2048,8192,0.05640000104904175
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,2048,5120,0.04949973424275716
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,2048,5120,0.07577173709869385
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,2048,5120,0.042029865582784015
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,2048,4096,0.039819733301798506
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,2048,4096,0.06047360102335612
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,2048,6144,0.04650986591974894
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,2048,3584,0.034436265627543136
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,2048,3584,0.05530986785888672
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,2048,4096,0.03568533261617025
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,2048,3072,0.030161066850026445
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,2048,3072,0.047602132956186934
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,2048,3584,0.03248959978421529
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,2048,2560,0.02555946707725525
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,2048,2560,0.04359039862950643
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,2048,3072,0.028497066100438433
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,2048,2560,0.026796799898147584
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,2048,2048,0.03744000196456909
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,2048,2048,0.02132586638132731
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,2048,1536,0.016642133394877114
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,2048,1536,0.032919466495513916
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,2048,1536,0.020257065693537392
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,2048,1024,0.012146133184432983
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,2048,1024,0.027755733331044512
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,2048,2048,0.023410133520762124
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,2048,768,0.010435199737548828
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,2048,768,0.026821333169937133
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,2048,1024,0.018221867084503175
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,2048,512,0.008347733815511068
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,2048,512,0.024460800488789878
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,2048,512,0.017094399531682333
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,2048,256,0.00678719977537791
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,2048,256,0.021011199553807577
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,2048,768,0.017062399784723917
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,2048,128,0.00673173318306605
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,2048,256,0.01731520096460978
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,2048,51200,0.2678943951924642
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,2048,128,0.020206934213638304
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,2048,64,0.005983999868233999
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,2048,32,0.00631573349237442
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,2048,128,0.017362133661905924
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,2048,64,0.019309866428375243
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,2048,32,0.018234666188557944
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,1536,65536,0.43589439392089846
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,1536,51200,0.34638827641805015
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,1536,65536,0.9088607788085937
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,1536,51200,0.7152405420939127
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,1536,16384,0.11210133234659832
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,1536,16384,0.23619413375854492
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,1536,16384,0.09705706437428793
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,1536,12288,0.0852725346883138
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,1536,12288,0.17900800704956055
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,2048,65536,0.33298454284667967
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,1536,10240,0.08301546573638915
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,1536,10240,0.15070080757141113
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,1536,12288,0.07684480349222819
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,1536,8192,0.06685439745585123
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,1536,8192,0.11473279794057209
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,1536,10240,0.07032533486684164
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,1536,7168,0.05748053391774496
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,1536,7168,0.0981877326965332
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,1536,8192,0.05727893511454264
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,1536,6144,0.0484224001566569
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,1536,6144,0.08325973351796469
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,1536,7168,0.05316799879074097
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,1536,5120,0.038839467366536456
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,1536,5120,0.06810133457183838
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,1536,6144,0.04670399824778239
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,1536,4096,0.031259733438491824
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,1536,4096,0.053521064917246494
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,1536,5120,0.04243733485539754
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,1536,3584,0.027853866418202717
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,1536,3584,0.05102293491363526
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,1536,4096,0.03656746546427409
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,1536,3072,0.024204800526301064
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,1536,3072,0.04321920077006022
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,1536,3072,0.030134399731953938
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,1536,3584,0.03383786678314209
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,1536,2560,0.020475733280181884
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,1536,2560,0.04104959964752197
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,1536,2048,0.016987733046213784
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,1536,2048,0.03552853266398112
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,1536,2560,0.02881173292795817
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,1536,2048,0.025569067398707075
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,1536,1536,0.013370666901270548
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,1536,1536,0.03147839903831482
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,1536,1024,0.009983999530474345
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,1536,1024,0.027276800076166792
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,1536,1024,0.021143466234207153
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,1536,768,0.008348799745241801
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,1536,1536,0.022887466351191203
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,1536,768,0.02717333237330119
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,1536,512,0.007189333438873291
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,1536,512,0.02318293253580729
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,1536,768,0.02016106645266215
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,1536,256,0.005870933334032694
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,1536,512,0.020272000630696615
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,1536,256,0.019879466295242308
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,1536,128,0.005075199902057648
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,1536,51200,0.2692010561625163
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,1536,128,0.018659200270970663
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,1536,64,0.005077333251635233
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,1536,256,0.02040533423423767
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,1536,64,0.01806933283805847
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,1536,32,0.005082666873931885
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,1536,32,0.016607999801635742
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,1024,65536,0.2940704027811686
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,1536,128,0.022310400009155275
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,1024,51200,0.23192532857259116
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,1024,65536,0.8381610870361328
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,1024,51200,0.6576138814290364
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,1024,16384,0.07639893690745035
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,1024,16384,0.219868803024292
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,1024,16384,0.07278079986572265
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,1024,12288,0.058176000912984215
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,1024,12288,0.1659114678700765
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,1536,65536,0.33709653218587243
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,1024,10240,0.04989759922027588
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,1024,10240,0.1372970740000407
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,1024,12288,0.05846933523813883
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,1024,8192,0.04005226691563924
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,1024,8192,0.1034442663192749
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,1024,10240,0.05339946746826172
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,1024,7168,0.035573331514994304
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,1024,7168,0.0874847968419393
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,1024,8192,0.042888534069061277
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,1024,6144,0.030561065673828124
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,1024,6144,0.07329280376434326
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,1024,7168,0.040022401014963786
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,1024,5120,0.026254934072494508
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,1024,5120,0.0586570660273234
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,1024,6144,0.034949334462483723
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,1024,4096,0.021451733509699502
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,1024,4096,0.048724265893300374
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,1024,5120,0.03263466755549113
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,1024,3584,0.019655466079711914
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,1024,3584,0.04497599999109904
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,1024,3584,0.025543467203776045
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,1024,3072,0.01727679967880249
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,1024,3072,0.03916800022125244
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,1024,4096,0.027672533194224042
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,1024,2560,0.015090133746465048
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,1024,2560,0.03706026474634806
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,1024,3072,0.023128533363342287
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,1024,2048,0.01288746694723765
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,1024,2048,0.03163733283678691
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,1024,2048,0.018616533279418944
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,1024,1536,0.010013866424560546
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,1024,2560,0.0215338667233785
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,1024,1536,0.029063467184702558
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,1024,1024,0.007912533481915791
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,1024,51200,0.19521600405375164
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,1024,1024,0.024613332748413087
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,1024,768,0.006795733173688252
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,1024,768,0.02335466742515564
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,1024,1024,0.01451520025730133
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,1024,512,0.005576533575852713
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,1024,768,0.013823999961217245
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,1024,512,0.020948266983032225
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,1024,256,0.004637866715590159
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,1024,256,0.018722132841746012
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,1024,512,0.014075733224550881
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,1024,1536,0.016225066781044007
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,1024,256,0.014101333419481912
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,1024,128,0.018011732896169027
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,1024,128,0.01529706617196401
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,1024,64,0.0038090666135152185
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,1024,64,0.01650773286819458
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,1024,32,0.003934933245182038
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,1024,32,0.015826132893562318
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,768,65536,0.238808536529541
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,1024,128,0.003982933362325033
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,768,65536,0.8171157201131185
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,768,51200,0.18713280359903972
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,768,51200,0.6442037582397461
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,768,16384,0.06062399943669637
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,768,16384,0.21105066935221353
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,1024,65536,0.2414741357167562
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,768,12288,0.045560534795125326
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,768,12288,0.15927252769470215
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,768,16384,0.07370666662851968
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,768,10240,0.039190399646759036
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,768,10240,0.13012800216674805
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,768,10240,0.0542464017868042
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,768,8192,0.03163626591364543
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,768,8192,0.09851840337117514
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,768,12288,0.05933546622594198
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,768,7168,0.027057067553202314
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,768,7168,0.08368213176727295
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,768,8192,0.0437450647354126
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,768,6144,0.024025599161783852
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,768,6144,0.06852693557739258
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,768,7168,0.039862398306528726
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,768,5120,0.020332799355189005
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,768,5120,0.05540586709976196
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,768,6144,0.03622293472290039
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,768,5120,0.03334933519363403
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,768,4096,0.01712533235549927
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,768,4096,0.04605973164240519
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,768,3584,0.015414399902025857
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,768,3584,0.043194667498270674
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,768,3584,0.026397866010665894
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,768,3072,0.013742933670679728
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,768,4096,0.028579199314117433
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,768,3072,0.03795733451843262
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,768,2560,0.011982933680216471
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,768,2560,0.0350602666536967
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,768,3072,0.023959465821584067
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,768,2560,0.022357332706451415
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,768,2048,0.030957865715026855
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,768,2048,0.019562667608261107
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,768,1536,0.008391466736793519
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,768,2048,0.009810133775075277
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,768,1536,0.02861013412475586
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,768,1024,0.00673173318306605
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,768,1024,0.023246934016545616
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,768,1536,0.017395200332005818
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,768,768,0.005942399799823761
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,768,1024,0.015343999862670899
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,768,768,0.023804799715677897
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,768,512,0.005050666630268097
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,768,512,0.020566399892171225
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,768,768,0.01492586632569631
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,768,256,0.003835733234882355
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,768,512,0.015292800466219583
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,768,256,0.018934400876363118
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,768,128,0.0034186666210492453
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,768,51200,0.19518826802571615
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,768,256,0.016156799594561257
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,768,64,0.003401600072781245
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,768,128,0.01622719963391622
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,768,64,0.01667520006497701
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,768,32,0.004624000191688538
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,768,32,0.01651093363761902
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,512,65536,0.17150400479634603
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,512,65536,0.7803040186564127
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,512,51200,0.13583359718322754
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,768,128,0.01725546717643738
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,512,51200,0.6160618464152019
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,512,16384,0.046133331457773846
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,512,16384,0.20353919665018716
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,512,16384,0.07171093622843425
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,512,12288,0.03533759911855062
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,512,12288,0.15373867352803547
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,768,65536,0.24210559527079262
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,512,10240,0.02958826621373494
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,512,10240,0.1251744031906128
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,512,12288,0.05723626613616943
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,512,8192,0.02288320064544678
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,512,8192,0.09289173285166422
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,512,10240,0.05236800114313761
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,512,7168,0.019747199614842732
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,512,7168,0.07729067007700602
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,512,8192,0.04092160065968831
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,512,6144,0.0171615997950236
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,512,6144,0.06253973245620728
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,512,7168,0.037178667386372884
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,512,5120,0.014777599771817525
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,512,5120,0.051481600602467856
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,512,6144,0.033445334434509276
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,512,4096,0.01258666714032491
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,512,5120,0.03171093265215556
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,512,4096,0.04174933433532715
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,512,51200,0.1934890588124593
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,512,3584,0.011320533355077107
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,512,3584,0.04064639806747437
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,512,3072,0.010221866766611735
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,512,3072,0.035300266742706296
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,512,4096,0.02680319945017497
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,512,2560,0.008775466680526733
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,512,3584,0.024395734071731567
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,512,2560,0.03330559929211934
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,512,2048,0.007602133353551229
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,512,2048,0.02911146680514018
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,512,3072,0.02186773419380188
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,512,1536,0.006457599997520447
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,512,2560,0.02059946656227112
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,512,1536,0.026895999908447266
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,512,1024,0.005190399785836538
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,512,1024,0.022721066077550253
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,512,1536,0.015761066476504007
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,512,768,0.004713599880536398
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,512,1024,0.0136245330174764
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,512,2048,0.018067200978597008
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,512,768,0.022805333137512207
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,512,512,0.003881600002447764
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,512,512,0.020660267273585
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,512,256,0.0034346667428811393
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,512,768,0.013265066345532737
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,512,256,0.01853546698888143
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,512,128,0.0030122667551040648
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,512,128,0.016613333423932394
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,512,512,0.014107732971509298
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,512,64,0.0030026666820049284
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,512,128,0.014071466525395713
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,512,32,0.003011200080315272
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,512,64,0.015128533045450846
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,512,32,0.014774399995803832
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,256,65536,0.11085226535797119
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,256,65536,0.7464458465576171
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,256,51200,0.084934401512146
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,512,256,0.013621333241462707
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,256,51200,0.588151486714681
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,256,16384,0.03672320048014323
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,512,65536,0.2407925287882487
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,256,16384,0.19920746485392254
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,256,12288,0.026632533470789595
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,256,12288,0.1496010621388753
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,256,12288,0.05698879957199097
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,256,10240,0.02066133419672648
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,256,16384,0.07165013154347738
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,256,10240,0.12208426793416341
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,256,8192,0.01737173398335775
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,256,8192,0.08870186805725097
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,256,10240,0.051958398024241126
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,256,8192,0.04021439949671428
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,256,7168,0.014334932963053385
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,256,7168,0.07240320046742757
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,256,6144,0.012900267044703165
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,256,6144,0.05774720112482706
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,256,7168,0.036432000001271565
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,256,5120,0.010854400197664897
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,256,5120,0.046122666200002035
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,256,6144,0.03336853186289469
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,256,4096,0.00918506681919098
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,256,4096,0.04006719986597697
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,256,5120,0.03107946713765462
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,256,3584,0.008335999647776286
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,256,4096,0.026743467648824053
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,256,3584,0.03724693457285563
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,256,3072,0.007853866616884867
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,256,3072,0.033268266916275026
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,256,3584,0.024378667275110878
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,256,51200,0.1934698740641276
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,256,2560,0.006726400057474773
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,256,2560,0.031032532453536987
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,256,3072,0.021884800990422566
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,256,2048,0.005914666752020518
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,256,2048,0.027005867163340254
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,256,1536,0.005427200098832448
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,256,1536,0.02564053336779277
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,256,2560,0.020270933707555137
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,256,1024,0.004387199878692627
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,256,2048,0.017754666010538735
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,256,1024,0.022533333301544188
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,256,768,0.003899733225504557
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,256,1536,0.015288533767064414
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,256,768,0.022499199708302817
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,256,512,0.003369600077470144
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,256,1024,0.013683199882507324
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,256,512,0.01976319948832194
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,256,768,0.013026133179664612
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,256,256,0.0030570665995279947
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,256,256,0.017127466201782227
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,256,512,0.01323946714401245
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,256,128,0.002571733295917511
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,256,256,0.013510400056838989
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,256,128,0.016457600394884746
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,256,64,0.0025727999707063037
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,256,32,0.002611200014750163
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,256,64,0.01527466674645742
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,256,32,0.01490133305390676
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,128,65536,0.09881599744160971
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,256,65536,0.24044373830159507
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,128,65536,0.7413397471110026
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,128,51200,0.07229973475138346
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,256,128,0.013718400398890176
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,128,51200,0.5816266377766927
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,128,16384,0.029700267314910888
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,128,16384,0.19474347432454425
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,128,12288,0.02097066640853882
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,128,12288,0.1451466719309489
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,128,16384,0.07044266859690348
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,128,10240,0.014881066481272378
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,128,10240,0.11791146596272786
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,128,12288,0.05579200188318888
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,128,8192,0.012503467003504434
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,128,8192,0.08512960275014242
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,128,10240,0.04912639856338501
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,128,7168,0.011012267072995503
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,128,7168,0.0693557341893514
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,128,8192,0.03898880084355672
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,128,6144,0.009581866860389709
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,128,51200,0.19022506078084309
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,128,7168,0.03541333278020223
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,128,6144,0.05462186733881632
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,128,5120,0.008653866251309712
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,128,5120,0.04524373213450114
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,128,4096,0.007156266768773396
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,128,4096,0.03883626858393351
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,128,6144,0.03211946686108907
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,128,3584,0.006751999755700429
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,128,5120,0.030115199089050294
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,128,65536,0.23648319244384766
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,128,3584,0.03701440095901489
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,128,3072,0.006347733239332835
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,128,4096,0.025527467330296833
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,128,3072,0.03288319905598958
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,128,2560,0.005437866846720377
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,128,2560,0.030833067496617635
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,128,2048,0.004702933132648468
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,128,2048,0.026768000920613606
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,128,2560,0.01945919990539551
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,128,1536,0.0042357335488001505
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,128,1536,0.024732800324757893
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,128,3072,0.021208532651265464
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,128,2048,0.016876800855000814
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,128,1024,0.0038250667353471124
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,128,1024,0.02146986722946167
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,128,768,0.0034133332471052804
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,128,1536,0.014419200023015341
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,128,768,0.022125866015752158
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,128,512,0.003020799905061722
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,128,1024,0.012903466820716858
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,128,768,0.012055466572443645
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,128,256,0.00258240004380544
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,128,256,0.017314134041468303
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,128,512,0.01248426636060079
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,128,256,0.012421333789825439
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,128,128,0.0026687999566396077
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,128,3584,0.023880533377329507
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,128,128,0.015873066584269204
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,128,64,0.0021930667261282604
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,128,32,0.0025600001215934755
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,128,64,0.01446399986743927
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,128,32,0.015240533153216043
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,4096,128,128,0.01272213359673818
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,64,65536,0.09346026579538981
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,64,51200,0.06967360178629557
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,64,16384,0.026268800099690754
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,64,16384,0.19386879603068036
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,64,51200,0.5786378860473633
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,128,512,0.018607999881108603
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,64,12288,0.017485866943995156
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,64,65536,0.7395264307657878
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,64,10240,0.014059733351071677
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,64,8192,0.011672533551851908
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,64,12288,0.14484160741170246
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,64,7168,0.010537599523862202
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,64,8192,0.08434133529663086
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,64,10240,0.11648106575012207
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,64,6144,0.05473493337631226
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,64,5120,0.0446943998336792
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,64,4096,0.006695466736952464
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,64,4096,0.038466131687164305
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,64,3584,0.006302933394908905
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,64,3584,0.03710720141728719
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,64,3072,0.005860266586144766
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,64,3072,0.03292160034179688
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,64,2560,0.005060266455014547
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,64,2560,0.030929066737492877
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,64,2048,0.004613333443800608
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,64,2048,0.0269813338915507
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,64,1536,0.0037952000896135964
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,64,6144,0.009176533420880635
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,64,1024,0.0034304000437259674
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,64,5120,0.00793280005455017
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,64,1024,0.0214026669661204
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,64,768,0.00311253344019254
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,64,768,0.021335466702779134
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,64,512,0.018911999464035035
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,64,1536,0.02472320000330607
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,64,256,0.0026837334036827086
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,64,256,0.01678933302561442
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,64,128,0.002164266755183538
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,64,512,0.003067733347415924
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,64,64,0.002327466756105423
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,64,128,0.015097600221633912
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,64,32,0.002183466653029124
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,64,64,0.014892799655596414
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,64,32,0.014456533392270408
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,32,65536,0.08921919663747152
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,32,51200,0.0701375961303711
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,32,65536,0.7326623916625976
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,64,7168,0.06810239950815836
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,32,51200,0.5777941385904948
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,32,16384,0.026501333713531493
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,32,12288,0.01595093309879303
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,32,16384,0.19182079633076984
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,32,10240,0.010851200421651204
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,32,12288,0.1440831979115804
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,32,8192,0.009529599547386169
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,32,10240,0.11890347003936767
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,32,7168,0.008682666222254436
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,32,6144,0.007842133442560833
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,32,8192,0.08634880383809408
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,32,5120,0.010577066739400228
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,32,4096,0.008954667051633199
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,32,6144,0.05542720158894857
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,32,3584,0.008247466882069905
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,32,5120,0.045269334316253663
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,32,4096,0.039019731680552165
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,32,3072,0.0073077330986658735
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,32,3584,0.03700586557388306
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,32,2560,0.0064735998709996535
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,32,3072,0.03381866614023844
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,32,2048,0.004610133171081543
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,32,2560,0.03096746603647868
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,32,1536,0.00384853333234787
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,32,1024,0.003483733286460241
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,32,2048,0.027270400524139406
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,32,1536,0.025043199459711712
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,32,768,0.0031413334111372627
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,32,512,0.002915200094381968
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,32,768,0.021258666117986044
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,32,256,0.002567466596762339
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,32,512,0.01913706660270691
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,32,128,0.0022698665658632917
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,32,128,0.016845866044362386
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,32,64,0.0021418665846188864
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,32,7168,0.06934506893157959
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,4096,32,32,0.002170666555563609
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,32,64,0.014849066734313965
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,32,32,0.013685333728790283
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,32,1024,0.021078399817148843
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,4096,32,256,0.017654399077097573
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,65536,16384,4.7907251993815105
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,65536,16384,2.603728993733724
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,65536,51200,8.284023539225261
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,65536,12288,3.8143305460611976
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,65536,12288,1.8974933624267578
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,65536,51200,15.941858927408854
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,65536,10240,3.268873596191406
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,65536,10240,1.5873483022054038
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,65536,10240,0.8529130935668945
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,65536,8192,2.6319552103678387
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,65536,8192,1.2778549194335938
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,65536,12288,0.986567497253418
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,65536,7168,2.2851211547851564
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,65536,7168,1.129314168294271
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,65536,16384,1.2820852915445964
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,65536,6144,1.9383829752604167
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,65536,6144,0.9072277069091796
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,65536,8192,0.707583999633789
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,65536,5120,1.5050827026367188
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,65536,5120,0.7823968251546224
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,65536,7168,0.6587061564127604
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,65536,4096,1.164696502685547
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,65536,4096,0.6445887883504231
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,65536,6144,0.5616565068562825
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,65536,3584,0.9537343978881836
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,65536,3584,0.5780127843221028
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,65536,4096,0.4561898549397786
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,65536,3072,0.8683594385782877
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,65536,3072,0.477341874440511
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,65536,5120,0.5086976051330566
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,65536,2560,0.6628672281901042
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,65536,2560,0.39213972091674804
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,65536,3584,0.4081376075744629
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,65536,2048,0.5335477193196614
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,65536,2048,0.32349866231282554
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,65536,2560,0.34098774592081704
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,65536,1536,0.43637866973876954
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,65536,1536,0.24878719647725425
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,65536,3072,0.3599221229553223
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,65536,1024,0.2860565185546875
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,65536,1024,0.19118080139160157
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,65536,1536,0.2592682679494222
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,65536,768,0.22107413609822593
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,65536,768,0.15485866864522296
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,65536,1024,0.25026027361551917
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,65536,512,0.16391785939534503
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,65536,512,0.13043733437856037
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,65536,2048,0.30454187393188475
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,65536,256,0.10867199897766114
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,65536,256,0.11255253156026204
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,65536,512,0.23871146837870277
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,65536,128,0.10308799743652344
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,65536,128,0.10143786271413166
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,65536,768,0.24157333374023438
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,65536,64,0.08644906679789224
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,65536,64,0.10794239838918049
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,65536,128,0.24123306274414064
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,65536,32,0.09290026823679606
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,65536,32,0.10678186416625976
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,65536,256,0.23909440040588378
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,51200,65536,8.67805887858073
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,51200,65536,16.728692626953126
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,51200,51200,12.669930013020833
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,51200,51200,6.672695414225261
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,51200,16384,3.99467519124349
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,51200,16384,2.0878154754638674
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,51200,16384,1.0225887934366862
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,51200,12288,2.9145098368326825
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,51200,12288,1.4887349446614584
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,51200,12288,0.7663722356160482
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,51200,10240,2.5630218505859377
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,51200,10240,1.290719985961914
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,51200,10240,0.6771221160888672
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,51200,8192,1.9902656555175782
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,51200,8192,0.9993312199910482
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,65536,51200,4.363151041666667
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,51200,7168,1.6343349456787108
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,51200,7168,0.8904725392659506
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,51200,8192,0.5541877110799154
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,51200,6144,1.3819498697916666
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,51200,6144,0.8352394739786784
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,51200,7168,0.5139424006144206
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,51200,5120,1.1804064432779948
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,51200,5120,0.6248074849446614
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,51200,5120,0.3996586799621582
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,51200,4096,0.9277983983357748
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,51200,4096,0.5068650563557943
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,51200,6144,0.43990612030029297
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,51200,3584,0.7724096298217773
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,51200,3584,0.45223251978556317
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,51200,51200,3.3974177042643228
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,51200,3072,0.6580064137776692
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,51200,3072,0.3904885292053223
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,51200,4096,0.35490986506144206
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,51200,2560,0.5520479838053386
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,51200,2560,0.33095146814982096
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,51200,3584,0.3171754519144694
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,51200,2048,0.4405152002970378
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,51200,2048,0.2673952102661133
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,51200,2048,0.23979093233744303
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,51200,1536,0.33665278752644856
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,51200,1536,0.22608213424682616
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,51200,3072,0.28192640940348307
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,51200,1024,0.2363488038380941
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,51200,1024,0.1687605381011963
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,51200,2560,0.2691669464111328
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,51200,768,0.198308261235555
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,51200,768,0.1403626600901286
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,51200,768,0.1899679978688558
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,51200,512,0.14476693471272786
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,51200,512,0.11203093528747558
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,51200,1024,0.2037226676940918
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,51200,256,0.09439146518707275
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,51200,256,0.09133546352386475
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,51200,512,0.19013439814249675
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,51200,128,0.0757258653640747
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,51200,128,0.0824885368347168
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,51200,256,0.18973546028137206
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,51200,64,0.06736533641815186
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,51200,64,0.08651519616444905
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,51200,32,0.07109866937001547
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,51200,32,0.0842847983042399
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,51200,128,0.19220800399780275
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,51200,1536,0.20797653198242189
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,16384,65536,4.779863484700521
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,16384,65536,2.887309773763021
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,16384,51200,2.2762377421061197
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,16384,51200,3.7496416727701822
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,16384,16384,1.1165162404378255
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,16384,16384,0.7597674687703451
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,16384,16384,0.37458985646565757
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,16384,12288,0.8216010411580404
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,16384,12288,0.5564490636189778
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,16384,12288,0.28357973098754885
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,16384,10240,0.6774805068969727
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,16384,10240,0.4615594546000163
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,16384,10240,0.25253440539042155
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,16384,8192,0.5477941513061524
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,16384,8192,0.37760534286499026
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,16384,8192,0.20752213795979818
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,16384,7168,0.4831690788269043
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,16384,7168,0.33128639856974285
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,16384,7168,0.19300907452901203
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,16384,51200,1.1594272613525392
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,16384,6144,0.412667719523112
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,16384,6144,0.27817920049031575
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,16384,5120,0.347657585144043
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,16384,5120,0.2467701276143392
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,16384,5120,0.15285654067993165
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,16384,4096,0.2713695844014486
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,16384,4096,0.1890058676401774
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,16384,6144,0.1683989365895589
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,16384,3584,0.24160213470458985
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,16384,3584,0.1665023962656657
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,16384,3584,0.11982080141703289
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,16384,4096,0.1328778664271037
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,16384,3072,0.20899413426717123
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,16384,3072,0.14726719856262208
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,16384,2560,0.17259413401285809
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,16384,2560,0.1251893361409505
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,51200,65536,4.360603841145833
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,16384,2048,0.1371840000152588
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,16384,2048,0.1019210656483968
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,16384,2560,0.1014954646428426
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,16384,1536,0.10698026816050213
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,16384,1536,0.08234879970550538
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,16384,3072,0.10511573155721028
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,16384,1024,0.07401493390401205
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,16384,1024,0.06320426861445108
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,16384,2048,0.08827839692433676
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,16384,768,0.057182931900024415
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,16384,768,0.058772265911102295
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,16384,768,0.06655253171920776
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,16384,512,0.04235519965489705
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,16384,512,0.05025920073191324
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,16384,1536,0.07703786691029867
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,16384,256,0.031223465998967487
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,16384,256,0.05525439977645874
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,16384,1024,0.07138773600260416
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,16384,128,0.025166932741800947
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,16384,128,0.03636480172475179
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,16384,128,0.05959039926528931
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,16384,64,0.021772799889246623
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,16384,256,0.05883626540501913
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,16384,64,0.035524264971415205
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,16384,32,0.02223680019378662
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,16384,32,0.035317333539326985
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,16384,512,0.06244479815165201
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,12288,65536,2.310004170735677
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,12288,65536,3.6126358032226564
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,12288,51200,2.7388565063476564
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,12288,51200,1.821396255493164
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,12288,16384,0.8194976170857748
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,12288,16384,0.5855050404866536
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,16384,65536,1.4740511576334634
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,12288,12288,0.5986623764038086
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,12288,12288,0.43725013732910156
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,12288,16384,0.2875594774881999
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,12288,10240,0.5676821390787761
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,12288,10240,0.40831785202026366
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,12288,12288,0.21852374076843262
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,12288,8192,0.4417450586954753
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,12288,8192,0.30110400517781577
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,12288,10240,0.19483413696289062
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,12288,7168,0.39353065490722655
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,12288,7168,0.2624255975087484
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,12288,8192,0.15874560674031574
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,12288,6144,0.3281749407450358
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,12288,6144,0.2396949291229248
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,12288,7168,0.14749867121378582
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,12288,5120,0.271943473815918
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,12288,5120,0.19552106857299806
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,12288,6144,0.1280138651529948
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,12288,4096,0.21247146924336752
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,12288,4096,0.15650986035664877
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,12288,5120,0.11678613026936849
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,12288,3584,0.18997227350870768
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,12288,3584,0.13940800031026204
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,12288,4096,0.10260586738586426
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,12288,3072,0.16087147394816081
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,12288,3072,0.121561598777771
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,12288,3584,0.09147626558939616
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,12288,2560,0.13747413953145343
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,12288,2560,0.10441919962565105
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,12288,3072,0.08231680393218994
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,12288,2048,0.10996159712473552
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,12288,2048,0.08680426279703776
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,12288,2560,0.07893439928690592
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,12288,1536,0.08547306855519612
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,12288,1536,0.07179306348164877
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,12288,1536,0.06424959897994995
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,12288,1024,0.059825066725413004
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,12288,1024,0.05714240074157715
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,12288,2048,0.07033812999725342
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,12288,768,0.04681280056635539
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,12288,768,0.053333334128061925
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,12288,768,0.05728213389714559
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,12288,512,0.03548159996668498
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,12288,512,0.04522133270899455
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,12288,1024,0.06184213161468506
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,12288,256,0.02500159939130147
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,12288,256,0.03549439907073974
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,12288,256,0.057530665397644044
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,12288,128,0.01962453325589498
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,12288,128,0.03097599943478902
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,12288,512,0.05756800174713135
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,12288,64,0.01683626572291056
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,12288,64,0.03057813247044881
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,12288,32,0.01713706652323405
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,12288,32,0.030266666412353517
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,12288,128,0.057158398628234866
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,10240,65536,2.0089664459228516
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,10240,65536,2.958220926920573
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,10240,51200,2.3292724609375
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,10240,51200,1.6049322764078777
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,12288,51200,0.860540771484375
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,10240,16384,0.6883978525797526
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,10240,16384,0.5118015925089519
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,10240,16384,0.26167999903361
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,10240,12288,0.5005407969156901
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,10240,12288,0.39283841451009116
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,12288,65536,1.0841227213541667
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,10240,10240,0.44667307535807294
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,10240,10240,0.33885014851888023
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,10240,12288,0.1987946669260661
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,10240,8192,0.36514987945556643
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,10240,8192,0.25758612950642906
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,10240,10240,0.17576853434244794
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,10240,7168,0.31437546412150064
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,10240,7168,0.22534292538960776
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,10240,8192,0.1447861353556315
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,10240,6144,0.271455987294515
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,10240,6144,0.2042304039001465
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,10240,7168,0.13339306513468424
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,10240,5120,0.21914560000101724
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,10240,5120,0.1636917273203532
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,10240,6144,0.11602986653645833
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,10240,4096,0.17775893211364746
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,10240,4096,0.1350858688354492
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,10240,51200,0.7862730662027995
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,10240,3584,0.15632319450378418
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,10240,3584,0.121124267578125
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,10240,5120,0.10601066748301188
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,10240,3072,0.1316800038019816
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,10240,3072,0.10683626333872479
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,10240,4096,0.09250986576080322
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,10240,2560,0.11218453248341878
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,10240,2560,0.09104639689127604
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,10240,3584,0.08450559775034586
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,10240,2048,0.0924021323521932
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,10240,2048,0.07877866427103677
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,10240,3072,0.07466773192087808
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,10240,1536,0.06923519770304362
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,10240,1536,0.06456000010172526
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,10240,2560,0.07291626930236816
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,10240,1024,0.049406933784484866
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,10240,1024,0.04960533380508423
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,10240,1024,0.056405333677927647
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,10240,768,0.03908160130182902
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,10240,2048,0.06567146778106689
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,10240,768,0.045651201407114664
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,10240,512,0.02890133261680603
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,10240,512,0.03931946754455566
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,10240,1536,0.05816106796264649
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,10240,256,0.02068693240483602
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,10240,256,0.034746666749318436
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,10240,512,0.05272106726964315
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,10240,128,0.016577066977818807
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,10240,128,0.027983999252319335
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,10240,256,0.05225280125935873
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,10240,64,0.01411626636981964
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,10240,64,0.027812266349792482
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,10240,32,0.014404267072677612
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,10240,32,0.027512532472610474
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,10240,128,0.05220479965209961
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,10240,768,0.05312533378601074
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,8192,65536,2.346105702718099
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,8192,65536,1.7270314534505207
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,8192,51200,1.3594154357910155
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,8192,51200,1.8280192057291669
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,8192,16384,0.5915967941284179
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,8192,16384,0.457150936126709
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,8192,16384,0.21381866137186684
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,8192,12288,0.39655679066975913
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,8192,12288,0.33651625315348305
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,10240,65536,0.9912010828653971
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,8192,10240,0.3342325210571289
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,8192,10240,0.28414932886759436
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,8192,10240,0.14526079495747884
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,8192,8192,0.2707669258117676
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,8192,8192,0.2194495995839437
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,8192,12288,0.16294827461242675
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,8192,7168,0.23424107233683267
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,8192,7168,0.19606720606486003
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,8192,7168,0.11061333020528157
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,8192,6144,0.2025514602661133
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,8192,6144,0.16347306569417316
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,8192,8192,0.11877653598785401
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,8192,5120,0.17518293062845866
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,8192,5120,0.1399509270985921
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,8192,5120,0.08833386898040771
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,8192,4096,0.1413109302520752
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,8192,4096,0.11582400004069011
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,8192,6144,0.09654080073038737
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,8192,3584,0.12268160184224446
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,8192,3584,0.10127466519673664
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,8192,3584,0.06736640135447183
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,8192,3072,0.10238080024719239
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,8192,3072,0.09401173591613769
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,8192,4096,0.07621119817097982
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,8192,2560,0.0871946652730306
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,8192,2560,0.07836373647054037
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,8192,2560,0.05642240047454834
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,8192,2048,0.07439893086751302
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,8192,2048,0.06730986436208089
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,8192,3072,0.058406400680541995
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,8192,1536,0.053559466203053796
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,8192,1536,0.05540266831715902
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,8192,1536,0.042396799723307295
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,8192,1024,0.037226665019989016
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,8192,1024,0.04328320026397705
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,8192,2048,0.049504001935323075
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,8192,768,0.029524266719818115
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,8192,768,0.040253865718841556
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,8192,768,0.03585919936498006
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,8192,1024,0.04043733278910319
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,8192,512,0.021939200162887574
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,8192,512,0.03510080178578694
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,8192,256,0.016569599509239197
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,8192,256,0.02890133261680603
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,8192,256,0.03610026836395264
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,8192,128,0.013684266805648803
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,8192,128,0.031854933500289916
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,8192,512,0.036339199542999266
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,8192,128,0.03667733271916707
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,8192,64,0.01188053290049235
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,8192,64,0.025847466786702473
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,8192,32,0.0120170662800471
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,8192,32,0.025123200813929242
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,7168,65536,1.583774948120117
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,7168,65536,1.9816469828287762
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,7168,51200,1.5816448211669922
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,7168,51200,1.246771240234375
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,8192,51200,0.6344789505004883
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,7168,16384,0.4693130811055501
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,7168,16384,0.41005652745564775
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,7168,16384,0.20488212903340658
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,7168,12288,0.35164054234822595
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,7168,12288,0.30788373947143555
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,8192,65536,0.8034058888753256
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,7168,10240,0.3083168029785156
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,7168,10240,0.26116266250610354
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,7168,12288,0.14868906339009602
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,7168,8192,0.24779839515686036
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,7168,8192,0.20289813677469887
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,7168,10240,0.13144853115081787
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,7168,7168,0.21830612818400064
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,7168,7168,0.17828267415364582
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,7168,8192,0.1061247984568278
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,7168,6144,0.18382186889648439
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,7168,6144,0.1504853407541911
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,7168,7168,0.09973333676656088
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,7168,5120,0.15291306177775066
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,7168,5120,0.12945386568705242
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,7168,5120,0.08093226750691732
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,7168,4096,0.12084480126698811
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,7168,4096,0.10643520355224609
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,7168,6144,0.08815146287282308
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,7168,3584,0.10714453061421711
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,7168,3584,0.09231999715169271
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,7168,4096,0.06969599723815918
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,7168,3072,0.09240960280100505
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,7168,3584,0.06229226589202881
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,7168,3072,0.08412586847941081
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,7168,2560,0.07643840312957764
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,7168,2560,0.07251199881235758
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,7168,2560,0.05304960012435913
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,7168,2048,0.06246933142344156
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,7168,2048,0.06202986637751261
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,7168,3072,0.055035734176635744
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,7168,1536,0.04734186728795369
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,7168,1536,0.049561599890391037
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,7168,1536,0.04152213335037232
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,7168,1024,0.03264213403065999
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,7168,1024,0.039112532138824464
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,7168,2048,0.046963198979695635
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,7168,768,0.025602134068806966
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,7168,768,0.037589331467946366
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,7168,768,0.03671893278757731
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,7168,512,0.019358932971954346
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,7168,512,0.03211200038592021
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,7168,1024,0.03949120044708252
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,7168,256,0.014111999670664468
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,7168,256,0.02664533257484436
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,7168,256,0.03672320048014323
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,7168,128,0.013265066345532737
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,7168,128,0.0247488001982371
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,7168,512,0.03697280089060466
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,7168,64,0.011572266618410747
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,7168,64,0.024650667111078897
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,7168,32,0.011677866180737812
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,7168,128,0.0370410680770874
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,7168,32,0.025014400482177734
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,6144,65536,1.4747029622395833
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,6144,65536,1.7568106333414715
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,6144,51200,1.3027904510498047
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,6144,51200,1.1453685760498047
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,7168,51200,0.5827637354532877
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,6144,16384,0.3990090688069662
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,6144,16384,0.40061333974202473
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,6144,16384,0.17267200152079265
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,6144,12288,0.3027690569559733
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,6144,12288,0.28023786544799806
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,7168,65536,0.7361536026000977
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,6144,10240,0.2520117282867432
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,6144,10240,0.23463145891825357
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,6144,10240,0.11579519907633465
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,6144,8192,0.20120320320129395
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,6144,8192,0.18146880467732746
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,6144,12288,0.12970560391743977
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,6144,7168,0.17953920364379883
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,6144,7168,0.16049599647521973
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,6144,7168,0.0893717368443807
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,6144,6144,0.1547221342722575
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,6144,6144,0.1401312033335368
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,6144,8192,0.09464213053385416
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,6144,5120,0.12833387056986492
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,6144,5120,0.11919679641723632
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,6144,5120,0.07274453639984131
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,6144,4096,0.10354133447011311
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,6144,4096,0.09965120156606039
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,6144,6144,0.07922879854838052
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,6144,3584,0.09188160101572672
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,6144,3584,0.08849706649780273
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,6144,3584,0.05538986523946127
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,6144,3072,0.0813269297281901
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,6144,3072,0.07965333461761474
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,6144,4096,0.061060265700022376
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,6144,2560,0.06798400084177653
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,6144,2560,0.0680021365483602
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,6144,3072,0.04861013491948445
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,6144,2048,0.0559445341428121
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,6144,2048,0.057911467552185056
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,6144,2560,0.046521600087483725
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,6144,1536,0.04267093340555827
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,6144,1536,0.0477567990620931
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,6144,1536,0.03743253151575725
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,6144,1024,0.029428267478942872
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,6144,1024,0.03909333149592082
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,6144,2048,0.041623465220133465
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,6144,768,0.023740800221761067
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,6144,768,0.03569066524505615
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,6144,1024,0.03502080043156942
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,6144,512,0.018051199118296304
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,6144,512,0.03086613416671753
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,6144,768,0.03291306694348653
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,6144,256,0.013040000200271606
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,6144,256,0.02644053300221761
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,6144,512,0.03336853186289469
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,6144,128,0.011917866269747416
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,6144,128,0.024571732680002848
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,6144,256,0.03260693351427714
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,6144,64,0.01039573351542155
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,6144,64,0.023886932929356893
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,6144,32,0.010405332843462626
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,6144,128,0.03294080098470052
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,6144,32,0.02320746580759684
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,5120,65536,1.564777628580729
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,5120,65536,1.3959573109944663
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,5120,51200,1.1900010426839194
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,5120,51200,1.0981578826904297
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,6144,51200,0.49969065984090166
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,5120,16384,0.37652053833007815
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,5120,16384,0.36354347864786785
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,5120,16384,0.1637407938639323
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,5120,12288,0.28399359385172523
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,5120,12288,0.2746538798014323
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,6144,65536,0.6325045267740885
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,5120,10240,0.23951892852783202
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,5120,10240,0.22309226989746095
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,5120,12288,0.12334187030792236
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,5120,8192,0.1894005298614502
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,5120,8192,0.17568000157674152
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,5120,10240,0.10994346936543782
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,5120,7168,0.1648416042327881
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,5120,7168,0.15413227081298828
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,5120,8192,0.08945813179016113
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,5120,6144,0.13813652992248535
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,5120,6144,0.1317951997121175
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,5120,7168,0.08409386475880941
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,5120,5120,0.1160703976949056
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,5120,5120,0.114356263478597
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,5120,5120,0.06669013500213623
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,5120,4096,0.09685226281483969
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,5120,4096,0.09610026677449544
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,5120,6144,0.0740885337193807
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,5120,3584,0.08450986544291178
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,5120,3584,0.08358613650004068
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,5120,3584,0.0511029322942098
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,5120,3072,0.07340799967447917
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,5120,3072,0.07534613609313964
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,5120,4096,0.056397867202758786
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,5120,2560,0.06220159928003947
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,5120,2560,0.06427626609802246
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,5120,3072,0.045723732312520346
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,5120,2560,0.04391680161158244
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,5120,2048,0.050635735193888344
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,5120,2048,0.05417173306147257
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,5120,1536,0.03856000105539958
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,5120,1536,0.044869331518809
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,5120,1536,0.034679468472798666
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,5120,1024,0.027211733659108478
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,5120,1024,0.03686293363571167
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,5120,2048,0.03935146729151408
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,5120,768,0.022316799561182658
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,5120,768,0.033149866263071696
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,5120,768,0.03091413378715515
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,5120,512,0.016778665781021117
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,5120,512,0.02911466757456462
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,5120,1024,0.03277653257052104
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,5120,256,0.0116565336783727
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,5120,256,0.02458239992459615
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,5120,256,0.031036800146102904
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,5120,128,0.008809600273768108
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,5120,512,0.03153813282648722
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,5120,128,0.020957867304484047
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,5120,51200,0.487718391418457
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,5120,64,0.007981866598129272
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,5120,64,0.02081706722577413
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,5120,32,0.008543999989827473
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,5120,32,0.020477867126464842
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,5120,128,0.03133760094642639
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,4096,65536,1.0617247899373372
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,4096,65536,1.1869173685709635
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,4096,51200,0.840550422668457
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,4096,51200,0.9327221552530924
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,4096,16384,0.26643733978271483
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,4096,16384,0.3059562683105469
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,4096,16384,0.140065066019694
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,4096,12288,0.20169599850972494
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,4096,12288,0.22959252993265786
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,5120,65536,0.6177663803100586
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,4096,10240,0.16563520431518555
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,4096,10240,0.18820373217264813
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,4096,12288,0.10806079705556233
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,4096,8192,0.13640534083048503
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,4096,8192,0.14842346509297688
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,4096,10240,0.09675626754760742
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,4096,7168,0.11788907051086425
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,4096,7168,0.13095146814982098
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,4096,7168,0.07484266757965088
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,4096,6144,0.10175466537475586
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,4096,6144,0.11557439963022868
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,4096,8192,0.07923946380615235
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,4096,5120,0.08664960066477458
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,4096,5120,0.09957653681437174
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,4096,5120,0.059444268544514976
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,4096,4096,0.0698527971903483
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,4096,4096,0.08285760084788005
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,4096,6144,0.06631893316904704
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,4096,3584,0.06118293205897013
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,4096,3584,0.07291626930236816
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,4096,3584,0.04501013358434041
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,4096,3072,0.05321066776911417
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,4096,3072,0.06412693262100219
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,4096,4096,0.04986240069071452
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,4096,2560,0.04554133415222168
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,4096,2560,0.053396264712015785
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,4096,2560,0.0375765323638916
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,4096,2048,0.037804798285166426
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,4096,2048,0.04716586669286092
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,4096,3072,0.03910613457361857
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,4096,1536,0.028282666206359865
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,4096,1536,0.04117653369903564
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,4096,1536,0.02810879945755005
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,4096,1024,0.02104426622390747
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,4096,1024,0.032620799541473386
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,4096,2048,0.03302399913469951
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,4096,768,0.017199999094009398
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,4096,768,0.03183466593424479
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,4096,1024,0.026679466168085735
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,4096,512,0.01218666632970174
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,4096,512,0.029286400477091475
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,4096,768,0.0234826664129893
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,4096,256,0.011665067076683045
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,4096,256,0.02472213308016459
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,4096,256,0.02390186587969462
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,4096,128,0.007969066500663757
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,4096,128,0.02064639925956726
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,4096,128,0.023946666717529298
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,4096,64,0.0066890666882197065
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,4096,64,0.019630932807922365
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,4096,32,0.00714026689529419
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,4096,32,0.020321067174275717
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,4096,512,0.023947733640670776
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,3584,65536,1.0047285079956054
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,3584,65536,1.1301386515299479
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,3584,51200,0.7415679931640625
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,3584,51200,0.8892960230509439
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,4096,51200,0.41303040186564133
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,3584,16384,0.24313066800435384
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,3584,16384,0.2936863899230957
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,3584,16384,0.11899519761403402
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,3584,12288,0.1835840066274007
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,3584,12288,0.22199573516845703
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,4096,65536,0.5222325325012207
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,3584,10240,0.1565077304840088
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,3584,10240,0.17990934054056804
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,3584,10240,0.0844661315282186
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,3584,8192,0.12532693545023602
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,3584,8192,0.13892159461975098
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,3584,12288,0.09320212999979655
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,3584,7168,0.10833813349405925
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,3584,7168,0.12525866826375326
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,3584,7168,0.06619733174641927
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,3584,6144,0.09208319981892904
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,3584,6144,0.1067733367284139
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,3584,8192,0.07040106455485026
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,3584,5120,0.07958613236745199
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,3584,5120,0.09173440138498942
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,3584,5120,0.05292479991912842
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,3584,4096,0.06144640048344931
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,3584,4096,0.07565226554870605
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,3584,6144,0.058457601070404056
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,3584,3584,0.05514133373896281
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,3584,3584,0.06520746548970541
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,3584,3584,0.04088426828384399
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,3584,3072,0.045620266596476236
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,3584,3072,0.05813759962717692
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,3584,4096,0.04487573305765788
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,3584,2560,0.040141868591308597
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,3584,2560,0.0492298682530721
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,3584,2560,0.03543039957682292
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,3584,2048,0.0319925328095754
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,3584,2048,0.043085865179697674
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,3584,3072,0.03702720006306966
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,3584,1536,0.025931733846664428
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,3584,1536,0.03715519905090332
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,3584,1536,0.027966932455698652
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,3584,1024,0.017574399709701538
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,3584,1024,0.031062400341033934
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,3584,2048,0.0313920001188914
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,3584,768,0.0141567995150884
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,3584,768,0.02865920066833496
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,3584,768,0.028731733560562134
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,3584,512,0.01053653359413147
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,3584,1024,0.026311467091242473
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,3584,512,0.024964267015457155
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,3584,256,0.00791786660750707
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,3584,256,0.022964266935984294
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,3584,256,0.028613332907358808
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,3584,128,0.007502933343251546
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,3584,51200,0.3351061185201009
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,3584,128,0.020222934087117513
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,3584,64,0.006691200037797292
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,3584,64,0.02107306718826294
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,3584,32,0.0071168000499407455
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,3584,32,0.019501866896947224
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,3584,128,0.027353600660959883
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,3584,512,0.028653866052627562
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,3072,65536,0.8389706929524741
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,3072,65536,1.0577866872151693
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,3072,51200,0.6974357604980469
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,3072,51200,0.8335061391194662
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,3072,16384,0.2044821262359619
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,3072,16384,0.27321386337280273
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,3072,16384,0.1175989309946696
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,3072,12288,0.15490026473999025
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,3072,12288,0.2069632053375244
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,3584,65536,0.4211562792460124
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,3072,10240,0.1451466719309489
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,3072,10240,0.17336853345235187
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,3072,12288,0.0916757345199585
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,3072,8192,0.1148085355758667
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,3072,8192,0.13700906435648602
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,3072,10240,0.08304853439331054
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,3072,7168,0.10067946910858154
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,3072,7168,0.12104106744130452
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,3072,7168,0.06500053405761719
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,3072,6144,0.08695573012034098
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,3072,6144,0.10410772959391276
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,3072,8192,0.06909653345743814
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,3072,5120,0.07562879721323648
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,3072,5120,0.08987733523050943
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,3072,5120,0.051592532793680826
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,3072,4096,0.05729813178380331
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,3072,4096,0.07244799931844076
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,3072,6144,0.05687146584192911
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,3072,3584,0.05029546817143758
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,3072,3584,0.06156266530354818
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,3072,3584,0.04012586673100789
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,3072,3072,0.044625067710876466
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,3072,3072,0.054136534531911217
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,3072,4096,0.04403306643168132
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,3072,2560,0.03780693213144938
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,3072,2560,0.04618560075759888
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,3072,3072,0.03596373399098714
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,3072,2048,0.03125653266906738
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,3072,2048,0.040433065096537275
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,3072,2560,0.0341536005338033
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,3072,1536,0.022923733790715536
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,3072,1536,0.035076268513997394
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,3072,1536,0.02762453357378642
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,3072,1024,0.015060266852378846
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,3072,1024,0.029982932408650714
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,3072,2048,0.030872533718744915
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,3072,1024,0.025677865743637084
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,3072,768,0.012241066495577494
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,3072,768,0.0268832008043925
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,3072,512,0.009958400328954061
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,3072,512,0.024651734034220378
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,3072,512,0.02847786744435628
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,3072,256,0.007605333129564922
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,3072,768,0.028084266185760497
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,3072,256,0.02826026678085327
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,3072,128,0.006759466727574666
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,3072,128,0.01918506622314453
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,3072,128,0.025015467405319215
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,3072,256,0.020862932999928793
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,3072,64,0.006293333570162455
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,3072,32,0.006333866715431213
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,3072,64,0.02003306746482849
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,3072,32,0.018641066551208497
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,2560,65536,0.7125834782918294
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,2560,65536,0.9941919962565103
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,2560,51200,0.5314933458964031
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,3072,51200,0.3342709223429362
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,2560,51200,0.8136128107706705
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,2560,16384,0.17329707145690917
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,2560,16384,0.2567829291025797
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,2560,16384,0.11379413604736328
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,2560,12288,0.13123093446095785
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,2560,12288,0.19345706303914387
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,3072,65536,0.4197706540425618
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,2560,10240,0.11465280055999756
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,2560,10240,0.16325759887695312
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,2560,12288,0.08815039793650309
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,2560,8192,0.09232213497161865
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,2560,8192,0.12541013558705646
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,2560,10240,0.07952533562978109
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,2560,7168,0.08112959861755371
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,2560,7168,0.11364266872406006
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,2560,8192,0.06540373166402182
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,2560,6144,0.06939840316772461
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,2560,6144,0.09451306660970052
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,2560,7168,0.05997226635615031
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,2560,5120,0.05931413173675537
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,2560,5120,0.08160106341044107
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,2560,6144,0.05278720060984293
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,2560,4096,0.04698559840520223
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,2560,5120,0.04750933249791463
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,2560,4096,0.06504640181859335
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,2560,3584,0.0404639999071757
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,2560,3584,0.057997866471608484
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,2560,3584,0.03582293192545573
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,2560,3072,0.035010135173797606
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,2560,3072,0.05149439970652262
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,2560,4096,0.03987840016682943
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,2560,2560,0.03171520034472148
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,2560,2560,0.045337601502736406
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,2560,3072,0.031753599643707275
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,2560,2048,0.02471359968185425
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,2560,2048,0.040148266156514484
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,2560,2560,0.03009919921557109
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,2560,1536,0.02132800022761027
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,2560,1536,0.034698665142059326
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,2560,2048,0.02635306715965271
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,2560,1024,0.014566399653752646
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,2560,1024,0.029128533601760865
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,2560,1024,0.02111039956410726
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,2560,768,0.012060800194740295
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,2560,768,0.027221333980560303
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,2560,1536,0.022664533058802287
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,2560,512,0.009649067123730978
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,2560,512,0.024264534314473472
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,2560,768,0.019447465737660728
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,2560,256,0.007122133175532024
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,2560,256,0.021619200706481934
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,2560,51200,0.32846508026123045
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,2560,128,0.0063381334145863845
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,2560,256,0.020010666052500407
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,2560,128,0.018263467152913413
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,2560,64,0.005436799923578898
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,2560,64,0.018053332964579262
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,2560,32,0.0055178667108217875
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,2560,128,0.019781333208084107
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,2560,32,0.01678826610247294
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,2560,512,0.019939200083414713
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,2048,65536,0.5088245391845703
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,2048,51200,0.41082026163736984
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,2048,65536,0.9170645395914713
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,2048,51200,0.7167711893717448
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,2048,16384,0.13061333497365316
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,2048,16384,0.2363978703816732
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,2048,16384,0.09297813574473063
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,2048,12288,0.09920746485392253
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,2048,12288,0.17992746035257975
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,2560,65536,0.4149482727050781
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,2048,10240,0.08424106438954672
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,2048,10240,0.1468949317932129
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,2048,12288,0.07313386599222818
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,2048,8192,0.06746346950531006
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,2048,8192,0.11539413134257
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,2048,10240,0.06674453417460123
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,2048,7168,0.060190931955973304
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,2048,7168,0.09973866939544677
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,2048,8192,0.055333332220713294
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,2048,6144,0.05157013336817423
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,2048,6144,0.08697600364685058
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,2048,7168,0.05060693422953287
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,2048,5120,0.043686401844024655
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,2048,5120,0.07183893521626791
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,2048,6144,0.044726399580637614
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,2048,4096,0.035554134845733644
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,2048,4096,0.05796053409576416
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,2048,5120,0.041247999668121337
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,2048,3584,0.03187839984893799
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,2048,3584,0.051064534982045495
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,2048,4096,0.0352288007736206
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,2048,3072,0.027799467245737713
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,2048,3072,0.046875735123952225
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,2048,3584,0.03153493404388428
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,2048,2560,0.02397759954134623
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,2048,2560,0.04129279851913452
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,2048,2560,0.02642773389816284
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,2048,2048,0.02030186653137207
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,2048,2048,0.03698773384094238
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,2048,3072,0.028753066062927247
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,2048,1536,0.015296000242233276
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,2048,1536,0.03278613289197286
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,2048,2048,0.022932267189025878
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,2048,1024,0.011549866199493409
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,2048,1024,0.028489599625269573
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,2048,1536,0.019887999693552653
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,2048,768,0.009639466802279156
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,2048,768,0.02637866735458374
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,2048,51200,0.25753599802652993
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,2048,512,0.00807360013326009
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,2048,512,0.022870399554570518
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,2048,1024,0.0182805339495341
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,2048,768,0.016555733482042947
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,2048,256,0.005901866654555003
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,2048,256,0.021383466323216756
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,2048,128,0.005123200019200643
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,2048,128,0.018639999628067016
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,2048,256,0.016960000991821288
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,2048,64,0.0046517332394917805
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,2048,512,0.01695573329925537
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,2048,32,0.005067733426888784
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,2048,128,0.01732800006866455
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,2048,64,0.016620799899101257
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,2048,32,0.017910399039586387
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,1536,65536,0.4011551856994629
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,1536,51200,0.31475518544514974
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,1536,65536,0.853382428487142
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,1536,51200,0.6719872156778972
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,1536,16384,0.10341760317484539
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,1536,16384,0.22358080546061196
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,1536,16384,0.09441706339518229
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,1536,12288,0.0788383960723877
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,1536,12288,0.1703839937845866
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,2048,65536,0.3213578542073568
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,1536,10240,0.0697429339090983
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,1536,10240,0.1387605349222819
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,1536,12288,0.07478293577829996
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,1536,8192,0.0547871987024943
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,1536,8192,0.10580693085988362
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,1536,10240,0.06756906509399414
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,1536,7168,0.0478986660639445
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,1536,7168,0.09292799631754557
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,1536,8192,0.055927467346191403
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,1536,6144,0.039961600303649904
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,1536,6144,0.07799786726633708
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,1536,7168,0.05179413159688314
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,1536,5120,0.03436799844106038
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,1536,5120,0.06475306749343872
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,1536,6144,0.04529813528060913
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,1536,4096,0.02672106623649597
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,1536,4096,0.05201706488927206
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,1536,5120,0.04201493263244629
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,1536,3584,0.025080533822377522
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,1536,3584,0.04694826602935791
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,1536,4096,0.03625919818878174
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,1536,3072,0.020932267109553017
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,1536,3072,0.04328960180282593
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,1536,3584,0.032858665784200033
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,1536,2560,0.01914560000101725
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,1536,2560,0.03851199944814046
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,1536,3072,0.030054400364557903
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,1536,2048,0.014825600385665893
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,1536,2048,0.03362346490224202
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,1536,2560,0.028394667307535808
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,1536,1536,0.012160000205039979
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,1536,1536,0.03075946569442749
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,1536,2048,0.025290666023890178
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,1536,1024,0.008785067001978557
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,1536,1024,0.025130667289098102
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,1536,1536,0.022323199113210044
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,1536,768,0.007478400071461995
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,1536,768,0.024762666225433348
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,1536,1024,0.020551466941833497
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,1536,512,0.00588266650835673
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,1536,512,0.02219946583112081
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,1536,51200,0.2587519963582357
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,1536,256,0.005134933193524678
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,1536,768,0.019679999351501463
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,1536,256,0.019716266791025797
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,1536,512,0.020020266373952232
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,1536,128,0.004752000172932943
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,1536,128,0.017008000612258913
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,1536,256,0.01986666719118754
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,1536,128,0.02130026618639628
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,1536,64,0.004413866500059763
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,1536,32,0.004649599889914194
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,1536,64,0.017015467087427773
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,1536,32,0.016634666919708253
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,1024,65536,0.31830291748046874
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,1024,65536,0.8052117029825846
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,1024,51200,0.25012480417887367
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,1024,51200,0.6390378952026368
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,1024,16384,0.08365546862284343
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,1024,16384,0.21359465916951498
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,1024,16384,0.0702677329381307
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,1024,12288,0.06382400194803874
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,1024,12288,0.16108907063802083
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,1536,65536,0.3250464121500651
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,1024,10240,0.054264533519744876
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,1024,10240,0.13161919911702474
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,1024,12288,0.055530667304992676
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,1024,8192,0.04319253365198771
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,1024,10240,0.05110506614049276
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,1024,8192,0.09917973677317302
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,1024,7168,0.03679360151290893
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,1024,7168,0.0841375986735026
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,1024,7168,0.03710826635360718
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,1024,6144,0.03121280074119568
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,1024,6144,0.06898773511250814
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,1024,8192,0.040779733657836915
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,1024,5120,0.025587199131647746
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,1024,5120,0.056814932823181154
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,1024,5120,0.031124265988667806
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,1024,6144,0.03300586740175883
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,1024,4096,0.021342933177947998
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,1024,4096,0.047386666138966874
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,1024,3584,0.01865066687266032
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,1024,3584,0.04166080156962077
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,1024,3584,0.02478933334350586
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,1024,3072,0.016591999928156534
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,1024,3072,0.039028267065684005
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,1024,4096,0.027203200260798137
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,1024,2560,0.013883733749389648
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,1024,2560,0.0352512001991272
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,1024,2560,0.020601600408554077
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,1024,2048,0.011889066298802693
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,1024,3072,0.022265599171320597
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,1024,2048,0.031515733400980635
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,1024,1536,0.009329066673914591
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,1024,1536,0.028632533550262452
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,1024,1536,0.01573973298072815
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,1024,1024,0.007516799867153168
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,1024,2048,0.018153599898020425
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,1024,1024,0.02468693256378174
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,1024,768,0.006517333288987477
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,1024,768,0.023373866081237794
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,1024,1024,0.014467199643452963
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,1024,512,0.005452799797058106
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,1024,768,0.013241599996884665
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,1024,512,0.02111146648724874
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,1024,256,0.004646400113900503
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,1024,256,0.018343466520309448
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,1024,512,0.014165332913398743
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,1024,128,0.003866666555404663
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,1024,256,0.014126933614412942
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,1024,128,0.016636799772580466
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,1024,64,0.003421866645415624
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,1024,128,0.014957867066065469
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,1024,64,0.015672533710797628
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,1024,32,0.0038015998899936674
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,1024,32,0.014698666334152222
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,768,65536,0.22008105913798012
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,1024,51200,0.18522559801737468
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,768,51200,0.17074880599975586
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,768,65536,0.7701109568277995
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,768,51200,0.6090186436971028
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,768,16384,0.05568000078201294
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,768,16384,0.19633280436197917
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,768,16384,0.07078080177307129
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,768,12288,0.042200533548990886
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,768,12288,0.14924052556355794
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,1024,65536,0.22938559850056967
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,768,10240,0.03520426750183105
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,768,10240,0.12286933263142903
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,768,12288,0.0570090651512146
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,768,8192,0.027939200401306152
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,768,10240,0.05234453280766806
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,768,8192,0.09281280040740966
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,768,7168,0.024258132775624594
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,768,7168,0.0795039971669515
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,768,7168,0.037829331556955975
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,768,6144,0.020881066719690956
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,768,6144,0.06502720117568969
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,768,8192,0.04156266848246257
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,768,5120,0.01742080052693685
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,768,5120,0.05323413213094076
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,768,5120,0.03243199984232585
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,768,6144,0.034322134653727215
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,768,4096,0.014659200112024942
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,768,4096,0.04497173229853312
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,768,3584,0.013474133610725404
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,768,3584,0.0401802659034729
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,768,4096,0.0280074675877889
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,768,3072,0.011529599626859028
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,768,3072,0.03737173477808635
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,768,3584,0.02595306634902954
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,768,2560,0.010028800368309021
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,768,2560,0.03331306576728821
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,768,51200,0.18567679723103842
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,768,2048,0.008758399883906047
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,768,2048,0.030216532945632934
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,768,2560,0.02222933371861776
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,768,1536,0.007293866574764251
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,768,1536,0.028626133998235066
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,768,3072,0.023561600844065347
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,768,1024,0.005937066674232483
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,768,2048,0.01936639944712321
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,768,1024,0.024091732501983643
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,768,768,0.005137066543102265
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,768,768,0.02275200088818868
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,768,1536,0.0169813334941864
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,768,512,0.0046079998215039575
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,768,1024,0.015735466281572977
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,768,512,0.015252266327540079
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,768,256,0.0038442666331926978
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,768,768,0.015014400084813436
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,768,256,0.018423465887705485
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,768,128,0.003402666747570038
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,768,128,0.016736000776290894
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,768,256,0.015929599603017174
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,768,64,0.003409066547950109
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,768,128,0.01623466710249583
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,768,64,0.01604480048020681
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,768,32,0.003443199892838796
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,768,32,0.015703466534614564
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,768,512,0.020787199338277183
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,512,65536,0.21514666875203453
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,512,51200,0.1683776060740153
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,512,65536,0.728546142578125
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,512,51200,0.5782709121704102
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,512,16384,0.053580800692240395
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,512,16384,0.19768640200297039
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,768,65536,0.2292309284210205
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,512,12288,0.04149440129597982
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,512,12288,0.14926400184631347
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,512,16384,0.06888533433278402
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,512,10240,0.0332042674223582
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,512,10240,0.12213013172149659
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,512,12288,0.05494079987208048
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,512,8192,0.028463999430338543
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,512,8192,0.09116160074869792
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,512,10240,0.0504586656888326
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,512,7168,0.025814400116602583
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,512,7168,0.07687679926554361
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,512,8192,0.03862186670303345
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,512,7168,0.03578026692072551
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,512,6144,0.0221450666586558
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,512,6144,0.06286826531092325
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,512,5120,0.01876266598701477
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,512,5120,0.05270613431930542
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,512,6144,0.03242026567459107
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,512,5120,0.030444800853729248
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,512,4096,0.014596266547838846
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,512,4096,0.043399465084075925
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,512,3584,0.013330133756001792
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,512,3584,0.039274664719899495
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,512,4096,0.02606186668078105
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,512,3584,0.023717333873113
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,512,3072,0.011241599917411804
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,512,3072,0.03704213301340739
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,512,2560,0.009661866227785747
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,512,2560,0.03303146759668986
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,512,3072,0.02185493310292562
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,512,2048,0.008257066706816356
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,512,2560,0.020257065693537392
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,512,2048,0.031173332532246904
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,512,1536,0.006951466699441274
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,512,1536,0.02727893392244975
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,512,2048,0.01776426633199056
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,512,1536,0.014891733725865683
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,512,1024,0.005494399865468343
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,512,1024,0.023652267456054688
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,512,768,0.005065600077311197
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,512,768,0.021571199099222817
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,512,1024,0.013713066776593527
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,512,512,0.004253866771856943
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,512,512,0.02062186598777771
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,512,768,0.012857600053151449
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,512,256,0.003420799970626831
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,512,256,0.018102399508158364
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,512,512,0.013699199755986533
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,512,128,0.0030165334542592366
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,512,256,0.013680000106493631
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,512,128,0.01593493322531382
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,512,64,0.002598399917284648
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,512,128,0.013684266805648803
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,512,64,0.01518186628818512
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,512,32,0.0030048000315825146
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,512,32,0.016421332955360413
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,256,65536,0.11466453075408936
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,256,65536,0.7082176208496094
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,512,51200,0.18330453236897787
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,256,51200,0.09043306509653727
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,256,51200,0.5563082377115885
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,256,16384,0.03754986524581909
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,256,16384,0.18962559700012208
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,512,65536,0.22766186396280924
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,256,12288,0.02837760051091512
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,256,12288,0.14233919779459636
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,256,16384,0.068832000096639
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,256,10240,0.02222506602605184
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,256,10240,0.11343146959940593
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,256,12288,0.05505173206329346
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,256,8192,0.017117865880330405
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,256,8192,0.08421546618143717
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,256,10240,0.05010666847229004
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,256,7168,0.015230933825174967
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,256,7168,0.06842772960662842
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,256,8192,0.03790826797485351
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,256,6144,0.013337600231170654
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,256,6144,0.05555946826934814
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,256,7168,0.035361067454020186
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,256,5120,0.011610666910807293
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,256,5120,0.04836479822794597
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,256,6144,0.032451200485229495
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,256,4096,0.00958079993724823
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,256,51200,0.1828170617421468
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,256,4096,0.041121065616607666
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,256,3584,0.009193600217501322
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,256,3584,0.037064532438913986
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,256,5120,0.030131200949350994
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,256,3072,0.008081066608428954
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,256,3072,0.03499946594238281
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,256,3584,0.023500800132751465
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,256,2560,0.007107200225194295
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,256,3072,0.021124267578125
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,256,2560,0.03134719928105672
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,256,2048,0.006355200211207073
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,256,2048,0.028892799218495684
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,256,4096,0.02569920023282369
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,256,1536,0.00555626650651296
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,256,2560,0.019777067502339683
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,256,1536,0.025769599278767902
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,256,1024,0.004704000055789947
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,256,2048,0.017744000752766928
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,256,1024,0.022740266720453896
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,256,768,0.004268800218900045
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,256,1024,0.013270399967829385
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,256,768,0.021452800432840983
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,256,65536,0.2283232053120931
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,256,512,0.0038922667503356934
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,256,768,0.012936533490816752
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,256,512,0.02029013236363729
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,256,256,0.003036800026893616
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,256,256,0.017170133193333943
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,256,256,0.01325759987036387
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,256,128,0.0026421333352724713
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,256,512,0.013297067085901896
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,256,128,0.01651093363761902
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,256,64,0.00271573339899381
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,256,32,0.0026677332818508146
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,256,64,0.015377066532770791
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,256,128,0.013342932860056559
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,128,65536,0.08944213390350342
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,256,32,0.014732799927393594
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,128,51200,0.07117013136545816
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,256,1536,0.015277866522471109
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,128,65536,0.6972351710001627
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,128,16384,0.029044266541798907
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,128,51200,0.5519765218098958
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,128,16384,0.1861471970876058
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,128,12288,0.02177066604296366
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,128,12288,0.13659200668334961
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,128,12288,0.053098666667938235
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,128,10240,0.015958399573961893
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,128,10240,0.1091274658838908
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,128,16384,0.06736853122711181
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,128,8192,0.012232533097267151
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,128,10240,0.04846293528874715
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,128,8192,0.07929386297861735
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,128,7168,0.010784000158309937
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,128,7168,0.06477333307266235
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,128,8192,0.03708053429921468
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,128,6144,0.009544533491134644
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,128,51200,0.1796533266703288
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,128,6144,0.05134933392206827
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,128,5120,0.008450133601824443
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,128,7168,0.03432106574376424
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,128,5120,0.0451690673828125
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,128,4096,0.007129600147406261
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,128,4096,0.038822400569915774
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,128,6144,0.03089386622111003
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,128,3584,0.006706133484840393
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,128,3584,0.03492586612701416
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,128,4096,0.024756266673405965
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,128,5120,0.028845866521199543
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,128,3072,0.006119466821352641
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,128,65536,0.2240778605143229
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,128,3072,0.03311893343925476
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,128,2560,0.005543466905752818
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,128,2560,0.030402133862177532
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,128,3072,0.02068159977595011
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,128,2048,0.005021866659323374
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,128,2048,0.028700800736745198
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,128,2560,0.019080533583958944
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,128,2048,0.016541866461435954
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,128,1536,0.004296533266703288
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,128,1536,0.0247925341129303
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,128,1024,0.0035360001027584078
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,128,1024,0.022612265745798745
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,128,1536,0.014129066467285156
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,128,768,0.0033834666013717652
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,128,3584,0.022762666145960488
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,128,1024,0.01276479959487915
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,128,512,0.0029834667841593427
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,128,768,0.021100799242655434
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,128,768,0.011975466211636861
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,128,512,0.01939199964205424
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,128,256,0.002614400039116542
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,128,256,0.017240534226099648
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,128,512,0.012242133418718975
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,128,256,0.012432000041007996
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,128,128,0.016743467251459757
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3841,128,128,0.012134400010108948
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,128,64,0.0025727999707063037
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,128,32,0.0026367999613285064
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,128,64,0.014538666605949402
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,128,32,0.014512000481287637
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,64,65536,0.08758506774902344
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,64,51200,0.06477973461151124
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,64,16384,0.025121066967646283
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,64,16384,0.18139732678731282
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,64,65536,0.6907061258951823
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,64,51200,0.5439552307128906
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,64,12288,0.014679466684659323
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,64,10240,0.011648000280062357
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,64,8192,0.010075733065605164
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,128,128,0.0025568000972270967
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,64,10240,0.10876479943593342
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,64,7168,0.009513599673906963
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,64,8192,0.07836693127950033
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,64,6144,0.00913706620534261
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,64,7168,0.06373013257980346
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,64,5120,0.007979733248551685
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,64,4096,0.006738133231798808
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,64,6144,0.050937600930531825
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,64,5120,0.0438261349995931
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,64,4096,0.03911360104878743
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,64,3072,0.005861333509286245
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,64,3072,0.033165866136550905
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,64,2560,0.005070933202902476
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,64,2560,0.030118399858474733
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,64,2048,0.004625066618124644
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,64,2048,0.028062933683395387
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,64,1536,0.003946666667858759
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,64,1536,0.02555519938468933
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,64,1024,0.003433600068092346
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,64,1024,0.022576000293095908
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,64,768,0.0030218665798505146
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,64,12288,0.13778133392333985
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,64,512,0.0030901332696278887
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,64,3584,0.006305066744486491
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,64,256,0.0026015999416510267
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,64,512,0.019735467433929444
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,64,3584,0.034993068377176924
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,64,128,0.0022304000953833262
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,64,256,0.016516266266504924
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,64,64,0.0022261333962281544
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,64,128,0.015343999862670899
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,64,64,0.014536533753077188
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,64,32,0.0022410665949185687
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,64,32,0.014535466829935709
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,32,65536,0.08274772961934408
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,32,51200,0.0659114678700765
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,32,16384,0.02439039945602417
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,32,16384,0.18138880729675294
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,32,12288,0.013364266355832419
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,32,51200,0.5440725326538086
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,32,65536,0.691099739074707
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,32,10240,0.010935466488202412
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,32,12288,0.13640106519063314
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,32,8192,0.00981119970480601
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,32,7168,0.009147733449935913
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,64,768,0.020834134022394816
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,32,8192,0.08089706897735596
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,32,6144,0.008351999521255492
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,32,7168,0.06576000054677328
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,32,5120,0.010423466563224792
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,32,6144,0.052269868055979406
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,32,4096,0.008975999553998311
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,32,5120,0.045186134179433186
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,32,3584,0.008326399823029835
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,32,4096,0.038967466354370116
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,32,3584,0.034959999720255534
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,32,2560,0.006673066814740498
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,32,2560,0.029473066329956055
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,32,2048,0.004655999938646952
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,32,10240,0.11159253120422363
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,32,2048,0.027271467447280883
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,32,3072,0.00726506660381953
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,32,1536,0.0038986665507157645
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,32,3072,0.03297066688537598
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,32,1536,0.02490239938100179
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,32,768,0.003036800026893616
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,32,1024,0.020822399854660036
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,32,768,0.02068159977595011
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,32,512,0.018747733036677042
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,32,256,0.0025792000194390613
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,32,256,0.01755946675936381
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,32,128,0.0022602667411168414
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,32,128,0.015877333283424378
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,32,64,0.002253866692384084
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,32,64,0.01576746702194214
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,32,32,0.0022111999491850535
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3841,32,32,0.015079466501871744
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,32,1024,0.0034933333595593774
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3841,32,512,0.002583466718594233
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,65536,16384,4.907144673665365
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,65536,16384,2.605901845296224
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,65536,51200,8.340435282389324
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,65536,12288,3.773580678304037
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,65536,12288,1.8927839914957683
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,65536,51200,16.140616861979165
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,65536,10240,3.1317568461100262
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,65536,10240,1.4813514709472657
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,65536,10240,0.8490400314331055
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,65536,8192,2.543190511067708
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,65536,8192,1.1691691080729165
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,65536,12288,0.9518677393595377
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,65536,7168,2.1664779663085936
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,65536,7168,1.0123189290364585
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,65536,16384,1.2978410085042318
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,65536,6144,1.7653087615966796
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,65536,6144,0.872215461730957
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,65536,8192,0.7055072148640951
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,65536,5120,1.315210723876953
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,65536,5120,0.7222741444905598
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,65536,7168,0.6585109074910481
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,65536,4096,1.0187690734863282
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,65536,4096,0.6067136128743489
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,65536,6144,0.5694592157999675
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,65536,3584,0.9539562861124674
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,65536,3584,0.5239690780639649
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,65536,5120,0.5159285227457683
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,65536,3072,0.7647562662760417
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,65536,3072,0.4531711896260579
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,65536,4096,0.46469974517822266
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,65536,2560,0.6397109349568685
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,65536,2560,0.39964478810628257
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,65536,3072,0.3674474716186523
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,65536,2048,0.5110389391581218
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,65536,2048,0.30671253204345705
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,65536,3584,0.41666132609049483
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,65536,1536,0.40053440729777023
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,65536,1536,0.24746452967325844
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,65536,2560,0.34697494506835935
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,65536,1024,0.2730229377746582
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,65536,1024,0.17784105936686198
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,65536,1024,0.2515946706136068
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,65536,768,0.21438934008280436
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,65536,768,0.14975253740946454
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,65536,1536,0.26098453203837074
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,65536,512,0.1586133321126302
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,65536,512,0.12545706431070963
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,65536,512,0.23853759765625
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,65536,256,0.10625066757202148
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,65536,256,0.10476373036702473
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,65536,768,0.2420405387878418
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,65536,128,0.08656746546427409
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,65536,128,0.09577279885609945
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,65536,2048,0.3093290646870931
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,65536,64,0.08037333488464356
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,65536,64,0.0947808027267456
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,65536,32,0.07973439693450927
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,65536,128,0.24436160723368325
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,65536,32,0.09433066844940186
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,65536,256,0.23636053403218588
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,51200,65536,7.895665995279948
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,51200,65536,14.986834716796874
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,51200,51200,11.332896931966145
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,51200,51200,6.257080586751302
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,51200,16384,3.6924779256184896
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,51200,16384,2.026428858439128
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,51200,16384,1.0079157511393229
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,51200,12288,2.6635978698730467
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,51200,12288,1.4594431559244792
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,51200,12288,0.7476191838582357
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,51200,10240,2.421891276041667
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,51200,10240,1.1950122833251953
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,51200,10240,0.6680447896321614
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,51200,8192,1.9114730834960938
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,51200,8192,0.9652245203653971
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,65536,51200,4.236476898193359
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,51200,7168,1.669673538208008
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,51200,7168,0.838037363688151
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,51200,8192,0.5501087824503581
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,51200,6144,1.420292282104492
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,51200,6144,0.6961845397949219
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,51200,7168,0.5148810704549154
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,51200,5120,1.048530133565267
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,51200,5120,0.6514922459920247
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,51200,5120,0.4047370592753093
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,51200,4096,0.8785578409830729
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,51200,4096,0.5441525141398112
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,51200,6144,0.44031680425008135
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,51200,3584,0.7462111790974935
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,51200,3584,0.41580801010131835
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,51200,3584,0.32051947911580403
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,51200,3072,0.6326847712198893
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,51200,3072,0.36017494201660155
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,51200,4096,0.36085119247436526
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,51200,2560,0.5289696057637532
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,51200,2560,0.3123999913533529
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,51200,51200,3.2518880208333334
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,51200,2048,0.42096961339314776
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,51200,2048,0.24667627016703286
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,51200,2560,0.27452907562255857
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,51200,1536,0.31135145823160804
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,51200,1536,0.19575467109680175
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,51200,3072,0.2860629399617513
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,51200,1024,0.21107093493143717
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,51200,1024,0.14191892941792805
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,51200,1024,0.2013792037963867
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,51200,768,0.16559359232584636
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,51200,768,0.12153493563334147
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,51200,1536,0.2069610595703125
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,51200,512,0.12342080275217693
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,51200,512,0.10151253541310627
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,51200,2048,0.24065920511881508
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,51200,256,0.08299626509348551
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,51200,256,0.0860970656077067
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,51200,512,0.18998506863911946
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,51200,128,0.07263786792755127
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,51200,128,0.07511253356933593
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,51200,768,0.18928747177124022
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,51200,256,0.188973871866862
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,51200,64,0.062295468648274746
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,51200,32,0.0642677346865336
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,51200,64,0.07762986818949381
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,51200,32,0.07796053091684976
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,51200,128,0.19192105929056805
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,16384,65536,2.867358907063802
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,16384,65536,4.696378580729166
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,16384,51200,3.6940714518229165
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,16384,51200,2.2457728068033855
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,16384,16384,1.1074644724527993
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,16384,16384,0.7469930648803711
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,16384,16384,0.36010878880818686
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,16384,12288,0.7840149561564128
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,16384,12288,0.5844682693481446
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,16384,12288,0.2723701477050781
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,16384,10240,0.66145814259847
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,16384,10240,0.45574080149332685
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,16384,10240,0.24295573234558104
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,16384,8192,0.5276970545450846
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,16384,8192,0.3495903968811035
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,16384,8192,0.20103252728780113
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,16384,7168,0.4585269292195638
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,16384,7168,0.2992629369099935
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,16384,51200,1.090170669555664
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,16384,6144,0.38626880645751954
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,16384,6144,0.28759466807047523
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,16384,7168,0.18738560676574706
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,16384,5120,0.3209151903788249
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,16384,5120,0.2210474650065104
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,16384,6144,0.1625343958536784
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,16384,4096,0.2535552024841309
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,16384,4096,0.18460480372111004
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,16384,5120,0.1495317300160726
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,16384,3584,0.22567359606424967
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,16384,3584,0.1577023983001709
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,16384,3584,0.11927786668141682
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,16384,3072,0.1945994695027669
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,16384,3072,0.13933226267496746
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,16384,4096,0.13228586514790852
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,16384,2560,0.16222933133443196
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,16384,2560,0.1168181339899699
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,16384,2560,0.10195093154907227
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,16384,2048,0.13138559659322102
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,16384,2048,0.09815786679585775
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,16384,3072,0.1057919979095459
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,16384,1536,0.10192000071207682
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,16384,1536,0.07994240125020345
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,16384,1536,0.07650559743245443
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,16384,1024,0.07095253467559814
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,16384,1024,0.062305064996083584
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,16384,2048,0.09028159777323405
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,16384,768,0.05494079987208048
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,16384,768,0.05650773445765177
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,51200,65536,4.281677754720052
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,16384,512,0.04546240170796712
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,16384,512,0.05575040181477865
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,16384,512,0.060185599327087405
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,16384,256,0.029262934128443403
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,16384,256,0.04126933415730794
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,16384,768,0.06541546583175659
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,16384,128,0.02679786682128906
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,16384,1024,0.07187519868214926
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,16384,256,0.057905066013336184
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,16384,128,0.036830933888753255
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,16384,64,0.022375466426213582
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,16384,64,0.037206399440765384
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,16384,32,0.023540266354878745
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,16384,32,0.037300264835357665
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,16384,128,0.059254399935404455
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,12288,65536,2.266802215576172
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,12288,65536,3.4454612731933594
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,12288,51200,1.7707135518391925
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,12288,51200,2.6799552917480467
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,12288,16384,0.794208017985026
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,12288,16384,0.5913941065470378
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,16384,65536,1.3812639872233072
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,12288,12288,0.5817002614339193
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,12288,12288,0.43633066813151045
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,12288,16384,0.2876490592956543
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,12288,10240,0.5321749369303386
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,12288,10240,0.40572268168131514
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,12288,12288,0.21908480326334634
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,12288,8192,0.41706988016764324
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,12288,8192,0.2882176081339518
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,12288,8192,0.16065173149108886
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,12288,7168,0.3678112030029297
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,12288,7168,0.2476341406504313
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,12288,10240,0.19537067413330078
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,12288,6144,0.31548480987548827
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,12288,6144,0.22137813568115233
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,12288,7168,0.14894827206929523
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,12288,5120,0.257098658879598
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,12288,5120,0.18396053314208985
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,12288,6144,0.13001706600189208
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,12288,4096,0.20540693600972496
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,12288,4096,0.15263892809549967
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,12288,4096,0.10505066712697346
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,12288,3584,0.17971307436625164
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,12288,3584,0.13284693559010824
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,12288,5120,0.11921280225118
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,12288,3072,0.15356480280558268
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,12288,3072,0.11545386314392089
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,12288,3072,0.08420693079630534
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,12288,2560,0.12885546684265137
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,12288,2560,0.09918506940205893
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,12288,3584,0.09355733394622803
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,12288,2048,0.10575466950734455
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,12288,2048,0.08567893505096436
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,12288,2048,0.07041172981262207
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,12288,1536,0.08111466566721598
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,12288,1536,0.07002986272176107
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,12288,2560,0.08045012950897217
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,12288,1024,0.05617493391036987
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,12288,1024,0.05439893404642741
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,12288,1024,0.06168959935506185
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,12288,768,0.04470613400141398
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,12288,768,0.05053973197937012
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,12288,1536,0.06374719937642416
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,12288,512,0.033725865681966144
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,12288,512,0.04314560095469157
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,12288,768,0.05727466742197672
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,12288,256,0.023561600844065347
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,12288,256,0.040516265233357746
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,12288,512,0.05741440057754517
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,12288,128,0.02103360096613566
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,12288,128,0.036789333820343016
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,12288,256,0.05674879948298136
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,12288,64,0.021833600600560506
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,12288,128,0.05681066513061524
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,12288,32,0.02105493346850077
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,12288,64,0.03538879950841268
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,12288,32,0.0348906675974528
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,10240,65536,1.98548583984375
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,10240,65536,2.832916259765625
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,10240,51200,2.2012054443359377
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,10240,51200,1.5420032501220704
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,12288,51200,0.8646154403686523
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,10240,16384,0.6265845616658529
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,10240,16384,0.500874678293864
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,10240,16384,0.24186239242553711
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,10240,12288,0.4742016156514485
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,10240,12288,0.3895754814147949
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,12288,65536,1.0869674682617188
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,10240,10240,0.4067509333292644
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,10240,10240,0.30659945805867517
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,10240,12288,0.18424320220947266
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,10240,8192,0.3314496040344238
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,10240,8192,0.24055360158284506
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,10240,10240,0.16336746215820314
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,10240,7168,0.2846570650736491
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,10240,7168,0.21611305872599282
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,10240,7168,0.12505066394805908
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,10240,6144,0.2416170597076416
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,10240,6144,0.18149333000183104
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,10240,8192,0.13549013137817384
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,10240,5120,0.2029205322265625
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,10240,5120,0.15456533432006836
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,10240,51200,0.7149098714192708
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,10240,4096,0.16205546061197917
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,10240,4096,0.12741440137227375
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,10240,6144,0.10958293279012044
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,10240,5120,0.100873597462972
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,10240,3584,0.14130345980326336
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,10240,3584,0.11144533157348632
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,10240,3072,0.12094293435414631
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,10240,3072,0.10021759668986004
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,10240,4096,0.088044802347819
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,10240,2560,0.10340692996978759
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,10240,2560,0.08453226884206136
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,10240,3584,0.07871146996816
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,10240,3072,0.07099093596140543
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,10240,2048,0.08579839865366617
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,10240,2048,0.0728384017944336
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,10240,1536,0.06340373357137044
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,10240,1536,0.059831468264261875
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,10240,2560,0.06832640171051026
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,10240,1024,0.04499200185139974
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,10240,1024,0.047277867794036865
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,10240,1536,0.05602026780446371
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,10240,768,0.03458986679712932
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,10240,768,0.04363413254419963
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,10240,1024,0.05382293462753296
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,10240,512,0.02576106588045756
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,10240,512,0.038159998257954915
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,10240,2048,0.0614303986231486
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,10240,256,0.018651733795801796
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,10240,256,0.03287786642710368
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,10240,768,0.05051626761754354
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,10240,128,0.01739306648572286
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,10240,128,0.029329067468643187
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,10240,512,0.050614400704701745
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,10240,64,0.01482133368651072
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,10240,128,0.05042773485183716
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,10240,64,0.027460267146428425
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,10240,32,0.014916266997655234
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,10240,32,0.027229867378870648
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,10240,256,0.05068480173746744
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,8192,65536,1.8076651255289715
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,8192,65536,2.4480809529622394
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,8192,51200,1.825549825032552
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,8192,51200,1.408892822265625
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,8192,16384,0.5667135874430339
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,8192,16384,0.4966304143269856
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,10240,65536,0.9137098948160807
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,8192,12288,0.43494612375895186
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,8192,12288,0.3651551882425944
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,8192,16384,0.21341867446899415
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,8192,10240,0.3505183855692545
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,8192,10240,0.2788821220397949
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,8192,12288,0.16352532704671224
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,8192,8192,0.2908543904622396
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,8192,8192,0.23890986442565917
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,8192,10240,0.1449717362721761
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,8192,7168,0.24659520785013833
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,8192,7168,0.19236480394999186
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,8192,7168,0.11139520009358723
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,8192,6144,0.2107861359914144
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,8192,6144,0.16618132591247559
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,8192,8192,0.12023039658864339
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,8192,5120,0.1769152005513509
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,8192,5120,0.1437130610148112
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,8192,5120,0.08978986740112305
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,8192,4096,0.14097280502319337
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,8192,4096,0.11829546292622883
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,8192,6144,0.09711466630299886
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,8192,3584,0.12043946584065754
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,8192,3584,0.10277012983957927
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,8192,3584,0.06811520258585611
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,8192,3072,0.10364586512247723
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,8192,3072,0.09271360238393148
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,8192,4096,0.07721386750539144
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,8192,2560,0.08936426639556885
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,8192,2560,0.08022080262502035
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,8192,2560,0.05681066513061524
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,8192,2048,0.07245439688364665
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,8192,2048,0.06859626770019531
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,8192,3072,0.05913706620534261
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,8192,1536,0.054934398333231604
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,8192,1536,0.05700266758600871
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,8192,1536,0.04233386516571045
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,8192,1024,0.038626134395599365
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,8192,1024,0.04221226771672566
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,8192,2048,0.04930880069732666
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,8192,768,0.03109440008799235
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,8192,768,0.039154132207234696
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,8192,1024,0.040481066703796385
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,8192,512,0.02073919971783956
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,8192,512,0.03389120101928711
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,8192,768,0.03529706796010335
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,8192,256,0.01593386630217234
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,8192,256,0.028572799762090047
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,8192,256,0.03556586503982544
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,8192,128,0.0131221334139506
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,8192,128,0.0297760009765625
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,8192,512,0.03608746528625488
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,8192,64,0.016804265975952148
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,8192,128,0.03626240094502767
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,8192,32,0.015287466843922935
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,8192,64,0.028475733598073323
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,8192,32,0.028014934062957762
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,7168,65536,1.5887743631998696
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,7168,65536,2.026584498087565
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,7168,51200,1.5335360209147135
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,7168,51200,1.2595488230387368
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,8192,51200,0.6355434417724609
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,7168,16384,0.45710614522298176
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,7168,16384,0.40812158584594727
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,7168,16384,0.2044480005900065
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,7168,12288,0.34475733439127604
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,7168,12288,0.3187978744506836
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,8192,65536,0.804531224568685
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,7168,10240,0.3037664095560709
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,7168,10240,0.255079460144043
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,7168,12288,0.14848532676696777
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,7168,8192,0.24704532623291015
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,7168,8192,0.20183146794637047
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,7168,10240,0.13029653231302898
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,7168,7168,0.2150965372721354
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,7168,7168,0.17451093991597494
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,7168,8192,0.10600746472676595
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,7168,6144,0.18491946856180827
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,7168,6144,0.15062185923258464
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,7168,7168,0.0994698683420817
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,7168,5120,0.15035732587178546
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,7168,5120,0.1287775993347168
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,7168,6144,0.08785066604614258
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,7168,4096,0.12020906607309978
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,7168,4096,0.10651840368906658
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,7168,5120,0.08100159962972006
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,7168,3584,0.10475733280181884
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,7168,3584,0.09444800217946371
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,7168,4096,0.06909759839375815
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,7168,3072,0.09088746706644693
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,7168,3072,0.08479999701182048
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,7168,3584,0.06181333462397257
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,7168,2560,0.07602559725443522
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,7168,2560,0.07404373486836752
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,7168,51200,0.5872341156005859
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,7168,2048,0.0634112000465393
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,7168,2048,0.06353066762288412
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,7168,2560,0.053071999549865724
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,7168,1536,0.04839893182118733
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,7168,1536,0.052475734551747644
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,7168,3072,0.054967466990152994
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,7168,1024,0.03360639810562134
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,7168,1024,0.04000106652577718
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,7168,2048,0.046838398774464926
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,7168,768,0.026779733101526898
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,7168,768,0.037324798107147214
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,7168,1024,0.0395082672437032
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,7168,512,0.019244800011316933
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,7168,512,0.0320032000541687
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,7168,1536,0.04132053454717
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,7168,256,0.014613333344459533
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,7168,256,0.02867199977238973
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,7168,768,0.036489601929982504
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,7168,128,0.01125759979089101
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,7168,128,0.023341866334279378
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,7168,512,0.03702293237050374
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,7168,64,0.010413866241772969
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,7168,64,0.022669865687688192
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,7168,32,0.010437333583831787
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,7168,32,0.022757333517074586
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,7168,128,0.036602667967478436
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,7168,256,0.03628480037053426
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,6144,65536,1.5175156911214194
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,6144,65536,1.8224202473958333
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,6144,51200,1.184008534749349
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,6144,51200,1.4141824086507162
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,6144,16384,0.42793919245402023
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,6144,16384,0.3959029197692871
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,7168,65536,0.7353205362955729
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,6144,12288,0.31378132502237954
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,6144,12288,0.309330145517985
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,6144,16384,0.17392320632934571
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,6144,10240,0.26473066012064617
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,6144,10240,0.2336906592051188
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,6144,12288,0.12958827018737792
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,6144,8192,0.20597546895345054
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,6144,8192,0.17832320531209309
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,6144,10240,0.11513386567433674
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,6144,7168,0.18671147028605145
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,6144,7168,0.16279786427815754
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,6144,7168,0.08912639617919922
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,6144,6144,0.15899306933085125
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,6144,6144,0.13782186508178712
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,6144,8192,0.09454507033030192
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,6144,5120,0.1264074643452962
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,6144,5120,0.11593706607818603
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,6144,6144,0.07899200121561686
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,6144,4096,0.10368959903717041
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,6144,4096,0.09818987051645914
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,6144,5120,0.07236800193786622
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,6144,3584,0.08818666934967041
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,6144,3584,0.0865546703338623
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,6144,3584,0.05484586556752523
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,6144,3072,0.07625919977823893
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,6144,3072,0.07813440163930258
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,6144,4096,0.06115946769714355
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,6144,2560,0.06582719882329305
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,6144,2560,0.06763520240783691
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,6144,3072,0.048613333702087404
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,6144,2048,0.052476799488067626
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,6144,2048,0.05648533503214518
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,6144,2560,0.046326398849487305
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,6144,1536,0.04168959856033325
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,6144,1536,0.04658453464508057
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,6144,2048,0.04159146547317505
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,6144,1024,0.028180267413457232
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,6144,1024,0.03727786540985108
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,6144,1536,0.036813867092132566
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,6144,768,0.023246934016545616
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,6144,768,0.034517331918080645
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,6144,768,0.03277653257052104
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,6144,512,0.01702186663945516
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,6144,512,0.029303467273712157
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,6144,1024,0.034832000732421875
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,6144,256,0.01284160017967224
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,6144,256,0.027666133642196656
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,6144,256,0.032595199346542356
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,6144,128,0.012880000472068786
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,6144,512,0.03294080098470052
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,6144,128,0.02598293423652649
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,6144,64,0.01165226697921753
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,6144,64,0.025575466950734455
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,6144,32,0.012446932991345723
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,6144,32,0.024855466683705647
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,6144,128,0.032968533039093015
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,5120,65536,1.3570677439371743
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,5120,65536,1.3070741017659506
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,5120,51200,0.988252830505371
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,5120,51200,1.032855478922526
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,6144,51200,0.49951467514038084
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,5120,16384,0.31747414271036783
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,5120,16384,0.3491082509358724
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,5120,16384,0.14595947265625
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,5120,12288,0.23753600120544432
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,5120,12288,0.2527445316314697
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,6144,65536,0.628925895690918
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,5120,10240,0.2098911921183268
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,5120,10240,0.2082538604736328
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,5120,12288,0.11220906575520832
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,5120,8192,0.17074774106343588
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,5120,8192,0.15780800183614094
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,5120,10240,0.1006773312886556
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,5120,7168,0.1480277379353841
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,5120,7168,0.1380234718322754
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,5120,8192,0.08305919965108235
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,5120,6144,0.12823893229166666
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,5120,6144,0.11960426966349284
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,5120,7168,0.07887360254923502
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,5120,5120,0.10502399603525799
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,5120,5120,0.10456746419270832
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,5120,6144,0.0702293316523234
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,5120,4096,0.08292373021443686
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,5120,4096,0.08644693692525228
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,5120,5120,0.06419519980748495
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,5120,3584,0.07167359987894693
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,5120,3584,0.07672426700592042
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,5120,4096,0.053667199611663816
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,5120,3072,0.062357334295908604
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,5120,3072,0.06776639620463053
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,5120,3584,0.048794666926066085
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,5120,2560,0.051939201354980466
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,5120,51200,0.4196938514709473
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,5120,2560,0.05792106787363688
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,5120,2048,0.042217600345611575
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,5120,2048,0.04949866533279419
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,5120,3072,0.043755733966827394
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,5120,1536,0.03208746711413066
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,5120,1536,0.04069120089213053
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,5120,1536,0.03428479830423991
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,5120,1024,0.02257279952367147
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,5120,1024,0.033030400673548385
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,5120,2048,0.03773973385492961
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,5120,768,0.01813439925511678
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,5120,768,0.031166932980219525
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,5120,2560,0.04202453295389812
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,5120,512,0.01383680005868276
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,5120,512,0.02706666588783264
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,5120,1024,0.03243946631749471
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,5120,256,0.009832533200581868
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,5120,512,0.031144533554712934
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,5120,256,0.030849067370096843
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,5120,128,0.009160533547401428
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,5120,128,0.021585067113240562
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,5120,256,0.022721066077550253
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,5120,64,0.008376533786455791
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,5120,768,0.031045333544413252
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,5120,64,0.020862932999928793
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,5120,32,0.008404266834259034
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,5120,32,0.02117439905802409
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,4096,65536,1.147526423136393
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,4096,65536,1.2165227254231772
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,5120,128,0.030972800652186078
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,4096,51200,0.9059466679890951
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,4096,51200,0.9638250350952149
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,4096,16384,0.2833621342976888
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,4096,16384,0.3247701327006022
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,5120,65536,0.524729601542155
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,4096,12288,0.2091829299926758
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,4096,16384,0.1407039960225423
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,4096,12288,0.24017173449198403
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,4096,10240,0.17587092717488606
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,4096,10240,0.1884640057881673
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,4096,10240,0.09614933331807454
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,4096,8192,0.13959147135416666
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,4096,8192,0.1458613395690918
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,4096,12288,0.10798079967498779
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,4096,7168,0.12458133697509766
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,4096,7168,0.12943360010782878
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,4096,8192,0.07939733664194742
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,4096,6144,0.10241280396779377
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,4096,6144,0.11221333344777425
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,4096,7168,0.07496533393859864
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,4096,5120,0.08839146296183267
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,4096,5120,0.09735466639200846
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,4096,5120,0.05892693201700846
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,4096,4096,0.07086186408996582
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,4096,4096,0.08190720081329346
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,4096,6144,0.06614506642023722
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,4096,3584,0.0603498657544454
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,4096,3584,0.0699402650197347
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,4096,3584,0.044240001837412515
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,4096,3072,0.05202986796696981
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,4096,3072,0.06196159919102987
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,4096,4096,0.05032853285471598
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,4096,2560,0.045073068141937254
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,4096,2560,0.05345173279444376
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,4096,2560,0.03723520040512085
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,4096,2048,0.03658239841461182
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,4096,2048,0.046301865577697755
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,4096,3072,0.039287467797597245
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,4096,1536,0.027381332715352376
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,4096,1536,0.03914133310317993
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,4096,1536,0.02818666696548462
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,4096,1024,0.01978666583697001
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,4096,1024,0.03309546709060669
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,4096,2048,0.03252373337745666
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,4096,768,0.016515200336774193
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,4096,768,0.031222399075826007
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,4096,1024,0.02670186758041382
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,4096,512,0.013273599743843078
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,4096,512,0.027849600712458296
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,4096,768,0.02351893385251363
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,4096,256,0.010797866185506185
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,4096,256,0.025678932666778564
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,4096,256,0.023918932676315306
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,4096,128,0.007479466497898102
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,4096,128,0.023822933435440063
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,4096,512,0.024305067459742227
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,4096,64,0.01011306643486023
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,4096,64,0.022802132368087768
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,4096,32,0.010431999961535137
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,4096,32,0.022670932610829673
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,4096,128,0.023869866132736207
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,3584,65536,0.9493951797485352
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,3584,65536,1.12237548828125
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,3584,51200,0.7399434407552083
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,3584,51200,0.9140352249145508
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,4096,51200,0.41368532180786133
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,3584,16384,0.23060480753580728
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,3584,16384,0.2860053380330404
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,3584,16384,0.11826666990915935
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,3584,12288,0.17502613067626954
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,3584,12288,0.21464746793111167
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,4096,65536,0.5207071940104167
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,3584,10240,0.15218559900919598
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,3584,10240,0.17754559516906737
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,3584,12288,0.09249493281046549
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,3584,8192,0.12368319829305012
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,3584,8192,0.13790507316589357
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,3584,10240,0.08380266825358072
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,3584,7168,0.10680747032165527
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,3584,7168,0.12251946926116944
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,3584,8192,0.06975680192311605
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,3584,6144,0.09187520345052083
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,3584,6144,0.10636373360951741
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,3584,7168,0.066102401415507
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,3584,5120,0.07552213668823242
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,3584,5120,0.09308160146077474
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,3584,6144,0.057897599538167324
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,3584,4096,0.06122879981994629
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,3584,4096,0.07573333581288656
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,3584,5120,0.05197120110193888
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,3584,3584,0.05409919818242391
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,3584,3584,0.0655455986658732
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,3584,4096,0.04464853207270304
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,3584,3072,0.04675519863764445
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,3584,3072,0.057543468475341794
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,3584,3584,0.04069226582845052
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,3584,2560,0.03908373514811198
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,3584,2560,0.049813334147135416
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,3584,3072,0.036526934305826826
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,3584,2048,0.03200640082359314
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,3584,2048,0.043435732523600265
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,3584,2560,0.03514773448308309
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,3584,1536,0.02467413345972697
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,3584,1536,0.037522133191426596
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,3584,2048,0.031274666388829545
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,3584,1024,0.01840959986050924
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,3584,51200,0.33578879038492837
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,3584,1024,0.03238079945246379
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,3584,1024,0.026470400889714557
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,3584,768,0.014740266402562461
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,3584,768,0.02893120050430298
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,3584,1536,0.028035199642181395
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,3584,512,0.011177600423494975
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,3584,512,0.025246934096018477
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,3584,256,0.008377599716186523
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,3584,256,0.024497065941492716
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,3584,512,0.028708267211914062
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,3584,128,0.0071274667978286745
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,3584,768,0.02792533238728841
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,3584,128,0.021038933595021566
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,3584,256,0.028518400589625043
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,3584,64,0.020823466777801513
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,3584,128,0.02796799937884013
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,3584,32,0.021193599700927733
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,3584,64,0.007924266656239827
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,3072,65536,0.8696309407552084
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,3584,32,0.008359466989835102
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,3072,51200,0.6585695902506511
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,3072,65536,1.0744341532389323
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,3072,51200,0.8466826756795248
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,3072,16384,0.21479679743448893
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,3072,16384,0.2789077440897624
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,3072,16384,0.11742613315582276
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,3072,12288,0.16247040430704754
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,3072,12288,0.21107306480407714
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,3584,65536,0.41876373291015623
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,3072,10240,0.13365866343180338
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,3072,10240,0.1644917329152425
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,3072,10240,0.08320213158925374
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,3072,8192,0.10643626848856609
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,3072,8192,0.12953279813130697
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,3072,12288,0.09163306554158529
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,3072,7168,0.0938741366068522
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,3072,7168,0.11415039698282878
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,3072,8192,0.0690720001856486
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,3072,6144,0.0771946668624878
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,3072,6144,0.09925866921742757
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,3072,7168,0.06496959924697876
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,3072,5120,0.06475199858347574
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,3072,5120,0.08601066271464029
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,3072,5120,0.05121493339538574
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,3072,4096,0.05187946557998657
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,3072,4096,0.06992426713307699
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,3072,6144,0.056966400146484374
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,3072,3584,0.04502293268839518
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,3072,3584,0.059734400113423666
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,3072,4096,0.04355200131734212
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,3072,3072,0.03940266768137614
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,3072,3072,0.05373866558074951
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,3072,3584,0.04028266668319702
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,3072,2560,0.03271573384602865
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,3072,2560,0.04551999966303508
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,3072,2560,0.034126933415730795
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,3072,2048,0.026902399460474652
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,3072,2048,0.04053440093994141
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,3072,3072,0.03617386817932129
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,3072,1536,0.02148053248723348
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,3072,1536,0.034935466448465985
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,3072,2048,0.03094826738039653
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,3072,1024,0.0147189329067866
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,3072,1024,0.029215999444325763
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,3072,1024,0.025681066513061523
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,3072,1536,0.0271999994913737
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,3072,768,0.011659733454386393
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,3072,768,0.027457066377003986
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,3072,512,0.009589333335558574
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,3072,512,0.02490880091985067
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,3072,768,0.027704532941182452
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,3072,256,0.007499733567237854
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,3072,256,0.021386667092641195
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,3072,512,0.027918932835261027
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,3072,128,0.006676266590754191
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,3072,256,0.02805226643880208
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,3072,128,0.01990506649017334
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,3072,64,0.005869866907596588
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,3072,64,0.019715199867884316
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,3072,32,0.005949866771697998
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,3072,128,0.025067732731501265
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,3072,32,0.018590933084487914
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,2560,65536,0.6053653081258138
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,2560,65536,0.9655306498209635
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,2560,51200,0.4774805386861165
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,2560,51200,0.7566826502482097
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,3072,51200,0.33502079645792643
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,2560,16384,0.15682986577351887
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,2560,16384,0.2463520050048828
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,2560,16384,0.09461973508199056
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,2560,12288,0.11860160032908122
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,2560,12288,0.1866154670715332
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,3072,65536,0.4203424135843913
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,2560,10240,0.1075818697611491
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,2560,10240,0.1557919979095459
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,2560,12288,0.07489066918691
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,2560,8192,0.08516373634338378
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,2560,8192,0.1196394681930542
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,2560,10240,0.06852160294850668
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,2560,7168,0.076146133740743
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,2560,7168,0.10580906867980958
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,2560,8192,0.057099731763203945
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,2560,6144,0.06375786860783896
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,2560,6144,0.09095359643300374
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,2560,7168,0.05311146577199301
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,2560,5120,0.053478399912516274
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,2560,5120,0.07782399654388428
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,2560,6144,0.04702506860097249
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,2560,4096,0.041414399941762284
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,2560,4096,0.061622401078542076
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,2560,5120,0.04205333391825358
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,2560,3584,0.03710079987843831
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,2560,3584,0.054080001513163244
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,2560,3584,0.03308586676915486
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,2560,3072,0.03217493295669556
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,2560,3072,0.04918826818466186
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,2560,4096,0.03659199873606364
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,2560,2560,0.026716800530751546
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,2560,2560,0.04319680134455363
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,2560,2560,0.02762453357378642
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,2560,3072,0.02916160027186076
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,2560,2048,0.022083199024200438
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,2560,2048,0.03898773193359375
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,2560,1536,0.017790933450063072
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,2560,1536,0.03295573393503825
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,2560,1536,0.02190720041592916
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,2560,1024,0.012900267044703165
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,2560,1024,0.027799467245737713
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,2560,2048,0.024346667528152465
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,2560,768,0.010215466221173603
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,2560,768,0.025128533442815144
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,2560,51200,0.25991039276123046
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,2560,512,0.007994666695594788
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,2560,1024,0.020206934213638304
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,2560,512,0.022762666145960488
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,2560,256,0.006685866912206013
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,2560,768,0.018542933464050292
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,2560,256,0.021636267503102623
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,2560,128,0.006222933530807495
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,2560,256,0.019088000059127808
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,2560,512,0.01909760038057963
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,2560,64,0.005073066552480062
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,2560,128,0.019207467635472618
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,2560,32,0.005486933390299479
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,2560,64,0.017147733767827352
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,2560,128,0.01902079979578654
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,2560,32,0.018518400192260743
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,2048,65536,0.562777582804362
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,2048,51200,0.45780372619628906
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,2048,65536,0.931109364827474
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,2048,51200,0.7334079742431641
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,2048,16384,0.14433280626932782
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,2048,16384,0.24472959836324057
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,2048,16384,0.09315520127614339
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,2048,12288,0.10987093448638915
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,2048,12288,0.18595946629842122
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,2560,65536,0.3233930587768555
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,2048,10240,0.09462400277455649
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,2048,10240,0.1523893356323242
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,2048,12288,0.07351786295572917
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,2048,8192,0.07733546892801921
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,2048,8192,0.1194922685623169
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,2048,10240,0.06645653247833253
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,2048,7168,0.06732266743977865
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,2048,7168,0.10456853707631428
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,2048,7168,0.051856001218159996
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,2048,6144,0.057962667942047116
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,2048,6144,0.0896778662999471
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,2048,8192,0.055181864897410074
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,2048,5120,0.04922346671422322
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,2048,5120,0.07436160246531168
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,2048,5120,0.04103253285090129
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,2048,4096,0.04042239983876546
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,2048,4096,0.059765334924062094
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,2048,6144,0.04495146671930949
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,2048,3584,0.035301331679026285
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,2048,3584,0.05374293327331543
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,2048,3584,0.03190720081329346
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,2048,3072,0.03068480094273885
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,2048,3072,0.04901333252588908
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,2048,4096,0.03460799853006999
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,2048,2560,0.026634667317072553
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,2048,2560,0.042668799559275314
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,2048,2560,0.026396799087524413
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,2048,2048,0.022550400098164877
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,2048,2048,0.03712746699651082
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,2048,3072,0.028140799204508467
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,2048,1536,0.017816533644994102
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,2048,1536,0.032943999767303465
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,2048,1536,0.01986133257548014
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,2048,1024,0.012902399897575379
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,2048,1024,0.028717867533365887
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,2048,2048,0.022702932357788086
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,2048,768,0.009853866696357728
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,2048,768,0.026499199867248534
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,2048,1024,0.01813653310139974
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,2048,512,0.008390399813652038
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,2048,512,0.0242741326491038
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,2048,512,0.017067732413609822
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,2048,768,0.016546133160591125
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,2048,256,0.006344533463319142
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,2048,256,0.0206058661142985
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,2048,256,0.016938666502634682
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,2048,128,0.005896533528963724
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,2048,128,0.018637865781784058
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,2048,128,0.017401599884033205
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,2048,64,0.005538133283456167
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,2048,64,0.01848640044530233
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,2048,32,0.005901866654555003
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,2048,32,0.017403733730316163
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,1536,65536,0.4236703872680664
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,1536,65536,0.8599050521850586
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,1536,51200,0.32700481414794924
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,2048,51200,0.2571477254231771
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,1536,51200,0.6840789159138997
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,1536,16384,0.10798186461130779
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,1536,16384,0.22624319394429526
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,1536,16384,0.09463146527608236
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,1536,12288,0.08198933601379395
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,1536,12288,0.1715274651845296
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,2048,65536,0.3210602760314941
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,1536,10240,0.06832213401794433
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,1536,10240,0.13565120697021485
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,1536,12288,0.07444159984588623
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,1536,8192,0.05362879832585653
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,1536,8192,0.10647146701812744
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,1536,10240,0.06792853673299154
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,1536,7168,0.04724693298339844
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,1536,7168,0.09175999959309897
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,1536,8192,0.05636906623840332
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,1536,6144,0.03930240074793498
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,1536,6144,0.07760426998138428
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,1536,7168,0.05036906798680624
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,1536,5120,0.03317973415056864
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,1536,5120,0.06495253245035806
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,1536,6144,0.04647253354390462
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,1536,4096,0.02642773389816284
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,1536,4096,0.051577599843343105
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,1536,5120,0.04209280014038086
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,1536,3584,0.023849600553512575
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,1536,3584,0.046649599075317384
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,1536,4096,0.0362335999806722
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,1536,3072,0.020809600750605263
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,1536,3072,0.04309759934743245
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,1536,3584,0.03283733328183492
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,1536,2560,0.018104533354441323
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,1536,2560,0.03784960110982259
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,1536,3072,0.030044800043106078
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,1536,2048,0.014629333217938741
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,1536,2048,0.03300053278605143
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,1536,2048,0.02529919942220052
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,1536,2560,0.02818560004234314
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,1536,1536,0.011847466230392456
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,1536,1536,0.0295360008875529
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,1536,1024,0.008808533350626629
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,1536,1024,0.025167999664942424
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,1536,1536,0.02253119945526123
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,1536,768,0.0071946665644645694
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,1536,51200,0.2582613309224447
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,1536,1024,0.02063466707865397
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,1536,512,0.005973333120346069
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,1536,512,0.0216757337252299
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,1536,768,0.019843200842539467
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,1536,256,0.0050794666012128195
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,1536,256,0.018653867642084758
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,1536,768,0.024064000447591147
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,1536,128,0.004907733201980591
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,1536,256,0.01986560026804606
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,1536,512,0.019922133286794028
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,1536,128,0.01839253306388855
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,1536,64,0.004281599819660187
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,1536,64,0.016720000902811685
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,1536,32,0.004389333228270212
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,1536,32,0.016747732957204185
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,1024,65536,0.2872746785481771
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,1024,65536,0.8096853256225586
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,1024,51200,0.21851305961608886
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,1536,128,0.020682666699091592
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,1024,51200,0.6232394536336263
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,1024,16384,0.07159146467844645
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,1024,16384,0.20723093350728355
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,1024,16384,0.06975680192311605
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,1024,12288,0.05458879868189494
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,1024,12288,0.15575146675109863
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,1536,65536,0.3243669191996256
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,1024,10240,0.04764053424199422
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,1024,10240,0.12844800154368083
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,1024,12288,0.05548799832661947
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,1024,8192,0.037481598059336346
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,1024,8192,0.09855573177337647
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,1024,10240,0.051001600424448644
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,1024,7168,0.03312533299128215
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,1024,7168,0.08416533470153809
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,1024,8192,0.04125653505325318
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,1024,6144,0.028056534131368
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,1024,6144,0.07057600021362305
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,1024,7168,0.03747200171152751
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,1024,5120,0.024399999777475992
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,1024,5120,0.05842133363087972
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,1024,6144,0.03368000189463298
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,1024,4096,0.02036693294843038
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,1024,4096,0.048052267233530684
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,1024,5120,0.03137279947598775
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,1024,3584,0.01798293391863505
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,1024,3584,0.04336853424708049
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,1024,4096,0.026420267422993977
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,1024,3072,0.015809067090352378
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,1024,3072,0.04088533322016398
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,1024,3584,0.024717867374420166
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,1024,2560,0.013416533668835958
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,1024,3072,0.022231467564900718
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,1024,2560,0.037020798524220785
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,1024,2048,0.011659733454386393
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,1024,2048,0.03297386765480041
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,1024,2560,0.02070186734199524
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,1024,1536,0.009683199723561605
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,1024,2048,0.018578133980433145
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,1024,1536,0.028889600435892743
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,1024,1024,0.007736533383528392
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,1024,51200,0.18433386484781902
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,1024,1024,0.02478826642036438
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,1024,768,0.0067775999506314594
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,1024,768,0.023177599906921385
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,1024,1024,0.014455466469128927
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,1024,1536,0.015773866573969522
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,1024,512,0.005500799914201101
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,1024,512,0.02118826707204183
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,1024,256,0.004682666560014089
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,1024,768,0.013649066289265951
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,1024,256,0.019107200702031455
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,1024,128,0.00420906643072764
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,1024,128,0.016708266735076905
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,1024,256,0.013688533504803976
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,1024,512,0.014098133643468222
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,1024,64,0.003902933249870936
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,1024,32,0.0038602667550245917
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,1024,64,0.016174933314323424
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,1024,128,0.014874666929244995
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,1024,32,0.01593280037244161
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,768,65536,0.21124374071756996
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,768,51200,0.16492373148600262
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,768,65536,0.749783452351888
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,768,51200,0.5885450363159179
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,768,16384,0.054169599215189615
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,768,16384,0.19483200709025067
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,768,16384,0.07026240030924479
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,768,12288,0.041509334246317545
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,1024,65536,0.22915520668029785
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,768,12288,0.14900479316711426
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,768,10240,0.034908799330393474
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,768,10240,0.12064426740010578
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,768,10240,0.05234026511510213
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,768,8192,0.027457066377003986
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,768,8192,0.09176959991455078
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,768,12288,0.056683735052744544
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,768,7168,0.02395840088526408
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,768,7168,0.07797973155975342
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,768,8192,0.04125119845072429
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,768,6144,0.020772266387939452
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,768,6144,0.06390719811121623
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,768,7168,0.03749866485595703
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,768,5120,0.017900800704956053
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,768,5120,0.05394346714019775
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,768,5120,0.03204693396886189
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,768,6144,0.03419626553853353
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,768,4096,0.014892799655596414
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,768,4096,0.04422080119450887
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,768,3584,0.013222400347391763
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,768,3584,0.03915946483612061
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,768,3584,0.025568000475565594
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,768,3072,0.011517866452534994
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,768,4096,0.02800106604894002
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,768,3072,0.037197868029276535
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,768,2560,0.010051199793815612
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,768,2560,0.032909866174062094
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,768,2560,0.02193066676457723
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,768,2048,0.00864533285299937
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,768,2048,0.030568534135818483
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,768,3072,0.023477333784103393
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,768,1536,0.007492266595363617
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,768,2048,0.019403733809789023
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,768,1536,0.027166932821273804
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,768,1024,0.005938133100668589
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,768,1024,0.024715733528137208
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,768,1536,0.016898133357365928
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,768,768,0.005081599950790406
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,768,768,0.02254826625188192
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,768,1024,0.01534293293952942
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,768,512,0.004691199958324432
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,768,512,0.02067306637763977
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,768,51200,0.1849727948506673
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,768,768,0.014932266871134438
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,768,256,0.01867093245188395
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,768,512,0.01532373329003652
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,768,256,0.015346133708953857
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,768,128,0.0034933333595593774
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,768,128,0.016198399662971496
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,768,64,0.0029898665845394133
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,768,64,0.015702399611473083
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,768,128,0.015784533818562825
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,768,32,0.0034815999368826545
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,768,32,0.01574720044930776
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,512,65536,0.1650880018870036
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,768,256,0.003913599997758865
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,512,65536,0.7335509618123373
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,512,51200,0.1301642656326294
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,512,51200,0.5838144302368165
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,512,16384,0.04495786825815837
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,512,16384,0.19565439224243164
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,768,65536,0.22861119906107583
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,512,12288,0.03837013244628906
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,512,12288,0.14897599220275878
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,512,16384,0.068940798441569
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,512,10240,0.03345706860224406
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,512,10240,0.11963626543680828
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,512,12288,0.05459520022074381
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,512,8192,0.02581546703974406
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,512,10240,0.05029546817143758
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,512,8192,0.09113066991170247
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,512,7168,0.022230400641759237
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,512,7168,0.07690133253733317
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,512,7168,0.03586453199386597
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,512,6144,0.018484266599019368
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,512,6144,0.062135465939839686
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,512,8192,0.03894613186518352
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,512,5120,0.015422933300336204
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,512,5120,0.05128320058186849
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,512,6144,0.03220053315162659
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,512,4096,0.012579199671745301
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,512,4096,0.04321599801381429
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,512,5120,0.029959466060002642
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,512,3584,0.01118293305238088
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,512,3584,0.03910613457361857
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,512,4096,0.0259552001953125
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,512,3072,0.010056533416112264
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,512,3072,0.03699733416239421
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,512,3584,0.02353386680285136
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,512,2560,0.008919466535250347
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,512,2560,0.032919466495513916
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,512,3072,0.021432532866795858
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,512,2048,0.00783679982026418
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,512,2048,0.03020693262418111
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,512,51200,0.18296000162760417
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,512,1536,0.00730560024579366
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,512,2048,0.01735466718673706
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,512,1536,0.02732906738917033
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,512,1024,0.005826133489608765
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,512,1024,0.02376000086466471
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,512,1536,0.015004799763361613
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,512,768,0.004676266511281332
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,512,768,0.022576000293095908
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,512,2560,0.019845332702000937
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,512,1024,0.01399679978688558
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,512,512,0.004227200150489807
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,512,768,0.012870400150616964
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,512,256,0.003435733417669932
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,512,512,0.0208512008190155
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,512,512,0.013247999548912048
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,512,128,0.003349333256483078
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,512,128,0.016055466731389363
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,512,256,0.013408000270525614
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,512,128,0.013637333114941915
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,512,64,0.014664533734321594
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,512,32,0.0030005333324273427
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,512,32,0.014479999740918478
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,256,65536,0.10821973482767741
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,512,256,0.017123200496037803
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,256,65536,0.7015040079752605
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,512,64,0.0032245332996050516
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,512,65536,0.22687466939290366
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,256,51200,0.5573631922403972
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,256,16384,0.036305065949757895
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,256,16384,0.1905621369679769
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,256,51200,0.08383786678314209
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,256,12288,0.0256223996480306
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,256,12288,0.13968639373779296
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,256,12288,0.0544981320699056
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,256,10240,0.01960106690724691
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,256,10240,0.11278080145517985
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,256,16384,0.06840746402740479
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,256,8192,0.01551466683546702
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,256,8192,0.08226453463236491
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,256,10240,0.05019520123799642
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,256,7168,0.01369493305683136
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,256,7168,0.06755626996358235
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,256,8192,0.03836053212483724
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,256,6144,0.011756799618403117
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,256,6144,0.05528000195821127
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,256,7168,0.034969600041707356
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,256,5120,0.010364799698193868
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,256,5120,0.04650026559829712
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,256,5120,0.02964479923248291
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,256,4096,0.008771199981371562
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,256,6144,0.03206186691919963
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,256,4096,0.03943893512090047
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,256,3584,0.007949866851170858
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,256,3584,0.03598933219909668
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,256,51200,0.1824512004852295
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,256,3072,0.007210666437943776
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,256,4096,0.025548799832661943
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,256,3584,0.023548799753189086
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,256,3072,0.034620801607767746
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,256,2560,0.006667733192443848
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,256,2560,0.030998400847117108
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,256,3072,0.021031467119852702
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,256,2048,0.005845333139101664
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,256,2048,0.028468267122904463
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,256,2560,0.01989013353983561
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,256,1536,0.0050901333491007485
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,256,1536,0.025834665695826216
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,256,2048,0.017372800906499227
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,256,1024,0.004222933451334635
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,256,1024,0.022707200050354003
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,256,1536,0.014929067095120749
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,256,768,0.003947733342647553
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,256,768,0.02173653244972229
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,256,65536,0.22681493759155275
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,256,512,0.003472000112136205
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,256,768,0.012796800335248312
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,256,512,0.01983786622683207
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,256,256,0.0034389334420363107
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,256,256,0.01744426687558492
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,256,512,0.01290880044301351
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,256,128,0.0031210665901501974
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,256,128,0.017385600010553996
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,256,128,0.012945066889127096
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,256,256,0.013333333532015481
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,256,64,0.002980266759792964
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,256,1024,0.013270399967829385
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,256,64,0.015761066476504007
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,256,32,0.014804266889890037
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,128,65536,0.08919573624928792
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,128,51200,0.0709013303120931
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,128,65536,0.6932074864705403
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,128,51200,0.5482922871907552
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,128,16384,0.028895999987920123
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,128,16384,0.18484373092651368
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,256,32,0.0030250666042168934
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,128,12288,0.020382932821909585
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,128,12288,0.13720213572184245
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,128,16384,0.06719573338826498
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,128,10240,0.014069333672523499
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,128,12288,0.0532970666885376
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,128,10240,0.10887573560078938
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,128,8192,0.011788800358772278
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,128,8192,0.0793183962504069
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,128,10240,0.04746880133946736
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,128,7168,0.010428800185521444
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,128,8192,0.037336532274882
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,128,7168,0.06410346825917562
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,128,6144,0.00918933351834615
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,128,6144,0.052597332000732425
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,128,7168,0.03410346508026123
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,128,5120,0.007968000074227651
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,128,6144,0.03131306568781535
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,128,5120,0.04488960107167562
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,128,51200,0.1791829268137614
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,128,4096,0.007110400001207988
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,128,4096,0.038754133383433025
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,128,3584,0.0063498665889104204
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,128,3584,0.03506666819254557
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,128,5120,0.028871466716130573
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,128,4096,0.024694399038950602
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,128,3072,0.005849599838256836
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,128,3072,0.03290773431460063
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,128,2560,0.005161599814891815
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,128,2560,0.029678932825724286
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,128,3584,0.023081600666046143
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,128,3072,0.020246400435765585
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,128,65536,0.22360746065775552
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,128,2048,0.02762986620267232
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,128,1536,0.00417493333419164
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,128,2560,0.018988800048828126
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,128,1536,0.024850134054819742
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,128,1024,0.003457066665093104
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,128,1024,0.02215253313382467
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,128,1536,0.014121599992116294
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,128,768,0.0029834667841593427
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,128,2048,0.01660693287849426
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,128,768,0.02070186734199524
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,128,1024,0.012857600053151449
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,128,512,0.0029504001140594482
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,128,512,0.019141334295272826
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,128,256,0.002666666607062022
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,128,512,0.012389333049456278
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,128,256,0.017208532492319743
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,128,768,0.012235732873280843
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,128,256,0.012005333105723064
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,128,2048,0.004665599763393402
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,128,128,0.015711999932924905
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,128,64,0.0022154666483402254
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,128,32,0.0021898667017618816
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,128,64,0.016023466984430947
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,128,32,0.014437333742777506
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,64,65536,0.080294402440389
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,64,51200,0.06417280038197835
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,64,65536,0.6875466664632162
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,64,16384,0.026293333371480303
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,64,51200,0.5458890914916992
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,128,128,0.0025653332471847535
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,64,12288,0.014984533190727234
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,64,16384,0.1823359966278076
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,64,10240,0.012101333340009053
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3840,128,128,0.012023466825485229
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,64,12288,0.13661227226257325
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,64,8192,0.012853333353996277
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,64,7168,0.01039466659228007
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,64,10240,0.10916906992594402
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,64,8192,0.0776309331258138
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,64,7168,0.06297599871953329
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,64,5120,0.007916800181070964
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,64,6144,0.05241599877675375
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,64,4096,0.007000533243020375
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,64,5120,0.04346453348795573
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,64,3584,0.00628053347269694
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,64,4096,0.039109333356221514
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,64,3072,0.005849599838256836
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,64,3584,0.03497386773427327
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,64,2560,0.005047466854254405
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,64,3072,0.033114665746688844
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,64,2048,0.004695466657479604
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,64,2560,0.029393066962560017
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,64,1536,0.003832533210515976
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,64,2048,0.027576533953348796
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,64,1024,0.0034122665723164878
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,64,1536,0.02469759980837504
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,64,768,0.0030752000709374744
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,64,1024,0.022352000077565513
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,64,512,0.0030037333567937215
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,64,768,0.0206496000289917
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,64,256,0.002569599946339925
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,64,512,0.01889066696166992
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,64,128,0.002205866575241089
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,64,256,0.016518400112787882
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,64,64,0.002164266755183538
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,64,128,0.01636373301347097
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,64,32,0.0021930667261282604
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,64,6144,0.009150933225949604
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,64,64,0.014578133821487427
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,64,32,0.015521066387494406
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,32,65536,0.07836373647054037
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,32,51200,0.062166400750478114
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,32,16384,0.02177919944127401
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,32,16384,0.1810080051422119
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,32,12288,0.01267519990603129
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,32,51200,0.5427402496337891
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,32,65536,0.6860554377237956
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,32,10240,0.00997759997844696
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,32,12288,0.13722559611002605
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,32,8192,0.009066667159398396
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,32,7168,0.008734933535257975
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,32,10240,0.11042133172353108
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,32,8192,0.07973759969075521
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,32,6144,0.007965866724650066
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,32,7168,0.06572479804356893
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,32,5120,0.010795733332633973
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,32,6144,0.05141866604487101
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,32,4096,0.00925973355770111
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,32,5120,0.045201067129770914
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,32,3584,0.008318933347860973
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,32,4096,0.03901760180791219
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,32,3072,0.007530666887760162
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,32,2560,0.006673066814740498
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,32,3072,0.032979200283686325
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,32,2048,0.004610133171081543
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,32,2560,0.02916693290074666
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,32,2048,0.02731093366940816
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,32,1536,0.024780799945195518
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,32,1536,0.004413866500059763
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,32,768,0.0030293333033720653
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,32,1024,0.020805333058039346
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,32,512,0.002614400039116542
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,32,512,0.018651733795801796
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,32,256,0.002593066543340683
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,32,256,0.016709333658218382
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,32,128,0.002194133400917053
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,32,768,0.02252053419748942
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,32,128,0.01567359964052836
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,32,64,0.0030346666773160298
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,32,64,0.014501333236694336
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,32,32,0.0022474666436513263
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,32,32,0.014247467120488485
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3840,32,3584,0.034886399904886886
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3840,32,1024,0.0033962666988372804
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,65536,16384,4.302188618977865
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,65536,16384,2.3824363708496095
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,65536,51200,7.428450012207032
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,65536,12288,3.4351412455240884
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,65536,12288,1.6967914581298829
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,65536,51200,14.066727701822916
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,65536,10240,2.8855145772298174
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,65536,10240,1.4648831685384116
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,65536,10240,0.7995082855224609
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,65536,8192,2.28245112101237
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,65536,8192,1.1934399922688803
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,65536,12288,0.9064714431762695
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,65536,7168,2.0003946940104167
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,65536,7168,0.9932512283325196
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,65536,16384,1.213217035929362
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,65536,6144,1.5912138621012368
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,65536,6144,0.8272981643676758
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,65536,8192,0.6554965337117513
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,65536,5120,1.3165803273518881
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,65536,5120,0.7001589457194011
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,65536,7168,0.6085653305053711
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,65536,4096,1.0877802530924479
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,65536,4096,0.6157973607381184
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,65536,6144,0.5183221181233724
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,65536,3584,0.9447231928507487
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,65536,3584,0.5033856074015299
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,65536,4096,0.4228255907694499
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,65536,3072,0.8204416275024414
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,65536,3072,0.470307191212972
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,65536,5120,0.47399040857950847
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,65536,2560,0.6682463963826497
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,65536,2560,0.37134081522623696
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,65536,3072,0.33590399424235023
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,65536,2048,0.5151786804199219
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,65536,2048,0.33266026178995767
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,65536,3584,0.3822751998901367
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,65536,2560,0.320086415608724
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,65536,1536,0.3870784123738607
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,65536,1536,0.23713067372639976
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,65536,1024,0.2686762809753418
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,65536,1024,0.17612694104512533
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,65536,1024,0.2367413361867269
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,65536,768,0.21098880767822265
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,65536,768,0.14820693333943685
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,65536,1536,0.24170026779174805
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,65536,512,0.15180800755818685
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,65536,512,0.1229045311609904
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,65536,768,0.22657599449157714
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,65536,256,0.10391786893208821
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,65536,256,0.1044810692469279
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,65536,512,0.2237610658009847
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,65536,128,0.0796778678894043
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,65536,128,0.08835840225219727
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,65536,256,0.22313706080118814
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,65536,64,0.07400746345520019
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,65536,64,0.09064106941223145
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,65536,32,0.0759722630182902
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,65536,32,0.09155413309733072
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,65536,128,0.22658987045288087
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,65536,2048,0.2826826731363932
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,51200,65536,7.359933980305989
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,51200,65536,13.906582641601563
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,51200,51200,5.823932902018229
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,51200,51200,10.689190673828126
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,51200,16384,3.4520533243815104
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,51200,16384,1.837051773071289
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,51200,16384,0.9542154947916666
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,51200,12288,2.5257237752278643
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,51200,12288,1.3361024220784505
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,51200,12288,0.712888526916504
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,51200,10240,2.41367670694987
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,51200,10240,1.1848981221516928
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,51200,10240,0.6301685333251953
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,51200,8192,1.9113824208577472
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,51200,8192,0.924565315246582
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,65536,51200,4.056453450520833
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,51200,7168,1.587925338745117
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,51200,7168,0.831116803487142
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,51200,8192,0.5162037213643391
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,51200,6144,1.3087168375651042
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,51200,6144,0.6796565373738607
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,51200,7168,0.4802474657694499
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,51200,5120,1.1017888387044272
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,51200,5120,0.6038581212361653
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,51200,5120,0.3724010785420736
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,51200,4096,0.8517621358235677
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,51200,4096,0.4881162643432617
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,51200,6144,0.4088490804036458
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,51200,3584,0.7032330830891926
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,51200,3584,0.4173162778218587
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,51200,3584,0.3005557378133138
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,51200,3072,0.6025781631469727
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,51200,3072,0.351419734954834
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,51200,4096,0.32773332595825194
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,51200,2560,0.5004799842834473
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,51200,2560,0.30083840688069663
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,51200,51200,3.1712181091308596
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,51200,2048,0.43396692276000975
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,51200,2048,0.2409066677093506
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,51200,2560,0.25151039759318033
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,51200,1536,0.3064607938130697
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,51200,1536,0.2016767978668213
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,51200,3072,0.2637813409169515
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,51200,1024,0.2085866610209147
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,51200,1024,0.14406399726867675
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,51200,2048,0.22418452898661295
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,51200,768,0.1652789274851481
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,51200,768,0.11983893712361653
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,51200,1536,0.19527039527893067
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,51200,512,0.12226346333821614
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,51200,512,0.10135680039723713
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,51200,1024,0.19084374109903973
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,51200,256,0.08146986961364747
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,51200,256,0.0860543966293335
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,51200,256,0.17676374117533367
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,51200,128,0.0705183982849121
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,51200,128,0.07483946482340495
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,51200,512,0.1784490744272868
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,51200,64,0.05961813529332479
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,51200,64,0.07604693571726481
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,51200,32,0.06292479832967122
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,51200,32,0.07609386444091797
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,51200,128,0.17981759707132977
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,51200,768,0.17928853034973144
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,16384,65536,2.71705805460612
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,16384,65536,4.588849894205729
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,16384,51200,3.525593566894531
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,16384,51200,2.146533457438151
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,16384,16384,1.0482805252075196
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,16384,16384,0.7261632283528645
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,16384,16384,0.351861317952474
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,16384,12288,0.7779392242431641
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,16384,12288,0.5471935907999674
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,16384,12288,0.2647701263427734
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,16384,10240,0.675539207458496
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,16384,10240,0.4604159990946452
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,16384,10240,0.23701759974161782
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,16384,8192,0.5398837407430013
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,16384,8192,0.3645375887552897
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,16384,8192,0.19361599286397296
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,16384,7168,0.4778666814168294
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,16384,7168,0.3088362693786621
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,16384,51200,1.0815957387288413
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,16384,6144,0.3760714530944824
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,16384,6144,0.24789439837137855
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,16384,7168,0.18032959302266438
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,16384,5120,0.33248427708943684
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,16384,5120,0.22127572695414224
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,16384,5120,0.14058879216512044
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,16384,4096,0.2561354637145996
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,16384,4096,0.17504960695902508
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,16384,6144,0.1553258736928304
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,16384,3584,0.22026027043660484
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,16384,3584,0.14805866877237955
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,16384,3584,0.11248640219370525
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,16384,3072,0.1856042702992757
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,16384,3072,0.13159680366516113
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,16384,4096,0.12312320073445637
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,16384,2560,0.16147839228312175
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,16384,2560,0.11398826440175373
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,16384,2560,0.09538773695627847
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,16384,2048,0.13058666388193768
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,16384,2048,0.09656746387481689
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,16384,3072,0.09914666811625163
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,16384,1536,0.0974783976872762
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,16384,1536,0.0762997309366862
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,16384,1536,0.07154346307118734
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,16384,1024,0.06808106899261475
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,16384,1024,0.06104106505711874
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,16384,2048,0.08297279675801596
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,16384,768,0.05517653226852417
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,16384,768,0.05490560134251913
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,16384,1024,0.06683200200398763
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,16384,512,0.03976106643676758
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,16384,512,0.04876586596171061
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,16384,768,0.058115200201670325
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,16384,256,0.028483200073242187
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,16384,256,0.03943146864573161
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,16384,256,0.054347733656565346
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,16384,128,0.02217386762301127
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,16384,128,0.034175999959309894
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,16384,512,0.05612159967422485
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,16384,64,0.020055466890335084
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,16384,64,0.03502080043156942
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,16384,32,0.02065599958101908
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,16384,32,0.03551679849624634
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,16384,128,0.05583039919535319
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,12288,65536,2.1296138763427734
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,12288,65536,3.210587819417318
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,12288,51200,2.5505386352539063
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,12288,51200,1.6309280395507812
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,51200,65536,4.104429880777994
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,12288,16384,0.726468276977539
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,12288,16384,0.5444021224975586
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,16384,65536,1.369922129313151
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,12288,12288,0.5552864074707031
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,12288,12288,0.4113792101542155
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,12288,16384,0.2824181238810221
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,12288,10240,0.5265994707743327
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,12288,10240,0.3392938613891602
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,12288,12288,0.21317760149637857
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,12288,8192,0.377567990620931
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,12288,8192,0.26151466369628906
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,12288,8192,0.1553599993387858
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,12288,7168,0.326911989847819
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,12288,7168,0.2442783991495768
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,12288,10240,0.1908031940460205
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,12288,6144,0.279315185546875
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,12288,6144,0.19872105916341146
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,12288,7168,0.14430932998657225
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,12288,5120,0.23218773206075033
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,12288,5120,0.17749759356180828
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,12288,6144,0.12480213642120361
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,12288,4096,0.18825066884358724
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,12288,4096,0.14423573811848958
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,12288,4096,0.09996693134307862
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,12288,3584,0.16585493087768555
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,12288,5120,0.11432212988535564
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,12288,3584,0.12616960207621258
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,12288,3072,0.1421120007832845
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,12288,3072,0.11401066780090333
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,12288,3072,0.080022398630778
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,12288,2560,0.1261034647623698
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,12288,2560,0.09641066392262777
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,12288,3584,0.08983999888102213
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,12288,2048,0.10160000324249267
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,12288,2048,0.08102506796518961
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,12288,2048,0.0680298646291097
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,12288,1536,0.07832427024841308
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,12288,1536,0.06468586524327596
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,12288,2560,0.07729492982228597
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,12288,1024,0.05238613287607828
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,12288,1024,0.051680000623067227
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,12288,1024,0.05890239874521891
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,12288,768,0.041230932871500654
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,12288,768,0.04710400104522705
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,12288,1536,0.06114346583684286
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,12288,512,0.030696533123652142
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,12288,512,0.04138773282368978
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,12288,768,0.05386773347854614
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,12288,256,0.02179093360900879
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,12288,256,0.03338666756947835
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,12288,512,0.05425066550572714
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,12288,128,0.017358932892481485
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,12288,256,0.05386879841486612
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,12288,128,0.029808000723520918
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,12288,64,0.015357866883277893
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,12288,64,0.03002133369445801
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,12288,32,0.015736533204714458
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,12288,32,0.02919253309567769
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,12288,128,0.05488106807072958
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,10240,65536,1.8459669748942058
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,10240,65536,2.615557352701823
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,10240,51200,2.071452840169271
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,10240,51200,1.4575658162434897
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,12288,51200,0.8512224197387696
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,10240,16384,0.5831797281901042
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,10240,16384,0.47096745173136395
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,10240,16384,0.2376479943593343
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,10240,12288,0.4377279917399089
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,10240,12288,0.353330135345459
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,12288,65536,1.0894026438395181
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,10240,10240,0.3803914705912272
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,10240,10240,0.30133012135823567
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,10240,12288,0.18098773956298828
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,10240,8192,0.2985973358154297
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,10240,8192,0.24810667037963868
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,10240,10240,0.16237653096516927
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,10240,7168,0.27111358642578126
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,10240,7168,0.22524800300598144
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,10240,8192,0.13195839722951253
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,10240,6144,0.24078933397928873
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,10240,6144,0.17562774022420247
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,10240,7168,0.1235477368036906
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,10240,5120,0.20173226992289223
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,10240,5120,0.15296106338500975
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,10240,6144,0.10641173521677653
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,10240,4096,0.1561855951944987
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,10240,4096,0.1227946678797404
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,10240,5120,0.098963197072347
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,10240,3584,0.14117973645528156
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,10240,3584,0.10870292981465657
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,10240,51200,0.7062613169352214
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,10240,3072,0.1190677324930827
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,10240,3072,0.09701546827952066
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,10240,3072,0.06922773520151773
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,10240,4096,0.08455680211385092
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,10240,2560,0.10509226322174073
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,10240,2560,0.0825984001159668
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,10240,2048,0.08081706364949545
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,10240,2048,0.07172053654988607
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,10240,3584,0.07738026777903238
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,10240,1536,0.06373120148976644
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,10240,1536,0.05864426692326864
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,10240,1536,0.05531200170516968
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,10240,1024,0.04440746704737346
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,10240,1024,0.0472213347752889
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,10240,2048,0.061034667491912845
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,10240,768,0.03394986788431804
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,10240,768,0.04335893392562866
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,10240,2560,0.06692799727121988
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,10240,512,0.025165865818659466
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,10240,512,0.03729493220647176
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,10240,1024,0.05372159878412882
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,10240,256,0.018403200308481853
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,10240,256,0.031143466631571453
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,10240,768,0.05004693269729614
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,10240,128,0.016169599692026772
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,10240,128,0.02746559977531433
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,10240,512,0.04997973442077637
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,10240,64,0.01376426617304484
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,10240,64,0.02714346647262573
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,10240,256,0.049466665585835776
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,10240,32,0.014097066720326743
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,10240,32,0.027723733584086103
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,10240,128,0.050179199377695716
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,8192,65536,1.6253611246744792
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,8192,65536,2.2094561258951826
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,8192,51200,1.7013429005940754
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,8192,51200,1.2978815714518228
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,8192,16384,0.517358938852946
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,8192,16384,0.44068374633789065
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,10240,65536,0.893397331237793
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,8192,12288,0.3802538553873698
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,8192,12288,0.31954240798950195
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,8192,16384,0.2091082731882731
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,8192,10240,0.32248748143514
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,8192,10240,0.2548362731933594
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,8192,12288,0.15945067405700683
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,8192,8192,0.2546229362487793
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,8192,8192,0.2061471939086914
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,8192,10240,0.14234986305236816
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,8192,7168,0.22240959803263344
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,8192,7168,0.17625279426574708
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,8192,7168,0.10843839645385742
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,8192,6144,0.1900448004404704
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,8192,6144,0.1501738707224528
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,8192,8192,0.116156800587972
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,8192,5120,0.15838932991027832
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,8192,5120,0.12592533429463704
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,8192,5120,0.08604479630788167
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,8192,4096,0.1267584005991618
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,8192,4096,0.10413546562194824
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,8192,6144,0.0940394639968872
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,8192,3584,0.11220479806264241
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,8192,3584,0.09347200393676758
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,8192,4096,0.07404159704844157
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,8192,3072,0.09605973561604818
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,8192,3072,0.08399573167165121
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,8192,3584,0.06586986780166626
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,8192,2560,0.08144106864929199
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,8192,2560,0.07291200160980224
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,8192,2560,0.05485546588897705
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,8192,2048,0.06644159952799479
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,8192,2048,0.06383999983469645
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,8192,3072,0.05683199961980184
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,8192,1536,0.05044159889221191
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,8192,1536,0.05123626788457235
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,8192,1536,0.0401802659034729
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,8192,1024,0.03500800132751465
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,8192,1024,0.04114453395207723
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,8192,2048,0.04730133215586345
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,8192,768,0.027381332715352376
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,8192,768,0.0370901346206665
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,8192,1024,0.03802880048751831
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,8192,512,0.020172800620396933
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,8192,768,0.03421226739883423
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,8192,512,0.03296746611595154
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,8192,256,0.014664533734321594
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,8192,256,0.027109332879384357
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,8192,256,0.034050134817759196
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,8192,128,0.012010666728019714
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,8192,128,0.023405865828196207
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,8192,128,0.03498239914576213
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,8192,64,0.010801066954930622
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,8192,64,0.025620265801747637
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,8192,32,0.011043199896812439
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,8192,32,0.023282132546106973
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,8192,512,0.03428266843159993
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,7168,65536,1.868346659342448
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,7168,65536,1.4927946726481118
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,7168,51200,1.3896213531494142
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,7168,51200,1.1737866719563803
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,8192,51200,0.6242432276407878
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,7168,16384,0.42938985824584963
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,7168,16384,0.3979381243387858
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,7168,16384,0.20203733444213867
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,7168,12288,0.3244757334391276
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,7168,12288,0.28916266759236653
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,8192,65536,0.7893194834391276
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,7168,10240,0.27308266957600913
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,7168,10240,0.23215039571126303
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,7168,12288,0.14573440551757813
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,7168,8192,0.22361493110656738
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,7168,8192,0.18978880246480306
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,7168,10240,0.12848320007324218
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,7168,7168,0.20776106516520182
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,7168,7168,0.15516160329182943
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,7168,8192,0.10469653606414794
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,7168,6144,0.1797994613647461
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,7168,6144,0.1365023930867513
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,7168,7168,0.09831466674804687
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,7168,5120,0.13521812756856283
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,7168,5120,0.11599893569946289
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,7168,6144,0.08684693177541097
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,7168,4096,0.11727360089619954
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,7168,4096,0.09728319644927978
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,7168,5120,0.07931733131408691
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,7168,3584,0.095524263381958
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,7168,3584,0.08572479883829752
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,7168,3584,0.06094826857248942
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,7168,3072,0.08271146615346273
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,7168,3072,0.07871359984079997
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,7168,4096,0.0674570639928182
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,7168,2560,0.07039999961853027
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,7168,51200,0.5840245564778646
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,7168,2560,0.06777706940968832
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,7168,2048,0.05798293352127075
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,7168,2048,0.059425067901611325
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,7168,3072,0.05395413239796957
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,7168,1536,0.043545599778493246
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,7168,1536,0.047549867630004884
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,7168,2048,0.04628906647364299
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,7168,1024,0.03035306731859843
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,7168,1024,0.0392192006111145
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,7168,2560,0.0519594669342041
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,7168,768,0.02408000032107035
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,7168,768,0.03517546653747559
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,7168,1024,0.03939626614252727
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,7168,768,0.035867734750111895
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,7168,512,0.018129066626230875
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,7168,1536,0.04075839916865031
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,7168,512,0.030894933144251506
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,7168,256,0.013244799772898355
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,7168,256,0.025254400571187337
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,7168,128,0.014178133010864258
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,7168,128,0.0248799999554952
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,7168,128,0.036287999153137206
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,7168,64,0.013294933239618936
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,7168,256,0.036235733826955156
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,7168,64,0.02472426692644755
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,7168,32,0.01363200048605601
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,7168,32,0.024903466304143272
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,7168,512,0.036600534121195474
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,6144,65536,1.5953706105550132
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,6144,65536,1.3818251291910806
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,6144,51200,1.0899188995361329
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,6144,51200,1.2774976094563804
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,6144,16384,0.4298837343851726
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,6144,16384,0.36504640579223635
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,7168,65536,0.7359776178995768
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,6144,12288,0.2946197191874186
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,6144,12288,0.2697087923685709
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,6144,16384,0.1708629290262858
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,6144,10240,0.2509781360626221
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,6144,10240,0.22137600580851235
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,6144,10240,0.11360533237457275
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,6144,8192,0.20267945925394693
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,6144,8192,0.17011307080586752
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,6144,12288,0.1264298677444458
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,6144,7168,0.1808362642923991
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,6144,7168,0.14861013094584147
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,6144,8192,0.09268266359965006
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,6144,6144,0.1519957383473714
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,6144,7168,0.08697280089060465
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,6144,6144,0.12976213296254474
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,6144,5120,0.12391359806060791
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,6144,5120,0.11266880035400391
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,6144,5120,0.07084906895955403
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,6144,4096,0.09813866615295411
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,6144,4096,0.094541867574056
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,6144,6144,0.07764480113983155
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,6144,3584,0.08794026374816895
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,6144,3584,0.08270400365193685
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,6144,4096,0.060022401809692386
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,6144,3072,0.07506133715311686
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,6144,3072,0.07619840304056803
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,6144,3584,0.05358186562856039
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,6144,2560,0.06393813292185466
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,6144,2560,0.06394666830698649
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,6144,2560,0.04539413452148437
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,6144,2048,0.052242132027943936
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,6144,2048,0.05274666547775268
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,6144,3072,0.047380268573760986
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,6144,1536,0.03989866574605306
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,6144,1536,0.04318933486938477
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,6144,1536,0.03668373425801595
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,6144,1024,0.025586134195327757
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,6144,1024,0.036136531829833986
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,6144,2048,0.041093333562215166
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,6144,768,0.02021119991938273
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,6144,768,0.03242239952087402
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,6144,1024,0.03537173271179199
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,6144,512,0.015289599696795145
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,6144,768,0.03210773269335429
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,6144,512,0.028828799724578857
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,6144,256,0.011281067132949829
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,6144,256,0.024836266040802003
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,6144,256,0.03216533263524373
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,6144,128,0.010779733459154766
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,6144,128,0.024979199965794882
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,6144,512,0.032561065753300984
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,6144,64,0.009566932916641235
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,6144,128,0.03253759940465291
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,6144,64,0.02068159977595011
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,6144,32,0.009916800260543823
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,6144,32,0.020623999834060668
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,5120,65536,1.15141970316569
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,5120,65536,1.2116522471110025
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,5120,51200,0.9133162816365561
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,5120,51200,0.9595765431722005
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,6144,51200,0.49707625706990555
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,5120,16384,0.3006208101908366
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,5120,16384,0.3131882667541504
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,5120,16384,0.1441215991973877
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,5120,12288,0.22857173283894858
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,5120,12288,0.2374549388885498
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,6144,65536,0.6240863800048828
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,5120,10240,0.18790507316589355
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,5120,10240,0.19144533475240072
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,5120,12288,0.11094186305999756
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,5120,8192,0.1539039929707845
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,5120,8192,0.14957332611083984
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,5120,10240,0.09924266338348389
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,5120,7168,0.13269440333048504
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,5120,7168,0.13076053460439047
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,5120,8192,0.08220266501108806
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,5120,6144,0.11671360333760579
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,5120,6144,0.11604800224304199
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,5120,7168,0.07787840366363526
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,5120,5120,0.09658453464508057
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,5120,5120,0.0995317300160726
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,5120,5120,0.06223680178324381
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,5120,4096,0.07832427024841308
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,5120,4096,0.08385813236236572
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,5120,6144,0.06909120082855225
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,5120,3584,0.0700160026550293
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,5120,3584,0.07295680046081543
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,5120,3584,0.04827733437220256
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,5120,3072,0.06053013404210409
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,5120,3072,0.0660693327585856
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,5120,4096,0.05276906490325928
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,5120,2560,0.05440746545791626
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,5120,2560,0.055693864822387695
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,5120,51200,0.41350399653116865
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,5120,2048,0.041121065616607666
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,5120,2048,0.04710719982783
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,5120,2560,0.041517865657806394
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,5120,1536,0.0339957316716512
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,5120,1536,0.03950933218002319
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,5120,3072,0.04324373404184977
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,5120,1024,0.022360533475875854
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,5120,1024,0.032995200157165526
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,5120,2048,0.037487999598185225
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,5120,768,0.01726400057474772
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,5120,768,0.029446399211883544
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,5120,1024,0.032459733883539836
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,5120,512,0.013268267114957174
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,5120,512,0.02691199978192647
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,5120,1536,0.03418773412704468
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,5120,256,0.009713066617647807
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,5120,256,0.02295573353767395
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,5120,256,0.030937600135803222
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,5120,128,0.008828799923261006
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,5120,512,0.03087573250134786
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,5120,128,0.02263573408126831
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,5120,64,0.007964799801508587
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,5120,64,0.021592533588409422
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,5120,32,0.00801813354094823
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,5120,128,0.030869332949320476
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,5120,32,0.02143893241882324
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,4096,65536,1.0275967915852866
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,4096,65536,1.1081685384114583
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,4096,51200,0.7845439910888672
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,5120,768,0.030852266152699787
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,4096,51200,0.8800202687581381
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,4096,16384,0.24779094060262047
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,4096,16384,0.29556798934936523
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,4096,16384,0.1364309310913086
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,4096,12288,0.18803839683532714
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,4096,12288,0.2149792035420736
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,5120,65536,0.5189119974772136
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,4096,10240,0.15964053471883138
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,4096,10240,0.1717898686726888
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,4096,10240,0.09377386569976806
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,4096,8192,0.1287349303563436
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,4096,8192,0.13082026640574138
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,4096,12288,0.10451200008392333
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,4096,7168,0.11457493305206298
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,4096,7168,0.11550186475118
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,4096,7168,0.07214293479919434
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,4096,6144,0.09747412999471029
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,4096,6144,0.10211733182271321
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,4096,8192,0.07675946553548177
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,4096,5120,0.07858666578928629
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,4096,5120,0.08836800257364909
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,4096,5120,0.055581867694854736
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,4096,4096,0.06303040186564127
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,4096,4096,0.07369813124338785
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,4096,6144,0.06263680060704549
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,4096,3584,0.05626879930496216
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,4096,3584,0.0639306664466858
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,4096,4096,0.046911998589833574
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,4096,3072,0.04772053162256877
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,4096,3072,0.057494401931762695
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,4096,3584,0.041628801822662355
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,4096,2560,0.040250666936238605
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,4096,2560,0.048853333791097006
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,4096,2560,0.03503893216451009
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,4096,2048,0.0325055996576945
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,4096,2048,0.04339306751887004
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,4096,3072,0.03639039993286133
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,4096,1536,0.024889600276947022
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,4096,1536,0.037033601601918535
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,4096,1536,0.026021333535512288
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,4096,1024,0.01805973251660665
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,4096,1024,0.03189546664555867
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,4096,2048,0.030101333061854047
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,4096,768,0.01404906709988912
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,4096,768,0.02903253237406413
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,4096,768,0.023462400833765665
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,4096,1024,0.02431679964065552
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,4096,512,0.010696533322334289
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,4096,512,0.025492266813913984
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,4096,256,0.008306133250395458
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,4096,256,0.02177600065867106
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,4096,256,0.023618133862813313
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,4096,128,0.00714026689529419
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,4096,128,0.020773333311080933
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,4096,512,0.023965867360432942
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,4096,64,0.006308266520500183
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,4096,64,0.02163413365681966
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,4096,128,0.02394346594810486
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,4096,32,0.006675200164318084
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,4096,32,0.019966934124628702
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,3584,65536,0.918561045328776
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,3584,65536,1.0555317560831705
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,3584,51200,0.6888618469238281
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,3584,51200,0.832485326131185
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,4096,51200,0.4035210609436035
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,3584,16384,0.2264469305674235
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,3584,16384,0.27240638732910155
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,3584,16384,0.11643093427022297
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,3584,12288,0.17191360791524252
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,3584,12288,0.2052992026011149
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,4096,65536,0.5114272117614747
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,3584,10240,0.15301547050476075
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,3584,10240,0.17102293968200682
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,3584,12288,0.0916693369547526
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,3584,8192,0.12159573237101237
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,3584,8192,0.12961173057556152
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,3584,10240,0.08297173182169595
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,3584,7168,0.10740266640981037
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,3584,7168,0.11664640108744304
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,3584,8192,0.06887253125508627
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,3584,6144,0.09013013044993082
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,3584,7168,0.06549439827601114
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,3584,6144,0.10117546717325847
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,3584,5120,0.07417386372884115
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,3584,5120,0.0873855988184611
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,3584,5120,0.051191465059916175
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,3584,4096,0.06010773181915283
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,3584,4096,0.07099093596140543
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,3584,6144,0.05666026671727499
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,3584,3584,0.053941333293914796
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,3584,3584,0.0626261313756307
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,3584,4096,0.043806934356689455
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,3584,3072,0.04603093465169271
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,3584,3072,0.05533653497695923
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,3584,3584,0.0398911992708842
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,3584,2560,0.03910613457361857
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,3584,2560,0.04850773413976033
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,3584,2560,0.03466026782989502
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,3584,2048,0.03214079936345418
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,3584,2048,0.04311360120773315
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,3584,3072,0.03632746537526448
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,3584,1536,0.024769065777460735
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,3584,1536,0.03646186590194702
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,3584,1536,0.027741867303848266
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,3584,1024,0.018210132916768394
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,3584,1024,0.031109333038330078
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,3584,2048,0.030932267506917317
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,3584,768,0.01458560029665629
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,3584,768,0.027346134185791016
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,3584,768,0.02797973354657491
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,3584,512,0.011161599556605022
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,3584,1024,0.026190932591756182
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,3584,512,0.025153066714604693
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,3584,256,0.008291199803352356
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,3584,256,0.023873066902160643
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,3584,256,0.02829866607983907
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,3584,128,0.006701866785685222
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,3584,128,0.020156800746917725
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,3584,512,0.028391466538111372
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,3584,64,0.006285866598288219
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,3584,64,0.01890773375829061
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,3584,32,0.00631573349237442
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,3584,32,0.021406932671864828
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,3584,128,0.025165865818659466
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,3072,65536,0.7918997446695963
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,3072,65536,0.9943637212117513
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,3584,51200,0.3312896092732748
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,3072,51200,0.5945632298787434
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,3072,51200,0.7736181259155274
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,3072,16384,0.18302292823791505
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,3072,16384,0.2504874706268311
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,3072,16384,0.11598186492919922
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,3072,12288,0.1389237403869629
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,3072,12288,0.18700052897135416
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,3584,65536,0.41534827550252273
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,3072,10240,0.12459200223286945
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,3072,10240,0.15243627230326334
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,3072,12288,0.09068693319956461
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,3072,8192,0.0980405330657959
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,3072,8192,0.11666666666666667
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,3072,10240,0.08202026685078939
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,3072,7168,0.08525119622548422
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,3072,7168,0.1029589335123698
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,3072,7168,0.06388053496678671
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,3072,6144,0.07200640042622884
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,3072,6144,0.0905344009399414
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,3072,8192,0.06840000152587891
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,3072,5120,0.0604362686475118
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,3072,5120,0.07815786997477213
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,3072,6144,0.05546773274739584
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,3072,4096,0.047144532203674316
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,3072,4096,0.062046933174133304
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,3072,5120,0.049556267261505124
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,3072,3584,0.04183679819107056
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,3072,3584,0.05457813342412313
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,3072,3584,0.039188265800476074
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,3072,3072,0.03595840136210124
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,3072,4096,0.043169065316518145
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,3072,3072,0.05026559829711914
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,3072,2560,0.030428800980250043
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,3072,2560,0.04359039862950643
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,3072,2560,0.03333013256390889
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,3072,2048,0.02483839988708496
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,3072,2048,0.0392576018969218
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,3072,3072,0.035225598017374675
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,3072,1536,0.019291732708613077
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,3072,1536,0.03309226632118225
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,3072,1536,0.02691199978192647
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,3072,1024,0.013493333260218301
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,3072,1024,0.029102933406829835
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,3072,1024,0.025543467203776045
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,3072,768,0.010821333527565003
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,3072,768,0.0288917342821757
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,3072,2048,0.030237867434819536
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,3072,512,0.00844586690266927
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,3072,512,0.024473599592844644
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,3072,768,0.026715733607610065
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,3072,256,0.007295999924341838
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,3072,256,0.020925867557525634
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,3072,256,0.026140799125035603
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,3072,512,0.024312533934911094
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,3072,128,0.005875200033187866
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,3072,128,0.01897066632906596
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,3072,64,0.005485866467158
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,3072,128,0.0243776003519694
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,3072,64,0.019602133830388387
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,3072,32,0.005541333556175232
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,3072,32,0.018974934021631876
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,2560,65536,0.5702912012736003
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,2560,65536,0.9008021036783853
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,3072,51200,0.33303467432657874
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,2560,51200,0.4504757245381673
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,2560,51200,0.715939203898112
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,2560,16384,0.14779094060262044
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,2560,16384,0.23238933881123863
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,2560,16384,0.09349653720855713
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,2560,12288,0.11272532939910888
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,2560,12288,0.17565973599751789
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,3072,65536,0.41681814193725586
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,2560,10240,0.10016746520996093
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,2560,10240,0.142138671875
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,2560,12288,0.0740938663482666
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,2560,8192,0.08185813426971436
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,2560,8192,0.10933866500854492
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,2560,10240,0.06778879960378012
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,2560,7168,0.07038186391194662
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,2560,7168,0.09623573621114095
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,2560,7168,0.05141119956970215
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,2560,6144,0.05989120006561279
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,2560,8192,0.0561621348063151
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,2560,6144,0.08405760129292807
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,2560,5120,0.049166933695475264
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,2560,5120,0.07091946601867676
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,2560,5120,0.04178026517232259
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,2560,4096,0.039128533999125165
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,2560,4096,0.057095468044281006
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,2560,6144,0.04549013376235962
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,2560,3584,0.03470720052719116
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,2560,3584,0.050113066037495935
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,2560,4096,0.03538346687952677
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,2560,3072,0.02997973362604777
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,2560,3584,0.03262719909350077
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,2560,3072,0.04615360101064046
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,2560,2560,0.025191466013590496
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,2560,2560,0.041205334663391116
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,2560,2560,0.02722240090370178
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,2560,2048,0.020959999163945517
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,2560,2048,0.03706560134887695
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,2560,3072,0.029045333464940388
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,2560,1536,0.016760534048080443
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,2560,1536,0.031660799185434976
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,2560,1536,0.021388800938924153
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,2560,1024,0.012085333466529846
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,2560,1024,0.02730453411738078
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,2560,2048,0.024309333165486655
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,2560,768,0.009563733140627544
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,2560,768,0.02479040026664734
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,2560,768,0.01858560045560201
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,2560,512,0.007561600208282471
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,2560,1024,0.020243199666341145
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,2560,512,0.022882133722305298
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,2560,512,0.018802134195963542
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,2560,256,0.006041599810123444
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,2560,256,0.020193066199620566
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,2560,256,0.018679465850194296
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,2560,128,0.005644799768924713
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,2560,128,0.01837973395983378
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,2560,128,0.018539732694625853
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,2560,64,0.005049600203831991
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,2560,64,0.01763413349787394
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,2560,32,0.005196799834569296
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,2560,32,0.01650773286819458
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,2048,65536,0.4861568133036296
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,2048,65536,0.8584458669026693
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,2048,51200,0.3854581197102865
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,2048,51200,0.6781845092773438
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,2560,51200,0.2553642590840658
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,2048,16384,0.12340266704559326
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,2048,16384,0.22055999437967935
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,2048,16384,0.09146347045898437
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,2048,12288,0.09361920356750489
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,2048,12288,0.16783253351847333
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,2560,65536,0.3186101277669271
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,2048,10240,0.08044373194376628
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,2048,10240,0.135641606648763
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,2048,12288,0.07208213011423746
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,2048,8192,0.06610986789067587
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,2048,8192,0.1027616024017334
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,2048,10240,0.06541759967803955
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,2048,7168,0.05613013505935669
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,2048,7168,0.08943466345469156
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,2048,8192,0.05426773230234781
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,2048,6144,0.0485919992129008
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,2048,6144,0.0782965342203776
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,2048,7168,0.04902506669362386
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,2048,5120,0.039648000399271646
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,2048,5120,0.06509013175964355
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,2048,5120,0.04009386698404948
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,2048,4096,0.031627732515335086
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,2048,4096,0.05338666836420695
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,2048,6144,0.04410880009333293
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,2048,3584,0.028143999973932902
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,2048,3584,0.04731733401616414
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,2048,4096,0.03381333351135254
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,2048,3072,0.02435520092646281
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,2048,3072,0.04420906702677409
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,2048,3584,0.030874667565027873
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,2048,2560,0.02132586638132731
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,2048,2560,0.03915946483612061
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,2048,3072,0.027318400144577027
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,2048,2048,0.01725013256072998
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,2048,2048,0.03522666692733765
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,2048,2560,0.025703465938568114
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,2048,2048,0.022691200176874794
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,2048,1536,0.013639466961224875
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,2048,1536,0.030483200152715044
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,2048,1024,0.010038399696350097
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,2048,1024,0.02658560077349345
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,2048,51200,0.25468692779541013
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,2048,1024,0.018195199966430663
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,2048,768,0.024755199750264488
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,2048,768,0.01657386620839437
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,2048,512,0.006720000008742015
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,2048,768,0.008426666259765625
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,2048,1536,0.01950719952583313
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,2048,256,0.005491200089454651
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,2048,256,0.01919040083885193
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,2048,512,0.016665599743525186
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,2048,128,0.005048533280690512
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,2048,128,0.017336533466974894
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,2048,256,0.016636799772580466
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,2048,64,0.004271999994913737
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,2048,64,0.01685333251953125
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,2048,128,0.017025067408879598
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,2048,32,0.004785066843032837
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,2048,512,0.02259520093599955
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,2048,32,0.019181867440541588
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,1536,65536,0.3700127919514974
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,1536,51200,0.2960277239481608
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,1536,65536,0.7916095733642579
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,1536,51200,0.6246805191040039
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,1536,16384,0.09304853280385336
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,1536,16384,0.2035424073537191
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,1536,16384,0.0921066681543986
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,1536,12288,0.07068479855855306
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,1536,12288,0.15465386708577472
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,2048,65536,0.31657705307006834
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,1536,10240,0.06844267050425211
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,1536,10240,0.12898133595784506
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,1536,10240,0.06628906726837158
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,1536,8192,0.05431146621704101
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,1536,8192,0.09799253145853679
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,1536,12288,0.07339093685150147
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,1536,7168,0.047364266713460286
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,1536,7168,0.08724160194396972
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,1536,7168,0.05008000135421753
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,1536,6144,0.0415829340616862
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,1536,6144,0.0729087988535563
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,1536,8192,0.05380906661351522
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,1536,5120,0.03392213185628255
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,1536,5120,0.0596501350402832
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,1536,6144,0.04442773262659709
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,1536,4096,0.027481599648793535
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,1536,4096,0.04982293446858724
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,1536,5120,0.040770133336385087
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,1536,3584,0.02482773264249166
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,1536,3584,0.044606932004292804
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,1536,4096,0.035069866975148516
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,1536,3072,0.021683200200398763
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,1536,3072,0.041338666280110674
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,1536,3584,0.0321941335995992
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,1536,2560,0.01842133402824402
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,1536,2560,0.035936001936594644
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,1536,3072,0.029257599512736005
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,1536,2048,0.015180800358454385
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,1536,2048,0.033539199829101564
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,1536,2560,0.027646932999293012
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,1536,1536,0.012131200234095255
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,1536,1536,0.02884693344434102
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,1536,1536,0.022359466552734374
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,1536,1024,0.008788266777992248
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,1536,2048,0.024754132827123007
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,1536,1024,0.02494293252627055
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,1536,768,0.007303466896216075
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,1536,768,0.02451733350753784
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,1536,51200,0.2546453317006429
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,1536,512,0.005942399799823761
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,1536,768,0.019398399194081626
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,1536,512,0.02188160022099813
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,1536,256,0.005087999999523163
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,1536,256,0.018691200017929076
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,1536,512,0.0197269340356191
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,1536,128,0.004227200150489807
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,1536,1024,0.02061333258946737
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,1536,128,0.016637866695721946
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,1536,64,0.003822933385769526
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,1536,64,0.016825600465138753
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,1536,32,0.003930666546026865
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,1536,128,0.020526933670043945
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,1536,32,0.015090133746465048
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,1024,65536,0.25647145907084146
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,1024,65536,0.7483733495076497
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,1536,256,0.019492266575495402
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,1024,51200,0.20146452585856117
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,1024,51200,0.5925386428833008
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,1024,16384,0.0642677346865336
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,1024,16384,0.1896458625793457
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,1024,16384,0.06853226820627847
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,1024,12288,0.04860906600952149
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,1024,12288,0.1448853333791097
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,1536,65536,0.32025171915690104
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,1024,10240,0.04268480141957601
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,1024,10240,0.11769386927286785
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,1024,12288,0.054554665088653566
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,1024,8192,0.032866134246190386
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,1024,8192,0.08757013479868571
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,1024,10240,0.05019946495691935
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,1024,7168,0.029357866446177168
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,1024,7168,0.07454400062561035
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,1024,8192,0.03968853155771891
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,1024,6144,0.024526933828989662
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,1024,6144,0.06313706636428833
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,1024,7168,0.035922133922576906
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,1024,5120,0.021358933051427206
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,1024,5120,0.0528160015741984
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,1024,6144,0.03253440062204997
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,1024,4096,0.01713706652323405
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,1024,4096,0.04495893319447835
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,1024,5120,0.030089600880940752
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,1024,3584,0.01566933294137319
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,1024,3584,0.040055465698242185
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,1024,4096,0.026425600051879883
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,1024,3072,0.013703466455141703
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,1024,3072,0.038610132535298665
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,1024,3584,0.02403306762377421
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,1024,2560,0.011901866396268208
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,1024,2560,0.033267199993133545
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,1024,3072,0.02188053329785665
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,1024,2048,0.010011733571688334
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,1024,2048,0.030979200204213457
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,1024,2560,0.020286933581034342
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,1024,1536,0.008419199784596761
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,1024,1536,0.02684906721115112
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,1024,2048,0.018186666568120322
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,1024,51200,0.18089493115743
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,1024,1024,0.006663466493288676
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,1024,1024,0.0230730672677358
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,1024,768,0.005957333246866862
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,1024,768,0.022707200050354003
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,1024,1024,0.014492799838383993
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,1024,512,0.004654933512210846
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,1024,768,0.013275733590126038
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,1024,512,0.013687466581662497
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,1024,256,0.0038719999293486277
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,1024,1536,0.015677866339683533
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,1024,256,0.017857066790262856
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,1024,128,0.003420799970626831
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,1024,128,0.016457600394884746
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,1024,256,0.01362773378690084
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,1024,64,0.0033674667278925574
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,1024,128,0.014590932925542196
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,1024,64,0.01630506714185079
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,1024,32,0.0034058667719364167
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,1024,512,0.02058346668879191
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,1024,32,0.014604799946149192
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,768,65536,0.2028554598490397
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,768,65536,0.7216906865437825
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,768,51200,0.1621514638264974
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,768,51200,0.5556650797526042
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,768,16384,0.0542143980662028
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,768,16384,0.1847317377726237
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,1024,65536,0.22447039286295573
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,768,12288,0.040695468584696456
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,768,12288,0.13852373758951825
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,768,16384,0.06880746682484945
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,768,10240,0.03357760111490886
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,768,10240,0.11136319637298583
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,768,12288,0.055794131755828855
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,768,8192,0.02563520073890686
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,768,8192,0.08238720099131266
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,768,10240,0.05077653328577677
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,768,7168,0.022536534070968627
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,768,7168,0.06964053312937418
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,768,8192,0.03992319901784261
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,768,7168,0.03696639935175578
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,768,6144,0.019128533204396565
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,768,6144,0.05740906794865926
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,768,5120,0.016375466187795003
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,768,5120,0.04869653383890788
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,768,5120,0.03129706581433614
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,768,4096,0.013863466183344521
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,768,4096,0.041161600748697916
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,768,6144,0.03336319923400879
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,768,3584,0.011829333504041036
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,768,3584,0.03849173386891683
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,768,3584,0.025139200687408447
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,768,4096,0.026953599850336712
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,768,3072,0.03694080114364624
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,768,3072,0.010664533575375874
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,768,2560,0.009480533003807069
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,768,2560,0.03309119939804077
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,768,3072,0.022663466135660806
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,768,2048,0.008059733112653096
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,768,2048,0.028819199403127032
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,768,2560,0.02110613385836283
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,768,1536,0.006748799979686737
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,768,2048,0.01887680093447367
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,768,1536,0.026746666431427
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,768,1024,0.0058773333827654515
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,768,1024,0.023129600286483764
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,768,1536,0.01656000018119812
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,768,768,0.004682666560014089
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,768,768,0.02145919998486837
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,768,1024,0.015847466389338174
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,768,512,0.0043381333351135256
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,768,512,0.020899200439453126
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,768,768,0.01490239997704824
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,768,256,0.0034645333886146545
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,768,51200,0.18122453689575196
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,768,256,0.0184170663356781
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,768,128,0.003420799970626831
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,768,128,0.015470932920773825
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,768,512,0.014962133765220643
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,768,256,0.015624533096949259
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,768,64,0.0031328000128269195
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,768,32,0.0030080000559488933
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,768,128,0.015313067038853965
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,768,64,0.015107199549674988
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,768,32,0.016212266683578492
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,512,65536,0.16415573755900065
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,512,51200,0.12931946913401288
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,512,65536,0.6817610422770183
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,512,51200,0.542897097269694
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,512,16384,0.041619201501210526
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,512,16384,0.18670825958251952
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,512,16384,0.0670517365137736
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,512,12288,0.031658667325973514
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,512,12288,0.14060373306274415
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,768,65536,0.2244490623474121
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,512,10240,0.02485760052998861
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,512,10240,0.11314880053202311
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,512,10240,0.047969067096710206
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,512,12288,0.053318401177724205
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,512,8192,0.019054933389027914
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,512,8192,0.07678186893463135
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,512,7168,0.016353066762288412
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,512,7168,0.06417493422826132
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,512,7168,0.035004798571268717
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,512,6144,0.014178133010864258
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,512,8192,0.0378325343132019
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,512,6144,0.053179732958475744
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,512,5120,0.012332800030708312
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,512,5120,0.04593066771825154
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,512,5120,0.029289599259694415
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,512,6144,0.03166933258374532
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,512,4096,0.010371200243631999
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,512,4096,0.039610668023427324
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,512,3584,0.009170132875442504
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,512,3584,0.03623786767323812
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,512,4096,0.025310933589935303
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,512,3584,0.02335360050201416
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,512,3072,0.008221866687138875
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,512,3072,0.03460586468378703
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,512,3072,0.02108373244603475
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,512,2560,0.007542400062084198
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,512,2560,0.030913066864013673
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,512,51200,0.17891519864400227
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,512,2048,0.006376533210277558
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,512,2048,0.028913066784540815
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,512,2560,0.019848533471425376
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,512,1536,0.00553706685702006
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,512,1536,0.026311467091242473
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,512,1024,0.0047189335028330484
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,512,2048,0.017373865842819212
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,512,1536,0.015271466970443726
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,512,768,0.003917866696914038
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,512,768,0.021604265769322714
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,512,1024,0.013660800457000733
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,512,512,0.003621333340803782
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,512,512,0.01895466645558675
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,512,768,0.01286079982916514
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,512,256,0.0030517332255840302
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,512,256,0.017735467354456583
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,512,512,0.012875733772913614
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,512,128,0.0029866665601730345
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,512,128,0.01621333360671997
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,512,256,0.013236266374588013
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,512,64,0.002661333233118057
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,512,1024,0.02288533250490824
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,512,64,0.014920533696810404
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,512,32,0.002865066627661387
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,512,32,0.016237866878509522
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,256,65536,0.11335786978403728
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,256,65536,0.6546944300333659
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,256,51200,0.08851093451182047
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,512,128,0.013643733660380044
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,256,51200,0.5206944147745769
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,256,16384,0.03596800168355306
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,256,16384,0.17780373891194662
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,512,65536,0.22278186480204262
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,256,12288,0.02712106704711914
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,256,12288,0.13319146633148193
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,256,16384,0.06720213095347086
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,256,10240,0.022127999862035116
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,256,10240,0.10770986874898274
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,256,12288,0.053657599290212
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,256,10240,0.04651840130488078
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,256,8192,0.01916159987449646
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,256,8192,0.07525333563486734
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,256,7168,0.01584106683731079
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,256,7168,0.06058986584345499
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,256,7168,0.03421226739883423
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,256,6144,0.014045866330464682
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,256,6144,0.05094079971313477
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,256,8192,0.03655253251393636
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,256,5120,0.012475732962290447
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,256,5120,0.04338773488998413
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,256,6144,0.031267199913660684
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,256,4096,0.007547733187675476
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,256,4096,0.03755520184834798
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,256,51200,0.17852479616800945
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,256,3584,0.007121066749095917
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,256,5120,0.029021867116292316
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,256,3584,0.0348416010538737
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,256,4096,0.0252074658870697
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,256,3072,0.0062613333264986675
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,256,3072,0.033062400420506795
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,256,2560,0.00588266650835673
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,256,2560,0.029422932863235475
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,256,3072,0.02063680092493693
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,256,2048,0.005047466854254405
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,256,2048,0.02686506708463033
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,256,2560,0.01943040092786153
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,256,1536,0.004649599889914194
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,256,1536,0.02479040026664734
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,256,3584,0.023171200354894003
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,256,1024,0.0038890667259693147
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,256,2048,0.01694719990094503
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,256,1024,0.022791467110315957
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,256,768,0.0035616000493367515
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,256,1536,0.01495680014292399
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,256,768,0.02129813234011332
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,256,1024,0.013379200299580892
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,256,512,0.0030613332986831666
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,256,65536,0.22302826245625815
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,256,512,0.018879999717076622
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,256,256,0.0025909334421157838
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,256,256,0.01705066760381063
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,256,128,0.0025759999950726825
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,256,256,0.012476799885431926
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,256,512,0.012930132945378623
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,256,128,0.014457600315411887
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,256,64,0.002550400048494339
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,256,64,0.016339199741681416
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,256,32,0.0027114666998386385
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,256,32,0.015853866934776306
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,128,65536,0.07788906892140707
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,128,65536,0.6512853622436523
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,128,51200,0.062059732278188076
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,256,768,0.012473600109418233
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,256,128,0.01286186675230662
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,128,51200,0.5161461194356283
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,128,16384,0.02590506672859192
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,128,16384,0.17225066820780438
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,128,16384,0.0654421329498291
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,128,12288,0.01599253316720327
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,128,12288,0.12726720174153644
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,128,10240,0.014443733294804893
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,128,10240,0.10212053457895916
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,128,12288,0.05198080142339071
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,128,8192,0.011610666910807293
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,128,8192,0.06981759866078695
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,128,10240,0.04461439847946167
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,128,7168,0.010406399766604107
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,128,7168,0.05728533267974854
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,128,51200,0.17581334114074706
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,128,8192,0.03565973440806071
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,128,6144,0.009287466605504353
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,128,6144,0.04726933240890503
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,128,5120,0.007944533228874206
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,128,5120,0.04188053210576375
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,128,7168,0.03301546573638916
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,128,4096,0.007065600156784058
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,128,6144,0.030425600210825604
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,128,4096,0.03704746564229329
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,128,3584,0.006333866715431213
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,128,65536,0.21912426948547364
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,128,3584,0.034525867303212485
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,128,5120,0.028036266565322876
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,128,3072,0.005838933090368906
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,128,3072,0.03298346598943074
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,128,2560,0.00508480022350947
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,128,4096,0.023962666591008507
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,128,2560,0.02881173292795817
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,128,2048,0.004646400113900503
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,128,3072,0.01984106699625651
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,128,2048,0.026840533812840777
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,128,1536,0.004196266829967499
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,128,2560,0.018979199727376304
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,128,1536,0.02324906587600708
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,128,2048,0.01620586713155111
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,128,1024,0.003416533271471659
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,128,3584,0.022216532627741495
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,128,1024,0.022278400262196858
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,128,1536,0.013967999815940857
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,128,768,0.02067306637763977
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,128,1024,0.012729600071907043
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,128,512,0.002993066608905792
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,128,768,0.011738666892051696
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,128,512,0.01893333395322164
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,128,256,0.0025781333446502686
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,128,256,0.0165802667538325
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,128,512,0.012106666962305706
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,128,128,0.0025685332715511323
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,128,256,0.012037332852681477
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,128,128,0.016455466548601784
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,128,64,0.002198400100072225
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,128,32,0.0026133333643277483
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3585,128,128,0.01200320025285085
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,128,64,0.014522666732470194
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,128,32,0.014528000354766845
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,64,65536,0.07792960007985433
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,64,51200,0.0599840005238851
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,64,16384,0.025545599063237508
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,64,16384,0.1704448064168294
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,64,12288,0.012903466820716858
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,64,51200,0.5105525334676106
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,64,65536,0.6448917388916016
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,128,768,0.0033941333492596946
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,64,12288,0.12849706808725994
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,64,10240,0.10234879652659098
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,64,7168,0.010856533050537109
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,64,7168,0.05614293416341146
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,64,6144,0.009613866607348125
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,64,6144,0.04751786788304647
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,64,5120,0.007946666578451793
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,64,5120,0.042005332310994466
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,64,4096,0.006828799843788147
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,64,4096,0.0370901346206665
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,64,3584,0.006251733501752217
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,64,3584,0.03471680084864299
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,64,3072,0.005897599955399831
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,64,3072,0.033030400673548385
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,64,10240,0.011246933539708456
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,64,2560,0.005067733426888784
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,64,2048,0.004621866842110952
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,64,2560,0.028810666004816694
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,64,8192,0.010414933164914448
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,64,1536,0.0038549333810806276
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,64,2048,0.027153066794077557
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,64,8192,0.06967039903004965
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,64,1024,0.003398400048414866
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,64,768,0.003985066711902618
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,64,1024,0.0225600004196167
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,64,512,0.0025909334421157838
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,64,768,0.020670932531356812
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,64,256,0.002585600068171819
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,64,1536,0.02536960045496623
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,64,512,0.01921066641807556
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,64,128,0.0024405332903067274
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,64,64,0.0022463999688625337
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,64,64,0.014571733276049294
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,64,128,0.015460266669591268
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,64,32,0.014681599537531533
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,32,65536,0.075709867477417
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,32,65536,0.644978141784668
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,64,256,0.016780799627304076
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,32,51200,0.05993386507034302
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,32,16384,0.022525866826375328
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,64,32,0.002584533393383026
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,32,12288,0.011593600114186604
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,32,16384,0.17009065945943197
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,32,12288,0.12770666281382242
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,32,51200,0.509661865234375
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,32,10240,0.010125866532325745
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,32,8192,0.00879039963086446
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,32,10240,0.11042986710866291
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,32,7168,0.00876693328221639
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,32,6144,0.008065066734949748
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,32,8192,0.07159146467844645
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,32,7168,0.05731946627298991
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,32,6144,0.049399467309316
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,32,5120,0.017842133839925133
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,32,4096,0.008851200342178345
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,32,5120,0.04322559833526611
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,32,3584,0.008029866715272267
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,32,3072,0.007090133428573608
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,32,3584,0.03349013328552246
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,32,2560,0.006678399940331777
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,32,3072,0.031850665807724
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,32,2048,0.004626133541266123
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,32,2560,0.02911679943402608
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,32,1536,0.0038954667747020722
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,32,2048,0.026741333802541095
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,32,1024,0.003390933324893316
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,32,1536,0.024808533986409507
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,32,768,0.0032170665760835014
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,32,1024,0.020722132921218873
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,32,768,0.02058666745821635
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,32,256,0.002526933451493581
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,32,512,0.018595200777053834
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,32,128,0.0021738665799299877
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,32,256,0.016812799374262492
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,32,64,0.0022304000953833262
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,32,128,0.014818132917086283
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,32,32,0.002161066730817159
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,32,64,0.014607999722162882
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,32,32,0.013429333766301474
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3585,32,4096,0.037373868624369304
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3585,32,512,0.0026549334327379864
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,65536,16384,4.320406595865885
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,65536,51200,7.4115030924479175
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,65536,16384,2.3837130228678385
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,65536,12288,3.353902943929037
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,65536,12288,1.7376127878824872
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,65536,51200,14.307129923502604
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,65536,10240,2.790497080485026
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,65536,10240,1.3961504618326823
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,65536,10240,0.7895882924397786
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,65536,8192,2.2308779398600262
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,65536,8192,1.0712287902832032
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,65536,12288,0.9029216130574544
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,65536,7168,1.9131924947102863
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,65536,7168,0.9605514526367187
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,65536,16384,1.1906356811523438
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,65536,6144,1.4826101938883463
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,65536,6144,0.826862907409668
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,65536,8192,0.6519722620646159
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,65536,5120,1.2277589162190756
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,65536,5120,0.670082155863444
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,65536,7168,0.6232842763264974
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,65536,4096,1.0233973185221354
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,65536,4096,0.5616416295369466
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,65536,6144,0.5339391708374024
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,65536,3584,0.8426197052001954
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,65536,3584,0.48400427500406906
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,65536,4096,0.4384042739868164
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,65536,3072,0.7275552113850912
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,65536,3072,0.45891841252644855
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,65536,5120,0.4807850519816081
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,65536,2560,0.5905386606852214
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,65536,2560,0.35134932200113933
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,65536,3584,0.3924031893412272
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,65536,2048,0.48603200912475586
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,65536,2048,0.3050389289855957
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,65536,2560,0.32695252100626626
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,65536,1536,0.36461334228515624
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,65536,1536,0.22587413787841798
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,65536,3072,0.34650665918986
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,65536,1024,0.256713596979777
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,65536,1024,0.17951679229736328
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,65536,1024,0.23605333964029945
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,65536,768,0.20073493321736655
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,65536,768,0.14397759437561036
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,65536,1536,0.24228906631469727
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,65536,512,0.1453802744547526
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,65536,512,0.11933333079020184
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,65536,512,0.22290773391723634
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,65536,256,0.09969173272450765
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,65536,256,0.0990336020787557
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,65536,2048,0.2891573270161947
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,65536,128,0.08675839900970458
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,65536,128,0.08670186996459961
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,65536,768,0.2251200040181478
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,65536,64,0.07384426593780517
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,65536,64,0.09014506340026855
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,65536,32,0.076146133740743
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,65536,32,0.0887829303741455
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,65536,128,0.22080532709757486
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,65536,256,0.2200064023335775
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,51200,65536,7.525350443522136
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,51200,65536,14.259220377604166
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,51200,51200,5.8965398152669275
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,51200,51200,10.959376017252605
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,51200,16384,3.4147659301757813
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,51200,16384,1.86136957804362
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,51200,16384,0.9381514867146811
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,51200,12288,2.485523223876953
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,51200,12288,1.3583658854166667
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,51200,12288,0.7060469309488933
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,51200,10240,2.1781654357910156
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,51200,10240,1.1016863505045573
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,51200,10240,0.6252288182576498
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,51200,8192,1.671670405069987
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,51200,8192,0.8473354975382487
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,65536,51200,3.9990954081217445
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,51200,7168,1.3338080088297526
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,51200,7168,0.784777577718099
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,51200,8192,0.5147829373677572
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,51200,6144,1.2246026357014974
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,51200,6144,0.6999637603759765
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,51200,7168,0.4803317387898763
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,51200,5120,0.9395989100138346
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,51200,5120,0.572921625773112
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,51200,6144,0.4139583905537923
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,51200,4096,0.7626303990681966
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,51200,4096,0.4406613349914551
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,51200,5120,0.38030614852905276
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,51200,3584,0.6497503916422527
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,51200,3584,0.40203307469685867
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,51200,4096,0.33913065592447916
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,51200,3072,0.5636501312255859
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,51200,3072,0.33934081395467125
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,51200,51200,3.087012227376302
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,51200,2560,0.48092158635457355
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,51200,2560,0.3030773480733236
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,51200,3584,0.30338134765625
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,51200,2048,0.38034025828043616
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,51200,2048,0.23684372901916503
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,51200,2560,0.2569429397583008
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,51200,1536,0.29347734451293944
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,51200,1536,0.1861226717631022
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,51200,2048,0.22613120079040527
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,51200,1024,0.20140906969706215
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,51200,1024,0.1355295976003011
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,51200,3072,0.26899305979410804
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,51200,768,0.15842986106872559
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,51200,768,0.11490240097045898
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,51200,1024,0.18939733505249023
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,51200,512,0.11548480192820232
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,51200,512,0.09876800378163655
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,51200,768,0.17758612632751464
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,51200,256,0.07869119644165039
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,51200,256,0.08116587003072104
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,51200,256,0.17634560267130533
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,51200,128,0.07036373615264893
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,51200,128,0.07476692994435628
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,51200,1536,0.19477866490681967
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,51200,64,0.060083198547363284
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,51200,64,0.0763434648513794
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,51200,32,0.06327999830245971
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,51200,32,0.07603200276692709
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,51200,128,0.17897814114888508
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,16384,65536,4.26452382405599
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,16384,65536,2.6324086507161457
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,51200,512,0.17794240315755208
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,16384,51200,3.3039563496907554
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,16384,51200,2.0256863911946614
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,16384,16384,1.008515230814616
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,16384,16384,0.713580830891927
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,16384,16384,0.3353962580362956
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,16384,12288,0.7037333170572917
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,16384,12288,0.521234130859375
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,16384,12288,0.25486507415771487
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,16384,10240,0.6168917338053386
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,16384,10240,0.4130912144978841
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,16384,10240,0.22767786979675292
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,16384,8192,0.4938517252604167
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,16384,8192,0.3287306785583496
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,16384,8192,0.18829545974731446
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,16384,7168,0.4278730710347493
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,16384,7168,0.286466121673584
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,16384,51200,1.0102837244669596
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,16384,6144,0.3641226768493652
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,16384,6144,0.26380799611409506
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,16384,7168,0.17589972813924154
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,16384,5120,0.29999465942382814
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,16384,5120,0.2048149267832438
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,16384,6144,0.15219839413960773
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,16384,4096,0.24074773788452147
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,16384,4096,0.16949973106384278
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,16384,5120,0.14063679377237956
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,16384,3584,0.21112747192382814
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,16384,3584,0.14563412666320802
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,16384,3584,0.11220373312632244
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,16384,3072,0.1818581263224284
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,16384,3072,0.1310645341873169
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,16384,4096,0.12333546479543049
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,16384,2560,0.15433386166890461
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,16384,2560,0.11006080309549968
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,51200,65536,3.9258346557617188
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,16384,2048,0.12200960318247478
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,16384,2048,0.09378666877746582
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,16384,2560,0.09416106541951498
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,16384,1536,0.09592959880828858
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,16384,1536,0.07442239920298258
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,16384,3072,0.09958826700846354
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,16384,1024,0.06643840074539184
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,16384,1024,0.06026560068130493
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,16384,2048,0.08314452966054281
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,16384,768,0.05185706615447998
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,16384,768,0.053780265649159754
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,16384,768,0.05878719886144003
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,16384,512,0.038771200180053714
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,16384,512,0.04666453202565511
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,16384,1024,0.06695786317189535
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,16384,1536,0.07048213481903076
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,16384,256,0.02701759934425354
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,16384,256,0.038541865348815915
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,16384,128,0.025009065866470337
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,16384,128,0.03546346823374431
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,16384,128,0.05518506765365601
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,16384,64,0.021074134111404418
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,16384,64,0.034363734722137454
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,16384,32,0.02146773338317871
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,16384,32,0.03420053323109944
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,16384,512,0.0552127997080485
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,16384,256,0.053602135181427
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,12288,65536,2.1394996643066406
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,12288,65536,3.229315185546875
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,12288,51200,1.7023211161295575
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,12288,51200,2.493939208984375
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,12288,16384,0.787280019124349
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,12288,16384,0.549624506632487
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,16384,65536,1.2951765696207682
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,12288,12288,0.5780693054199219
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,12288,12288,0.43200213114420577
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,12288,16384,0.26299734115600587
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,12288,10240,0.45711145401000974
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,12288,10240,0.3439178784688314
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,12288,12288,0.20227200190226235
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,12288,8192,0.3775456110636393
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,12288,8192,0.25540800094604493
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,12288,10240,0.1786058743794759
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,12288,7168,0.3180287996927897
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,12288,7168,0.22229866981506347
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,12288,8192,0.14835626284281414
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,12288,6144,0.27829440434773767
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,12288,6144,0.1974826653798421
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,12288,7168,0.13666772842407227
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,12288,5120,0.22452373504638673
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,12288,5120,0.16354880332946778
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,12288,6144,0.1205141305923462
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,12288,4096,0.18467733065287273
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,12288,4096,0.13854613304138183
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,12288,5120,0.10894292990366619
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,12288,3584,0.15736106236775715
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,12288,3584,0.11757653554280598
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,12288,4096,0.09676053524017333
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,12288,3072,0.13959466616312663
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,12288,3072,0.10685439904530843
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,12288,3584,0.08525013128916423
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,12288,2560,0.1143232027689616
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,12288,2560,0.09021440347035727
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,12288,3072,0.07787413597106933
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,12288,2048,0.09555413722991943
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,12288,2048,0.07912960052490234
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,12288,2560,0.073198930422465
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,12288,1536,0.07111679712931315
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,12288,1536,0.06343040068944296
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,12288,2048,0.06511360009511312
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,12288,1024,0.05104853312174479
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,12288,1024,0.05151146650314331
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,12288,1536,0.05840959946314493
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,12288,768,0.038891732692718506
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,12288,768,0.04654186566670736
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,12288,1024,0.05691200097401937
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,12288,512,0.02990613381067912
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,12288,512,0.03989760080973308
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,12288,512,0.053274667263031004
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,12288,256,0.0211082657178243
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,12288,256,0.03296533425649007
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,12288,768,0.05263893206914266
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,12288,128,0.016976000865300496
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,12288,128,0.02932586669921875
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,12288,128,0.052381865183512365
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,12288,64,0.014991999665896098
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,12288,64,0.02887360056241353
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,12288,32,0.015359999736150107
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,12288,256,0.052589865525563564
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,12288,32,0.029176533222198486
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,10240,65536,2.694550323486328
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,10240,65536,1.8461920420328777
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,10240,51200,2.081595738728841
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,10240,51200,1.454303995768229
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,12288,51200,0.7861034393310546
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,10240,16384,0.6257013320922852
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,10240,16384,0.4808181444803874
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,10240,16384,0.2375648021697998
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,10240,12288,0.4589664141337077
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,10240,12288,0.35770772298177084
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,12288,65536,0.9870389302571615
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,10240,10240,0.3909354527791341
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,10240,10240,0.2968768119812012
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,10240,12288,0.18209600448608398
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,10240,8192,0.31474453608194985
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,10240,8192,0.22872427304585777
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,10240,10240,0.16261760393778485
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,10240,7168,0.27256854375203454
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,10240,7168,0.21021013259887694
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,10240,8192,0.13374932607014972
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,10240,6144,0.23491840362548827
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,10240,6144,0.174834140141805
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,10240,7168,0.12449599901835125
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,10240,5120,0.19215359687805175
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,10240,5120,0.14629012743632
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,10240,6144,0.10875733693440755
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,10240,4096,0.1597376028696696
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,10240,4096,0.11946986516316731
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,10240,5120,0.09976853529612223
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,10240,3584,0.13402560551961262
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,10240,3584,0.10689706802368164
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,10240,51200,0.7054335912068684
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,10240,3072,0.1144917329152425
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,10240,3072,0.09668906529744467
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,10240,4096,0.08673173586527506
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,10240,2560,0.09764586289723715
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,10240,2560,0.08074453671773275
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,10240,2560,0.06731200218200684
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,10240,2048,0.07900266647338867
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,10240,2048,0.0710261344909668
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,10240,3584,0.07743146419525146
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,10240,1536,0.0605621337890625
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,10240,3072,0.06905813217163086
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,10240,1536,0.058051200707753506
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,10240,1024,0.042530135313669844
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,10240,1024,0.04602133433024089
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,10240,1024,0.053556267420450845
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,10240,768,0.033445334434509276
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,10240,768,0.04239999850591024
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,10240,1536,0.0550816019376119
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,10240,512,0.023881600300470988
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,10240,512,0.037282133102416994
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,10240,2048,0.06039679845174154
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,10240,256,0.018402133385340372
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,10240,256,0.03112000028292338
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,10240,768,0.04969173272450765
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,10240,128,0.016537599762280784
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,10240,128,0.02810666759808858
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,10240,256,0.04921600023905436
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,10240,64,0.01370133360226949
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,10240,64,0.028264532486597698
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,10240,32,0.013665067156155905
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,10240,32,0.026345600684483845
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,10240,128,0.049437868595123294
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,10240,512,0.04935146570205688
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,8192,65536,1.5685461680094401
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,8192,65536,2.0237354278564452
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,8192,51200,1.2442133585611979
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,8192,51200,1.5779295603434245
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,8192,16384,0.5251296043395997
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,8192,16384,0.4227274576822917
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,10240,65536,0.8945685068766276
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,8192,12288,0.35973119735717773
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,8192,12288,0.3092106819152832
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,8192,16384,0.1910538673400879
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,8192,10240,0.3107391993204752
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,8192,10240,0.24778347015380858
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,8192,10240,0.13096213340759277
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,8192,8192,0.24986133575439454
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,8192,8192,0.19274454116821288
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,8192,12288,0.14707093238830565
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,8192,7168,0.21945707003275552
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,8192,7168,0.17133866945902507
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,8192,7168,0.10023573239644368
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,8192,6144,0.1878005345662435
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,8192,6144,0.14773653348286947
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,8192,8192,0.10747626622517903
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,8192,5120,0.15474987030029297
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,8192,5120,0.12520853678385419
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,8192,5120,0.08041706879933676
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,8192,4096,0.12209386825561523
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,8192,4096,0.10383573373158772
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,8192,6144,0.08798720041910807
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,8192,3584,0.10924800237019856
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,8192,3584,0.09182720184326172
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,8192,3584,0.06082239945729574
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,8192,3072,0.09413973490397134
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,8192,3072,0.08375893433888754
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,8192,4096,0.0682965358098348
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,8192,2560,0.07813653151194254
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,8192,2560,0.07134186426798503
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,8192,3072,0.05273386637369791
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,8192,2048,0.06402026812235515
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,8192,2048,0.06264640092849731
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,8192,2560,0.05066239833831787
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,8192,1536,0.04845120112101237
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,8192,1536,0.04988480011622111
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,8192,1536,0.038286932309468585
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,8192,1024,0.03341439962387085
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,8192,1024,0.04073813358942668
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,8192,2048,0.04354986747105916
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,8192,768,0.02615999976793925
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,8192,768,0.03685973485310872
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,8192,768,0.03280213276545207
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,8192,512,0.019639466206232706
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,8192,512,0.03226026693979899
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,8192,512,0.032596266269683837
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,8192,256,0.014167466759681701
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,8192,256,0.0269813338915507
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,8192,1024,0.03636906544367473
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,8192,128,0.013892267147699991
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,8192,128,0.025066665808359784
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,8192,256,0.03260266582171122
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,8192,64,0.011640533804893494
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,8192,64,0.0247381329536438
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,8192,32,0.012014933427174886
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,8192,32,0.024264534314473472
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,8192,128,0.03341546853383382
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,7168,65536,1.8848096211751302
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,7168,65536,1.498467254638672
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,7168,51200,1.4362037658691407
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,7168,51200,1.1830516815185548
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,8192,51200,0.5572490692138672
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,7168,16384,0.4431370735168457
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,7168,16384,0.3867168108622233
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,7168,16384,0.20156799952189125
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,7168,12288,0.3342645327250163
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,7168,12288,0.29337921142578127
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,8192,65536,0.699017588297526
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,7168,10240,0.2775978724161784
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,7168,10240,0.24821972846984863
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,7168,12288,0.1464906692504883
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,7168,8192,0.22571306228637694
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,7168,8192,0.18853119214375813
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,7168,10240,0.12898773352305096
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,7168,7168,0.18881386121114094
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,7168,7168,0.155569060643514
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,7168,8192,0.10450879732767741
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,7168,6144,0.1721567948659261
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,7168,6144,0.13451520601908368
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,7168,7168,0.09825493494669596
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,7168,5120,0.13939627011617023
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,7168,5120,0.11526719729105632
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,7168,5120,0.07919253508249918
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,7168,4096,0.11084372997283935
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,7168,4096,0.09706133206685384
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,7168,6144,0.08663787047068278
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,7168,3584,0.09778453509012858
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,7168,3584,0.08479359944661459
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,7168,3584,0.060770134131113686
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,7168,3072,0.08407253424326579
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,7168,3072,0.07864426771799723
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,7168,4096,0.06753173669179281
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,7168,2560,0.0694528023401896
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,7168,2560,0.0673472007115682
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,7168,3072,0.053971199194590244
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,7168,2048,0.05755306482315063
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,7168,2048,0.05805333455403646
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,7168,2560,0.05173013210296631
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,7168,1536,0.04343786636988322
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,7168,1536,0.04739733139673869
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,7168,1536,0.0412714680035909
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,7168,1024,0.029956267277399702
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,7168,1024,0.03869760036468506
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,7168,2048,0.045697065194447835
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,7168,768,0.023769599199295045
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,7168,768,0.034995198249816895
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,7168,1024,0.03949546813964844
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,7168,512,0.017467733224232992
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,7168,512,0.03110613425572713
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,7168,768,0.035843201478322345
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,7168,256,0.013247999548912048
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,7168,256,0.026039467255274458
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,7168,256,0.03581226666768392
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,7168,128,0.012078932921091716
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,7168,128,0.022997333606084188
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,7168,512,0.0363317330678304
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,7168,128,0.03617386817932129
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,7168,64,0.010032000144322713
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,7168,32,0.010827733079592387
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,7168,64,0.02438186605771383
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,7168,51200,0.5805215835571289
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,6144,65536,1.5274144490559896
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,7168,32,0.022887466351191203
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,6144,65536,1.345357894897461
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,6144,51200,1.2008405049641928
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,6144,51200,1.0570784250895182
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,6144,16384,0.35755627950032554
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,6144,16384,0.3417813301086426
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,6144,16384,0.16985173225402833
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,6144,12288,0.2692981402079264
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,6144,12288,0.26507946650187175
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,7168,65536,0.7394858678181966
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,6144,10240,0.24868373870849608
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,6144,10240,0.21956799825032552
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,6144,10240,0.11340266863505047
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,6144,8192,0.1980768044789632
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,6144,8192,0.16997866630554198
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,6144,12288,0.12662506898244222
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,6144,7168,0.17056427001953126
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,6144,7168,0.14655893643697102
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,6144,7168,0.08731520175933838
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,6144,6144,0.14495573043823243
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,6144,6144,0.12853333155314128
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,6144,8192,0.09272212982177734
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,6144,5120,0.12152960300445556
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,6144,5120,0.11107199986775715
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,6144,5120,0.07067306836446127
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,6144,4096,0.0964085340499878
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,6144,4096,0.09266666571299234
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,6144,6144,0.07770026524861653
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,6144,3584,0.08338773250579834
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,6144,3584,0.08146879673004151
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,6144,3584,0.05357973178227743
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,6144,3072,0.07177920341491699
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,6144,3072,0.07489706675211588
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,6144,4096,0.05982720057169596
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,6144,2560,0.06144320170084635
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,6144,2560,0.0635744015375773
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,6144,2560,0.04548693497975667
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,6144,2048,0.050162132581075033
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,6144,3072,0.04821653366088867
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,6144,2048,0.05317226648330688
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,6144,1536,0.038446935017903645
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,6144,1536,0.042922667662302655
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,6144,1536,0.036298668384552
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,6144,1024,0.026523733139038087
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,6144,1024,0.03585066795349121
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,6144,2048,0.04033279816309611
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,6144,768,0.020567466815312706
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,6144,768,0.033225599924723306
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,6144,1024,0.03461546500523885
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,6144,512,0.015752533078193666
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,6144,512,0.02885546684265137
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,6144,768,0.03215893308321635
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,6144,256,0.01125333309173584
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,6144,512,0.0321941335995992
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,6144,256,0.025524266560872394
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,6144,128,0.009246933460235595
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,6144,128,0.021923200289408366
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,6144,256,0.03211733301480611
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,6144,128,0.03215679923693339
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,6144,64,0.00841813286145528
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,6144,32,0.008385066191355388
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,6144,64,0.021525333325068154
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,6144,32,0.021026132504145305
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,5120,65536,1.2050421396891275
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,5120,65536,1.2271573384602865
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,5120,51200,0.9601909637451171
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,5120,51200,0.9659039815266928
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,6144,51200,0.4932863871256511
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,5120,16384,0.29623359044392905
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,5120,16384,0.31684907277425134
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,5120,16384,0.14342506726582843
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,5120,12288,0.2246784051259359
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,5120,12288,0.2359999974568685
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,6144,65536,0.6257226943969727
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,5120,10240,0.19966079394022623
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,5120,10240,0.19653654098510742
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,5120,12288,0.10997546513875325
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,5120,8192,0.16524906158447267
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,5120,8192,0.1487328052520752
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,5120,10240,0.09823893706003825
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,5120,7168,0.14162240028381348
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,5120,7168,0.13012693723042806
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,5120,7168,0.07681066989898681
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,5120,6144,0.12027626832326252
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,5120,6144,0.1136405309041341
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,5120,8192,0.08130773703257242
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,5120,5120,0.0970367987950643
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,5120,5120,0.09889600276947022
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,5120,5120,0.061810131867726645
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,5120,4096,0.07968106269836425
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,5120,4096,0.08401599725087484
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,5120,6144,0.06844800313313802
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,5120,3584,0.06870826880137125
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,5120,3584,0.07203413645426432
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,5120,3584,0.04754453500111898
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,5120,4096,0.05230933427810669
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,5120,3072,0.05932693481445313
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,5120,3072,0.0664405345916748
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,5120,2560,0.04906986554463704
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,5120,2560,0.055239466826121006
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,5120,2560,0.04146133263905843
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,5120,2048,0.041833599408467606
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,5120,2048,0.047924268245697024
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,5120,3072,0.043244798978169754
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,5120,1536,0.031241599718729657
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,5120,1536,0.0392031987508138
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,5120,2048,0.03754986524581909
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,5120,1024,0.02190293272336324
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,5120,1024,0.03331200083096822
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,5120,1536,0.033982932567596436
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,5120,768,0.01753173271814982
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,5120,768,0.029478400945663452
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,5120,1024,0.032603732744852704
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,5120,512,0.0127018670241038
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,5120,512,0.027858134110768634
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,5120,768,0.03038826584815979
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,5120,256,0.00962666670481364
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,5120,256,0.023099732398986817
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,5120,512,0.030703999598821002
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,5120,128,0.00877333382765452
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,5120,128,0.02262293299039205
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,5120,256,0.030510934193929036
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,5120,128,0.03023253281911214
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,5120,64,0.0075552001595497135
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,5120,32,0.007890133559703827
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,5120,64,0.021234132846196494
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,5120,32,0.02113386591275533
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,4096,65536,0.9520469029744467
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,4096,65536,1.1065322875976562
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,4096,51200,0.7770336151123047
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,5120,51200,0.41469332377115886
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,4096,51200,0.8699573516845703
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,4096,16384,0.24497599601745607
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,4096,16384,0.283623472849528
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,4096,16384,0.11690879662831624
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,4096,12288,0.1853973388671875
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,4096,12288,0.21078720092773437
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,5120,65536,0.5187679926554363
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,4096,10240,0.15589866638183594
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,4096,10240,0.16958826382954914
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,4096,12288,0.09142399628957112
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,4096,8192,0.12712106704711915
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,4096,8192,0.1294495979944865
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,4096,10240,0.0825119972229004
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,4096,7168,0.11031253337860107
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,4096,7168,0.11642346382141114
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,4096,8192,0.06791573365529378
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,4096,6144,0.09497600396474203
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,4096,6144,0.10093013445536296
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,4096,7168,0.0642741322517395
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,4096,5120,0.07725226879119873
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,4096,5120,0.08852372964223226
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,4096,6144,0.056745600700378415
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,4096,4096,0.062108798821767175
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,4096,4096,0.07295893033345541
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,4096,5120,0.05069226821263632
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,4096,3584,0.05381866693496704
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,4096,3584,0.06392000118891397
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,4096,4096,0.04242773453394572
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,4096,3072,0.04666879971822103
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,4096,3072,0.055644798278808597
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,4096,3584,0.037939198811848956
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,4096,2560,0.03890453179677327
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,4096,2560,0.048769064744313556
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,4096,3072,0.034061865011850996
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,4096,2048,0.031829333305358885
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,4096,2048,0.04288640022277832
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,4096,2560,0.032509867350260416
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,4096,1536,0.024574933449427287
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,4096,1536,0.03647786776224772
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,4096,2048,0.028060799837112425
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,4096,1024,0.017627733945846557
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,4096,1024,0.030203733841578168
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,4096,1536,0.024039467175801597
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,4096,768,0.014108799894650779
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,4096,768,0.02938773234685262
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,4096,1024,0.022745599349339805
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,4096,512,0.010481066505114238
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,4096,512,0.024780799945195518
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,4096,768,0.023527467250823976
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,4096,256,0.007993599772453308
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,4096,256,0.0221781333287557
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,4096,512,0.023574399948120116
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,4096,128,0.007096533477306366
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,4096,128,0.01950613260269165
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,4096,128,0.023577600717544556
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,4096,64,0.0063391998410224915
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,4096,256,0.02363413373629252
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,4096,32,0.0067658667763074234
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,4096,32,0.018924800554911296
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,3584,65536,0.9012832005818685
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,4096,64,0.019291732708613077
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,3584,65536,1.0826026916503906
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,3584,51200,0.6918368021647135
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,4096,51200,0.33300692240397134
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,3584,51200,0.8435157140096029
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,3584,16384,0.22678933143615723
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,3584,16384,0.2717130661010742
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,3584,16384,0.11656213601430257
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,3584,12288,0.17081386248270672
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,3584,12288,0.20633066495259605
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,4096,65536,0.4149685223897298
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,3584,10240,0.1425312042236328
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,3584,10240,0.1679338614145915
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,3584,10240,0.0825376033782959
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,3584,8192,0.11462293465932209
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,3584,12288,0.09116693337758383
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,3584,8192,0.12889280319213867
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,3584,7168,0.09938560326894125
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,3584,7168,0.11365653673807781
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,3584,7168,0.06452159881591797
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,3584,6144,0.08465493520100911
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,3584,6144,0.09989759922027588
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,3584,8192,0.06901866594950358
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,3584,5120,0.0717898686726888
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,3584,5120,0.08685759703318277
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,3584,5120,0.05064853429794312
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,3584,4096,0.057450668017069495
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,3584,4096,0.07093119621276855
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,3584,6144,0.0564799984296163
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,3584,3584,0.050477866331736246
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,3584,3584,0.06137173175811768
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,3584,4096,0.0437120000521342
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,3584,3072,0.043952000141143796
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,3584,3584,0.040148266156514484
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,3584,3072,0.05485973358154297
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,3584,2560,0.037357866764068604
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,3584,2560,0.047457067171732585
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,3584,3072,0.03601919809977214
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,3584,2048,0.030296534299850464
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,3584,2048,0.04326719840367635
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,3584,2560,0.03442453145980835
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,3584,1536,0.024022400379180908
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,3584,1536,0.03597226540247599
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,3584,2048,0.030961066484451294
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,3584,1024,0.017190400759379068
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,3584,1024,0.03076266646385193
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,3584,1536,0.027515733242034913
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,3584,768,0.013665067156155905
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,3584,768,0.02909119923909505
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,3584,1024,0.026419200499852497
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,3584,512,0.010404266913731893
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,3584,512,0.026693334182103474
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,3584,768,0.027734400828679402
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,3584,256,0.009311999877293904
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,3584,256,0.024165334304173787
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,3584,512,0.02845226724942525
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,3584,128,0.008731733759244282
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,3584,128,0.021280000607172646
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,3584,256,0.02770026723543803
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,3584,64,0.008138666550318401
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,3584,64,0.022913066546122234
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,3584,32,0.008455466230710347
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,3584,128,0.025576533873875935
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,3584,32,0.02094506621360779
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,3072,65536,0.753111457824707
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,3072,65536,0.9893194834391276
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,3072,51200,0.5795136133829752
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,3072,51200,0.8037962595621744
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,3584,51200,0.33141867319742835
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,3072,16384,0.19295466740926107
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,3072,16384,0.260590934753418
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,3072,16384,0.11593706607818603
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,3072,12288,0.1449120044708252
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,3072,12288,0.18973546028137206
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,3584,65536,0.41477120717366533
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,3072,10240,0.12573440074920655
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,3072,10240,0.15099520683288575
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,3072,12288,0.09020906289418539
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,3072,8192,0.1010858694712321
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,3072,8192,0.1177514632542928
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,3072,10240,0.08204586505889892
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,3072,7168,0.08786666393280029
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,3072,7168,0.10292479991912842
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,3072,7168,0.06325546503067017
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,3072,6144,0.07448000113169352
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,3072,6144,0.08988800048828124
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,3072,8192,0.06789013544718424
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,3072,5120,0.060595198472340905
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,3072,5120,0.07800853252410889
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,3072,5120,0.05020053386688232
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,3072,4096,0.048519468307495116
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,3072,4096,0.06325973272323608
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,3072,6144,0.054257067044576016
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,3072,3584,0.04215039809544881
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,3072,3584,0.05383466482162476
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,3072,3584,0.03919680118560791
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,3072,3072,0.03626240094502767
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,3072,3072,0.04947199821472168
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,3072,4096,0.042871467272440594
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,3072,2560,0.030741333961486816
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,3072,2560,0.04317973454793294
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,3072,3072,0.03512959877649943
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,3072,2048,0.02473919987678528
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,3072,2560,0.03369599978129069
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,3072,2048,0.03813120126724243
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,3072,1536,0.019370667139689126
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,3072,1536,0.03288853367169698
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,3072,1536,0.026800000667572023
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,3072,1024,0.013699199755986533
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,3072,1024,0.027635200818379717
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,3072,2048,0.030153600374857585
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,3072,768,0.011193600296974183
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,3072,768,0.02619839906692505
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,3072,1024,0.025234133005142212
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,3072,512,0.008339200417200725
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,3072,768,0.02391466697057088
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,3072,512,0.023475199937820435
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,3072,512,0.02410986622174581
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,3072,256,0.007454933226108551
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,3072,256,0.02190399964650472
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,3072,256,0.02424959937731425
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,3072,128,0.005858133236567179
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,3072,128,0.018879999717076622
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,3072,64,0.005667200187842051
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,3072,128,0.023987199862798056
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,3072,32,0.0054837331175804135
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,3072,64,0.01717546582221985
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,3072,32,0.01758506695429484
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,2560,65536,0.5888863881429036
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,2560,65536,0.9140085220336914
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,2560,51200,0.4647978782653809
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,2560,51200,0.7223562876383464
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,3072,51200,0.3329440116882324
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,2560,16384,0.1488661289215088
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,2560,16384,0.23479146957397462
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,3072,65536,0.4169002532958984
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,2560,12288,0.1121951977411906
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,2560,16384,0.09370559851328532
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,2560,12288,0.17801814079284667
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,2560,10240,0.09898666540781656
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,2560,10240,0.14640746116638184
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,2560,10240,0.0676736036936442
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,2560,8192,0.07984853585561116
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,2560,8192,0.11115413506825764
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,2560,12288,0.0735584020614624
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,2560,7168,0.07064426740010579
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,2560,7168,0.09882240295410157
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,2560,8192,0.05602026780446371
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,2560,6144,0.05945493380228678
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,2560,6144,0.08598399957021077
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,2560,7168,0.051520001888275144
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,2560,5120,0.049872001012166337
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,2560,5120,0.07287786801656088
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,2560,5120,0.041560534636179605
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,2560,6144,0.04517120122909546
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,2560,4096,0.040447998046875
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,2560,4096,0.057752533753712976
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,2560,3584,0.03503893216451009
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,2560,3584,0.05147306521733602
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,2560,3584,0.03222186764081319
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,2560,3072,0.03065920074780782
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,2560,3072,0.04732799927393595
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,2560,4096,0.03546666701634725
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,2560,2560,0.02640213370323181
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,2560,2560,0.04177600145339966
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,2560,3072,0.029037866989771528
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,2560,2048,0.021641600131988525
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,2560,2048,0.03814506530761719
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,2560,2560,0.027085866530736285
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,2560,1536,0.01727573275566101
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,2560,1536,0.0324234664440155
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,2560,2048,0.02399786710739136
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,2560,1024,0.012467199563980102
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,2560,1024,0.026964267094930012
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,2560,1536,0.0210591991742452
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,2560,768,0.010223999619483948
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,2560,1024,0.02026346723238627
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,2560,768,0.02531519929567973
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,2560,512,0.007976533472537994
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,2560,512,0.023150932788848878
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,2560,768,0.018521600961685182
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,2560,512,0.018914133310317993
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,2560,256,0.005932799975077311
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,2560,256,0.02039466698964437
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,2560,128,0.005534933507442474
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,2560,128,0.018689066171646118
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,2560,256,0.018245333433151247
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,2560,64,0.0050901333491007485
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,2560,64,0.01875306765238444
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,2560,128,0.018641066551208497
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,2560,32,0.005449600021044413
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,2560,32,0.017080533504486083
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,2560,51200,0.25520213445027673
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,2048,65536,0.4689205487569173
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,2048,51200,0.3697930653889974
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,2048,65536,0.8491445541381836
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,2048,51200,0.6765066782633464
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,2048,16384,0.11851413249969482
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,2048,16384,0.21658347447713217
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,2048,16384,0.09118719895680746
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,2048,12288,0.09002559979756673
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,2048,12288,0.1650378704071045
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,2560,65536,0.3172565460205078
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,2048,10240,0.08251519997914633
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,2048,10240,0.13367786407470703
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,2048,12288,0.07140693664550782
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,2048,8192,0.06492693424224853
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,2048,8192,0.10263786315917969
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,2048,10240,0.06522666613260905
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,2048,7168,0.0577023983001709
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,2048,7168,0.09027520020802816
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,2048,8192,0.05405439933141073
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,2048,6144,0.047571198145548506
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,2048,6144,0.07830719947814942
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,2048,7168,0.049048535029093426
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,2048,5120,0.040294400850931805
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,2048,5120,0.06451413234074911
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,2048,6144,0.043365331490834554
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,2048,4096,0.031754666566848756
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,2048,4096,0.052868266900380455
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,2048,5120,0.039944533507029215
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,2048,3584,0.02832000056902568
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,2048,3584,0.04705173174540202
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,2048,4096,0.034227200349171955
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,2048,3072,0.024341332912445068
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,2048,3072,0.04380266666412354
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,2048,3584,0.0305461327234904
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,2048,2560,0.021066667636235555
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,2048,2560,0.03794879913330078
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,2048,3072,0.02754453420639038
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,2048,2048,0.017478400468826295
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,2048,2048,0.03514240185419719
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,2048,2560,0.02598080039024353
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,2048,1536,0.013310933113098144
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,2048,1536,0.029038933912913005
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,2048,1536,0.01939520041147868
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,2048,1024,0.010016000270843506
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,2048,1024,0.025542400280634564
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,2048,2048,0.02268480062484741
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,2048,768,0.008301866551240284
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,2048,768,0.02476159930229187
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,2048,1024,0.018169599771499633
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,2048,768,0.016548267006874083
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,2048,512,0.006708266834417979
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,2048,512,0.023105067014694215
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,2048,256,0.005456000069777171
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,2048,256,0.020040533939997354
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,2048,256,0.016565333803494772
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,2048,512,0.016961065928141277
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,2048,128,0.005053866902987162
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,2048,128,0.01665066679318746
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,2048,64,0.004281599819660187
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,2048,64,0.016680532693862916
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,2048,128,0.016547200083732606
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,2048,32,0.004670933385690053
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,2048,32,0.0165802667538325
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,1536,65536,0.3693322817484538
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,1536,65536,0.7904117584228516
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,1536,51200,0.28268906275431316
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,2048,51200,0.2550645351409912
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,1536,51200,0.6246239980061848
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,1536,16384,0.09186879793802896
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,1536,16384,0.20303893089294434
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,1536,16384,0.09215573469797769
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,1536,12288,0.06958719889322916
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,1536,12288,0.1547658602396647
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,2048,65536,0.31865173975626626
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,1536,10240,0.06945386727650961
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,1536,10240,0.1309066692988078
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,1536,12288,0.0731754700342814
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,1536,8192,0.052833068370819095
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,1536,8192,0.09740373293558756
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,1536,10240,0.06670826276143392
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,1536,7168,0.04806079864501953
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,1536,7168,0.08542933464050292
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,1536,8192,0.0537226676940918
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,1536,6144,0.03983146746953328
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,1536,6144,0.07251199881235758
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,1536,7168,0.049586133162180586
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,1536,5120,0.03426880041758219
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,1536,5120,0.05970986684163412
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,1536,5120,0.04102506637573242
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,1536,4096,0.027511467536290485
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,1536,6144,0.04451839923858643
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,1536,4096,0.049831465880076094
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,1536,3584,0.025364265839258833
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,1536,3584,0.045227734247843425
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,1536,3584,0.03227519989013672
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,1536,3072,0.022181334098180135
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,1536,3072,0.04124586582183838
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,1536,4096,0.03528106609980265
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,1536,2560,0.018170666694641114
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,1536,2560,0.036714665095011395
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,1536,3072,0.029521065950393676
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,1536,2048,0.015513599912325541
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,1536,2560,0.027592533826828004
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,1536,2048,0.03384000062942505
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,1536,1536,0.012177067001660664
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,1536,1536,0.028804266452789308
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,1536,1536,0.022361600399017335
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,1536,1024,0.008515200018882752
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,1536,1024,0.0248416006565094
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,1536,2048,0.02488213380177816
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,1536,768,0.007214933137098948
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,1536,768,0.023707733551661173
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,1536,1024,0.020961066087086998
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,1536,512,0.005883733431498209
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,1536,512,0.02060799996058146
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,1536,512,0.0194048007329305
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,1536,256,0.005006933212280273
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,1536,256,0.01881706714630127
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,1536,256,0.019378133614857993
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,1536,128,0.004227200150489807
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,1536,768,0.019509333372116088
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,1536,128,0.01649386684099833
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,1536,128,0.020107734203338622
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,1536,64,0.003828266759713491
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,1536,64,0.016637866695721946
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,1536,32,0.004042666653792063
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,1536,32,0.016537599762280784
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,1024,65536,0.25435733795166016
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,1536,51200,0.25431146621704104
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,1024,65536,0.745092264811198
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,1024,51200,0.19670400619506836
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,1024,51200,0.5915637334187825
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,1024,16384,0.0664959987004598
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,1024,16384,0.18976000150044758
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,1024,16384,0.0676746686299642
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,1024,12288,0.050385065873463954
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,1024,12288,0.1441482702891032
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,1536,65536,0.31997013092041016
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,1024,10240,0.04185173511505127
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,1024,10240,0.11663466294606525
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,1024,12288,0.054478931427001956
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,1024,8192,0.03427306811014812
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,1024,8192,0.0868565320968628
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,1024,10240,0.049449598789215087
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,1024,7168,0.02895893255869548
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,1024,7168,0.07469013532002768
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,1024,8192,0.03915199836095174
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,1024,6144,0.025008000930150348
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,1024,6144,0.06276479959487916
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,1024,7168,0.036253865559895834
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,1024,5120,0.020645334323247274
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,1024,5120,0.051451734701792394
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,1024,5120,0.030052266518274945
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,1024,4096,0.017282134294509886
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,1024,6144,0.032307199637095135
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,1024,4096,0.04337706565856934
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,1024,3584,0.015438933173815408
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,1024,3584,0.0392576018969218
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,1024,3584,0.023884799083073935
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,1024,3072,0.013521066308021546
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,1024,3072,0.037045331796010335
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,1024,4096,0.02598506609598796
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,1024,2560,0.011891200145085653
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,1024,2560,0.03293226758639018
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,1024,2560,0.020643200476964316
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,1024,3072,0.021843200922012328
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,1024,2048,0.010089600086212158
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,1024,2048,0.030959999561309813
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,1024,1536,0.008393599589665731
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,1024,1536,0.026769065856933595
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,1024,1536,0.015780267119407655
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,1024,1024,0.006691200037797292
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,1024,2048,0.01780479947725932
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,1024,1024,0.023596799373626708
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,1024,768,0.0056415999929110205
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,1024,768,0.022715733448664347
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,1024,51200,0.1804159959157308
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,1024,1024,0.014065066973368326
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,1024,512,0.004806399842103322
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,1024,512,0.020619734128316244
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,1024,768,0.013662933309872945
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,1024,256,0.003996799886226654
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,1024,256,0.01759786605834961
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,1024,128,0.003458133339881897
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,1024,128,0.016215466459592185
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,1024,256,0.013559466600418091
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,1024,64,0.003390933324893316
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,1024,512,0.013665067156155905
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,1024,32,0.00342399999499321
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,1024,64,0.016578132907549538
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,1024,128,0.014504533012708029
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,1024,32,0.014890666802724203
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,768,65536,0.21227092742919923
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,768,51200,0.16631253560384113
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,768,65536,0.7122410456339519
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,768,51200,0.5957482655843098
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,768,16384,0.05356266498565674
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,768,16384,0.1852469285329183
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,768,16384,0.06849599679311116
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,768,12288,0.04119146664937337
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,768,12288,0.13985172907511395
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,1024,65536,0.22346773147583007
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,768,10240,0.034167468547821045
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,768,10240,0.11370879809061687
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,768,12288,0.05503893295923869
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,768,8192,0.027511467536290485
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,768,8192,0.08249066670735677
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,768,10240,0.05021013418833414
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,768,7168,0.023592533667882283
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,768,7168,0.07084906895955403
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,768,7168,0.0362496018409729
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,768,6144,0.02076693375905355
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,768,8192,0.039443198839823404
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,768,6144,0.05903786818186442
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,768,5120,0.0176746666431427
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,768,5120,0.04983786741892497
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,768,6144,0.03300693432490031
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,768,4096,0.015083733201026916
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,768,4096,0.042377599080403644
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,768,5120,0.03099520007769267
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,768,3584,0.013343999783198038
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,768,3584,0.0390720009803772
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,768,4096,0.026794666051864625
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,768,3072,0.012276267011960346
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,768,3584,0.02472106615702311
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,768,3072,0.03695253531138103
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,768,2560,0.010440533359845478
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,768,2560,0.03355413277943929
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,768,3072,0.022682666778564453
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,768,51200,0.18080107371012372
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,768,2048,0.008753066261609394
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,768,2048,0.030140799283981324
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,768,1536,0.0076330666740735366
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,768,1536,0.027151999870936076
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,768,2560,0.021423999468485513
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,768,1024,0.006387199958165486
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,768,1024,0.024320000410079957
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,768,2048,0.019057067235310872
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,768,768,0.005494399865468343
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,768,1536,0.01656000018119812
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,768,768,0.022040534019470214
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,768,512,0.0042805333932240805
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,768,512,0.01984000007311503
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,768,768,0.014442666371663412
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,768,256,0.0036661334335803984
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,768,1024,0.015615999698638916
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,768,512,0.014871467153231302
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,768,256,0.017685333887736
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,768,128,0.003401600072781245
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,768,128,0.016561067104339598
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,768,64,0.003031466652949651
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,768,128,0.015258666872978211
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,768,64,0.017322667439778647
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,768,32,0.0033930666744709016
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,768,256,0.01511146624883016
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,512,65536,0.16367360750834148
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,512,65536,0.6813034693400065
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,512,51200,0.1263797362645467
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,768,32,0.014471466342608133
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,512,51200,0.5425333023071289
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,512,16384,0.041281068325042726
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,512,16384,0.18229227066040038
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,768,65536,0.2235029379526774
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,512,12288,0.03095039923985799
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,512,12288,0.13555946350097656
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,512,16384,0.06685333251953125
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,512,10240,0.024785067637761435
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,512,10240,0.10863040288289387
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,512,12288,0.053532799084981285
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,512,8192,0.018605866034825645
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,512,8192,0.07673172950744629
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,512,10240,0.04777066707611084
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,512,7168,0.016277333100636802
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,512,8192,0.037412265936533615
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,512,7168,0.06480853160222372
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,512,6144,0.0143477330605189
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,512,6144,0.05363733371098837
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,512,7168,0.03429439862569173
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,512,5120,0.012102400263150532
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,512,5120,0.04616853396097819
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,512,6144,0.031319467226664226
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,512,4096,0.010059733192125957
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,512,4096,0.039556264877319336
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,512,5120,0.028867199023564653
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,512,3584,0.009082667032877604
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,512,3584,0.036600534121195474
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,512,3584,0.023107200860977173
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,512,3072,0.008065066734949748
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,512,3072,0.03481706778208415
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,512,4096,0.025226666529973345
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,512,51200,0.1788042704264323
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,512,2560,0.007472000022729237
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,512,2560,0.030987733602523805
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,512,2048,0.006413866579532623
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,512,2048,0.02887786626815796
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,512,3072,0.021028266350428263
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,512,1536,0.00552106648683548
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,512,1536,0.024851200977961223
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,512,2048,0.017313067118326822
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,512,1024,0.0045525332291920986
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,512,2560,0.019554134209950766
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,512,1536,0.014855466286341348
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,512,1024,0.022473599513371786
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,512,768,0.003937066594759623
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,512,768,0.02064853310585022
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,512,512,0.003437866767247518
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,512,512,0.020139733950297035
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,512,768,0.012922666470209756
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,512,256,0.0030378667016824085
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,512,256,0.017069866259892784
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,512,512,0.012842667102813721
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,512,1024,0.013729066650072733
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,512,128,0.0030261332790056865
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,512,256,0.013246933619181315
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,512,128,0.016669867436091106
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,512,64,0.002593066543340683
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,512,128,0.013287466764450074
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,512,32,0.003005866706371307
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,512,32,0.0146506667137146
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,256,65536,0.09571946461995443
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,256,65536,0.6553173065185547
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,256,51200,0.07610666751861572
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,512,64,0.01864853302637736
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,256,51200,0.5199722607930501
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,512,65536,0.22220053672790527
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,256,16384,0.030418133735656737
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,256,16384,0.17648000717163087
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,256,12288,0.024372265736262003
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,256,12288,0.13087999820709229
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,256,16384,0.06679573059082031
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,256,10240,0.019851734240849815
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,256,12288,0.053097601731618246
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,256,10240,0.10418773492177327
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,256,8192,0.015454933047294617
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,256,8192,0.07358293533325196
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,256,10240,0.04655893246332805
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,256,7168,0.013285332918167114
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,256,7168,0.06042559941609701
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,256,8192,0.03627306620279948
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,256,7168,0.03377279837926229
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,256,6144,0.012084266543388367
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,256,6144,0.04974933465321858
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,256,5120,0.010002133250236512
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,256,5120,0.043188265959421795
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,256,51200,0.17815680503845216
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,256,4096,0.0076106667518615724
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,256,4096,0.038276267051696775
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,256,6144,0.030949334303538006
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,256,5120,0.028924800952275592
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,256,3584,0.03465706507364909
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,256,3584,0.022707200050354003
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,256,3072,0.006275199850400289
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,256,4096,0.024872533480326333
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,256,3072,0.03289493322372437
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,256,2560,0.005854933460553487
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,256,2560,0.030635732412338256
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,256,65536,0.22190400759379067
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,256,3584,0.007052800059318543
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,256,2048,0.026897066831588747
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,256,2560,0.019386667013168334
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,256,1536,0.004582400123278299
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,256,1536,0.02500586708386739
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,256,2048,0.016952532529830932
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,256,3072,0.020946133136749267
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,256,1024,0.003832533210515976
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,256,1024,0.022420267264048256
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,256,768,0.0036608000596364343
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,256,1536,0.014919466773668923
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,256,768,0.021335466702779134
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,256,1024,0.013274666666984559
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,256,512,0.0036746665835380556
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,256,768,0.012427733341852824
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,256,512,0.01882879932721456
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,256,2048,0.005203199883302053
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,256,256,0.002630399912595749
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,256,128,0.002598399917284648
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,256,256,0.01705706715583801
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,256,128,0.015742933750152587
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,256,256,0.012819199760754903
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,256,512,0.012552533547083536
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,256,32,0.002605866640806198
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,256,64,0.016219733158747356
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,256,128,0.012859732906023661
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,128,65536,0.0847978671391805
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,256,32,0.014628266294797262
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,128,51200,0.06310506661732992
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,128,65536,0.6463946660359701
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,128,51200,0.5144970575968425
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,128,16384,0.025165865818659466
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,256,64,0.0021909333765506743
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,128,16384,0.17435520490010578
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,128,12288,0.01839146614074707
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,128,12288,0.1275264024734497
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,128,16384,0.06546346743901571
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,128,10240,0.014139733711878457
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,128,10240,0.10087573528289795
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,128,10240,0.04480640093485515
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,128,8192,0.01202880044778188
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,128,8192,0.06914666493733725
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,128,12288,0.05180906852086385
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,128,7168,0.010466133554776508
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,128,7168,0.05903253157933554
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,128,8192,0.03542506694793701
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,128,51200,0.17553280194600424
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,128,6144,0.00915839970111847
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,128,6144,0.04791253407796224
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,128,5120,0.00807360013326009
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,128,5120,0.0414463996887207
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,128,7168,0.033086933692296344
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,128,4096,0.006908800204594929
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,128,6144,0.030032000939051312
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,128,65536,0.21866453488667809
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,128,3584,0.006345599889755249
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,128,4096,0.03830826679865519
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,128,3584,0.03449493249257406
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,128,3072,0.0059562668204307554
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,128,5120,0.02800533374150594
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,128,4096,0.024312533934911094
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,128,3072,0.03308266599973043
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,128,2560,0.00508480022350947
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,128,2560,0.029576534032821657
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,128,2048,0.00469760000705719
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,128,3072,0.020245333512624107
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,128,2560,0.018572799364725747
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,128,3584,0.02222933371861776
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,128,2048,0.026872533559799194
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,128,1536,0.004274133344491323
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,128,1536,0.02449173331260681
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,128,1024,0.0034624000390370687
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,128,2048,0.016174933314323424
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,128,1024,0.02166293263435364
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,128,1536,0.014174933234850565
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,128,768,0.003421866645415624
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,128,1024,0.012489599982897441
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,128,512,0.003031466652949651
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,128,768,0.020829866329828896
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,128,512,0.018577067057291667
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,128,768,0.012000000476837159
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,128,256,0.002586666742960612
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,128,256,0.016809600591659545
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,128,512,0.012146133184432983
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,128,256,0.012020267049471537
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3584,128,128,0.012079999844233195
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,128,64,0.002233600119749705
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,128,64,0.01495146652062734
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,128,32,0.0025962665677070618
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,128,32,0.014686933159828186
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,64,65536,0.07633919715881347
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,64,65536,0.6452704111735026
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,128,128,0.002598399917284648
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,64,51200,0.06165653467178345
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,64,16384,0.023890133698781332
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,128,128,0.015894400080045064
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,64,12288,0.013410133123397828
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,64,16384,0.1714058717091878
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,64,10240,0.011327999830245971
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,64,12288,0.12775039672851562
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,64,8192,0.012314666310946147
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,64,51200,0.5111477216084798
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,64,10240,0.10033066272735595
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,64,6144,0.009145599603652955
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,64,8192,0.06846079826354981
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,64,7168,0.05647786855697632
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,64,6144,0.04726826747258504
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,64,4096,0.006683733562628429
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,64,5120,0.0414250651995341
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,64,3584,0.006346666812896728
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,64,4096,0.036823467413584395
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,64,3072,0.00550186683734258
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,64,3584,0.033297065893809
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,64,3072,0.031154133876164752
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,64,2560,0.005050666630268097
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,64,2048,0.0046528001626332605
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,64,2560,0.028806400299072266
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,64,1536,0.003881600002447764
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,64,2048,0.026910932858784993
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,64,1536,0.024425599972407022
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,64,1024,0.02114026745160421
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,64,768,0.0029813334345817565
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,64,768,0.02060799996058146
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,64,512,0.0026208000878492994
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,64,512,0.018630399306615194
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,64,256,0.0025888000925381976
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,64,7168,0.010950400431950887
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,64,5120,0.007959466675917308
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,64,128,0.0026367999613285064
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,64,256,0.016532267133394875
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,64,64,0.0023829333484172823
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,64,128,0.016380799810091655
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,64,32,0.0027530667682488757
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,64,32,0.01529706617196401
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,32,65536,0.07339306672414145
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,32,51200,0.05823573271433512
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,64,1024,0.0034624000390370687
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,32,16384,0.02239146629969279
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,32,65536,0.6434261322021484
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,32,12288,0.011191466450691223
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,32,51200,0.5105685234069824
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,32,16384,0.16915200551350912
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,32,10240,0.009620267152786254
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,32,12288,0.12733866373697916
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,32,8192,0.008744532863299053
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,32,10240,0.10154346625010173
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,32,7168,0.00846613347530365
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,32,8192,0.0705578645070394
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,32,6144,0.00790293316046397
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,32,7168,0.056683735052744544
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,64,64,0.014814933141072592
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,32,6144,0.049318401018778484
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,32,4096,0.009121066331863404
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,32,3584,0.008066133161385854
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,32,5120,0.04255040089289348
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,32,4096,0.03709866603215535
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,32,3072,0.007457066575686138
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,32,3584,0.0328981339931488
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,32,2560,0.006466133395830791
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,32,3072,0.03282559911410014
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,32,2048,0.0046858668327331545
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,32,2560,0.02891626755396525
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,32,1536,0.003908266623814901
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,32,2048,0.02686400016148885
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,32,1024,0.00346666673819224
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,32,1536,0.024173865715662636
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,32,1024,0.022065067291259767
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,32,768,0.0029898665845394133
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,32,512,0.0025994665920734406
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,32,768,0.020662399133046468
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,32,256,0.0026186667382717133
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,32,512,0.01878933310508728
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,32,128,0.0023562667270501454
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,32,256,0.017079466581344606
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,32,64,0.0022272000710169475
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,32,128,0.015101866920789084
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,32,32,0.0021557333568731946
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,32,64,0.014613333344459533
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3584,32,32,0.01360426644484202
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3584,32,5120,0.010575999816258747
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,65536,16384,4.147433725992839
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,65536,51200,6.984395853678386
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,65536,16384,2.1927594502766925
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,65536,12288,3.0922581990559896
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,65536,51200,13.34308064778646
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,65536,12288,1.64345703125
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,65536,10240,2.633592478434245
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,65536,10240,1.3403050740559896
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,65536,10240,0.7550090789794922
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,65536,8192,2.0342943827311197
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,65536,8192,1.0314282735188802
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,65536,12288,0.843878428141276
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,65536,7168,1.7395103454589844
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,65536,7168,0.9110890706380209
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,65536,16384,1.1294058481852214
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,65536,6144,1.4746283213297526
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,65536,6144,0.8219071706136069
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,65536,8192,0.6139743804931641
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,65536,5120,1.2513632456461588
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,65536,5120,0.6754037221272786
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,65536,7168,0.5700565338134765
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,65536,4096,0.954101308186849
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,65536,4096,0.5265749295552571
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,65536,6144,0.48634878794352215
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,65536,3584,0.8145376205444336
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,65536,3584,0.47352854410807294
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,65536,5120,0.44278507232666015
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,65536,3072,0.7121738433837891
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,65536,3072,0.4327232042948405
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,65536,4096,0.3924565315246582
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,65536,2560,0.5912810643513997
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,65536,2560,0.34048426946004234
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,65536,2560,0.29475733439127605
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,65536,2048,0.4660799980163574
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,65536,2048,0.2965717315673828
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,65536,3584,0.35402774810791016
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,65536,1536,0.36221014658610023
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,65536,1536,0.22077013651529948
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,65536,3072,0.3139103889465332
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,65536,1024,0.25054933230082194
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,65536,1024,0.18753600120544434
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,65536,1024,0.21942933400472006
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,65536,768,0.19486506779988605
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,65536,768,0.1549407958984375
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,65536,1536,0.2235040028889974
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,65536,512,0.1603882630666097
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,65536,512,0.12295573552449543
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,65536,768,0.2076863924662272
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,65536,256,0.10431573390960694
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,65536,256,0.10273173650105794
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,65536,512,0.20365333557128906
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,65536,128,0.0831712007522583
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,65536,128,0.09518933296203613
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,65536,2048,0.2610880057017008
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,65536,64,0.08007360299428304
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,65536,64,0.09940799872080484
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,65536,128,0.20579093297322593
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,65536,32,0.08370986779530844
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,65536,32,0.09744533697764078
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,65536,256,0.20613546371459962
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,51200,65536,7.058053080240886
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,51200,65536,13.505172729492188
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,51200,51200,10.17573750813802
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,51200,51200,5.484904479980469
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,51200,16384,3.1393163045247396
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,51200,16384,1.7232170104980469
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,51200,16384,0.8993824005126954
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,51200,12288,2.3758176167805987
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,51200,12288,1.2874229431152344
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,51200,12288,0.6718357086181641
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,51200,10240,1.9885237375895184
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,51200,10240,1.063329060872396
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,51200,10240,0.5951103846232096
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,51200,8192,1.5630079905192056
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,51200,8192,0.8760159810384115
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,65536,51200,3.771289571126302
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,51200,7168,1.2560223897298177
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,51200,7168,0.7132256189982097
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,51200,8192,0.4873023986816406
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,51200,6144,1.2026570638020835
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,51200,6144,0.625933837890625
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,51200,7168,0.45022185643513996
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,51200,5120,0.940945053100586
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,51200,5120,0.5775519688924153
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,51200,6144,0.38505493799845375
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,51200,4096,0.7183210372924804
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,51200,4096,0.4481216112772624
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,51200,5120,0.350656000773112
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,51200,3584,0.6336597442626953
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,51200,3584,0.3909258524576823
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,51200,51200,2.8971839904785157
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,51200,3072,0.5490048090616862
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,51200,3072,0.3425984064737956
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,51200,4096,0.3085205396016439
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,51200,2560,0.4744736035664876
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,51200,2560,0.28168748219807943
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,51200,3584,0.2796213467915853
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,51200,2048,0.38276265462239584
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,51200,2048,0.231278928120931
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,51200,3072,0.24546346664428711
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,51200,2560,0.23437760670979818
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,51200,1536,0.18649919827779132
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,51200,1536,0.30577065149943033
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,51200,1024,0.20046720504760743
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,51200,1024,0.13907306989034016
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,51200,2048,0.2091754595438639
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,51200,768,0.16086187362670898
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,51200,768,0.11320213476816814
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,51200,1024,0.1779402732849121
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,51200,512,0.11299413045247395
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,51200,512,0.10046186447143554
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,51200,1536,0.18219839731852214
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,51200,256,0.0815392017364502
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,51200,256,0.08465706507364909
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,51200,768,0.16749439239501954
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,51200,128,0.06136639912923177
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,51200,128,0.06813653310139975
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,51200,512,0.1663050651550293
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,51200,64,0.05324906508127848
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,51200,64,0.07218560377756754
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,51200,32,0.056169601281483975
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,51200,32,0.06980266571044921
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,51200,128,0.16864213943481446
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,16384,65536,3.933436838785807
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,16384,65536,2.476714579264323
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,51200,256,0.164683739344279
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,16384,51200,1.8949556986490887
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,16384,51200,3.083716328938802
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,16384,16384,0.9517205556233724
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,16384,16384,0.6412565231323242
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,16384,16384,0.32664426167805993
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,16384,12288,0.673361078898112
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,16384,12288,0.47092587153116866
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,16384,12288,0.24722347259521485
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,16384,10240,0.554966417948405
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,16384,10240,0.4046815872192383
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,16384,10240,0.22012906074523925
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,16384,8192,0.44826879501342776
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,16384,8192,0.31983146667480467
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,16384,8192,0.18065600395202636
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,16384,7168,0.39306987126668297
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,16384,7168,0.29883947372436526
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,16384,51200,0.9909013112386068
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,16384,6144,0.3593407948811849
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,16384,6144,0.24697386423746742
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,16384,7168,0.1674890677134196
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,16384,5120,0.3094208081563314
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,16384,5120,0.201637331644694
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,16384,6144,0.14495786031087238
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,16384,4096,0.24595093727111816
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,16384,4096,0.1660352071126302
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,16384,5120,0.13224960168202718
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,16384,3584,0.2220245361328125
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,16384,3584,0.14584107398986818
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,16384,4096,0.11636906464894611
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,16384,3072,0.17996800740559896
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,16384,3072,0.1258794705073039
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,16384,3584,0.1047935962677002
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,16384,2560,0.15291412671407062
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,16384,2560,0.1065941333770752
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,16384,2560,0.08841173648834229
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,16384,2048,0.12281280358632404
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,16384,2048,0.09190080165863038
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,16384,3072,0.0921183983484904
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,16384,1536,0.09702293078104654
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,16384,1536,0.07626667022705078
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,51200,65536,3.7904683430989587
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,16384,1024,0.06544959942499796
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,16384,1024,0.060610131422678626
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,16384,1024,0.06268479824066162
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,16384,768,0.05158400138219198
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,16384,768,0.0528277317682902
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,16384,1536,0.06548373301823934
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,16384,512,0.03927786747614543
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,16384,512,0.047277867794036865
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,16384,2048,0.07743146419525146
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,16384,256,0.02791999975840251
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,16384,256,0.03946453332901001
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,16384,768,0.05338240067164103
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,16384,128,0.023507199684778848
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,16384,128,0.03385920127232869
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,16384,512,0.05274026791254679
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,16384,64,0.01997013290723165
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,16384,64,0.03421226739883423
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,16384,32,0.020284799734751384
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,16384,128,0.05384853283564249
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,16384,32,0.033817601203918454
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,16384,256,0.05248106718063354
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,12288,65536,1.9972437540690104
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,12288,65536,3.089985148111979
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,12288,51200,1.603158442179362
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,12288,51200,2.3917633056640626
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,12288,16384,0.739784558614095
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,12288,16384,0.5266677220662435
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,16384,65536,1.264352035522461
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,12288,12288,0.5319744110107422
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,12288,12288,0.3874762535095215
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,12288,16384,0.2572533289591471
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,12288,10240,0.4666847864786784
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,12288,10240,0.32155628204345704
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,12288,12288,0.19619626998901368
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,12288,8192,0.3675029436747233
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,12288,8192,0.25821119944254556
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,12288,10240,0.17416426340738933
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,12288,7168,0.3300586700439453
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,12288,7168,0.2455296039581299
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,12288,8192,0.14291839599609374
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,12288,6144,0.2863210678100586
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,12288,6144,0.1919648011525472
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,12288,6144,0.1148096005121867
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,12288,5120,0.22605013847351074
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,12288,5120,0.16042346954345704
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,12288,7168,0.13265706698099772
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,12288,4096,0.18018453915913898
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,12288,4096,0.13656320571899414
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,12288,4096,0.09190719922383626
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,12288,3584,0.16315520604451497
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,12288,3584,0.11755092938741048
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,12288,5120,0.1049184004465739
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,12288,3072,0.13583893775939943
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,12288,3072,0.10264639854431153
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,12288,3584,0.08264959653218587
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,12288,2560,0.1134335994720459
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,12288,2560,0.08744426568349203
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,12288,3072,0.07337493101755778
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,12288,2048,0.09233280022939047
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,12288,2048,0.07918186982472739
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,12288,2048,0.06360213359196981
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,12288,1536,0.07427199681599936
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,12288,1536,0.060583468278249106
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,12288,2560,0.07037226359049478
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,12288,1024,0.05128320058186849
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,12288,1024,0.05194453398386637
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,12288,1024,0.05631680091222128
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,12288,768,0.040232535203297934
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,12288,768,0.045398398240407305
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,12288,1536,0.058090666929880776
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,12288,512,0.028505599498748778
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,12288,512,0.039162667592366536
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,12288,768,0.05266453425089518
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,12288,256,0.020804266134897866
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,12288,256,0.03415786822636922
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,12288,512,0.05228586594263712
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,12288,128,0.016107733050982155
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,12288,128,0.028479999303817748
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,12288,128,0.051768533388773595
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,12288,64,0.014564266800880432
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,12288,64,0.028617600599924724
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,12288,32,0.014936533570289613
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,12288,32,0.02828693389892578
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,12288,256,0.051804800828297935
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,10240,65536,1.7411018371582032
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,10240,65536,2.5206944783528646
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,10240,51200,1.9700106302897136
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,10240,51200,1.3711872100830078
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,12288,51200,0.7735328038533529
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,10240,16384,0.5776394526163737
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,10240,16384,0.44722773234049484
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,10240,16384,0.2309845288594564
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,10240,12288,0.43723198572794597
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,10240,12288,0.3411648114522298
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,12288,65536,0.9678421020507812
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,10240,10240,0.3591957410176595
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,10240,10240,0.2715264002482096
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,10240,12288,0.17572693824768065
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,10240,8192,0.29664427439371743
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,10240,8192,0.2214421272277832
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,10240,10240,0.15601280530293782
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,10240,7168,0.2512330691019694
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,10240,7168,0.19280212720235188
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,10240,8192,0.128002134958903
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,10240,6144,0.22547200520833335
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,10240,6144,0.1757599989573161
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,10240,7168,0.11875519752502442
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,10240,5120,0.182860803604126
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,10240,5120,0.13715945879618327
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,10240,5120,0.09425493081410727
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,10240,4096,0.14749226570129395
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,10240,4096,0.1138698657353719
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,10240,6144,0.10326826572418213
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,10240,3584,0.13514879544576008
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,10240,3584,0.10365333557128906
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,10240,51200,0.6907178878784179
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,10240,3072,0.11343786716461182
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,10240,3072,0.09051946798960367
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,10240,3584,0.07466666698455811
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,10240,2560,0.09559786319732666
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,10240,2560,0.07830506960550944
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,10240,4096,0.08218026955922444
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,10240,2048,0.07487146854400635
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,10240,2048,0.06906346480051676
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,10240,2048,0.05759679873784383
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,10240,1536,0.05788053274154663
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,10240,1536,0.055610664685567225
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,10240,3072,0.06640533208847046
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,10240,1024,0.04129493236541748
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,10240,1024,0.04525866508483887
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,10240,2560,0.06423786481221518
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,10240,768,0.03274346590042114
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,10240,768,0.040541867415110275
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,10240,768,0.045978665351867676
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,10240,512,0.023504000902175904
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,10240,512,0.03516906499862671
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,10240,1024,0.04993173281351725
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,10240,256,0.017374932765960693
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,10240,256,0.02977706591288249
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,10240,256,0.04542826811472575
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,10240,128,0.013422933220863343
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,10240,512,0.045994667212168376
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,10240,128,0.027059199412663777
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,10240,64,0.012422399719556172
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,10240,64,0.024829866488774617
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,10240,32,0.012546133001645407
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,10240,32,0.02476159930229187
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,10240,1536,0.05198613405227661
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,10240,128,0.045713067054748535
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,8192,65536,1.4623850504557292
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,8192,65536,1.9343616485595703
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,8192,51200,1.425389862060547
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,8192,51200,1.1686283111572267
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,8192,16384,0.4425546646118164
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,8192,16384,0.4054538726806641
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,8192,16384,0.18536960283915202
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,8192,12288,0.3338720003763834
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,8192,12288,0.297981866200765
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,10240,65536,0.872867202758789
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,8192,10240,0.276802126566569
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,8192,10240,0.24272212982177735
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,8192,12288,0.14131627082824708
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,8192,8192,0.226198403040568
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,8192,8192,0.19408853848775226
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,8192,10240,0.12604906558990478
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,8192,7168,0.19756906827290852
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,8192,7168,0.17367466290791828
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,8192,7168,0.09662079811096191
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,8192,6144,0.18635199864705404
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,8192,6144,0.14356266657511393
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,8192,8192,0.10367146333058674
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,8192,5120,0.15363413492838543
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,8192,5120,0.11998720169067383
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,8192,6144,0.08416319688161214
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,8192,4096,0.1209450642267863
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,8192,4096,0.10596586863199871
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,8192,5120,0.07674986521402995
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,8192,3584,0.11066880226135253
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,8192,3584,0.09035413265228272
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,8192,3584,0.058429865042368564
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,8192,3072,0.0914346694946289
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,8192,3072,0.08048746585845948
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,8192,4096,0.06537386576334635
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,8192,2560,0.07705173492431641
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,8192,2560,0.06873599688212076
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,8192,3072,0.05134933392206827
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,8192,2048,0.06217919985453287
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,8192,2048,0.06243840058644613
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,8192,2560,0.04942506551742554
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,8192,1536,0.047413333257039385
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,8192,1536,0.04892799854278564
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,8192,1536,0.03750400145848592
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,8192,1024,0.03492693503697713
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,8192,1024,0.04092586835225423
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,8192,2048,0.04320319890975952
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,8192,768,0.025244800249735515
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,8192,768,0.03497493267059326
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,8192,768,0.0321237325668335
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,8192,512,0.01898346741994222
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,8192,512,0.031115732590357464
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,8192,1024,0.035895466804504395
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,8192,256,0.014702933033307395
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,8192,256,0.027003733317057292
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,8192,512,0.03214613397916158
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,8192,128,0.011218133568763732
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,8192,128,0.023562665780385336
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,8192,256,0.03215786616007487
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,8192,64,0.010440533359845478
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,8192,64,0.024449066321055094
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,8192,32,0.010820266604423524
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,8192,128,0.032527999083201094
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,8192,32,0.023655466238657632
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,7168,65536,1.4044811248779296
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,7168,65536,1.7274463653564454
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,7168,51200,1.3723349253336588
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,7168,51200,1.1065216064453125
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,8192,51200,0.5412192026774089
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,7168,16384,0.4134805361429851
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,7168,16384,0.36041599909464517
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,7168,16384,0.16285227139790853
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,7168,12288,0.31128320693969724
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,7168,12288,0.2719029426574707
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,8192,65536,0.6862880071004231
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,7168,10240,0.27285760243733725
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,7168,10240,0.2198784033457438
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,7168,12288,0.12433493137359619
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,7168,8192,0.21038400332132973
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,7168,8192,0.1700149377187093
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,7168,10240,0.11210560003916423
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,7168,7168,0.18727253278096517
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,7168,7168,0.1526869297027588
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,7168,8192,0.09196586608886718
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,7168,6144,0.15422719319661457
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,7168,6144,0.1293503999710083
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,7168,7168,0.08630932966868082
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,7168,5120,0.13223573366800945
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,7168,5120,0.11101226806640625
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,7168,5120,0.06937920252482097
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,7168,4096,0.1058026631673177
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,7168,4096,0.09324053128560385
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,7168,6144,0.07546986738840739
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,7168,3584,0.0893951972325643
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,7168,3584,0.08418666521708171
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,7168,3584,0.052868266900380455
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,7168,3072,0.07733226617177327
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,7168,3072,0.07483200232187906
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,7168,4096,0.058398934205373135
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,7168,2560,0.0676800012588501
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,7168,2560,0.06561066706975302
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,7168,2560,0.04532693227132161
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,7168,2048,0.054883201917012535
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,7168,2048,0.05778239965438843
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,7168,3072,0.04698346853256226
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,7168,1536,0.04082133372624715
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,7168,1536,0.047326934337615964
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,7168,1536,0.03620586792627971
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,7168,1024,0.030070400238037108
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,7168,1024,0.03910826841990153
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,7168,2048,0.040559999148050946
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,7168,768,0.024230400721232094
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,7168,768,0.03498986562093099
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,7168,768,0.03242986599604289
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,7168,512,0.018812799453735353
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,7168,1024,0.03486613432566325
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,7168,512,0.029605333010355634
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,7168,256,0.013260799646377563
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,7168,256,0.025073067347208662
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,7168,256,0.03222080071767171
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,7168,128,0.013337600231170654
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,7168,128,0.022850133975346885
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,7168,512,0.03260693351427714
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,7168,64,0.011732266346613566
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,7168,128,0.03249599933624268
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,7168,32,0.011927466591199238
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,7168,64,0.022778666019439696
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,7168,32,0.023367466529210408
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,7168,51200,0.47196801503499347
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,6144,65536,1.2874591827392579
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,6144,51200,1.111068852742513
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,6144,65536,1.5120970408121743
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,6144,51200,1.0262922922770181
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,6144,16384,0.35719572703043617
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,6144,16384,0.3555232048034668
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,6144,16384,0.16674772898356122
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,6144,12288,0.2715925216674805
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,6144,12288,0.25991573333740237
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,7168,65536,0.5929023742675781
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,6144,10240,0.23078932762145996
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,6144,10240,0.19718720118204752
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,6144,12288,0.12368853092193603
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,6144,8192,0.186735995610555
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,6144,8192,0.16117653846740723
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,6144,10240,0.10981439749399821
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,6144,7168,0.15960319836934406
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,6144,7168,0.1380234718322754
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,6144,7168,0.08476373354593912
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,6144,6144,0.13428692817687987
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,6144,6144,0.11947946548461914
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,6144,8192,0.0902346690495809
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,6144,5120,0.11352960268656413
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,6144,5120,0.10212586720784504
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,6144,5120,0.0679263989130656
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,6144,4096,0.09106667041778564
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,6144,4096,0.08626773357391357
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,6144,6144,0.07491093476613363
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,6144,3584,0.07789866924285889
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,6144,3584,0.07827306588490804
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,6144,4096,0.058014933268229166
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,6144,3072,0.06900266806284586
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,6144,3072,0.06903573671976725
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,6144,3584,0.052269868055979406
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,6144,2560,0.0571232000986735
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,6144,2560,0.05785919825236002
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,6144,2560,0.04489066600799561
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,6144,2048,0.04628693262736003
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,6144,2048,0.050337068239847814
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,6144,3072,0.04627093474070231
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,6144,1536,0.036510932445526126
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,6144,1536,0.04129386742909749
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,6144,2048,0.039869864781697587
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,6144,1024,0.024886399507522583
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,6144,1024,0.03464320103327433
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,6144,1536,0.03545920054117839
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,6144,768,0.020450133085250854
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,6144,768,0.03154453237851461
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,6144,768,0.03175573348999024
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,6144,512,0.014607999722162882
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,6144,512,0.02728533347447713
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,6144,1024,0.03468586603800456
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,6144,256,0.010867200295130412
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,6144,256,0.02495573361714681
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,6144,512,0.03176213304201762
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,6144,128,0.010020266969998677
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,6144,128,0.022871466477711995
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,6144,256,0.031377067168553666
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,6144,64,0.009148800373077392
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,6144,128,0.03171520034472148
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,6144,32,0.009553066889444987
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,6144,64,0.021204266945521036
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,6144,32,0.022472532590230306
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,5120,65536,1.1510079701741538
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,5120,65536,1.15022824605306
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,5120,51200,0.9041333516438803
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,5120,51200,0.9183680216471354
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,6144,51200,0.4831669489542644
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,5120,16384,0.27928425470987955
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,5120,16384,0.2911967913309733
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,5120,16384,0.14000959396362306
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,5120,12288,0.21133759816487632
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,5120,12288,0.21949547131856284
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,6144,65536,0.6133909225463867
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,5120,10240,0.17957226435343426
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,5120,10240,0.1759658654530843
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,5120,12288,0.10639572938283284
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,5120,8192,0.14291307131449382
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,5120,8192,0.14125866889953614
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,5120,10240,0.09547733465830485
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,5120,7168,0.1279520034790039
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,5120,7168,0.1275157372156779
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,5120,8192,0.07931199868520102
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,5120,6144,0.10875306924184162
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,5120,6144,0.11170559724171955
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,5120,7168,0.07471146583557128
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,5120,5120,0.09267840385437012
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,5120,5120,0.0947007973988851
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,5120,5120,0.05969599882761637
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,5120,4096,0.0749023993810018
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,5120,4096,0.07728532950083414
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,5120,6144,0.0662496010462443
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,5120,3584,0.06555840174357096
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,5120,3584,0.07058346271514893
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,5120,4096,0.05108160177866618
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,5120,3072,0.05734719832738241
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,5120,3584,0.04694720109303792
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,5120,3072,0.06127573251724243
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,5120,2560,0.04881706635157267
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,5120,2560,0.05165760119756063
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,5120,2560,0.04054613510767619
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,5120,2048,0.03896213372548421
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,5120,2048,0.046214401721954346
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,5120,3072,0.04246400197347005
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,5120,1536,0.029634133974711103
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,5120,1536,0.0384554664293925
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,5120,1536,0.033421866099039715
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,5120,1024,0.02123840053876241
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,5120,1024,0.03269973397254944
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,5120,2048,0.03683626651763916
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,5120,768,0.01685653328895569
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,5120,768,0.028938666979471846
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,5120,51200,0.40486825307210283
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,5120,512,0.012702932953834534
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,5120,512,0.026409600178400678
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,5120,768,0.030369067192077638
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,5120,256,0.0095551997423172
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,5120,256,0.023084799448649086
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,5120,1024,0.0321343998114268
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,5120,128,0.00840106705824534
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,5120,128,0.020756266514460244
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,5120,512,0.03030719955762227
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,5120,64,0.007098666826883952
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,5120,128,0.030155734221140547
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,5120,32,0.007515733440717061
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,5120,256,0.03009919921557109
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,5120,64,0.020513067642847695
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,5120,32,0.022612265745798745
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,4096,65536,0.9294549306233725
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,4096,65536,1.0495221455891928
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,4096,51200,0.780079968770345
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,4096,51200,0.8234517415364584
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,4096,16384,0.23745387395222983
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,4096,16384,0.26779092152913414
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,4096,16384,0.11267200311024983
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,4096,12288,0.18026026089986164
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,4096,12288,0.20117546717325846
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,5120,65536,0.5071733474731446
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,4096,10240,0.15648214022318524
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,4096,10240,0.1614496072133382
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,4096,12288,0.08808000087738037
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,4096,8192,0.12565866311391194
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,4096,8192,0.13013652960459393
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,4096,10240,0.07941333452860513
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,4096,7168,0.10915733178456624
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,4096,7168,0.11780800024668377
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,4096,8192,0.06564160188039145
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,4096,6144,0.09302186965942383
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,4096,6144,0.0990997314453125
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,4096,7168,0.06198720137278239
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,4096,5120,0.075326935450236
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,4096,5120,0.08322772979736329
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,4096,6144,0.05461333195368449
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,4096,4096,0.06455466747283936
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,4096,4096,0.0686090628306071
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,4096,5120,0.04857066472371419
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,4096,3584,0.05282026529312134
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,4096,3584,0.06226346492767334
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,4096,3584,0.037239468097686766
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,4096,3072,0.04553279876708984
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,4096,3072,0.05646933317184448
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,4096,4096,0.04205013513565063
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,4096,2560,0.039136000474294025
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,4096,2560,0.046087467670440675
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,4096,3072,0.033403734366099044
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,4096,2048,0.031395200888315836
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,4096,2048,0.042370132605234784
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,4096,2560,0.03170773386955261
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,4096,1536,0.024743467569351196
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,4096,1536,0.03563626607259114
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,4096,1536,0.023948800563812257
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,4096,1024,0.017661867539087932
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,4096,1024,0.030693332354227703
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,4096,2048,0.02768320043881734
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,4096,768,0.013868799805641175
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,4096,768,0.027661865949630736
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,4096,1024,0.02274880011876424
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,4096,512,0.01046720047791799
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,4096,512,0.02474986712137858
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,4096,768,0.02352213263511658
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,4096,256,0.007994666695594788
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,4096,256,0.021896533171335855
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,4096,512,0.023119999965031942
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,4096,128,0.006704000135262807
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,4096,128,0.020721065998077392
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,4096,128,0.021782400210698445
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,4096,64,0.005906133353710175
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,4096,256,0.02313279906908671
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,4096,64,0.019641600052515664
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,4096,32,0.0065290664633115125
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,4096,32,0.018577067057291667
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,3584,65536,0.8311232248942056
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,3584,65536,0.9948437372843424
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,4096,51200,0.3213376045227051
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,3584,51200,0.6570079803466797
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,3584,51200,0.7772245407104492
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,3584,16384,0.21065813700358071
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,3584,16384,0.25356906255086265
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,3584,16384,0.11277440388997395
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,3584,12288,0.1605077266693115
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,3584,12288,0.18969920476277669
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,4096,65536,0.4020511945088704
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,3584,10240,0.13991999626159668
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,3584,10240,0.15575253168741862
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,3584,12288,0.08831893603006999
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,3584,8192,0.11348053614298503
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,3584,8192,0.1250762701034546
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,3584,10240,0.07965333461761474
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,3584,7168,0.0980949322382609
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,3584,7168,0.11262293656667073
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,3584,8192,0.06653120120366415
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,3584,6144,0.0841877301534017
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,3584,6144,0.0967093308766683
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,3584,7168,0.06191786527633667
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,3584,5120,0.06967786947886148
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,3584,5120,0.0812000036239624
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,3584,5120,0.04941013256708781
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,3584,4096,0.05642666816711426
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,3584,4096,0.06332906484603881
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,3584,6144,0.05413866837819418
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,3584,3584,0.04985066652297974
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,3584,3584,0.05653866529464722
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,3584,3584,0.03934719959894816
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,3584,3072,0.043287467956542966
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,3584,3072,0.05014400084813436
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,3584,4096,0.04285973310470581
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,3584,2560,0.036713600158691406
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,3584,2560,0.04367573261260986
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,3584,3072,0.03504213492075602
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,3584,2048,0.02752000093460083
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,3584,2048,0.03934719959894816
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,3584,2560,0.03374826510747274
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,3584,1536,0.021356799205144248
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,3584,1536,0.033445334434509276
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,3584,1536,0.027270400524139406
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,3584,2048,0.029924267530441286
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,3584,1024,0.01497706671555837
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,3584,1024,0.028883200883865357
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,3584,768,0.011904000242551168
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,3584,768,0.026765867074330645
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,3584,768,0.026233599583307905
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,3584,1024,0.026425600051879883
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,3584,512,0.009769599636395771
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,3584,512,0.024906667073567708
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,3584,512,0.027812266349792482
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,3584,256,0.007514666517575581
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,3584,256,0.020677334070205687
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,3584,256,0.02765546639760335
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,3584,128,0.006306133170922597
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,3584,128,0.01941439906756083
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,3584,128,0.024663466215133666
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,3584,64,0.006010666489601135
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,3584,64,0.01934400002161662
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,3584,32,0.006045866509278615
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,3584,32,0.0200981338818868
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,3072,65536,0.7022453308105469
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,3072,65536,0.9136693318684896
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,3584,51200,0.32184534072875975
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,3072,51200,0.5303530693054199
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,3072,51200,0.7272970835367839
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,3072,16384,0.17768747011820477
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,3072,16384,0.23554347356160482
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,3072,16384,0.1121077299118042
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,3072,12288,0.1346677303314209
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,3072,12288,0.174071470896403
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,3584,65536,0.4025568008422852
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,3072,10240,0.11036693255106608
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,3072,10240,0.1377781391143799
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,3072,12288,0.08731626669565837
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,3072,8192,0.09061546325683593
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,3072,8192,0.11166933377583821
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,3072,10240,0.07903467019399008
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,3072,7168,0.07821760177612305
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,3072,7168,0.09978666305541992
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,3072,8192,0.06574293375015258
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,3072,6144,0.06781760056813559
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,3072,6144,0.08659733136494954
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,3072,7168,0.06051626602808634
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,3072,5120,0.055338664849599206
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,3072,5120,0.07179626623789469
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,3072,5120,0.0477514664332072
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,3072,4096,0.04395413398742676
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,3072,6144,0.052655998865763345
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,3072,4096,0.05775466759999594
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,3072,3584,0.03885226647059123
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,3072,3584,0.053325867652893065
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,3072,3584,0.03819839954376221
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,3072,3072,0.03345280090967814
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,3072,3072,0.04749546845753987
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,3072,4096,0.04196586608886719
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,3072,2560,0.029156267642974854
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,3072,2560,0.0412501335144043
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,3072,2560,0.03260693351427714
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,3072,2048,0.023982934157053628
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,3072,2048,0.039510401089986165
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,3072,3072,0.034031999111175534
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,3072,1536,0.018926932414372762
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,3072,1536,0.0326474666595459
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,3072,2048,0.0293013334274292
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,3072,1024,0.014174933234850565
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,3072,1024,0.02855786681175232
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,3072,1536,0.025997867186864216
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,3072,768,0.01123520036538442
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,3072,768,0.026009599367777508
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,3072,1024,0.025231999158859254
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,3072,768,0.023525333404541014
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,3072,512,0.00885973374048869
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,3072,512,0.02460800011952718
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,3072,256,0.007143466671307881
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,3072,256,0.020837332804997763
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,3072,512,0.023079466819763184
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,3072,128,0.0061152001221974695
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,3072,256,0.02323733369509379
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,3072,128,0.01952106753985087
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,3072,64,0.0055189331372578945
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,3072,64,0.019359999895095827
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,3072,32,0.005871999760468801
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,3072,32,0.01843199928601583
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,3072,128,0.023588265975316366
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,2560,65536,0.5589610417683919
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,2560,65536,0.8460778554280599
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,2560,51200,0.42888425191243484
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,2560,51200,0.6698357264200847
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,3072,51200,0.3218154589335124
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,2560,16384,0.139192533493042
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,2560,16384,0.21973973910013833
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,2560,16384,0.08910079797108969
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,2560,12288,0.10553812980651855
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,2560,12288,0.1664138634999593
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,3072,65536,0.40498666763305663
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,2560,10240,0.09438186486562093
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,2560,10240,0.13177599906921386
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,2560,12288,0.07041280269622803
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,2560,8192,0.07512746651967367
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,2560,8192,0.10553812980651855
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,2560,10240,0.06460800170898437
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,2560,7168,0.06678079764048259
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,2560,7168,0.09398826758066813
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,2560,8192,0.05338773330052694
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,2560,6144,0.05559573173522949
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,2560,6144,0.08028586705525717
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,2560,7168,0.049259734153747556
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,2560,5120,0.04644586642583211
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,2560,5120,0.06666239897410074
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,2560,5120,0.04009279807408651
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,2560,4096,0.03734399875005086
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,2560,4096,0.054626135031382236
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,2560,6144,0.04350613355636597
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,2560,3584,0.0334666649500529
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,2560,3584,0.049466665585835776
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,2560,3584,0.03141653339068095
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,2560,3072,0.02903040051460266
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,2560,3072,0.04530666669209798
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,2560,4096,0.03511679967244466
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,2560,2560,0.024699733654658
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,2560,2560,0.039408000310262044
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,2560,3072,0.027882667382558186
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,2560,2048,0.020665599902470907
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,2560,2560,0.02682346701622009
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,2560,2048,0.03612693150838216
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,2560,1536,0.016292267044385276
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,2560,1536,0.030925865968068438
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,2560,1536,0.021107200781504312
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,2560,1024,0.01167039970556895
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,2560,1024,0.02677653431892395
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,2560,2048,0.023678932587305704
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,2560,768,0.009619200229644775
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,2560,768,0.024921600023905435
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,2560,1024,0.01980479955673218
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,2560,512,0.007559466858704884
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,2560,512,0.022528000672658286
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,2560,768,0.01818986733754476
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,2560,256,0.005888000130653381
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,2560,256,0.01998400092124939
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,2560,256,0.018230400482813516
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,2560,512,0.018678400913874307
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,2560,128,0.005443199972311656
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,2560,128,0.01871573328971863
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,2560,64,0.004644266764322917
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,2560,64,0.01689173380533854
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,2560,32,0.005062399804592133
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,2560,32,0.016532267133394875
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,2048,65536,0.45372800827026366
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,2560,128,0.018279467026392618
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,2048,65536,0.7947733561197917
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,2048,51200,0.3535029411315918
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,2560,51200,0.24457066853841147
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,2048,51200,0.6283701578776042
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,2048,16384,0.11565759976704915
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,2048,16384,0.20447680155436196
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,2048,16384,0.08773120244344076
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,2048,12288,0.08778986930847169
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,2048,12288,0.15466559727986653
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,2560,65536,0.3047605196634928
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,2048,10240,0.08038187026977539
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,2048,10240,0.12376213073730469
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,2048,12288,0.06917226314544678
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,2048,8192,0.06320106585820516
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,2048,8192,0.09867626825968424
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,2048,10240,0.06275413433710733
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,2048,7168,0.055379199981689456
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,2048,7168,0.088210129737854
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,2048,7168,0.04762453238169352
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,2048,6144,0.04706133206685384
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,2048,6144,0.07402026653289795
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,2048,8192,0.0515936017036438
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,2048,5120,0.03843626578648885
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,2048,5120,0.05910506645838419
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,2048,5120,0.038711468378702804
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,2048,4096,0.030959999561309813
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,2048,4096,0.049483732382456465
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,2048,6144,0.042294398943583174
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,2048,3584,0.028117332855860395
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,2048,3584,0.04692586660385132
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,2048,3584,0.03012266755104065
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,2048,3072,0.02468693256378174
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,2048,3072,0.04200746615727742
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,2048,4096,0.033766400814056394
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,2048,2560,0.02082560062408447
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,2048,2560,0.03705919981002807
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,2048,2560,0.024913066625595094
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,2048,2048,0.01725119948387146
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,2048,3072,0.026573866605758667
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,2048,2048,0.034807467460632326
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,2048,1536,0.01325866679350535
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,2048,1536,0.029330132404963176
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,2048,1536,0.0190720001856486
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,2048,1024,0.010070400436719258
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,2048,1024,0.025049599011739095
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,2048,2048,0.02211839954058329
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,2048,768,0.00831573357184728
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,2048,768,0.02360960046450297
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,2048,1024,0.01824959913889567
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,2048,512,0.006333866715431213
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,2048,768,0.016221867005030314
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,2048,512,0.021248000860214233
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,2048,256,0.005662933488686879
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,2048,256,0.018926932414372762
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,2048,256,0.016498133540153503
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,2048,128,0.004642133414745331
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,2048,512,0.016574933131535848
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,2048,128,0.016692266861597697
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,2048,128,0.016536532839139303
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,2048,64,0.004276266694068909
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,2048,64,0.017268266280492148
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,2048,32,0.004637866715590159
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,2048,32,0.016612266500790916
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,1536,65536,0.3418944040934245
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,1536,65536,0.7344309488932292
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,1536,51200,0.2725290616353353
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,1536,51200,0.5807519912719726
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,2048,51200,0.24514986673990885
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,1536,16384,0.08671039740244547
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,1536,16384,0.1889461358388265
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,1536,16384,0.08794240156809488
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,1536,12288,0.06580053170522054
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,1536,12288,0.14300692876180013
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,2048,65536,0.3049205462137858
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,1536,10240,0.05433386564254761
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,1536,10240,0.11325546900431316
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,1536,12288,0.06997120380401611
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,1536,8192,0.04445866743723552
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,1536,8192,0.08969066937764486
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,1536,10240,0.06318720181783041
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,1536,7168,0.038184531529744464
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,1536,7168,0.07877653439839681
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,1536,8192,0.05181440114974976
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,1536,6144,0.03333439826965332
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,1536,6144,0.06563839912414551
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,1536,7168,0.048130134741465255
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,1536,5120,0.027505065997441607
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,1536,5120,0.05528533458709717
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,1536,6144,0.04280853271484375
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,1536,4096,0.022921599944432578
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,1536,4096,0.046284798781077066
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,1536,5120,0.03865813414255778
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,1536,3584,0.020396800835927327
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,1536,3584,0.04325759808222453
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,1536,4096,0.03408426841100057
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,1536,3072,0.018252799908320107
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,1536,3072,0.03924586772918701
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,1536,3584,0.03091946641604106
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,1536,2560,0.015708800156911215
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,1536,2560,0.034967466195424395
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,1536,3072,0.02874666651089986
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,1536,2048,0.013030399878819784
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,1536,2048,0.03218666712443034
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,1536,2560,0.02693120042483012
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,1536,1536,0.010614400108655293
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,1536,1536,0.028161066770553588
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,1536,1536,0.02146240075429281
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,1536,1024,0.008332799871762593
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,1536,1024,0.024906667073567708
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,1536,51200,0.24470507303873695
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,1536,768,0.007155199845631917
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,1536,768,0.022706133127212525
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,1536,1024,0.01991893251736959
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,1536,2048,0.02355946699778239
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,1536,512,0.005921066800753275
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,1536,512,0.022587732474009196
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,1536,256,0.005086933573087057
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,1536,768,0.018772266308466592
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,1536,256,0.019740800062815346
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,1536,128,0.004311466713746389
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,1536,128,0.01748159925142924
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,1536,512,0.01900906761487325
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,1536,256,0.018637865781784058
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,1536,64,0.003894400099913279
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,1536,32,0.004211199780305227
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,1536,64,0.016992000738779704
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,1536,32,0.016615466276804606
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,1536,128,0.019399466117223103
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,1024,65536,0.24628480275472006
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,1024,51200,0.19241493542989094
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,1024,65536,0.6970453262329102
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,1024,51200,0.5805418650309245
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,1024,16384,0.06243306795756022
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,1024,16384,0.17706027030944824
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,1024,16384,0.06447466611862182
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,1024,12288,0.0466869314511617
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,1024,12288,0.13319253126780192
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,1536,65536,0.30690666834513347
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,1024,10240,0.03870613177617391
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,1024,10240,0.10442773501078288
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,1024,12288,0.05147413412729899
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,1024,8192,0.03127786715825399
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,1024,8192,0.08131306966145833
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,1024,10240,0.04642453193664551
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,1024,7168,0.026634667317072553
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,1024,7168,0.07022613684336344
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,1024,7168,0.034135464827219644
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,1024,6144,0.022827732563018798
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,1024,6144,0.057714132467905674
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,1024,8192,0.037036800384521486
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,1024,5120,0.019562667608261107
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,1024,5120,0.04849066734313965
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,1024,5120,0.02924799919128418
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,1024,4096,0.016525866587956746
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,1024,6144,0.03133333325386047
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,1024,4096,0.0415338675181071
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,1024,3584,0.014605866869290671
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,1024,4096,0.025253333648045856
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,1024,3584,0.038977066675821945
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,1024,3072,0.012936533490816752
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,1024,3072,0.03564266761144002
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,1024,3584,0.023119999965031942
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,1024,2560,0.010820266604423524
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,1024,3072,0.02103253404299418
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,1024,2560,0.03089386622111003
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,1024,2048,0.009276800354321798
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,1024,2048,0.029686399300893146
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,1024,2560,0.01948053240776062
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,1024,2048,0.017359999815622966
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,1024,1536,0.00798933357000351
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,1024,1536,0.02587839961051941
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,1024,1024,0.006308266520500183
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,1024,1024,0.022768000761667885
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,1024,1536,0.015292800466219583
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,1024,1024,0.01448853313922882
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,1024,768,0.005457066496213277
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,1024,768,0.02165013353029887
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,1024,512,0.004659200211366018
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,1024,512,0.020886399348576865
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,1024,768,0.013307733337084451
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,1024,256,0.003908266623814901
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,1024,512,0.013660800457000733
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,1024,256,0.01760853330294291
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,1024,128,0.0033973333736260734
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,1024,128,0.01551466683546702
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,1024,256,0.013237333297729493
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,1024,64,0.0030218665798505146
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,1024,128,0.014155733585357665
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,1024,64,0.016266666849454246
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,1024,32,0.003458133339881897
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,1024,51200,0.1702474594116211
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,1024,32,0.014965333541234336
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,768,65536,0.20930132865905762
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,768,51200,0.1628917376200358
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,768,65536,0.6619274775187175
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,768,51200,0.5260202725728352
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,768,16384,0.05395946502685547
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,768,16384,0.17493546803792318
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,768,16384,0.06536746819814046
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,768,12288,0.04154133399327596
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,768,12288,0.13162879943847655
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,1024,65536,0.21107734044392906
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,768,10240,0.03354239861170451
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,768,10240,0.10156479676564534
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,768,12288,0.05292799870173136
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,768,8192,0.027111466725667315
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,768,8192,0.07839146455128988
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,768,10240,0.047244799137115476
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,768,7168,0.023685334126154582
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,768,7168,0.06628053188323975
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,768,8192,0.03784106572469075
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,768,6144,0.021134932835896812
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,768,6144,0.055845332145690915
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,768,7168,0.035070931911468504
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,768,5120,0.017922133207321167
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,768,5120,0.04721920092900594
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,768,6144,0.03246506651242574
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,768,4096,0.014999467134475707
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,768,4096,0.04096320072809855
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,768,5120,0.0300981342792511
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,768,3584,0.013484799861907959
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,768,3584,0.03839786847432454
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,768,4096,0.02637973427772522
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,768,3072,0.012070399522781373
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,768,3072,0.035027201970418295
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,768,3584,0.024321067333221435
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,768,2560,0.010327466328938802
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,768,2560,0.03116160035133362
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,768,3072,0.022207999229431154
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,768,2048,0.008826667070388794
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,768,2048,0.028961066404978437
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,768,2048,0.01853546698888143
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,768,1536,0.007575466732184092
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,768,1536,0.02477653423945109
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,768,51200,0.1703381379445394
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,768,1024,0.0061589335401852924
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,768,1024,0.022843732436498006
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,768,1536,0.016145066420237223
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,768,768,0.005510400235652924
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,768,1024,0.015339733163515726
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,768,2560,0.020746666193008422
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,768,512,0.0046623999873797095
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,768,512,0.02118826707204183
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,768,768,0.0145578662554423
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,768,256,0.004249600072701773
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,768,256,0.01889280080795288
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,768,512,0.014891733725865683
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,768,128,0.003923200070858002
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,768,256,0.015265066425005594
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,768,128,0.016923733552296958
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,768,64,0.003435733417669932
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,768,128,0.015326933066050211
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,768,64,0.016827734311421712
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,768,32,0.018514132499694823
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,512,65536,0.1454922676086426
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,512,65536,0.6364501317342122
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,768,32,0.0034613333642482757
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,768,768,0.021262933810551964
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,512,51200,0.11643520196278889
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,512,51200,0.5069258689880372
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,512,16384,0.039190399646759036
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,512,16384,0.17062400182088217
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,768,65536,0.21068479220072428
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,512,12288,0.029392000039418536
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,512,12288,0.1266869306564331
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,512,16384,0.06367679834365844
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,512,10240,0.023774933815002442
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,512,10240,0.09910293420155844
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,512,12288,0.050817068417867026
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,512,8192,0.01876373291015625
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,512,8192,0.07084693113962809
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,512,10240,0.04485546747843425
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,512,7168,0.016151466965675355
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,512,8192,0.03537173271179199
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,512,7168,0.060627198219299315
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,512,6144,0.014177067081133523
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,512,6144,0.051082666714986166
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,512,7168,0.03308373292287191
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,512,5120,0.012329600254694621
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,512,6144,0.030456533034642536
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,512,5120,0.04322346846262614
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,512,4096,0.010403199990590414
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,512,4096,0.03728426694869995
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,512,5120,0.028411734104156493
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,512,3584,0.009588266412417095
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,512,4096,0.02442453304926554
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,512,3584,0.03509013255437215
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,512,3072,0.008409600456555684
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,512,3072,0.033062400420506795
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,512,3584,0.02268480062484741
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,512,2560,0.007584000130494435
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,512,3072,0.0206496000289917
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,512,2560,0.02889813383420308
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,512,2048,0.006611200173695882
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,512,2048,0.027752532561620073
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,512,2560,0.01902079979578654
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,512,1536,0.005895466605822245
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,512,2048,0.016923733552296958
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,512,1536,0.024795732895533242
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,512,1024,0.004692266881465912
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,512,1024,0.023141332467397056
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,512,1536,0.014542933305104574
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,512,768,0.004249600072701773
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,512,1024,0.013702399532000222
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,512,768,0.012556800246238708
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,512,512,0.0038602667550245917
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,512,512,0.019869865973790486
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,512,512,0.012885333100954691
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,512,256,0.003453866640726725
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,512,51200,0.16820160547892254
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,512,256,0.018781866629918417
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,512,128,0.0030730667213598887
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,512,256,0.012853333353996277
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,512,128,0.01644480029741923
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,512,64,0.002979200085004171
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,512,768,0.020651733875274657
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,512,32,0.0030239999294281008
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,512,128,0.012888532876968384
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,512,32,0.014841600259145101
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,256,65536,0.09629013538360595
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,256,65536,0.6092021306355794
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,256,51200,0.07530346711476645
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,256,51200,0.48337386449178055
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,512,64,0.015608533223470052
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,256,16384,0.030598400036493938
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,256,16384,0.1606229305267334
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,512,65536,0.2101813316345215
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,256,12288,0.022366933027903237
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,256,12288,0.12022933165232341
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,256,16384,0.06347200075785318
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,256,10240,0.015617066621780395
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,256,10240,0.09202986558278402
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,256,12288,0.05084480047225952
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,256,8192,0.013392000397046407
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,256,8192,0.06668266455332438
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,256,10240,0.04361813465754191
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,256,7168,0.01160533328851064
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,256,7168,0.055409065882364904
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,256,7168,0.03300159970919291
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,256,6144,0.010566400488217671
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,256,6144,0.047381333510080975
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,256,8192,0.034976001580556235
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,256,5120,0.009478400150934856
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,256,5120,0.041227734088897704
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,256,51200,0.16826240221659344
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,256,4096,0.008366933465003968
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,256,6144,0.030450133482615153
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,256,4096,0.0365994652112325
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,256,3584,0.007565866907437642
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,256,3584,0.03431893189748128
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,256,4096,0.0243562658627828
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,256,3072,0.0067669332027435304
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,256,3072,0.03243839939435323
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,256,3584,0.02268480062484741
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,256,2560,0.005936000247796377
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,256,5120,0.028220800558725993
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,256,2560,0.02921813329060872
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,256,65536,0.2095424016316732
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,256,3072,0.020243199666341145
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,256,2048,0.00553173323472341
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,256,2048,0.027698133389155073
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,256,1536,0.004949333270390829
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,256,1536,0.023754666248957314
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,256,2048,0.0165802667538325
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,256,1024,0.003853866706291834
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,256,1536,0.01446613371372223
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,256,1024,0.02108906706174215
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,256,768,0.003487999985615412
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,256,768,0.02070186734199524
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,256,1024,0.013726933797200521
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,256,512,0.003058133274316788
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,256,512,0.01925333340962728
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,256,768,0.012473600109418233
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,256,256,0.0027413333455721537
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,256,2560,0.018972800175348917
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,256,512,0.012854400277137756
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,256,128,0.0026208000878492994
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,256,128,0.01630826691786448
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,256,256,0.012533332904179892
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,256,128,0.012444800138473511
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,256,64,0.002586666742960612
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,256,32,0.0026581334571043652
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,256,64,0.01639893352985382
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,256,32,0.01776960094769796
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,256,256,0.01770346760749817
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,128,65536,0.6075957616170247
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,128,51200,0.059938132762908936
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,128,51200,0.47899414698282883
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,128,65536,0.08139733473459879
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,128,16384,0.023829332987467446
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,128,16384,0.1615104039510091
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,128,12288,0.015005866686503092
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,128,12288,0.11872426668802898
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,128,12288,0.04882239898045858
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,128,10240,0.0132533331712087
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,128,16384,0.06206826766331991
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,128,10240,0.08889813423156738
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,128,8192,0.011932800213495891
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,128,8192,0.06375786860783896
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,128,8192,0.03423360188802083
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,128,7168,0.010384000341097514
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,128,7168,0.05356906652450562
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,128,10240,0.04163306554158529
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,128,6144,0.009149866302808125
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,128,6144,0.046112000942230225
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,128,51200,0.16477440198262533
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,128,5120,0.00798399994770686
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,128,5120,0.03914346694946289
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,128,7168,0.03208213249842326
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,128,4096,0.007129600147406261
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,128,4096,0.035284264882405596
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,128,65536,0.20584746996561684
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,128,5120,0.02712213397026062
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,128,3584,0.0067658667763074234
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,128,6144,0.028853332996368407
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,128,3584,0.03297706643740336
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,128,4096,0.023913600047429404
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,128,3072,0.006342400113741558
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,128,3072,0.03088853359222412
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,128,2560,0.005514666438102722
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,128,2560,0.026936533053716023
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,128,2048,0.004637866715590159
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,128,3584,0.02145919998486837
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,128,3072,0.019367466370264687
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,128,2048,0.026784000794092815
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,128,1536,0.004281599819660187
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,128,1536,0.024025599161783852
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,128,1024,0.003443199892838796
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,128,2048,0.015686399737993875
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,128,2560,0.018227199713389076
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,128,1536,0.013643733660380044
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,128,1024,0.021201066176096597
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,128,768,0.00340693344672521
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,128,512,0.003083733220895131
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,128,768,0.02063573400179545
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,128,1024,0.012866133451461792
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,128,768,0.011563733220100403
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,128,512,0.019299199183781944
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,128,512,0.011627733707427979
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,128,256,0.016578132907549538
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,128,128,0.0025450666745503742
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,128,128,0.015165866414705912
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,128,64,0.00257493332028389
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,128,128,0.012063999970753986
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3329,128,256,0.011637333035469054
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,128,32,0.002595199892918269
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,128,64,0.014788267016410828
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,128,32,0.01470080018043518
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,64,65536,0.07181119918823242
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,64,51200,0.05722346703211466
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,64,16384,0.021949867407480873
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,64,51200,0.4761237462361653
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,64,65536,0.5995253245035808
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,64,16384,0.15869867006937663
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,64,12288,0.012583466370900473
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,64,10240,0.010840533177057902
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,64,12288,0.11737279891967774
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,64,8192,0.009541333715120951
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,64,10240,0.08832746346791585
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,128,256,0.00264533335963885
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,64,7168,0.008930133779843648
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,64,6144,0.009176533420880635
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,64,7168,0.05295039812723795
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,64,8192,0.06369173526763916
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,64,6144,0.04529386758804321
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,64,5120,0.007976533472537994
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,64,4096,0.006741333504517872
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,64,3584,0.006371200084686279
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,64,4096,0.035010135173797606
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,64,5120,0.03912746508916219
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,64,3072,0.005829333265622457
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,64,2560,0.00535999983549118
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,64,3072,0.03113066752751668
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,64,2560,0.02689066727956136
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,64,2048,0.00476800004641215
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,64,2048,0.026446932554244997
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,64,1536,0.0042346666256586705
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,64,1024,0.003491200009981791
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,64,1536,0.022779732942581177
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,64,768,0.003052799900372823
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,64,1024,0.021718400716781616
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,64,768,0.02035306692123413
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,64,512,0.0030645333230495454
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,64,256,0.0025909334421157838
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,64,512,0.01869759956995646
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,64,128,0.0022133332987626392
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,64,256,0.018062933286031087
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,64,128,0.015209600329399109
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,64,64,0.0022656001150608064
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,64,32,0.0025898667673269907
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,64,64,0.014965333541234336
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,64,32,0.016012799739837647
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,64,3584,0.03306666612625122
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,32,51200,0.05695679982503256
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,32,65536,0.6016138712565104
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,32,16384,0.020447999238967896
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,32,51200,0.47636906305948895
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,32,12288,0.010889599720637005
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,32,16384,0.15737600326538087
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,32,12288,0.12072746753692627
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,32,10240,0.009562666217486065
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,32,8192,0.00862506628036499
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,32,10240,0.09027946790059407
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,32,7168,0.008558932940165203
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,32,8192,0.06448426644007364
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,32,65536,0.07108480135599772
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,32,6144,0.013184000054995218
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,32,7168,0.05362666845321655
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,32,5120,0.011289599537849426
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,32,6144,0.04783573150634766
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,32,4096,0.009193600217501322
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,32,3584,0.008737066388130188
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,32,4096,0.03504213492075602
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,32,5120,0.04119253158569336
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,32,3584,0.033181866010030106
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,32,3072,0.007692799965540569
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,32,2560,0.007096533477306366
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,32,2048,0.0046965335806210835
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,32,3072,0.030908799171447753
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,32,2560,0.028924800952275592
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,32,1536,0.004011733333269755
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,32,2048,0.026272000869115193
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,32,1536,0.022694400946299233
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,32,768,0.0031648000081380212
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,32,768,0.020060799519220986
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,32,512,0.0027274665733178455
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,32,1024,0.021499733130137123
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,32,256,0.0026047999660174055
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,32,512,0.018633600076039633
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,32,128,0.0024906667570273083
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,32,64,0.0022890667120615644
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,32,128,0.014453333616256715
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,32,32,0.0021717332303524016
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,32,64,0.014542933305104574
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,32,32,0.013752532998720804
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3329,32,1024,0.0034304000437259674
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3329,32,256,0.017180800437927246
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,65536,16384,3.9070762634277343
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,65536,51200,6.793325805664063
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,65536,16384,2.0736202239990233
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,65536,12288,3.020299784342448
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,65536,12288,1.5001599629720053
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,65536,51200,12.818016560872397
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,65536,10240,2.561781311035156
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,65536,10240,1.263606389363607
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,65536,10240,0.7411381403605144
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,65536,8192,2.0679957071940107
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,65536,8192,1.057055981953939
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,65536,12288,0.8304543813069662
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,65536,7168,1.7324095408121747
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,65536,7168,0.9294837315877279
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,65536,16384,1.1220767974853516
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,65536,6144,1.4098239898681642
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,65536,6144,0.7542474746704102
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,65536,8192,0.61572691599528
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,65536,5120,1.1700672149658202
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,65536,5120,0.6631893157958985
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,65536,7168,0.5713866551717122
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,65536,4096,0.9037525177001953
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,65536,4096,0.5175498644510905
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,65536,6144,0.49164905548095705
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,65536,3584,0.7881376266479492
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,65536,3584,0.45349973042805986
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,65536,5120,0.4532928148905436
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,65536,3072,0.6791818618774415
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,65536,3072,0.4084938685099284
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,65536,4096,0.40421546300252276
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,65536,2560,0.5618357340494792
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,65536,2560,0.3412800153096517
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,65536,3584,0.3618037223815918
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,65536,2048,0.4577269236246745
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,65536,2048,0.3002559979756673
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,65536,2560,0.30536746978759766
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,65536,1536,0.34957974751790366
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,65536,1536,0.21365653673807777
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,65536,3072,0.32288853327433265
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,65536,1024,0.24292160669962565
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,65536,1024,0.1626431941986084
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,65536,1024,0.21931519508361816
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,65536,768,0.19016960461934407
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,65536,768,0.1367263952891032
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,65536,1536,0.22854399681091309
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,65536,512,0.13840640385945638
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,65536,512,0.11375146706899006
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,65536,512,0.21078933080037437
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,65536,256,0.09497066338857016
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,65536,256,0.0955509344736735
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,65536,768,0.20842132568359376
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,65536,128,0.08304426670074463
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,65536,128,0.08382186889648438
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,65536,2048,0.2685994784037272
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,65536,64,0.0689237356185913
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,65536,64,0.08489387035369873
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,65536,32,0.07101120154062907
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,65536,32,0.0842570702234904
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,65536,128,0.20942079226175941
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,65536,256,0.20985706647237143
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,51200,65536,7.354859924316406
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,51200,65536,14.360089111328126
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,51200,51200,5.700369262695313
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,51200,51200,10.93217061360677
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,51200,16384,3.3283137003580725
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,51200,16384,1.7277450561523438
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,51200,16384,0.877294921875
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,51200,12288,2.49410883585612
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,51200,12288,1.3248074849446616
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,51200,12288,0.6603957494099935
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,51200,10240,2.0652725219726564
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,51200,10240,1.0330613454182944
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,51200,10240,0.5869173049926758
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,51200,8192,1.5700852711995443
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,51200,8192,0.8205535888671875
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,65536,51200,3.6447583516438806
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,51200,7168,1.2814037322998046
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,51200,7168,0.6891221364339193
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,51200,8192,0.48665812810262044
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,51200,6144,1.0454943974812827
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,51200,6144,0.6317237218221028
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,51200,7168,0.4493674596150716
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,51200,5120,0.8514848073323569
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,51200,5120,0.554578145345052
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,51200,6144,0.38922026952107747
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,51200,4096,0.7000778834025065
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,51200,4096,0.417688528696696
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,51200,5120,0.3509152094523112
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,51200,3584,0.6248010635375977
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,51200,3584,0.36506560643513997
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,51200,51200,2.8341524759928385
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,51200,3072,0.5314112027486165
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,51200,3072,0.31952107747395836
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,51200,3584,0.28224639892578124
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,51200,2560,0.43352638880411787
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,51200,2560,0.2606602668762207
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,51200,4096,0.3181941350301107
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,51200,2048,0.3602858543395996
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,51200,2048,0.23048426310221354
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,51200,2560,0.23928960164388022
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,51200,1536,0.2683327992757162
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,51200,1536,0.1708288033803304
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,51200,2048,0.21221440633138022
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,51200,1024,0.19206612904866535
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,51200,1024,0.131605331103007
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,51200,3072,0.25277759234110514
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,51200,768,0.14788907368977863
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,51200,768,0.10934186776479085
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,51200,1024,0.17610559463500977
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,51200,512,0.1099082628885905
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,51200,512,0.09477866490681966
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,51200,768,0.1654965400695801
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,51200,256,0.0746560017267863
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,51200,256,0.07839679718017578
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,51200,256,0.16422719955444337
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,51200,128,0.05932373205820719
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,51200,128,0.06936960220336914
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,51200,1536,0.1810922622680664
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,51200,64,0.05457706848780314
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,51200,64,0.07034666538238525
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,51200,128,0.16670400301615398
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,51200,32,0.05462400118509928
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,51200,32,0.06864853700002035
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,16384,65536,2.4977066040039064
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,51200,512,0.16544960339864095
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,16384,65536,4.003757985432943
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,16384,51200,1.9721322377522785
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,16384,51200,3.2251253763834633
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,16384,16384,0.9711594899495444
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,16384,16384,0.6417952219645182
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,16384,16384,0.31014506022135413
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,16384,12288,0.6997247695922851
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,16384,12288,0.5090282758076985
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,16384,12288,0.2363466739654541
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,16384,10240,0.5991114934285482
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,16384,10240,0.4097589174906413
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,16384,10240,0.21057066917419434
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,16384,8192,0.46931413014729817
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,16384,8192,0.329203192392985
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,16384,8192,0.17436800003051758
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,16384,7168,0.4036320050557454
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,16384,7168,0.2864341417948405
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,16384,51200,0.922212282816569
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,16384,6144,0.3420607884724935
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,16384,6144,0.23340266545613608
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,16384,7168,0.16175786654154461
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,16384,5120,0.28703254063924155
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,16384,5120,0.19243626594543456
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,16384,5120,0.128984530766805
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,16384,4096,0.2290463924407959
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,16384,4096,0.15849812825520831
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,16384,6144,0.13984853426615398
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,16384,3584,0.19950720469156902
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,16384,3584,0.13904639879862468
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,16384,3584,0.10175039768218994
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,16384,3072,0.1723573366800944
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,16384,3072,0.12096107006072998
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,16384,4096,0.11395520369211834
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,16384,2560,0.14602452913920086
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,16384,2560,0.10326186815897624
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,16384,2560,0.08758827050526938
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,16384,2048,0.11820053259531657
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,16384,3072,0.09106986522674561
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,16384,2048,0.0893450657526652
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,16384,1536,0.0904362678527832
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,16384,1536,0.06994346777598062
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,16384,1536,0.06571306784947714
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,16384,1024,0.06279999812444051
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,16384,1024,0.05804053147633871
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,16384,2048,0.0758997360865275
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,16384,768,0.048681600888570147
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,16384,768,0.05075200001398722
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,16384,1024,0.06006079912185669
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,16384,512,0.03678400119145711
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,16384,512,0.044531198342641194
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,16384,768,0.05172160069147745
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,16384,256,0.025486934185028075
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,16384,256,0.03726079861323039
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,16384,512,0.05085973342259725
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,16384,128,0.021028266350428263
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,16384,128,0.033037867148717245
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,16384,256,0.05011839866638183
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,16384,64,0.01958400011062622
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,16384,64,0.03290773431460063
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,16384,32,0.02018666664759318
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,16384,32,0.03325546582539876
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,16384,128,0.05118933518727621
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,12288,65536,3.194139607747396
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,12288,65536,2.0762762705485027
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,12288,51200,2.4362411499023438
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,12288,51200,1.6115114847819012
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,51200,65536,3.669290669759115
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,12288,16384,0.729414431254069
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,12288,16384,0.5156415939331055
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,16384,65536,1.1848084767659506
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,12288,12288,0.5473269144694011
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,12288,12288,0.40092267990112307
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,12288,12288,0.19609707196553547
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,12288,10240,0.45007041295369465
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,12288,10240,0.35051094690958656
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,12288,16384,0.25880746841430663
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,12288,8192,0.36787732442220056
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,12288,8192,0.25379625956217444
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,12288,10240,0.1753973325093587
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,12288,7168,0.3184618631998698
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,12288,7168,0.220633602142334
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,12288,8192,0.1445269266764323
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,12288,6144,0.27380587259928385
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,12288,6144,0.1905941327412923
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,12288,7168,0.13408212661743163
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,12288,5120,0.22321279843648276
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,12288,5120,0.1585770606994629
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,12288,6144,0.11725653012593587
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,12288,4096,0.17890772819519044
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,12288,4096,0.13030186494191487
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,12288,5120,0.10649279753367107
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,12288,3584,0.15495893160502117
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,12288,3584,0.11751680374145508
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,12288,4096,0.09411306381225586
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,12288,3072,0.13407252629597982
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,12288,3072,0.10392213662465413
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,12288,3584,0.08414080142974853
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,12288,2560,0.11154239972432453
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,12288,2560,0.08756159941355388
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,12288,3072,0.07512106895446777
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,12288,2048,0.09147946834564209
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,12288,2048,0.07791786988576253
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,12288,2048,0.06334506670633952
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,12288,1536,0.07029546896616617
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,12288,1536,0.06155093510945638
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,12288,2560,0.07083733081817627
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,12288,1024,0.04901866515477498
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,12288,1024,0.05181013345718384
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,12288,1024,0.05589973529179891
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,12288,768,0.03849493265151978
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,12288,768,0.049507200717926025
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,12288,1536,0.05784853299458822
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,12288,512,0.02956693371136983
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,12288,512,0.04500480095545451
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,12288,768,0.051874132951100674
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,12288,256,0.024771199623743693
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,12288,512,0.05190399885177612
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,12288,256,0.03711680173873901
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,12288,128,0.021066667636235555
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,12288,128,0.033720533053080246
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,12288,128,0.05145066579182943
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,12288,64,0.019003732999165853
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,12288,64,0.03378453254699707
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,12288,32,0.020232532421747842
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,12288,32,0.03315733273824056
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,12288,256,0.05186560153961182
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,10240,65536,2.5761067708333334
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,10240,65536,1.7587018330891926
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,10240,51200,2.002085367838542
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,10240,51200,1.3621631622314454
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,12288,51200,0.7734218597412109
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,10240,16384,0.6127253214518229
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,10240,16384,0.45744854609171554
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,10240,16384,0.212610133488973
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,10240,12288,0.44948479334513347
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,10240,12288,0.34185492197672523
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,12288,65536,0.9744757334391275
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,10240,10240,0.36511360804239906
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,10240,10240,0.28533547719319663
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,10240,12288,0.16372373898824055
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,10240,8192,0.2957344055175781
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,10240,8192,0.2150165398915609
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,10240,10240,0.14537173906962078
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,10240,7168,0.2574965318044027
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,10240,7168,0.18887252807617189
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,10240,8192,0.12054719924926757
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,10240,6144,0.21873706181844077
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,10240,6144,0.1635434627532959
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,10240,7168,0.11145599683125813
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,10240,5120,0.18176852862040202
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,10240,5120,0.13533120155334472
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,10240,6144,0.09815359910329183
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,10240,4096,0.14425600369771321
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,10240,4096,0.11202452977498371
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,10240,5120,0.09008959929148355
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,10240,3584,0.12582080364227294
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,10240,3584,0.09915200074513754
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,10240,3584,0.06984639962514241
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,10240,3072,0.1085045337677002
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,10240,3072,0.08768746852874756
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,10240,4096,0.0783349355061849
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,10240,2560,0.09141653378804525
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,10240,2560,0.07485973040262858
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,10240,51200,0.620142936706543
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,10240,2048,0.0731007973353068
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,10240,2048,0.06795732975006104
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,10240,3072,0.06257493495941162
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,10240,1536,0.05645653406778971
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,10240,1536,0.05317759911219279
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,10240,2048,0.054695467154184975
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,10240,1024,0.03900800148646037
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,10240,2560,0.06023039817810059
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,10240,1024,0.044863998889923096
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,10240,768,0.030434133609135945
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,10240,768,0.03935679992039998
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,10240,768,0.04582080046335856
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,10240,512,0.02217386762301127
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,10240,512,0.034306132793426515
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,10240,1024,0.04858666658401489
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,10240,256,0.01629973351955414
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,10240,256,0.028716800610224406
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,10240,256,0.044531198342641194
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,10240,128,0.015275733669598899
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,10240,128,0.0266485333442688
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,10240,512,0.04536853233973186
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,10240,64,0.012851199507713318
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,10240,1536,0.0498250683148702
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,10240,32,0.013698133826255798
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,10240,64,0.026080000400543212
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,10240,32,0.025727999210357667
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,10240,128,0.044785066445668535
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,8192,65536,1.5077013651529947
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,8192,65536,1.9182591756184895
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,8192,51200,1.1811178843180339
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,8192,51200,1.5055957794189454
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,8192,16384,0.458515199025472
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,8192,16384,0.3904266675313314
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,8192,16384,0.18589226404825848
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,8192,12288,0.3463871955871582
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,8192,12288,0.30754025777180993
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,10240,65536,0.7794890721638997
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,8192,10240,0.30088214874267577
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,8192,10240,0.2549184004465739
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,8192,10240,0.12691413561503093
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,8192,8192,0.2404874642690023
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,8192,8192,0.1881290594736735
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,8192,12288,0.14246826171875
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,8192,7168,0.21589760780334472
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,8192,7168,0.16760746637980145
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,8192,8192,0.10474879741668701
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,8192,6144,0.1807967980702718
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,8192,6144,0.14369066556294757
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,8192,7168,0.09737813472747803
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,8192,5120,0.15171839396158854
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,8192,5120,0.11973439852396647
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,8192,5120,0.07820693651835123
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,8192,4096,0.12009173234303791
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,8192,4096,0.1000480016072591
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,8192,6144,0.08509439627329508
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,8192,3584,0.10372587045033772
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,8192,3584,0.09186986287434896
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,8192,4096,0.06625599861145019
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,8192,3072,0.09082773526509604
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,8192,3072,0.0812000036239624
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,8192,3584,0.058874666690826416
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,8192,2560,0.07542293071746826
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,8192,2560,0.06981226603190104
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,8192,2560,0.04934826691945394
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,8192,2048,0.06109866698582968
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,8192,2048,0.06184106667836507
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,8192,3072,0.051028267542521155
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,8192,1536,0.04807680050532023
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,8192,1536,0.04920639991760254
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,8192,1536,0.0370901346206665
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,8192,1024,0.03347413142522176
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,8192,1024,0.039851733048756915
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,8192,2048,0.042906665802001955
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,8192,768,0.026165332396825152
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,8192,768,0.03649066686630249
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,8192,1024,0.03583893378575643
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,8192,512,0.01976213256518046
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,8192,512,0.03153280019760132
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,8192,768,0.032340266307195026
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,8192,256,0.014762666821479798
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,8192,256,0.026964267094930012
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,8192,256,0.032219733794530234
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,8192,128,0.011907200018564861
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,8192,128,0.02445546587308248
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,8192,128,0.03255253235499064
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,8192,512,0.03174933393796285
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,8192,64,0.010788266857465107
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,8192,32,0.011264000336329143
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,8192,64,0.024877866109212242
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,8192,32,0.023513599236806234
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,7168,65536,1.421407953898112
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,7168,65536,1.903489049275716
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,7168,51200,1.4659348805745442
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,7168,51200,1.1314069112141927
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,8192,51200,0.542794672648112
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,7168,16384,0.4269290606180827
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,7168,16384,0.39869120915730794
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,7168,16384,0.1634218692779541
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,7168,12288,0.32370452880859374
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,7168,12288,0.2949664115905762
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,8192,65536,0.6858399709065754
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,7168,10240,0.2592746734619141
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,7168,10240,0.23223360379536948
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,7168,12288,0.12580373287200927
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,7168,8192,0.20822399457295737
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,7168,8192,0.16724692980448405
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,7168,10240,0.11267840067545573
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,7168,7168,0.18282559712727864
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,7168,7168,0.15092159907023112
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,7168,7168,0.08663893540700277
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,7168,6144,0.1553599993387858
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,7168,6144,0.12913493315378827
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,7168,8192,0.09315199851989746
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,7168,5120,0.12799572944641113
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,7168,5120,0.10766719977060954
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,7168,5120,0.06949439843495687
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,7168,4096,0.10067626635233562
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,7168,4096,0.09059092998504639
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,7168,6144,0.07632213433583578
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,7168,3584,0.08974613348642985
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,7168,3584,0.08268906275431315
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,7168,3584,0.05266666809717814
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,7168,3072,0.07678613662719727
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,7168,3072,0.073744002978007
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,7168,4096,0.05934613148371378
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,7168,2560,0.06373546520868936
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,7168,2560,0.061933867136637365
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,7168,2560,0.04514026641845703
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,7168,2048,0.05172266562779745
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,7168,2048,0.053736531734466554
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,7168,3072,0.046988801161448164
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,7168,1536,0.040139734745025635
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,7168,1536,0.04346133470535278
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,7168,51200,0.4711445490519206
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,7168,1024,0.027112533648808796
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,7168,1024,0.03654719988505046
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,7168,2048,0.040345601240793866
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,7168,768,0.02171199917793274
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,7168,768,0.0331007997194926
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,7168,1024,0.035069866975148516
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,7168,512,0.01618773341178894
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,7168,512,0.0291594664255778
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,7168,1536,0.03625066677729289
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,7168,256,0.012312533458073933
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,7168,256,0.02477333347002665
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,7168,256,0.032023467620213825
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,7168,128,0.009960533181826273
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,7168,128,0.027514666318893433
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,7168,512,0.03253226677576701
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,7168,64,0.013262933492660523
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,7168,64,0.026746666431427
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,7168,32,0.014146133263905843
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,7168,768,0.032553599278132124
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,7168,32,0.026973867416381837
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,7168,128,0.0321941335995992
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,6144,65536,1.2712053934733072
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,6144,65536,1.5334964752197267
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,6144,51200,1.0120384216308593
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,6144,51200,1.1567274729410806
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,6144,16384,0.4048181215922038
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,6144,16384,0.32874558766682943
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,6144,16384,0.1657002607981364
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,6144,12288,0.2679807980855306
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,6144,12288,0.24532373746236166
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,7168,65536,0.5894154866536458
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,6144,10240,0.22597014109293617
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,6144,10240,0.19625919659932453
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,6144,10240,0.10943360328674316
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,6144,8192,0.1807360013326009
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,6144,8192,0.15441385904947919
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,6144,12288,0.12377920150756835
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,6144,7168,0.161082665125529
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,6144,7168,0.137553071975708
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,6144,7168,0.08439146677652995
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,6144,6144,0.1359221299489339
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,6144,6144,0.11899519761403402
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,6144,8192,0.089956267674764
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,6144,5120,0.10997119744618733
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,6144,5120,0.10078293482462566
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,6144,5120,0.06755413214365641
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,6144,4096,0.0877013365427653
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,6144,4096,0.08529386520385743
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,6144,6144,0.07446293036142984
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,6144,3584,0.07865386803944906
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,6144,3584,0.07759146690368653
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,6144,3584,0.052118400732676186
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,6144,3072,0.06796159744262695
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,6144,3072,0.06721599896748862
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,6144,4096,0.057608532905578616
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,6144,2560,0.05639253457387289
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,6144,2560,0.05586986541748047
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,6144,3072,0.04676266511281331
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,6144,2048,0.04642239809036255
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,6144,2560,0.04469226598739624
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,6144,2048,0.048978133996327715
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,6144,1536,0.03452800114949544
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,6144,1536,0.04124586582183838
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,6144,1536,0.03548906644185384
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,6144,1024,0.024461867411931355
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,6144,1024,0.03679466644922892
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,6144,2048,0.03981120189030965
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,6144,768,0.020077866315841675
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,6144,768,0.033012266953786215
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,6144,1024,0.03423573176066081
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,6144,512,0.016642133394877114
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,6144,512,0.029230932394663494
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,6144,768,0.031648000081380204
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,6144,256,0.012924800316492716
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,6144,256,0.02696853280067444
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,6144,512,0.03171199957529704
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,6144,128,0.011175466577212016
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,6144,128,0.02353066603342692
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,6144,256,0.031727999448776245
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,6144,64,0.010417067011197408
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,6144,64,0.02393066684405009
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,6144,32,0.010835199554761251
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,6144,128,0.031568000713984176
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,6144,32,0.023179733753204347
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,5120,65536,1.203653335571289
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,5120,65536,1.13841921488444
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,5120,51200,0.921394157409668
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,5120,51200,0.9000480016072592
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,6144,51200,0.4825440088907878
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,5120,16384,0.2852383931477865
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,5120,16384,0.2940832138061523
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,5120,16384,0.1388586680094401
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,5120,12288,0.2160714626312256
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,5120,12288,0.22125226656595864
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,6144,65536,0.6143914540608724
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,5120,10240,0.18427519798278807
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,5120,10240,0.17979092597961427
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,5120,12288,0.1060693343480428
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,5120,8192,0.14614933331807453
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,5120,8192,0.14182292620340983
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,5120,10240,0.09558186531066895
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,5120,7168,0.1313215970993042
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,5120,7168,0.1265226682027181
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,5120,8192,0.07913386821746826
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,5120,6144,0.11129066944122315
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,5120,7168,0.07438720067342122
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,5120,6144,0.10963199933369953
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,5120,5120,0.09361920356750489
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,5120,5120,0.09303253491719564
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,5120,5120,0.05925120115280151
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,5120,4096,0.07542080084482829
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,5120,4096,0.07810986836751302
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,5120,6144,0.0659605344136556
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,5120,3584,0.06719253063201905
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,5120,3584,0.07036159833272299
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,5120,4096,0.05133333206176758
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,5120,3072,0.05699626604715983
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,5120,3584,0.046868268648783365
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,5120,3072,0.06166613499323527
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,5120,2560,0.04804799954096477
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,5120,2560,0.051742935180664064
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,5120,2560,0.04085119962692261
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,5120,2048,0.03908160130182902
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,5120,2048,0.046299731731414794
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,5120,3072,0.04225600163141886
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,5120,1536,0.031030400594075518
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,5120,1536,0.03947413365046183
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,5120,51200,0.4024810791015625
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,5120,1024,0.02105493346850077
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,5120,1024,0.0326581339041392
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,5120,1536,0.03333226641019185
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,5120,768,0.01758613387743632
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,5120,768,0.02914453347524007
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,5120,1024,0.03239360054334005
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,5120,512,0.012411733468373615
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,5120,512,0.026425600051879883
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,5120,2048,0.036651734511057535
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,5120,256,0.009749333063761394
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,5120,256,0.024475733439127602
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,5120,512,0.030528000990549725
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,5120,128,0.00881173312664032
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,5120,128,0.020836265881856282
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,5120,768,0.03002026677131653
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,5120,64,0.0074976002176602675
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,5120,64,0.021396267414093017
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,5120,32,0.007913599908351897
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,5120,128,0.030409600337346392
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,5120,32,0.019352533419926963
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,4096,65536,0.9022752126057944
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,4096,65536,1.0471477508544922
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,5120,256,0.030213334163029987
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,4096,51200,0.7360703786214192
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,4096,51200,0.8144682566324869
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,4096,16384,0.23074026107788087
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,4096,16384,0.26604159673055017
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,4096,16384,0.11176106929779053
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,4096,12288,0.17500799496968586
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,4096,12288,0.1985856056213379
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,5120,65536,0.5082005182902019
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,4096,10240,0.1517866611480713
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,4096,10240,0.15850133895874025
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,4096,12288,0.08732799688975015
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,4096,8192,0.12220586935679119
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,4096,8192,0.12711466948191324
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,4096,10240,0.07842026551564535
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,4096,7168,0.1078111966451009
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,4096,7168,0.11395413080851238
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,4096,7168,0.06145493189493815
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,4096,6144,0.0918229341506958
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,4096,6144,0.09863359928131103
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,4096,8192,0.06538986762364705
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,4096,5120,0.075054931640625
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,4096,5120,0.08454293409983317
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,4096,5120,0.04831573168436686
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,4096,4096,0.060862934589385985
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,4096,4096,0.06995519797007242
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,4096,6144,0.05400853157043457
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,4096,3584,0.05414933363596598
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,4096,3584,0.061818667252858485
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,4096,4096,0.04137920141220093
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,4096,3584,0.037110400199890134
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,4096,3072,0.046727466583251956
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,4096,3072,0.053887999057769774
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,4096,2560,0.03925866683324178
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,4096,2560,0.04707839886347453
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,4096,2560,0.03127040068308513
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,4096,2048,0.03204586704572042
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,4096,2048,0.041723732153574625
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,4096,3072,0.03302719990412394
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,4096,1536,0.025098667542139692
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,4096,1536,0.035420799255371095
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,4096,2048,0.0272053341070811
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,4096,1024,0.017978666226069133
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,4096,1024,0.030923734108606975
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,4096,1536,0.02355946699778239
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,4096,768,0.014485333363215128
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,4096,768,0.02722240090370178
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,4096,1024,0.022709333896636964
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,4096,512,0.011487999558448791
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,4096,512,0.025851732492446898
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,4096,768,0.023169066508611044
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,4096,256,0.008301866551240284
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,4096,256,0.02393066684405009
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,4096,256,0.02309653361638387
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,4096,128,0.007092266778151194
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,4096,128,0.020618667205174766
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,4096,512,0.023600000143051147
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,4096,64,0.007965866724650066
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,4096,64,0.02109760046005249
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,4096,32,0.00842133363087972
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,4096,32,0.020811732610066733
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,3584,65536,0.8780949274698893
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,4096,128,0.02353066603342692
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,3584,65536,1.0054815928141276
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,3584,51200,0.6754463831583659
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,3584,51200,0.7925194422403972
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,4096,51200,0.32063999176025393
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,3584,16384,0.22213546435038248
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,3584,16384,0.26119786898295083
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,3584,16384,0.11294399897257487
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,3584,12288,0.16793813705444335
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,3584,12288,0.1968191941579183
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,4096,65536,0.40070400238037107
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,3584,10240,0.14040746688842773
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,3584,10240,0.15589866638183594
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,3584,12288,0.08818026383717856
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,3584,8192,0.11482026576995849
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,3584,8192,0.12607786655426026
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,3584,10240,0.07971200148264566
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,3584,7168,0.09769492944081624
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,3584,7168,0.11158400376637775
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,3584,7168,0.061936000982920326
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,3584,6144,0.08319359620412191
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,3584,6144,0.09552319844563803
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,3584,8192,0.06685653527577719
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,3584,5120,0.07092586358388266
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,3584,5120,0.08031040032704671
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,3584,5120,0.04927999973297119
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,3584,4096,0.05733760197957357
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,3584,4096,0.06561280091603597
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,3584,6144,0.0544106682141622
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,3584,3584,0.0498741348584493
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,3584,3584,0.05717973311742147
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,3584,3584,0.03883519967397054
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,3584,3072,0.04145919879277547
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,3584,3072,0.05050133466720581
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,3584,4096,0.043332266807556155
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,3584,2560,0.036184533437093096
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,3584,2560,0.043321601549784344
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,3584,2560,0.03385066588719686
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,3584,2048,0.029553065697352093
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,3584,2048,0.03934933344523112
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,3584,3072,0.035266133149464925
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,3584,1536,0.02137706677118937
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,3584,1536,0.03349119822184245
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,3584,1536,0.02717120051383972
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,3584,1024,0.015059199929237366
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,3584,1024,0.0288917342821757
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,3584,2048,0.030055467287699384
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,3584,768,0.012107732892036437
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,3584,768,0.026977066198984784
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,3584,768,0.025778132677078246
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,3584,512,0.00997226635615031
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,3584,512,0.024475733439127602
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,3584,1024,0.02600533366203308
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,3584,256,0.007541333138942718
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,3584,256,0.021362133820851645
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,3584,256,0.027356799443562823
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,3584,128,0.006332799792289734
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,3584,512,0.02637653350830078
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,3584,128,0.019700266917546592
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,3584,64,0.005929600199063619
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,3584,64,0.018552533785502114
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,3584,32,0.006279466549555461
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,3584,32,0.018598399559656777
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,3072,65536,0.6897834777832031
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,3584,128,0.023885866006215416
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,3072,65536,0.9185120264689127
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,3072,51200,0.5389429092407226
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,3072,51200,0.731118901570638
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,3584,51200,0.3218282699584961
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,3072,16384,0.178330659866333
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,3072,16384,0.23421972592671714
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,3072,16384,0.11193066438039143
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,3072,12288,0.1348042647043864
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,3072,12288,0.17552000681559246
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,3584,65536,0.4038858731587728
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,3072,10240,0.11840319633483887
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,3072,10240,0.13995733261108398
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,3072,12288,0.08703253269195557
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,3072,8192,0.09469652970631917
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,3072,8192,0.11325333118438721
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,3072,10240,0.07907413641611735
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,3072,7168,0.08243839740753174
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,3072,7168,0.10072533289591472
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,3072,7168,0.060190931955973304
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,3072,6144,0.06924053033192953
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,3072,8192,0.06589013338088989
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,3072,6144,0.08745173613230386
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,3072,5120,0.05690666834513346
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,3072,5120,0.07208213011423746
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,3072,5120,0.0477567990620931
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,3072,4096,0.04564479986826579
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,3072,4096,0.05842666625976563
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,3072,6144,0.05225813388824463
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,3072,3584,0.04012159903844197
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,3072,3584,0.053574399153391516
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,3072,3584,0.03794986804326375
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,3072,3072,0.03516586621602376
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,3072,3072,0.047915732860565184
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,3072,4096,0.0416757345199585
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,3072,2560,0.029410133759180706
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,3072,2560,0.041010133425394696
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,3072,2560,0.032289065917332965
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,3072,2048,0.024495999018351235
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,3072,2048,0.03719360033671061
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,3072,3072,0.03377813498179118
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,3072,1536,0.019308799505233766
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,3072,1536,0.03313493331273397
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,3072,51200,0.3183648109436035
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,3072,1024,0.013402666648228964
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,3072,1024,0.027383466561635334
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,3072,2048,0.028880000114440918
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,3072,768,0.010827733079592387
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,3072,1536,0.025969066222508747
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,3072,768,0.025655466318130492
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,3072,512,0.008211199939250947
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,3072,512,0.022809600830078124
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,3072,1024,0.025174399216969807
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,3072,512,0.02355626622835795
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,3072,256,0.006810666620731353
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,3072,256,0.021332265933354695
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,3072,128,0.0059114664793014525
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,3072,128,0.01862186590830485
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,3072,256,0.023373866081237794
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,3072,64,0.005463466544946035
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,3072,128,0.02352213263511658
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,3072,64,0.018517333269119262
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,3072,32,0.0054954667886098225
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,3072,32,0.017284266153971353
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,3072,768,0.023457066218058268
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,2560,65536,0.5592629114786785
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,2560,65536,0.8469269434611002
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,2560,51200,0.448685868581136
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,2560,51200,0.6680522918701172
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,2560,16384,0.1404703934987386
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,2560,16384,0.21943999926249186
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,2560,16384,0.0887285312016805
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,2560,12288,0.10689173539479573
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,2560,12288,0.1663690725962321
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,3072,65536,0.40227413177490234
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,2560,10240,0.09198079903920492
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,2560,10240,0.13161706924438477
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,2560,10240,0.0640992005666097
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,2560,8192,0.07440746625264485
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,2560,12288,0.07050346533457438
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,2560,8192,0.10645866394042969
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,2560,7168,0.06560426553090414
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,2560,7168,0.09743999640146891
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,2560,7168,0.049109331766764325
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,2560,6144,0.05671573479970297
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,2560,6144,0.08314133485158284
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,2560,8192,0.052774401505788174
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,2560,5120,0.04665066798528035
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,2560,5120,0.06802879969278972
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,2560,6144,0.04416213432947795
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,2560,4096,0.03895893494288127
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,2560,5120,0.03989439805348714
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,2560,4096,0.05570559899012247
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,2560,3584,0.036297599474589035
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,2560,3584,0.05147093137105306
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,2560,3584,0.03160640001296997
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,2560,3072,0.03154453237851461
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,2560,3072,0.046461868286132815
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,2560,4096,0.034251733620961504
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,2560,2560,0.026530132691065474
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,2560,2560,0.04112319946289063
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,2560,3072,0.028085333108901978
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,2560,2048,0.02230613430341085
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,2560,2048,0.037239468097686766
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,2560,2560,0.026422399282455444
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,2560,1536,0.017704532543818156
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,2560,1536,0.03163733283678691
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,2560,1536,0.020641066630681357
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,2560,1024,0.012568533420562744
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,2560,1024,0.027323732773462932
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,2560,2048,0.023483733336130776
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,2560,768,0.01030293305714925
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,2560,768,0.02469759980837504
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,2560,1024,0.019825067122777304
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,2560,512,0.00804373323917389
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,2560,512,0.023082667589187623
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,2560,768,0.018183465798695883
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,2560,256,0.006470400094985962
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,2560,256,0.020782933632532755
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,2560,512,0.01823893388112386
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,2560,256,0.0184714674949646
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,2560,128,0.0054848000407218935
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,2560,128,0.020056533813476562
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,2560,128,0.018209065993626913
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,2560,64,0.00591786652803421
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,2560,64,0.019038933515548705
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,2560,32,0.006278400123119354
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,2560,32,0.018538665771484376
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,2048,65536,0.4453440030415853
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,2048,65536,0.8119402567545573
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,2048,51200,0.3663701375325521
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,2048,51200,0.628118387858073
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,2560,51200,0.24340693155924478
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,2048,16384,0.11580373446146648
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,2048,16384,0.20396052996317543
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,2048,16384,0.08744213581085206
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,2048,12288,0.08790079752604166
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,2048,12288,0.15510613123575848
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,2560,65536,0.3037226676940918
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,2048,10240,0.0785312016805013
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,2048,10240,0.12296853065490723
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,2048,12288,0.06870293617248535
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,2048,8192,0.06309119860331217
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,2048,8192,0.09872106711069742
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,2048,10240,0.06291093428929648
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,2048,7168,0.053914666175842285
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,2048,7168,0.08769280115763346
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,2048,8192,0.05112640062967936
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,2048,6144,0.04613440036773682
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,2048,6144,0.07385919888814291
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,2048,7168,0.047017598152160646
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,2048,5120,0.0384768009185791
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,2048,5120,0.0598965326944987
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,2048,5120,0.037649067242940266
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,2048,4096,0.031352533896764116
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,2048,4096,0.049803733825683594
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,2048,6144,0.041870931784311935
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,2048,3584,0.02754773298899333
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,2048,3584,0.04560106595357259
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,2048,3584,0.030088533957799275
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,2048,3072,0.02401919960975647
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,2048,3072,0.04134399890899658
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,2048,4096,0.03282986680666606
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,2048,2560,0.02103360096613566
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,2048,2560,0.03711893161137898
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,2048,3072,0.02641493280728658
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,2048,2048,0.017067732413609822
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,2048,2048,0.033019733428955075
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,2048,2048,0.021868799130121866
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,2048,1536,0.01328000028928121
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,2048,1536,0.028844799598058062
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,2048,2560,0.02472533384958903
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,2048,1024,0.009983999530474345
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,2048,1024,0.026629332701365156
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,2048,1536,0.019041067361831664
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,2048,768,0.008318933347860973
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,2048,768,0.022742400566736855
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,2048,768,0.01622719963391622
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,2048,51200,0.24497386614481606
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,2048,512,0.02097066640853882
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,2048,1024,0.018075732390085857
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,2048,256,0.0054837331175804135
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,2048,256,0.019859200716018675
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,2048,512,0.016130133469899496
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,2048,128,0.005107200145721436
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,2048,128,0.01665493349234263
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,2048,512,0.006583466629187266
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,2048,64,0.004248533149560293
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,2048,64,0.016780799627304076
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,2048,128,0.016269866625467935
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,2048,32,0.004604800045490265
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,2048,256,0.016323199868202208
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,2048,32,0.016643200318018594
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,1536,65536,0.34949121475219724
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,1536,51200,0.2717535972595215
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,1536,65536,0.73787841796875
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,1536,51200,0.5874517440795899
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,1536,16384,0.0890186627705892
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,1536,16384,0.19075946807861327
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,1536,16384,0.08800000349680583
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,1536,12288,0.06791040102640787
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,1536,12288,0.14476586977640787
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,2048,65536,0.30347092946370446
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,1536,10240,0.05639359951019287
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,1536,10240,0.11463466485341389
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,1536,12288,0.07002346515655518
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,1536,8192,0.04600533246994019
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,1536,10240,0.06375146706899007
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,1536,8192,0.09007039864857992
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,1536,7168,0.04027413527170817
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,1536,7168,0.07900799910227457
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,1536,7168,0.047203199068705244
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,1536,6144,0.034612266222635905
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,1536,6144,0.0657482663790385
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,1536,8192,0.051641599337259925
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,1536,5120,0.02916693290074666
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,1536,5120,0.05561920007069906
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,1536,6144,0.04240959882736206
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,1536,4096,0.02419840097427368
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,1536,5120,0.03824319839477539
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,1536,4096,0.04668586651484172
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,1536,3584,0.021703465779622396
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,1536,3584,0.043399465084075925
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,1536,4096,0.03373653491338094
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,1536,3072,0.019236266613006592
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,1536,3072,0.03988800048828125
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,1536,3584,0.030500266949335737
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,1536,2560,0.01648319959640503
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,1536,2560,0.03461013237635295
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,1536,3072,0.02804800073305766
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,1536,2048,0.013380266229311624
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,1536,2048,0.03230186700820923
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,1536,2560,0.02683200041453044
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,1536,1536,0.010873599847157796
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,1536,1536,0.029128533601760865
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,1536,2048,0.023502933979034423
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,1536,1024,0.008368000388145447
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,1536,1024,0.025616000096003216
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,1536,1024,0.01949333349863688
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,1536,1536,0.02138026754061381
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,1536,768,0.022639999787012734
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,1536,768,0.018959999084472656
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,1536,512,0.006251733501752217
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,1536,512,0.02094399929046631
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,1536,768,0.007260799904664357
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,1536,256,0.00522986650466919
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,1536,256,0.019093332688013713
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,1536,512,0.019398399194081626
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,1536,128,0.004215466479460398
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,1536,256,0.018773333231608073
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,1536,51200,0.24339839617411294
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,1536,64,0.003905066599448522
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,1536,64,0.01842986742655436
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,1536,32,0.004706133405367533
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,1536,32,0.01855573256810506
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,1536,128,0.016611199577649435
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,1024,65536,0.23986345926920571
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,1536,128,0.019477333625157675
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,1024,51200,0.18564693133036297
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,1024,65536,0.6968522389729818
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,1024,51200,0.5532063802083333
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,1024,16384,0.06099520126978556
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,1024,16384,0.17796692848205567
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,1024,16384,0.06425813436508179
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,1024,12288,0.04644053379694621
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,1024,12288,0.13418240547180177
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,1536,65536,0.3048522631327311
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,1024,10240,0.04002666473388672
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,1024,10240,0.10479786396026611
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,1024,12288,0.05127786795298258
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,1024,8192,0.031421866019566855
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,1024,8192,0.08190080324808756
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,1024,10240,0.04612693389256795
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,1024,7168,0.027906133731206255
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,1024,7168,0.0697760025660197
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,1024,8192,0.03776106834411621
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,1024,6144,0.023861332734425863
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,1024,6144,0.05873920122782389
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,1024,7168,0.034195200602213545
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,1024,5120,0.021242666244506835
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,1024,5120,0.04963093201319377
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,1024,6144,0.03086613416671753
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,1024,4096,0.016937599579493205
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,1024,4096,0.042744533220926924
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,1024,5120,0.029059199492136638
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,1024,3584,0.015633066495259605
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,1024,3584,0.039589333534240725
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,1024,4096,0.025153066714604693
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,1024,3072,0.01376106639703115
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,1024,3584,0.023091200987497965
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,1024,3072,0.03583893378575643
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,1024,2560,0.01213759978612264
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,1024,2560,0.03304959932963054
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,1024,2560,0.01930560072263082
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,1024,2048,0.009879466891288758
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,1024,3072,0.020641066630681357
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,1024,2048,0.030931200583775836
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,1024,1536,0.008378666639328004
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,1024,1536,0.026811732848485308
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,1024,1536,0.014935466647148132
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,1024,1024,0.006791466474533081
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,1024,2048,0.016988799969355265
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,1024,1024,0.02391786575317383
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,1024,768,0.00584853341182073
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,1024,768,0.02257919907569885
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,1024,1024,0.014125866691271463
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,1024,512,0.005108266572157542
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,1024,768,0.013335466384887695
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,1024,512,0.019802665710449217
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,1024,51200,0.16904640197753906
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,1024,256,0.004221866528193155
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,1024,256,0.01692053278287252
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,1024,128,0.00346666673819224
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,1024,128,0.01643519997596741
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,1024,512,0.013271466890970866
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,1024,256,0.013330133756001792
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,1024,64,0.003416533271471659
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,1024,32,0.0034474665919939675
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,1024,64,0.016345600287119545
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,1024,128,0.014095999797185264
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,1024,32,0.016720000902811685
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,768,65536,0.21010026931762696
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,768,65536,0.6635029474894206
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,768,51200,0.1615008036295573
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,768,51200,0.5249109268188477
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,768,16384,0.05282346804936727
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,768,16384,0.17342613538106283
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,1024,65536,0.20986560185750328
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,768,12288,0.04059413274129232
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,768,12288,0.13036906719207764
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,768,16384,0.06525973478953043
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,768,10240,0.03330346743265788
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,768,10240,0.10280426343282063
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,768,12288,0.05262613296508789
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,768,8192,0.026652799050013228
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,768,10240,0.046828798453013104
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,768,8192,0.07760426998138428
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,768,7168,0.023721599578857423
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,768,7168,0.06658346652984619
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,768,7168,0.03445440133412679
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,768,6144,0.02085226575533549
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,768,6144,0.05531626542409261
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,768,8192,0.03712639808654785
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,768,5120,0.017795199155807497
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,768,5120,0.047332266966501876
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,768,5120,0.029681066672007244
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,768,4096,0.014862933754920959
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,768,4096,0.041155199209849044
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,768,6144,0.03175999919573466
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,768,3584,0.01334826648235321
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,768,3584,0.03758613268534343
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,768,4096,0.026014933983484905
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,768,3072,0.011917866269747416
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,768,3072,0.03494186798731486
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,768,3584,0.024312533934911094
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,768,2560,0.010385066270828247
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,768,2560,0.030884265899658203
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,768,3072,0.02191466689109802
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,768,2048,0.008853333195050557
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,768,2048,0.02905600070953369
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,768,2048,0.018169599771499633
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,768,1536,0.007562666634718577
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,768,1536,0.026587732632954914
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,768,2560,0.02063680092493693
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,768,1024,0.006164266665776571
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,768,1024,0.02346773346265157
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,768,51200,0.16936319669087727
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,768,768,0.005551999807357788
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,768,1536,0.01616213321685791
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,768,768,0.021421867609024047
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,768,512,0.004588800172011057
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,768,512,0.019875200589497884
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,768,768,0.01458026667435964
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,768,256,0.00383146678407987
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,768,256,0.01853013237317403
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,768,512,0.014582399527231851
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,768,1024,0.015348266561826071
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,768,128,0.00340693344672521
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,768,128,0.016416000326474507
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,768,256,0.014945066968599954
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,768,64,0.003091199944416682
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,768,128,0.014917332927385965
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,768,64,0.015572266777356467
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,768,32,0.00347626656293869
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,768,32,0.016428800423940022
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,512,65536,0.1461685339609782
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,512,51200,0.11462079683939616
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,512,65536,0.6373568216959635
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,512,51200,0.505838934580485
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,512,16384,0.039046398798624676
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,512,16384,0.16981226603190105
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,512,16384,0.06338346799214681
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,512,12288,0.02949333389600118
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,512,12288,0.12728959719340008
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,768,65536,0.21011306444803873
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,512,10240,0.02369920015335083
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,512,10240,0.09731840292612712
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,512,10240,0.04461013476053874
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,512,8192,0.019093332688013713
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,512,8192,0.07133653163909912
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,512,12288,0.050621867179870605
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,512,7168,0.016243199507395424
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,512,7168,0.06028159856796265
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,512,8192,0.03505066633224487
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,512,6144,0.014500266313552857
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,512,6144,0.051069867610931394
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,512,7168,0.0330186665058136
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,512,5120,0.012115200360616047
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,512,5120,0.0433354655901591
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,512,6144,0.029938133557637532
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,512,4096,0.01051200032234192
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,512,4096,0.037198932965596516
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,512,5120,0.027718400955200194
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,512,3584,0.009264000256856282
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,512,3584,0.03536426623662313
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,512,4096,0.024338134129842124
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,512,3072,0.00872213343779246
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,512,3072,0.03290773431460063
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,512,3584,0.0222762664159139
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,512,2560,0.007546666761239369
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,512,2560,0.02974613308906555
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,512,3072,0.02029866576194763
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,512,51200,0.16766079266866046
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,512,2048,0.007057066758473713
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,512,2560,0.018651733795801796
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,512,2048,0.028572799762090047
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,512,1536,0.006190933287143707
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,512,1536,0.0251093327999115
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,512,2048,0.016532267133394875
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,512,1024,0.0046623999873797095
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,512,1024,0.022844799359639487
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,512,768,0.004445866743723551
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,512,1024,0.013318399588267008
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,512,1536,0.014590932925542196
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,512,768,0.021203200022379555
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,512,512,0.003972266614437103
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,512,512,0.019653334220250448
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,512,256,0.0033642667035261786
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,512,768,0.01283519963423411
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,512,256,0.01766080061594645
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,512,128,0.00311253344019254
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,512,256,0.012813867131868998
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,512,512,0.012993066509564718
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,512,128,0.015434666474660238
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,512,64,0.002749866743882497
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,512,128,0.012890666723251343
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,512,64,0.015594666202863058
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,512,32,0.014495999614397685
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,256,65536,0.09301973183949788
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,256,65536,0.610856564839681
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,256,51200,0.07406079769134521
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,512,32,0.003102933367093404
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,256,51200,0.48431787490844724
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,256,16384,0.029693865776062013
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,256,16384,0.1645962715148926
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,512,65536,0.20843092600504556
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,256,12288,0.0232586661974589
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,256,12288,0.12121813297271729
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,256,16384,0.06305600007375081
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,256,10240,0.01887680093447367
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,256,10240,0.09172800381978354
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,256,12288,0.05012906789779663
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,256,8192,0.014972800016403198
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,256,8192,0.06643413305282593
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,256,10240,0.0440661350886027
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,256,7168,0.013583999872207642
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,256,7168,0.055713065465291346
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,256,8192,0.034958934783935545
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,256,6144,0.011640533804893494
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,256,6144,0.047186132272084555
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,256,7168,0.032808534304300946
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,256,5120,0.010098133484522502
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,256,5120,0.04239893356959025
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,256,6144,0.029688533147176104
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,256,4096,0.008022400240103405
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,256,5120,0.02757226626078288
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,256,4096,0.035974399248758955
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,256,3584,0.007605333129564922
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,256,3584,0.03440213203430176
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,256,51200,0.16720959345499675
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,256,4096,0.024307199319203696
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,256,3072,0.006826666494210561
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,256,3072,0.03127359946568807
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,256,3584,0.021873066822687783
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,256,2560,0.00606826643149058
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,256,2560,0.028618667523066205
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,256,2048,0.005425066749254862
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,256,2048,0.02672106623649597
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,256,2560,0.01856000026067098
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,256,1536,0.004681600133577982
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,256,3072,0.019937066237131755
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,256,1536,0.024102399746576943
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,256,2048,0.016581333676973977
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,256,1024,0.003853866706291834
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,256,1024,0.020999467372894286
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,256,768,0.0034304000437259674
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,256,768,0.020614399512608846
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,256,1024,0.01321386694908142
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,256,512,0.0030559999247392017
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,256,512,0.018705066045125326
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,256,768,0.012357333302497863
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,256,1536,0.014476799964904785
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,256,256,0.002657066782315572
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,256,512,0.01241600016752879
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,256,128,0.002548266698916753
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,256,256,0.016914133230845133
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,256,256,0.01244160036245982
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,256,64,0.0026101333399613695
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,256,128,0.015132799744606018
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,256,65536,0.20822399457295737
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,256,32,0.002605866640806198
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,256,64,0.015054933230082192
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,128,65536,0.07918720245361328
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,256,32,0.014617600043614707
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,128,51200,0.06033066511154175
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,128,65536,0.602952512105306
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,128,51200,0.4791989326477051
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,128,16384,0.023937066396077476
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,128,16384,0.16236693064371746
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,256,128,0.012924800316492716
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,128,12288,0.016852267583211265
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,128,16384,0.061375999450683595
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,128,12288,0.11831040382385254
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,128,12288,0.0487125317255656
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,128,10240,0.014132266243298849
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,128,10240,0.08882133165995279
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,128,8192,0.012517333030700684
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,128,51200,0.16422293980916342
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,128,10240,0.04086079994837443
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,128,7168,0.010403199990590414
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,128,8192,0.06510719855626425
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,128,7168,0.0534666657447815
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,128,6144,0.009258666634559631
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,128,6144,0.04556586742401123
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,128,65536,0.20457173983256022
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,128,5120,0.00836906631787618
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,128,8192,0.033819735050201416
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,128,5120,0.03944746653238933
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,128,4096,0.007100800176461537
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,128,4096,0.035019731521606444
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,128,6144,0.028830933570861816
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,128,3584,0.006308266520500183
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,128,5120,0.026800000667572023
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,128,7168,0.03218666712443034
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,128,3584,0.03496319850285848
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,128,3072,0.00588266650835673
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,128,3072,0.03202773332595825
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,128,4096,0.023178666830062866
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,128,2560,0.005534933507442474
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,128,2560,0.02763413389523824
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,128,3584,0.021317332983016968
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,128,2048,0.0046304002404212955
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,128,2048,0.02680640021959941
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,128,1536,0.004244266450405121
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,128,2560,0.017785600821177163
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,128,3072,0.019692800442377725
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,128,1536,0.02290880084037781
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,128,2048,0.01569066643714905
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,128,1024,0.0035434665779272715
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,128,1024,0.021447465817133585
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,128,768,0.0033941333492596946
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,128,1536,0.01362559994061788
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,128,512,0.0030762667457262674
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,128,768,0.02027413249015808
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,128,1024,0.012446932991345723
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,128,256,0.0025770666698614756
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,128,512,0.01882560054461161
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,128,512,0.011598933736483257
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,128,128,0.0025568000972270967
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,128,256,0.011646933356920878
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,128,128,0.014713600277900696
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,128,64,0.0022314667701721193
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,128,128,0.011661866307258606
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,128,32,0.0026335999369621276
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,128,64,0.01479573349157969
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,128,32,0.014574933052062988
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,64,65536,0.07130346298217774
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,64,51200,0.05812373161315918
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,64,51200,0.4779328028361003
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,64,16384,0.022821333010991415
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,128,256,0.016862932840983072
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,64,16384,0.15959253311157226
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3328,128,768,0.011608533064524333
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,64,12288,0.1168394645055135
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,64,10240,0.01146986683209737
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,64,10240,0.08800426324208578
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,64,8192,0.011817600329717
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,64,7168,0.010973866780598958
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,64,8192,0.06227306524912516
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,64,6144,0.009139200051625569
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,64,6144,0.045208533604939774
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,64,7168,0.05323520104090372
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,64,5120,0.007924266656239827
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,64,4096,0.006764799853165944
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,64,4096,0.03489919900894165
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,64,3584,0.006291200220584869
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,64,5120,0.040057599544525146
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,64,65536,0.605781364440918
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,64,3072,0.00582826683918635
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,64,3072,0.030927999814351397
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,64,2560,0.00521066685517629
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,64,2560,0.028566400210062664
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,64,2048,0.004780800143877665
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,64,12288,0.012429866194725036
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,64,2048,0.026734934250513716
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,64,1536,0.004394666850566864
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,64,1024,0.0034272000193595886
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,64,768,0.0030858665704727173
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,64,1536,0.0240064005057017
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,64,512,0.002926933268706004
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,64,768,0.020692267020543418
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,64,256,0.002624000112215678
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,64,512,0.018747733036677042
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,64,128,0.00223786657055219
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,64,256,0.01758293310801188
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,64,64,0.002254933367172877
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,64,128,0.01464959979057312
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,64,32,0.002234666546185811
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,64,64,0.014784000317255654
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,64,32,0.013896532853444419
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,32,65536,0.07269439697265626
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,64,3584,0.03288319905598958
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,32,51200,0.05664639870325724
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,32,65536,0.6025599797566732
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,32,16384,0.01931519905726115
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,64,1024,0.0220960001150767
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,32,16384,0.1574677308400472
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,32,12288,0.010956799983978272
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,32,10240,0.009725866715113322
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,32,12288,0.11748800277709961
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,32,8192,0.008429867029190064
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,32,10240,0.08960533142089844
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,32,8192,0.06442559957504272
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,32,7168,0.008459732929865519
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,32,7168,0.05382399956385294
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,32,6144,0.046263468265533444
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,32,5120,0.00997226635615031
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,32,5120,0.039900799592336014
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,32,4096,0.008819199601809184
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,32,4096,0.03524586757024129
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,32,3584,0.007899733384450276
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,32,51200,0.47609812418619796
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,32,3584,0.033199999729792276
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,32,3072,0.0071487997968991595
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,32,2560,0.006345599889755249
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,32,3072,0.030836266279220582
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,32,2048,0.004636799792448679
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,32,2560,0.026753065983454387
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,32,1536,0.0038421332836151125
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,32,2048,0.02648959954579671
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,32,1536,0.022674133380254112
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,32,1024,0.003385599950949351
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,32,768,0.002985599885384242
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,32,1024,0.020784000555674233
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,32,768,0.02062826752662659
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,32,512,0.018573866287867228
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,32,256,0.0026634665826956432
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,32,256,0.016506666938463845
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,32,128,0.002384000023206075
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,32,64,0.002367999901374181
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,32,128,0.015000533064206442
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,32,32,0.002186666677395503
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,32,64,0.014712533354759217
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3328,32,32,0.013778133193651834
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,32,6144,0.008087466657161712
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3328,32,512,0.0026176000634829206
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,65536,16384,3.8849525451660156
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,65536,51200,6.542762756347656
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,65536,16384,2.0464757283528643
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,65536,12288,2.8998667399088545
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,65536,51200,12.537565104166667
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,65536,12288,1.4941791534423827
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,65536,10240,2.4327713012695313
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,65536,10240,1.3236160278320312
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,65536,10240,0.6971818923950195
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,65536,8192,1.8023146311442058
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,65536,8192,1.0409034729003905
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,65536,12288,0.7833632151285808
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,65536,7168,1.6221621195475262
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,65536,7168,0.9137866973876954
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,65536,16384,1.0598154703776042
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,65536,6144,1.3441738128662108
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,65536,6144,0.7837845484415691
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,65536,8192,0.5715626398722331
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,65536,5120,1.1377471923828124
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,65536,5120,0.6737450917561849
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,65536,7168,0.5277503967285156
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,65536,4096,0.8722496032714844
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,65536,4096,0.5109386761983236
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,65536,6144,0.4520682652791341
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,65536,3584,0.8141951878865561
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,65536,3584,0.4720128059387207
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,65536,4096,0.36646718978881837
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,65536,3072,0.6759594599405925
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,65536,3072,0.3944714546203613
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,65536,5120,0.41572478612263997
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,65536,2560,0.5795658747355144
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,65536,2560,0.3556672096252441
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,65536,3584,0.3305205345153809
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,65536,2048,0.4693269411722819
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,65536,2048,0.2796778678894043
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,65536,2560,0.27696746190389
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,65536,1536,0.35678399403889977
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,65536,1536,0.21371413866678873
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,65536,3072,0.2906538645426432
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,65536,1024,0.23602879842122398
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,65536,1024,0.15720213254292806
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,65536,1024,0.2048543930053711
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,65536,768,0.1840544064839681
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,65536,768,0.13214826583862305
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,65536,1536,0.21070399284362792
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,65536,512,0.13662293752034504
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,65536,512,0.11149226824442546
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,65536,768,0.19420480728149414
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,65536,256,0.0924565315246582
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,65536,256,0.09869866371154785
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,65536,2048,0.2448842684427897
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,65536,128,0.07765973409016927
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,65536,128,0.08139946460723876
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,65536,512,0.19341227213541667
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,65536,64,0.06639039913813273
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,65536,64,0.08373119831085205
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,65536,32,0.06981226603190104
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,65536,32,0.08395413557688394
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,65536,128,0.19093972841898602
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,65536,256,0.19258453051249186
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,51200,65536,6.614989725748698
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,51200,65536,12.781644694010417
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,51200,51200,9.674600219726562
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,51200,51200,5.207878621419271
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,51200,16384,3.003919982910156
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,51200,16384,1.5689695994059245
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,51200,16384,0.8357909520467123
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,51200,12288,2.225424957275391
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,51200,12288,1.1872287750244142
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,51200,12288,0.6270624160766601
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,51200,10240,1.8652384440104168
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,51200,10240,1.0193717320760092
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,51200,10240,0.5535989125569661
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,51200,8192,1.4669994354248046
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,51200,8192,0.8061407725016275
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,65536,51200,3.4918495178222657
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,51200,7168,1.1581738789876304
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,51200,7168,0.7400010426839192
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,51200,8192,0.45221652984619143
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,51200,6144,1.063878377278646
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,51200,6144,0.6292330423990886
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,51200,7168,0.4173834800720215
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,51200,5120,0.8836042404174804
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,51200,5120,0.49997866948445635
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,51200,6144,0.3574933369954427
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,51200,4096,0.731104024251302
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,51200,4096,0.44022614161173507
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,51200,5120,0.32576001485188805
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,51200,3584,0.6266442616780599
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,51200,3584,0.3659274737040202
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,51200,51200,2.695459238688151
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,51200,3072,0.5594229380289714
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,51200,3072,0.3381120045979818
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,51200,4096,0.2898634592692057
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,51200,2560,0.46759573618570965
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,51200,2560,0.29414825439453124
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,51200,3584,0.2579605261484782
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,51200,2048,0.3611200014750163
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,51200,2048,0.23507413864135743
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,51200,3072,0.22828052838643392
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,51200,1536,0.29459091822306316
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,51200,1536,0.19258774121602376
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,51200,2560,0.2188704013824463
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,51200,2048,0.19421332677205402
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,51200,1024,0.20548052787780763
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,51200,1024,0.13874239921569825
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,51200,768,0.16113279660542806
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,51200,768,0.11713066895802815
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,51200,1536,0.1691871960957845
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,51200,512,0.11777066389719645
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,51200,512,0.09481600125630697
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,51200,768,0.15525867144266764
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,51200,256,0.07690133253733317
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,51200,256,0.07789119879404703
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,51200,256,0.1541866620381673
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,51200,128,0.05677866538365682
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,51200,128,0.06487253506978354
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,51200,1024,0.16592960357666015
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,51200,64,0.04977173407872518
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,51200,64,0.06690133412679036
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,51200,32,0.052634668350219724
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,51200,32,0.06705919901529947
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,51200,512,0.15395520528157552
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,51200,128,0.15568960507710775
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,16384,65536,2.3031893412272137
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,16384,65536,3.735973358154297
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,16384,51200,1.8011039733886718
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,16384,51200,2.8396682739257812
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,16384,16384,0.9075530370076498
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,16384,16384,0.5698453267415364
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,16384,16384,0.30268799463907875
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,16384,12288,0.6417205174763997
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,16384,12288,0.46140480041503906
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,16384,12288,0.22975467046101888
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,16384,10240,0.5899893442789714
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,16384,10240,0.37865705490112306
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,16384,10240,0.20443520545959473
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,16384,8192,0.46887680689493816
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,16384,8192,0.31277653376261394
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,16384,8192,0.16762986183166503
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,16384,7168,0.4143274625142415
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,16384,7168,0.2598965326944987
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,16384,7168,0.15552959442138672
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,16384,6144,0.34362665812174475
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,16384,6144,0.233573325475057
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,16384,51200,0.9223509470621745
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,16384,5120,0.2862410545349121
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,16384,5120,0.1955466588338216
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,16384,6144,0.1342090606689453
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,16384,4096,0.2254474639892578
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,16384,4096,0.15260373751322429
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,16384,5120,0.1237877368927002
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,16384,3584,0.20082240104675292
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,16384,3584,0.1392650604248047
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,16384,3584,0.09830719629923503
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,16384,3072,0.17593919436136882
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,16384,3072,0.12010666529337566
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,16384,4096,0.10824853579203289
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,16384,2560,0.14277760187784833
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,16384,2560,0.10203839937845867
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,16384,2560,0.08179306983947754
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,16384,2048,0.11544746557871502
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,16384,2048,0.08701866467793783
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,16384,3072,0.08553813298543295
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,16384,1536,0.08868160247802734
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,16384,1536,0.06964373588562012
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,51200,65536,3.4843050638834634
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,16384,1024,0.05989973147710165
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,16384,1024,0.05787839889526367
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,16384,1024,0.05754133462905884
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,16384,768,0.04721493323644002
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,16384,768,0.051134932041168216
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,16384,1536,0.060296531518300375
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,16384,512,0.03628373146057129
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,16384,512,0.044498133659362796
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,16384,512,0.049009064833323165
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,16384,256,0.025763199726740522
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,16384,768,0.04921280145645142
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,16384,256,0.03816320101420085
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,16384,128,0.021091200908025107
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,16384,128,0.03113600015640259
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,16384,128,0.04976746638615926
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,16384,64,0.01775146722793579
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,16384,2048,0.07068053086598715
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,16384,64,0.03141653339068095
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,16384,32,0.01835626761118571
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,16384,32,0.03351680040359497
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,16384,256,0.048582398891448976
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,12288,65536,1.9042954762776692
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,12288,65536,2.8911712646484373
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,12288,51200,2.260492706298828
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,12288,51200,1.484445826212565
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,12288,16384,0.6723967870076497
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,12288,16384,0.5013440132141114
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,16384,65536,1.1536362965901694
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,12288,12288,0.5044128100077312
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,12288,12288,0.36832958857218423
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,12288,16384,0.23374080657958984
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,12288,10240,0.4482560157775879
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,12288,10240,0.33884798685709633
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,12288,12288,0.17838080724080402
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,12288,8192,0.35945920944213866
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,12288,8192,0.24058240254720054
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,12288,10240,0.15863787333170573
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,12288,7168,0.3152096112569173
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,12288,7168,0.21293867429097496
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,12288,8192,0.1301983992258708
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,12288,6144,0.2679072062174479
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,12288,6144,0.18766613006591798
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,12288,7168,0.12134186426798503
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,12288,5120,0.22159892717997232
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,12288,5120,0.15785387357076008
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,12288,6144,0.10528533458709717
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,12288,4096,0.174561071395874
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,12288,4096,0.12429760297139485
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,12288,5120,0.09658133188883464
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,12288,3584,0.15866133371988933
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,12288,3584,0.11405119895935059
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,12288,3584,0.074945068359375
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,12288,3072,0.13505600293477377
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,12288,3072,0.09686613082885742
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,12288,4096,0.08419199784596762
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,12288,2560,0.11182186603546143
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,12288,2560,0.08598399957021077
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,12288,2560,0.06407466729482016
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,12288,2048,0.08997973601023355
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,12288,3072,0.06691413720448812
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,12288,2048,0.07413547039031983
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,12288,1536,0.06866346995035807
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,12288,1536,0.060762667655944826
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,12288,1536,0.05309226512908936
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,12288,1024,0.0484661340713501
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,12288,1024,0.05130879878997803
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,12288,2048,0.057801600297292074
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,12288,768,0.03903466860453288
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,12288,768,0.04427200158437093
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,12288,1024,0.051907201608022056
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,12288,512,0.029206399122873945
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,12288,512,0.03811946709950765
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,12288,768,0.04815680185953776
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,12288,256,0.020154666900634766
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,12288,256,0.03104426662127177
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,12288,256,0.0475658655166626
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,12288,128,0.017806933323542277
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,12288,128,0.030067199468612672
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,12288,128,0.04767466783523559
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,12288,64,0.015333333611488342
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,12288,64,0.02913493315378825
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,12288,32,0.01574613352616628
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,12288,32,0.029365332921346028
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,12288,512,0.04790293375651042
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,10240,65536,1.6002452850341797
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,10240,65536,2.3306912740071613
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,10240,51200,1.8101173400878907
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,10240,51200,1.2768341064453126
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,12288,51200,0.690838368733724
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,10240,16384,0.5457898457845052
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,10240,16384,0.4275552113850911
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,10240,16384,0.20919466018676758
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,10240,12288,0.4073919932047526
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,10240,12288,0.3245173454284668
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,12288,65536,0.8683904012044271
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,10240,10240,0.33218453725179037
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,10240,10240,0.25749333699544275
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,10240,12288,0.15983786582946777
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,10240,8192,0.2777386665344238
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,10240,8192,0.2107818603515625
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,10240,10240,0.14185279210408527
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,10240,7168,0.23638827006022134
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,10240,7168,0.18021119435628255
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,10240,8192,0.11724053223927815
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,10240,6144,0.2177504062652588
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,10240,6144,0.16484373410542805
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,10240,7168,0.10864426294962566
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,10240,5120,0.17560426394144696
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,10240,5120,0.13484692573547363
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,10240,6144,0.09486400286356608
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,10240,4096,0.14302186965942382
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,10240,4096,0.11002346674601238
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,10240,5120,0.08656000296274821
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,10240,3584,0.12602880001068115
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,10240,3584,0.10039573510487873
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,10240,51200,0.615185038248698
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,10240,3072,0.10857919851938884
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,10240,3072,0.08591893513997396
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,10240,3584,0.0674506664276123
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,10240,2560,0.08870399792989095
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,10240,2560,0.07551466623942057
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,10240,4096,0.07481706937154134
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,10240,2048,0.07554026444753012
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,10240,2048,0.06568426688512166
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,10240,3072,0.06074879964192709
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,10240,1536,0.05561279853185018
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,10240,1536,0.051785600185394284
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,10240,2048,0.05363306601842245
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,10240,1024,0.03834986686706543
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,10240,2560,0.05896533330281576
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,10240,1024,0.043798398971557614
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,10240,768,0.03179200092951457
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,10240,768,0.03846826553344727
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,10240,768,0.04434773524602254
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,10240,512,0.022350933154424033
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,10240,512,0.0343018651008606
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,10240,1024,0.04763840039571126
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,10240,256,0.016768000523249307
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,10240,256,0.02921813329060872
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,10240,512,0.044536534945170084
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,10240,128,0.013639466961224875
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,10240,256,0.044248533248901364
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,10240,1536,0.04925546646118164
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,10240,64,0.011648000280062357
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,10240,128,0.04437439839045207
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,10240,32,0.011659733454386393
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,10240,64,0.023220266898473105
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,10240,32,0.025500800212224322
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,10240,128,0.024732800324757893
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,8192,65536,1.357532755533854
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,8192,65536,1.801582972208659
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,8192,51200,1.4248096466064453
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,8192,51200,1.0941749572753907
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,8192,16384,0.4139871915181478
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,8192,16384,0.36025066375732423
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,8192,16384,0.18134400049845378
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,8192,12288,0.3155914624532064
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,8192,12288,0.270087464650472
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,10240,65536,0.7754037221272786
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,8192,10240,0.2944586753845215
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,8192,10240,0.22540480295817056
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,8192,12288,0.13871893882751465
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,8192,8192,0.24029760360717772
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,8192,8192,0.180294402440389
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,8192,10240,0.12421013514200847
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,8192,7168,0.2082047939300537
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,8192,7168,0.158515199025472
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,8192,7168,0.0958570639292399
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,8192,6144,0.17737174034118652
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,8192,6144,0.13987627029418945
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,8192,8192,0.10202879905700683
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,8192,5120,0.14720212618509929
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,8192,5120,0.11936960220336915
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,8192,5120,0.07522026697794595
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,8192,4096,0.1170570691426595
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,8192,4096,0.09605226516723633
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,8192,6144,0.08319359620412191
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,8192,3584,0.10260586738586426
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,8192,3584,0.09056533177693685
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,8192,3584,0.05693759918212891
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,8192,3072,0.08841280142466226
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,8192,3072,0.0803669293721517
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,8192,4096,0.0637941320737203
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,8192,2560,0.07469546794891357
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,8192,2560,0.06927253405253092
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,8192,2560,0.04841386477152507
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,8192,2048,0.060641066233317054
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,8192,3072,0.05013866821924845
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,8192,2048,0.058422398567199704
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,8192,1536,0.047430400053660074
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,8192,1536,0.04816106557846069
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,8192,1536,0.03709333340326945
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,8192,1024,0.03330133358637492
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,8192,1024,0.039163732528686525
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,8192,2048,0.04193280140558879
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,8192,768,0.02558506727218628
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,8192,768,0.034687999884287515
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,8192,768,0.03089066743850708
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,8192,512,0.018652800718943277
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,8192,512,0.03110613425572713
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,8192,512,0.030868266026179
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,8192,256,0.014470400412877402
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,8192,256,0.027242666482925414
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,8192,1024,0.03545813163121541
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,8192,128,0.011706667145093282
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,8192,128,0.02445866664250692
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,8192,256,0.030297599236170453
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,8192,64,0.00995733340581258
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,8192,64,0.024546132485071818
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,8192,32,0.010016000270843506
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,8192,32,0.022741333643595377
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,8192,128,0.031677865982055665
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,7168,65536,1.3098720550537108
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,7168,65536,1.6229024251302082
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,7168,51200,1.29203732808431
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,7168,51200,1.0427775700887045
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,8192,51200,0.5366602579752604
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,7168,16384,0.3866101264953613
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,7168,16384,0.34449494679768883
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,7168,16384,0.15892267227172852
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,7168,12288,0.29272212982177737
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,7168,12288,0.24937920570373534
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,8192,65536,0.6797568003336589
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,7168,10240,0.2679445266723633
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,7168,10240,0.2069397290547689
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,7168,12288,0.12225813070933025
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,7168,8192,0.20346666971842448
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,7168,8192,0.1584565321604411
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,7168,10240,0.11013973553975423
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,7168,7168,0.1820031960805257
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,7168,7168,0.148525873819987
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,7168,8192,0.09059092998504639
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,7168,6144,0.156330664952596
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,7168,6144,0.12889066537221272
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,7168,7168,0.08510933717091879
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,7168,5120,0.12942720254262288
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,7168,5120,0.10922773679097493
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,7168,6144,0.07388479709625244
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,7168,4096,0.10206720034281414
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,7168,4096,0.08735360304514567
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,7168,5120,0.06742933591206869
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,7168,3584,0.09068160057067871
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,7168,3584,0.0817525307337443
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,7168,3584,0.05185279846191406
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,7168,3072,0.07954986890157065
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,7168,4096,0.05726079940795899
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,7168,3072,0.07080000241597494
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,7168,2560,0.06341973145802816
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,7168,2560,0.061626664797465
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,7168,2560,0.04458133379618327
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,7168,2048,0.05529919862747192
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,7168,2048,0.05144319931666056
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,7168,3072,0.04573546648025513
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,7168,1536,0.04269760052363078
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,7168,1536,0.041971198717753094
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,7168,1536,0.03569386800130208
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,7168,1024,0.0265610675017039
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,7168,1024,0.035903998215993244
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,7168,2048,0.040106666088104245
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,7168,768,0.023492266734441124
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,7168,768,0.03255253235499064
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,7168,768,0.031850665807724
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,7168,512,0.016275200247764587
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,7168,512,0.028600533803304035
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,7168,1024,0.03463360071182251
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,7168,256,0.01204159955183665
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,7168,256,0.024774400393168132
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,7168,256,0.03179306586583455
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,7168,128,0.010396800438563029
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,7168,128,0.02295573353767395
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,7168,512,0.03174826701482137
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,7168,64,0.009190400441487629
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,7168,128,0.031804800033569336
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,7168,32,0.00940053363641103
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,7168,64,0.021245867013931274
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,7168,32,0.02358400026957194
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,6144,65536,1.1684064229329427
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,6144,65536,1.4059925079345703
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,6144,51200,1.0515701293945312
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,6144,51200,0.939848518371582
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,7168,51200,0.4644447962443034
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,6144,16384,0.3290015856424967
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,6144,16384,0.3010282516479492
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,6144,16384,0.14178239504496257
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,6144,12288,0.24944960276285805
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,6144,12288,0.22133226394653321
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,7168,65536,0.5849760055541993
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,6144,10240,0.22178026835123696
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,6144,10240,0.1877845287322998
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,6144,12288,0.10763946374257405
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,6144,8192,0.17962239583333334
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,6144,8192,0.14751893679300945
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,6144,10240,0.0960927963256836
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,6144,7168,0.1566677411397298
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,6144,7168,0.13092479705810547
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,6144,8192,0.07882346312204996
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,6144,6144,0.135973326365153
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,6144,6144,0.11442240079243977
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,6144,7168,0.07545813719431559
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,6144,5120,0.11255679925282795
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,6144,5120,0.10002559820810955
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,6144,6144,0.06602453390757243
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,6144,4096,0.08895466327667237
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,6144,4096,0.07975146770477295
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,6144,5120,0.05974506537119547
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,6144,3584,0.07662826379140218
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,6144,3584,0.0750485340754191
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,6144,4096,0.050543999671936034
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,6144,3072,0.06691839694976806
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,6144,3072,0.06478186845779418
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,6144,3584,0.04540266593297322
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,6144,2560,0.05619946718215942
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,6144,2560,0.057302399476369226
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,6144,3072,0.040439466635386154
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,6144,2048,0.045509334405263266
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,6144,2048,0.04772479931513469
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,6144,2560,0.03885973294576009
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,6144,1536,0.03448959986368815
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,6144,1536,0.0404309352238973
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,6144,2048,0.03460906744003296
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,6144,1024,0.024090667565663658
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,6144,1024,0.034456535180409746
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,6144,1536,0.031231999397277832
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,6144,768,0.019517866770426433
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,6144,768,0.03136639992396037
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,6144,1024,0.030410667260487873
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,6144,512,0.01453013320763906
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,6144,512,0.027028266588846845
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,6144,768,0.02808000048001607
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,6144,256,0.010820266604423524
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,6144,256,0.023180800676345825
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,6144,512,0.02767146627108256
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,6144,128,0.009553066889444987
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,6144,256,0.02765226761500041
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,6144,128,0.02173759937286377
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,6144,64,0.008339200417200725
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,6144,64,0.02100480000178019
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,6144,32,0.008755200107892354
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,6144,32,0.021527467171351115
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,6144,128,0.02759573260943095
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,5120,65536,1.1189717610677083
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,6144,51200,0.40552533467610674
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,5120,65536,1.0596138636271157
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,5120,51200,0.8478783925374349
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,5120,51200,0.8354421615600586
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,5120,16384,0.2664586702982584
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,5120,16384,0.27469011942545574
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,5120,16384,0.1375658671061198
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,5120,12288,0.20019413630167643
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,5120,12288,0.1972981293996175
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,6144,65536,0.5076661427815755
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,5120,10240,0.1654314676920573
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,5120,10240,0.16330773035685223
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,5120,12288,0.10445226828257244
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,5120,8192,0.13527466456095377
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,5120,8192,0.13034559885660807
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,5120,10240,0.09410239855448405
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,5120,7168,0.11784106890360516
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,5120,7168,0.1175317366917928
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,5120,8192,0.07811840375264487
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,5120,6144,0.10254826545715331
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,5120,6144,0.1054357369740804
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,5120,7168,0.07389333248138427
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,5120,5120,0.08542613188425699
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,5120,5120,0.09228266874949137
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,5120,6144,0.06462613344192505
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,5120,4096,0.07025067011515299
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,5120,4096,0.07529493172963461
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,5120,5120,0.05780800183614095
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,5120,3584,0.06125760078430176
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,5120,3584,0.06821440060933431
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,5120,4096,0.0499125321706136
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,5120,3072,0.05400853157043457
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,5120,3072,0.0582751989364624
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,5120,3584,0.04630293448766072
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,5120,2560,0.045330135027567546
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,5120,2560,0.05355199972788492
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,5120,2560,0.04025919834772746
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,5120,2048,0.03762453397115072
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,5120,2048,0.04575893481572469
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,5120,3072,0.04180479844411214
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,5120,1536,0.02865920066833496
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,5120,1536,0.037970133622487384
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,5120,2048,0.036408531665802005
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,5120,1024,0.0211082657178243
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,5120,1024,0.031137067079544067
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,5120,1536,0.032935466368993124
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,5120,768,0.016979199647903443
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,5120,1024,0.0316266675790151
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,5120,768,0.031060266494750976
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,5120,512,0.01241386632124583
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,5120,512,0.028632533550262452
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,5120,51200,0.40106665293375654
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,5120,256,0.011681066950162251
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,5120,256,0.024807467063268026
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,5120,256,0.029281065861384077
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,5120,128,0.00780266672372818
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,5120,512,0.029803733030954998
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,5120,128,0.020258132616678873
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,5120,64,0.0067093332608540845
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,5120,64,0.020489599307378134
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,5120,32,0.007102933526039123
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,5120,32,0.02076586683591207
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,5120,128,0.02943893273671468
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,5120,768,0.02987733284632365
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,4096,65536,0.8763359705607096
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,4096,65536,0.9534581502278646
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,4096,51200,0.7284159978230794
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,4096,51200,0.7514336268107097
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,4096,16384,0.21346027056376138
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,4096,16384,0.2453162670135498
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,4096,16384,0.11050453186035156
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,4096,12288,0.16164053281148275
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,4096,12288,0.17723840077718098
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,5120,65536,0.504197343190511
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,4096,10240,0.15068799654642742
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,4096,10240,0.15154773394266766
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,4096,12288,0.08579626878102621
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,4096,8192,0.11926080385843914
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,4096,8192,0.12029759883880616
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,4096,10240,0.07772586345672608
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,4096,7168,0.10548266569773357
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,4096,7168,0.10805439949035645
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,4096,8192,0.06427520116170247
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,4096,6144,0.08847040335337321
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,4096,6144,0.09667519728342691
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,4096,7168,0.05980053345362345
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,4096,5120,0.074125870068868
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,4096,5120,0.083078400293986
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,4096,5120,0.04736959934234619
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,4096,6144,0.052585601806640625
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,4096,4096,0.05987199942270914
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,4096,4096,0.06548906564712524
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,4096,3584,0.05271253188451132
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,4096,3584,0.060652800401051844
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,4096,3584,0.036294400691986084
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,4096,3072,0.04587946732838948
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,4096,3072,0.05226133267084757
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,4096,4096,0.04046613375345866
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,4096,2560,0.038729600111643475
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,4096,2560,0.047192533810933426
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,4096,3072,0.03221653302510579
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,4096,2048,0.031395200888315836
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,4096,2048,0.041282133261362715
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,4096,2560,0.030877866347630817
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,4096,1536,0.024769065777460735
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,4096,1536,0.03483839829762776
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,4096,1536,0.023082667589187623
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,4096,1024,0.018407466014226277
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,4096,1024,0.029151999950408937
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,4096,2048,0.02683626612027486
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,4096,768,0.014407466848691305
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,4096,768,0.027669332424799603
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,4096,1024,0.022270933787027995
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,4096,512,0.010834133625030518
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,4096,512,0.0256223996480306
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,4096,768,0.023085866371790567
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,4096,256,0.009649067123730978
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,4096,256,0.02277440031369527
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,4096,256,0.022728532552719116
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,4096,512,0.023143466313680014
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,4096,128,0.006759466727574666
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,4096,128,0.01861013372739156
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,4096,64,0.005919999877611796
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,4096,64,0.018626133600870766
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,4096,32,0.006326400240262349
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,4096,128,0.020679465929667153
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,4096,32,0.018636800845464072
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,3584,65536,0.7765631993611654
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,3584,65536,0.9281034469604492
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,3584,51200,0.6228736241658528
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,3584,51200,0.7346719741821289
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,4096,51200,0.31811841328938806
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,3584,16384,0.20402879714965821
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,3584,16384,0.23881600697835287
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,3584,16384,0.11158933639526367
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,3584,12288,0.15382933616638184
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,3584,12288,0.17361706097920734
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,4096,65536,0.39849812189737954
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,3584,10240,0.130404265721639
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,3584,10240,0.14018559455871582
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,3584,12288,0.0871445337931315
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,3584,8192,0.10550613403320312
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,3584,8192,0.11139413515726726
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,3584,10240,0.07909440199534098
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,3584,7168,0.08946773211161295
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,3584,7168,0.10040533542633057
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,3584,8192,0.06607679923375448
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,3584,6144,0.0771071990331014
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,3584,6144,0.08893226782480876
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,3584,7168,0.060994132359822595
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,3584,5120,0.06309546629587809
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,3584,5120,0.07671573162078857
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,3584,5120,0.04855999946594238
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,3584,4096,0.05061653455098471
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,3584,4096,0.06013973156611124
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,3584,6144,0.0530784010887146
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,3584,3584,0.04548906485239665
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,3584,3584,0.05488213300704956
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,3584,4096,0.04181119998296102
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,3584,3072,0.03907306591669719
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,3584,3072,0.049159467220306396
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,3584,3584,0.03844266732533773
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,3584,2560,0.03327999909718831
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,3584,2560,0.043714133898417155
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,3584,3072,0.034764798482259114
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,3584,2560,0.03301546573638916
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,3584,2048,0.02718720038731893
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,3584,2048,0.03873279889424642
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,3584,1536,0.021740800142288207
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,3584,1536,0.0328981339931488
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,3584,1536,0.026837333043416338
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,3584,1024,0.015531733632087708
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,3584,1024,0.028853332996368407
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,3584,2048,0.029742934306462604
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,3584,768,0.011735467116038005
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,3584,768,0.02680319945017497
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,3584,51200,0.31933120091756184
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,3584,512,0.00957546631495158
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,3584,1024,0.02601813276608785
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,3584,512,0.02330026626586914
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,3584,256,0.007548800110816956
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,3584,256,0.021180800596872964
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,3584,256,0.02355626622835795
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,3584,128,0.007101866602897644
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,3584,128,0.019321600596110024
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,3584,512,0.023779199520746867
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,3584,64,0.006670933465162914
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,3584,64,0.018569600582122803
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,3584,32,0.007090133428573608
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,3584,128,0.023944532871246337
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,3584,32,0.019859200716018675
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,3584,768,0.0240447998046875
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,3072,65536,0.6333717346191406
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,3072,65536,0.851691754659017
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,3072,51200,0.5538357416788737
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,3072,51200,0.6668170928955078
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,3072,16384,0.1599733352661133
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,3072,16384,0.21611199378967286
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,3072,16384,0.089246932665507
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,3072,12288,0.12217600345611572
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,3072,12288,0.1579541365305583
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,3584,65536,0.4012127876281738
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,3072,10240,0.11280106703440348
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,3072,10240,0.1351093292236328
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,3072,12288,0.07101866404215494
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,3072,8192,0.09251519838968912
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,3072,8192,0.10649387041727702
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,3072,10240,0.06517333189646403
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,3072,7168,0.07971733411153158
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,3072,7168,0.09512320359547934
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,3072,8192,0.053540265560150145
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,3072,6144,0.06897599697113037
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,3072,6144,0.08507839838663736
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,3072,7168,0.0494485338528951
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,3072,5120,0.056075731913248696
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,3072,5120,0.07155199845631918
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,3072,6144,0.04421759843826294
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,3072,4096,0.045798401037851974
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,3072,4096,0.056201601028442384
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,3072,5120,0.04009600083033244
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,3072,3584,0.039827199776967366
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,3072,3584,0.052142934004465735
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,3072,4096,0.034679468472798666
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,3072,3072,0.035545599460601804
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,3072,3072,0.04723306496938069
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,3072,3584,0.03177599906921387
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,3072,2560,0.029509333769480388
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,3072,2560,0.04155626694361369
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,3072,3072,0.02910826603571574
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,3072,2048,0.024964267015457155
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,3072,2048,0.03691946665445964
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,3072,2560,0.02738879919052124
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,3072,1536,0.019845332702000937
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,3072,1536,0.031464533011118574
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,3072,1536,0.02226026654243469
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,3072,2048,0.024375466505686443
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,3072,1024,0.013182933131853739
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,3072,1024,0.028475733598073323
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,3072,768,0.010790399710337321
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,3072,768,0.0247978667418162
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,3072,768,0.01984000007311503
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,3072,512,0.008416000008583068
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,3072,512,0.02263360023498535
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,3072,1024,0.021362133820851645
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,3072,256,0.0062613333264986675
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,3072,256,0.020563199122746786
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,3072,256,0.01944213310877482
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,3072,512,0.019483733177185058
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,3072,128,0.005909333129723867
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,3072,128,0.018849066893259683
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,3072,128,0.019427200158437095
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,3072,64,0.005494399865468343
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,3072,64,0.018308266003926595
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,3072,32,0.005523199836413065
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,3072,32,0.01796906590461731
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,2560,65536,0.5249909400939942
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,3072,51200,0.2431626637776693
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,2560,65536,0.7955071767171223
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,2560,51200,0.40929492314656574
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,2560,51200,0.626207987467448
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,2560,16384,0.133076270421346
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,2560,16384,0.2038368066151937
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,2560,16384,0.08720746835072836
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,2560,12288,0.10115946928660076
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,2560,12288,0.14770347277323406
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,3072,65536,0.3025898615519206
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,2560,10240,0.08379093011220297
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,2560,10240,0.12260053157806397
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,2560,12288,0.06917760372161866
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,2560,8192,0.06884693304697673
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,2560,8192,0.09762880007425943
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,2560,10240,0.06342399915059407
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,2560,7168,0.06027413209279379
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,2560,7168,0.08767146269480387
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,2560,8192,0.05219626824061076
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,2560,6144,0.052602668603261314
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,2560,6144,0.07627200285593669
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,2560,7168,0.04789760112762451
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,2560,5120,0.04385493199030559
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,2560,5120,0.06543573141098022
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,2560,6144,0.04357440074284871
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,2560,4096,0.03637866576512654
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,2560,4096,0.051268267631530764
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,2560,5120,0.03941440184911092
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,2560,3584,0.032075732946395874
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,2560,3584,0.04851093292236328
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,2560,3584,0.03094826738039653
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,2560,4096,0.033769599596659344
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,2560,3072,0.028165332476298016
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,2560,3072,0.04277546803156535
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,2560,2560,0.024680533011754355
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,2560,2560,0.03911573489507039
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,2560,2560,0.02637973427772522
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,2560,2048,0.02071040074030558
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,2560,2048,0.034780800342559814
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,2560,3072,0.027191466093063353
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,2560,1536,0.01564586659272512
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,2560,1536,0.031057065725326537
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,2560,2048,0.023127466440200806
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,2560,1024,0.011666133006413778
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,2560,1024,0.02768320043881734
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,2560,1536,0.020753065745035805
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,2560,1024,0.01988160014152527
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,2560,768,0.009689600268999735
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,2560,768,0.025538132588068647
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,2560,512,0.009071999788284301
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,2560,512,0.022754132747650146
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,2560,512,0.018268799781799315
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,2560,256,0.00726506660381953
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,2560,768,0.018322134017944337
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,2560,256,0.02103466590245565
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,2560,128,0.0050687998533248905
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,2560,128,0.01808746655782064
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,2560,256,0.018244266510009766
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,2560,64,0.004735999802748362
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,2560,128,0.018177066246668497
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,2560,32,0.004810666541258494
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,2560,64,0.017332265774408974
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,2560,32,0.01658560037612915
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,2048,65536,0.41388479868570965
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,2048,65536,0.7391722361246745
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,2048,51200,0.3339765230814616
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,2048,51200,0.581551996866862
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,2560,51200,0.24019947052001953
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,2048,16384,0.10631253719329833
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,2048,16384,0.18954346974690756
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,2048,16384,0.08595200379689535
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,2048,12288,0.0806943972905477
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,2048,12288,0.13876372973124187
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,2560,65536,0.30020373662312827
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,2048,10240,0.06877760092417398
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,2048,10240,0.11410133043924968
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,2048,12288,0.06795946756998697
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,2048,8192,0.054890668392181395
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,2048,8192,0.08905173142751058
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,2048,10240,0.061774933338165285
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,2048,7168,0.048750933011372885
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,2048,7168,0.0785973310470581
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,2048,8192,0.049643735090891525
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,2048,6144,0.04160746733347575
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,2048,6144,0.06762026945749919
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,2048,7168,0.04574933449427287
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,2048,5120,0.03523840109507243
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,2048,5120,0.05711679855982462
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,2048,5120,0.03711466789245606
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,2048,4096,0.029191466172536214
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,2048,4096,0.047378134727478025
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,2048,6144,0.04162559906641643
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,2048,3584,0.02603626648585002
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,2048,3584,0.0462442676226298
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,2048,4096,0.03224853277206421
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,2048,3072,0.022690133253733317
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,2048,3072,0.0412224014600118
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,2048,3584,0.02930026650428772
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,2048,2560,0.019853866100311278
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,2048,2560,0.036637866497039796
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,2048,3072,0.02640746633211772
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,2048,2048,0.01725013256072998
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,2048,2048,0.0327839990456899
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,2048,2560,0.024685867627461753
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,2048,1536,0.0132832000652949
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,2048,1536,0.029174399375915528
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,2048,1536,0.01862506667772929
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,2048,2048,0.021473066012064616
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,2048,1024,0.026817067464192705
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,2048,1024,0.010033067067464192
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,2048,768,0.008489599823951722
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,2048,768,0.02425280014673869
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,2048,1024,0.017992534240086875
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,2048,512,0.0075541332364082335
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,2048,512,0.022299732764561972
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,2048,512,0.016198399662971496
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,2048,768,0.016158933440844216
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,2048,256,0.00592853327592214
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,2048,256,0.01977919936180115
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,2048,128,0.00462719996770223
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,2048,128,0.017065600554148356
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,2048,256,0.01618346671263377
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,2048,64,0.004264533519744873
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,2048,128,0.016135467092196147
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,2048,64,0.01653333306312561
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,2048,32,0.004296533266703288
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,2048,32,0.016614400347073875
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,2048,51200,0.24065492947896322
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,1536,65536,0.3423765182495117
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,1536,51200,0.2678528149922689
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,1536,65536,0.6913205464680989
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,1536,51200,0.5685834884643555
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,1536,16384,0.08852266470591227
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,1536,16384,0.18000853856404622
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,1536,16384,0.0657482663790385
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,1536,12288,0.06749119758605956
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,1536,12288,0.13025920391082763
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,2048,65536,0.3024042765299479
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,1536,10240,0.0569546659787496
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,1536,10240,0.10804693698883057
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,1536,12288,0.05312639872233073
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,1536,8192,0.046012798945109054
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,1536,8192,0.08220160007476807
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,1536,10240,0.04789653221766154
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,1536,7168,0.03973226547241211
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,1536,7168,0.07266133626302083
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,1536,8192,0.03944000005722046
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,1536,6144,0.03388479948043823
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,1536,6144,0.06250346501668294
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,1536,7168,0.03640213410059611
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,1536,5120,0.028295467297236126
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,1536,5120,0.05321280161539713
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,1536,6144,0.033048532406489056
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,1536,4096,0.023077332973480226
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,1536,4096,0.0433514674504598
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,1536,5120,0.03099520007769267
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,1536,3584,0.021117866039276123
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,1536,3584,0.04138666788736979
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,1536,4096,0.027281065781911213
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,1536,3072,0.01855573256810506
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,1536,3072,0.03699839909871419
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,1536,3584,0.02518506646156311
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,1536,2560,0.01620693306128184
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,1536,2560,0.03471253315607707
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,1536,3072,0.023496532440185548
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,1536,2048,0.013651200135548911
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,1536,2048,0.030744532744089764
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,1536,2560,0.02167146603266398
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,1536,2048,0.019883733987808228
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,1536,1536,0.011013333002726238
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,1536,1536,0.02767360011736552
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,1536,1024,0.007969066500663757
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,1536,1024,0.024733867247899374
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,1536,1024,0.016571733355522155
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,1536,768,0.006939733525117238
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,1536,768,0.022792534033457438
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,1536,51200,0.1688800017038981
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,1536,1536,0.017767467101415
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,1536,512,0.005913599828879039
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,1536,512,0.021237333615620933
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,1536,768,0.015708800156911215
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,1536,256,0.0047189335028330484
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,1536,256,0.018776534001032512
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,1536,128,0.004187733431657155
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,1536,128,0.016810667514801026
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,1536,512,0.015740799903869628
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,1536,64,0.0038463999827702843
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,1536,128,0.0169322669506073
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,1536,64,0.016561067104339598
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,1536,256,0.015659733613332113
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,1536,32,0.016702934106191
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,1024,65536,0.24466026624043785
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,1024,65536,0.6502773284912109
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,1536,32,0.0038773333032925926
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,1024,51200,0.18638827006022135
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,1024,51200,0.5165919939676921
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,1024,16384,0.059623467922210696
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,1024,16384,0.16651520729064942
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,1024,16384,0.06318613290786743
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,1024,12288,0.04528106848398845
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,1024,12288,0.12120426495869954
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,1536,65536,0.20968106587727867
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,1024,10240,0.04099306662877401
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,1024,10240,0.09928960005442301
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,1024,12288,0.0504693349202474
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,1024,8192,0.03067306677500407
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,1024,8192,0.0741055965423584
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,1024,10240,0.04438399871190389
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,1024,7168,0.027636265754699706
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,1024,7168,0.06346133152643839
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,1024,7168,0.03301333387692769
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,1024,6144,0.024104533592859904
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,1024,6144,0.05601066748301188
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,1024,8192,0.035877335071563723
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,1024,5120,0.020406399170557657
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,1024,5120,0.04773759841918945
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,1024,5120,0.028085333108901978
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,1024,4096,0.01712426741917928
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,1024,4096,0.040278398990631105
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,1024,6144,0.029959466060002642
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,1024,3584,0.01581760048866272
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,1024,3584,0.03793280124664307
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,1024,4096,0.024778666098912557
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,1024,3072,0.013598933815956116
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,1024,3072,0.034984532992045084
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,1024,3584,0.022243199745814006
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,1024,2560,0.011635200182596842
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,1024,2560,0.031592533985773726
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,1024,3072,0.020426666736602782
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,1024,2048,0.010062932968139648
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,1024,2048,0.02900480031967163
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,1024,2560,0.018990933895111084
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,1024,1536,0.008371200164159138
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,1024,1536,0.025293866793314617
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,1024,2048,0.01691840092341105
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,1024,1024,0.0067775999506314594
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,1024,1536,0.01447466711203257
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,1024,1024,0.024182399113972984
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,1024,768,0.00591893345117569
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,1024,768,0.022755199670791627
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,1024,51200,0.16521600087483723
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,1024,512,0.005042133231957754
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,1024,512,0.021014400323232017
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,1024,768,0.012989866733551025
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,1024,256,0.0046186665693918865
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,1024,256,0.018542933464050292
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,1024,512,0.013184000054995218
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,1024,128,0.0037994667887687682
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,1024,128,0.017402666807174682
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,1024,256,0.012859732906023661
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,1024,64,0.003409066547950109
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,1024,64,0.01651413341363271
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,1024,128,0.013688533504803976
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,1024,32,0.003487999985615412
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,1024,32,0.015757866700490317
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,768,65536,0.185915740331014
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,1024,1024,0.014111999670664468
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,768,51200,0.14822079340616862
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,768,65536,0.6288170496622721
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,768,51200,0.5019296010335286
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,768,16384,0.04786026477813721
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,768,16384,0.15869332949320475
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,768,16384,0.06377919912338256
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,768,12288,0.036245334148406985
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,768,12288,0.11405759652455646
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,1024,65536,0.20599786440531412
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,768,10240,0.02983466585477193
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,768,10240,0.0922869364420573
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,768,12288,0.05129493474960327
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,768,8192,0.023027199506759643
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,768,8192,0.06760426362355551
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,768,10240,0.044742401440938315
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,768,7168,0.02000853419303894
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,768,7168,0.058628265062967935
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,768,8192,0.0362773338953654
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,768,7168,0.03417493502298991
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,768,6144,0.01742080052693685
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,768,6144,0.050330666700998936
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,768,5120,0.014886400103569031
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,768,5120,0.044640000661214194
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,768,6144,0.03136853377024333
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,768,4096,0.012743467092514038
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,768,4096,0.03713493347167969
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,768,5120,0.029312000672022505
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,768,3584,0.01092906693617503
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,768,3584,0.036380799611409505
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,768,4096,0.025560534000396727
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,768,3072,0.010115200281143188
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,768,3072,0.03277440071105957
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,768,3584,0.02387946645418803
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,768,2560,0.008519466718037922
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,768,3072,0.021491199731826782
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,768,2560,0.030666667222976684
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,768,2560,0.020166399081548055
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,768,2048,0.007613866527875264
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,768,2048,0.02839786609013875
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,768,1536,0.006730666756629944
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,768,1536,0.025421865781148273
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,768,51200,0.16616746584574382
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,768,1024,0.005127466718355815
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,768,2048,0.01775040030479431
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,768,1024,0.022428800662358604
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,768,768,0.004660266637802124
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,768,768,0.02058773239453634
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,768,1024,0.015217066804567973
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,768,512,0.0038293334345022834
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,768,768,0.014479999740918478
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,768,512,0.018705066045125326
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,768,256,0.0033887999753157297
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,768,1536,0.01606933375199636
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,768,256,0.01758506695429484
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,768,128,0.0030570665995279947
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,768,512,0.014454399545987448
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,768,128,0.015923200050989787
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,768,64,0.002977066735426585
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,768,128,0.014472533265749613
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,768,64,0.015290666619936624
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,768,32,0.003036800026893616
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,768,256,0.014983466267585755
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,768,32,0.014685866236686707
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,512,65536,0.16665172576904297
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,512,51200,0.13124053478240966
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,512,65536,0.5893376032511394
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,512,51200,0.46833171844482424
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,512,16384,0.045124268531799315
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,512,16384,0.1588223934173584
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,768,65536,0.20647892951965333
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,512,12288,0.03408533334732056
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,512,12288,0.11404906908671061
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,512,16384,0.06206826766331991
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,512,10240,0.02656853397687276
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,512,10240,0.08984746932983398
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,512,10240,0.04245973428090413
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,512,8192,0.021144533157348634
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,512,8192,0.06460053523381551
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,512,12288,0.04996266762415568
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,512,7168,0.018242132663726807
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,512,7168,0.05603839953740438
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,512,7168,0.03247893253962199
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,512,6144,0.016039466857910155
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,512,8192,0.03456639846165975
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,512,6144,0.04923733472824097
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,512,5120,0.01388159990310669
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,512,5120,0.04324800173441569
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,512,5120,0.02723840077718099
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,512,4096,0.011264000336329143
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,512,6144,0.028875732421875
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,512,4096,0.03703253269195557
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,512,3584,0.010386133193969726
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,512,3584,0.03503893216451009
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,512,4096,0.02395520011583964
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,512,3584,0.02214720050493876
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,512,3072,0.009553066889444987
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,512,3072,0.03288319905598958
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,512,2560,0.008251733581225077
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,512,2560,0.030039467414220172
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,512,3072,0.01948480010032654
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,512,2048,0.007123200098673503
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,512,2560,0.018643200397491455
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,512,2048,0.026811732848485308
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,512,1536,0.006296533346176148
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,512,1536,0.02395840088526408
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,512,51200,0.16475200653076172
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,512,2048,0.0165802667538325
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,512,1024,0.005096533397833506
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,512,1536,0.014524799585342408
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,512,1024,0.022731733322143555
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,512,768,0.004638933142026265
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,512,768,0.020823466777801513
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,512,512,0.004129066566626231
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,512,512,0.0190720001856486
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,512,768,0.012549333771069846
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,512,256,0.0034143999218940735
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,512,256,0.01693120002746582
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,512,512,0.01285866697629293
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,512,128,0.0030261332790056865
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,512,1024,0.01393066644668579
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,512,128,0.01675093372662862
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,512,64,0.003005866706371307
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,512,64,0.01479680041472117
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,512,32,0.0029813334345817565
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,512,32,0.014827733238538107
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,512,128,0.012731732924779257
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,256,65536,0.096888534228007
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,256,65536,0.5656927744547526
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,256,51200,0.07731413046518962
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,512,256,0.012423466642697651
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,256,51200,0.45078506469726565
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,256,16384,0.031147734324137373
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,256,16384,0.15174506505330404
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,256,16384,0.0621450662612915
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,512,65536,0.20488319396972657
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,256,12288,0.021835732460021972
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,256,12288,0.10813759962717692
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,256,12288,0.048293332258860275
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,256,10240,0.018950400749842326
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,256,10240,0.08454399903615316
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,256,8192,0.017077332735061644
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,256,8192,0.06248106559117635
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,256,8192,0.03379093408584595
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,256,7168,0.015191466609636942
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,256,10240,0.041229867935180665
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,256,7168,0.050436266263326014
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,256,51200,0.16398293177286785
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,256,6144,0.01341759959856669
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,256,6144,0.045075198014577225
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,256,5120,0.012165333827336628
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,256,5120,0.039420799414316816
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,256,6144,0.029240532716115313
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,256,4096,0.008547199765841167
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,256,5120,0.027166932821273804
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,256,7168,0.03217173417409261
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,256,4096,0.034723198413848876
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,256,3584,0.007547733187675476
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,256,3584,0.0329258660475413
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,256,3072,0.006670933465162914
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,256,3072,0.029732267061869305
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,256,4096,0.023537067572275798
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,256,2560,0.0059114664793014525
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,256,65536,0.20445013046264648
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,256,3072,0.01943999926249186
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,256,2560,0.02797973354657491
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,256,2048,0.00572266678015391
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,256,2048,0.026446932554244997
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,256,1536,0.004667733112970988
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,256,1536,0.022797866662343343
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,256,3584,0.02153279980023702
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,256,1024,0.003945599993069967
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,256,2048,0.016123732924461363
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,256,2560,0.018631466229756675
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,256,768,0.0035071998834609987
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,256,1024,0.021492266654968263
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,256,1536,0.014452266693115234
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,256,512,0.003033600002527237
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,256,768,0.021444267034530638
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,256,1024,0.013244799772898355
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,256,512,0.019057067235310872
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,256,256,0.0026335999369621276
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,256,256,0.016638933618863424
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,256,512,0.012427733341852824
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,256,128,0.00258240004380544
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,256,256,0.012060800194740295
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,256,128,0.014896000425020853
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,256,64,0.002605866640806198
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,256,32,0.0025898667673269907
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,256,128,0.012428800264994305
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,256,64,0.014693333705266317
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,256,32,0.016028799613316855
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,128,65536,0.0680021365483602
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,128,51200,0.05575466553370158
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,128,65536,0.5607402801513672
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,128,51200,0.4438666661580403
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,128,16384,0.02218986749649048
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,128,16384,0.14922666549682617
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,256,768,0.012743467092514038
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,128,12288,0.015176533659299215
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,128,12288,0.10317013263702393
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,128,16384,0.060619731744130455
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,128,10240,0.014518400033315023
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,128,10240,0.08068586985270182
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,128,12288,0.04705280065536499
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,128,8192,0.011849600076675414
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,128,8192,0.057202132542928066
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,128,10240,0.03987306753794352
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,128,7168,0.010657067100207012
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,128,7168,0.048681600888570147
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,128,7168,0.03094293276468913
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,128,6144,0.00920746624469757
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,128,8192,0.032849067449569704
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,128,51200,0.160697603225708
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,128,5120,0.00837546686331431
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,128,5120,0.03908053239186605
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,128,65536,0.20049386024475097
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,128,5120,0.026336000363032026
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,128,4096,0.0071381335457166035
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,128,4096,0.033539199829101564
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,128,3584,0.006708266834417979
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,128,6144,0.028546132644017536
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,128,3584,0.033003733555475874
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,128,3072,0.006213333209355672
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,128,4096,0.022770132621129355
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,128,6144,0.043986133734385174
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,128,3072,0.02985600034395854
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,128,2560,0.0054848000407218935
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,128,2560,0.02878613273302714
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,128,2048,0.004706133405367533
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,128,2048,0.02595626711845398
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,128,3072,0.018978132804234823
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,128,2560,0.017756799856821694
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,128,3584,0.02063039938608805
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,128,1536,0.022825600703557332
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,128,2048,0.01548906664053599
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,128,1024,0.0034400001168251038
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,128,1024,0.02146773338317871
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,128,1536,0.013673599561055502
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,128,768,0.0034304000437259674
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,128,1024,0.012566399574279786
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,128,768,0.02065066695213318
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,128,512,0.003014400104681651
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,128,768,0.011655466755231221
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,128,512,0.01880000034968058
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,128,512,0.011453866958618164
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,128,256,0.002733866622050603
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,128,256,0.017129600048065186
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,128,1536,0.0042357335488001505
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,128,128,0.002567466596762339
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,128,256,0.011601066589355469
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,128,64,0.0025290665527184803
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3073,128,128,0.011653332908948263
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,128,32,0.002552533398071925
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,128,64,0.014703999956448874
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,128,32,0.015892266233762106
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,64,65536,0.06748373508453369
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,64,51200,0.05324800014495849
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,64,65536,0.5556693394978841
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,64,16384,0.01648853321870168
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,64,51200,0.4418933232625325
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,64,12288,0.01213759978612264
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,128,128,0.01690559983253479
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,64,16384,0.14646933873494467
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,64,10240,0.010939733187357584
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,64,8192,0.009607467055320739
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,64,7168,0.009050666292508443
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,64,12288,0.10273919900258381
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,64,10240,0.07893013159434001
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,64,8192,0.05551040172576904
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,64,6144,0.008437333504358928
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,64,7168,0.048604798316955564
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,64,4096,0.0067071999112765
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,64,5120,0.03930986722310384
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,64,3584,0.0063498665889104204
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,64,3584,0.03326399922370911
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,64,3072,0.005902933577696482
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,64,3072,0.029180800914764403
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,64,2560,0.00516480008761088
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,64,2560,0.027932800849278766
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,64,2048,0.004645333190759023
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,64,2048,0.025337600708007814
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,64,1536,0.004198400179545084
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,64,5120,0.00796693315108617
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,64,1024,0.0034175999462604523
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,64,1536,0.023116799195607503
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,64,1024,0.020702934265136717
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,64,4096,0.033319467306137086
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,64,768,0.003065599997838338
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,64,512,0.002930133293072383
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,64,256,0.002525866776704788
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,64,768,0.019900800784428914
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,64,128,0.002183466653029124
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,64,512,0.018990933895111084
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,64,128,0.015044266978899637
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,64,64,0.002203733225663503
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,64,32,0.002436266591151555
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,64,64,0.014503467082977294
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,32,65536,0.061950933933258054
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,64,32,0.014617600043614707
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,32,51200,0.05003093481063843
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,32,65536,0.5531541188557942
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,32,16384,0.015436800320943198
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,32,51200,0.44000959396362305
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,64,6144,0.04314133326212565
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,32,12288,0.010172800223032633
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,32,10240,0.009075199564297993
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,32,16384,0.1444394588470459
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,32,12288,0.10166079998016357
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,32,8192,0.008225066463152568
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,32,7168,0.008035199840863545
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,32,10240,0.08117547035217285
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,32,8192,0.05663040081659952
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,64,256,0.017146666844685875
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,32,7168,0.050240000089009605
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,32,5120,0.01064639985561371
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,32,6144,0.04498453140258789
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,32,5120,0.039451734224955244
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,32,3584,0.008232533435026805
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,32,4096,0.0329258660475413
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,32,3072,0.007175466914971669
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,32,3584,0.031726932525634764
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,32,2560,0.006682666639486949
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,32,3072,0.029614933331807453
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,32,2048,0.0044938668608665465
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,32,2560,0.027034666140874224
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,32,1536,0.0038389332592487337
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,32,2048,0.02482773264249166
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,32,1024,0.003385599950949351
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,32,1536,0.023218133052190146
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,32,768,0.0030005333324273427
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,32,1024,0.021153066555658975
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,32,512,0.00264533335963885
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,32,768,0.01964906652768453
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,32,256,0.0025759999950726825
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,32,512,0.018526933590571084
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,32,256,0.016521599888801575
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,32,128,0.002186666677395503
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,32,128,0.015830399592717488
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,32,64,0.002186666677395503
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,32,6144,0.007517866790294647
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,32,32,0.0022240000466505687
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,32,64,0.014199466506640116
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3073,32,4096,0.008909866213798523
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3073,32,32,0.014957867066065469
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,65536,16384,3.6048021952311196
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,65536,51200,6.272643025716146
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,65536,16384,1.9023754119873046
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,65536,12288,2.75054931640625
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,65536,51200,12.02716064453125
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,65536,12288,1.4912799835205077
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,65536,10240,2.5238548278808595
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,65536,10240,1.153100840250651
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,65536,10240,0.7001077016194661
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,65536,8192,1.862054443359375
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,65536,8192,0.9180362701416016
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,65536,12288,0.7816191991170247
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,65536,7168,1.5993855794270835
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,65536,7168,0.8171733220418295
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,65536,16384,1.0319552103678384
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,65536,6144,1.305612818400065
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,65536,6144,0.6819786707560221
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,65536,8192,0.5812277475992839
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,65536,5120,1.0635605494181315
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,65536,5120,0.5897013346354167
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,65536,7168,0.5400597254435222
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,65536,4096,0.8485909144083659
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,65536,4096,0.4934837341308594
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,65536,6144,0.46706132888793944
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,65536,3584,0.7075647989908854
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,65536,3584,0.41301333109537763
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,65536,5120,0.427890141805013
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,65536,3072,0.6362773259480794
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,65536,3072,0.36476907730102537
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,65536,4096,0.38500585556030276
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,65536,2560,0.5231381416320801
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,65536,2560,0.30721918741861975
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,65536,3072,0.30385173161824547
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,65536,2048,0.4183082580566406
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,65536,2048,0.2473973274230957
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,65536,2560,0.28854506810506186
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,65536,1536,0.3174933433532715
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,65536,1536,0.20084266662597655
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,65536,3584,0.34289172490437825
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,65536,1024,0.21406720479329427
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,65536,1024,0.14719573656717938
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,65536,1024,0.20380694071451821
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,65536,768,0.16693013509114582
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,65536,768,0.1206549326578776
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,65536,1536,0.21331307093302407
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,65536,512,0.12521386941274007
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,65536,512,0.10310400327046712
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,65536,512,0.19373332659403483
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,65536,256,0.08477333386739096
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,65536,256,0.08563626607259114
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,65536,768,0.1948256015777588
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,65536,128,0.07135146458943685
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,65536,128,0.08049066861470541
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,65536,2048,0.2513525327046712
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,65536,64,0.06542719999949137
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,65536,64,0.07856853008270263
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,65536,32,0.06573013464609781
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,65536,32,0.07851626873016357
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,65536,256,0.19180266062418622
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,65536,128,0.19643200238545735
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,51200,65536,6.34454091389974
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,51200,65536,12.101938883463543
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,51200,51200,9.43389383951823
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,51200,51200,4.889850870768229
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,51200,16384,2.833577473958333
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,51200,16384,1.562270991007487
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,51200,16384,0.8141578674316406
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,51200,12288,2.126688003540039
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,51200,12288,1.140441640218099
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,51200,12288,0.6159104029337565
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,51200,10240,1.8102549235026042
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,51200,10240,0.9005568186442057
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,51200,10240,0.547221310933431
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,51200,8192,1.4011455535888673
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,51200,8192,0.724998410542806
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,65536,51200,3.3414154052734375
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,51200,7168,1.133898671468099
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,51200,7168,0.6936874389648438
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,51200,8192,0.4524256070454915
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,51200,6144,0.9972031911214193
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,51200,6144,0.5560672124226888
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,51200,7168,0.4236597379048665
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,51200,5120,0.7987125396728516
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,51200,5120,0.5079541206359863
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,51200,6144,0.36367359161376955
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,51200,4096,0.6281120300292968
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,51200,4096,0.3912864049275716
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,51200,5120,0.3341514587402344
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,51200,3584,0.5465738932291667
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,51200,3584,0.3293696085611979
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,51200,4096,0.29768107732137045
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,51200,3072,0.4763338724772136
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,51200,3072,0.2843541463216146
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,51200,51200,2.5959882100423175
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,51200,2560,0.4009322802225749
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,51200,2560,0.24489067395528158
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,51200,3584,0.2670986811319987
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,51200,2048,0.3199743906656901
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,51200,2048,0.2016138712565104
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,51200,2560,0.22430720329284667
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,51200,1536,0.24701333045959473
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,51200,1536,0.15747307141621908
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,51200,3072,0.2360981305440267
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,51200,1024,0.17095573743184406
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,51200,1024,0.11834560235341389
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,51200,2048,0.1984831968943278
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,51200,768,0.13285760084788006
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,51200,768,0.0991701364517212
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,51200,1024,0.1632693290710449
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,51200,512,0.09765333334604899
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,51200,512,0.08469546635945638
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,51200,1536,0.1697717348734538
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,51200,256,0.06693759759267172
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,51200,256,0.07135679721832275
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,51200,768,0.15446826616923015
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,51200,128,0.06010666688283285
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,51200,128,0.06606826782226563
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,51200,256,0.152783997853597
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,51200,128,0.1558677355448405
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,51200,64,0.05097493330637613
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,51200,32,0.0520032008488973
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,51200,64,0.06628479957580566
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,51200,32,0.06525013446807862
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,51200,512,0.15402026176452638
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,16384,65536,2.3884096781412762
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,16384,65536,3.8930880228678384
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,16384,51200,3.0278976440429686
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,16384,51200,1.8342763264973958
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,16384,16384,0.8724255879720053
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,16384,16384,0.624780782063802
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,16384,16384,0.305020809173584
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,16384,12288,0.6532970428466797
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,16384,12288,0.434770139058431
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,16384,12288,0.23173227310180664
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,16384,10240,0.5359861373901367
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,16384,10240,0.35467306772867835
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,16384,10240,0.2070613384246826
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,16384,8192,0.4292906761169434
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,16384,8192,0.27811412811279296
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,16384,8192,0.17077226638793946
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,16384,7168,0.36848427454630533
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,16384,7168,0.2530282656351725
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,16384,51200,0.9141045252482096
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,16384,6144,0.3181173324584961
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,16384,6144,0.2151807943979899
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,16384,7168,0.16011306444803874
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,16384,5120,0.2618133385976156
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,16384,5120,0.18274666468302408
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,16384,5120,0.12713599999745687
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,16384,4096,0.20908479690551757
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,16384,4096,0.14478079477945965
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,16384,6144,0.13925973574320477
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,16384,3584,0.18654613494873046
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,16384,3584,0.13371200561523439
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,16384,3584,0.10312533378601074
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,16384,3072,0.1611189365386963
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,16384,3072,0.11487039724985759
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,16384,4096,0.11256639957427979
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,16384,2560,0.13277653058369954
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,16384,2560,0.09822826385498047
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,16384,2560,0.08552959760030111
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,16384,2048,0.10744000275929769
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,16384,2048,0.08281493186950684
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,16384,3072,0.09117226600646973
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,16384,1536,0.08387946287790934
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,16384,1536,0.06653759876887003
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,51200,65536,3.3428426106770837
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,16384,1024,0.06080106496810913
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,16384,1024,0.07019200325012206
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,16384,1024,0.05775466759999594
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,16384,768,0.04479680061340332
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,16384,768,0.05539946556091309
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,16384,1536,0.06162346601486206
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,16384,512,0.038950399557749434
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,16384,512,0.04925119876861572
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,16384,2048,0.07283733685811361
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,16384,256,0.030136533578236896
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,16384,256,0.043353601296742754
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,16384,768,0.049634134769439696
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,16384,128,0.019407999515533448
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,16384,128,0.03234133323033651
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,16384,128,0.05006826718648275
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,16384,64,0.017707733313242595
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,16384,64,0.03260800043741862
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,16384,512,0.04921066761016846
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,16384,32,0.018155733744303383
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,16384,32,0.031066666046778362
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,16384,256,0.04830933411916097
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,12288,65536,1.7958922068277996
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,12288,65536,2.7309524536132814
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,12288,51200,2.0602773030598955
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,12288,51200,1.410601552327474
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,12288,16384,0.6028810501098633
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,12288,16384,0.4442432085673015
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,16384,65536,1.16035525004069
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,12288,12288,0.46262187957763673
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,12288,12288,0.33344106674194335
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,12288,16384,0.2339637279510498
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,12288,10240,0.3902581214904785
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,12288,10240,0.27777493794759117
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,12288,12288,0.18026879628499348
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,12288,8192,0.32133547465006507
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,12288,8192,0.24102293650309242
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,12288,10240,0.159772793451945
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,12288,7168,0.27298774719238283
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,12288,7168,0.20408213933308922
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,12288,8192,0.13269440333048504
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,12288,6144,0.2366741339365641
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,12288,6144,0.17112213770548504
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,12288,7168,0.12194026311238607
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,12288,5120,0.19164586067199707
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,12288,5120,0.1431477387746175
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,12288,6144,0.10751360257466633
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,12288,4096,0.15637013117472331
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,12288,4096,0.11420053641001385
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,12288,4096,0.08677226702372233
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,12288,3584,0.137444273630778
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,12288,3584,0.10483840306599934
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,12288,5120,0.09762773513793946
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,12288,3072,0.11597653230031331
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,12288,3072,0.09189653396606445
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,12288,3584,0.07714666525522867
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,12288,3072,0.06800639629364014
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,12288,2560,0.09952106475830078
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,12288,2560,0.08019093672434488
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,12288,2048,0.0792458693186442
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,12288,2048,0.06940266291300455
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,12288,2048,0.05762453476587931
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,12288,1536,0.061093334356943765
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,12288,1536,0.05556159814198812
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,12288,2560,0.0645962675412496
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,12288,1024,0.0433781345685323
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,12288,1024,0.046801066398620604
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,12288,1024,0.050988801320393885
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,12288,768,0.03404693206151326
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,12288,768,0.041212801138559976
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,12288,1536,0.05268266598383585
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,12288,512,0.024151466290156045
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,12288,512,0.03654826482137044
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,12288,512,0.04785919984181722
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,12288,256,0.01802133321762085
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,12288,256,0.030990932385126752
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,12288,768,0.04738346735636394
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,12288,128,0.017339734236399333
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,12288,128,0.02853333353996277
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,12288,128,0.04699519872665405
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,12288,64,0.014874666929244995
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,12288,256,0.047068798542022706
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,12288,64,0.026843732595443724
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,12288,32,0.014903466900189719
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,12288,32,0.026833067337671917
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,10240,65536,1.712503433227539
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,10240,65536,2.4791425069173174
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,10240,51200,1.8943477630615235
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,10240,51200,1.3159957885742188
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,12288,51200,0.6925888061523438
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,10240,16384,0.5766165415445964
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,10240,16384,0.4634943962097168
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,10240,16384,0.21005120277404785
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,10240,12288,0.43384428024291993
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,10240,12288,0.34047892888387044
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,12288,65536,0.873628807067871
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,10240,10240,0.34019947052001953
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,10240,10240,0.25045760472615564
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,10240,12288,0.16151679356892903
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,10240,8192,0.2893024126688639
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,10240,8192,0.20615679423014321
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,10240,10240,0.14394240379333495
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,10240,7168,0.23816426595052084
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,10240,7168,0.17614720662434896
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,10240,8192,0.11867626508076985
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,10240,6144,0.20424106915791831
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,10240,6144,0.15535252888997395
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,10240,7168,0.11038400332132976
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,10240,5120,0.16669972737630206
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,10240,5120,0.13086079756418864
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,10240,6144,0.09676693280537924
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,10240,4096,0.13547520637512206
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,10240,4096,0.1051594654719035
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,10240,51200,0.6131050745646159
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,10240,3584,0.11552213033040364
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,10240,3584,0.09665706952412924
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,10240,5120,0.0874613364537557
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,10240,3072,0.10054612954457601
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,10240,3072,0.08427093029022217
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,10240,3584,0.06794346968332926
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,10240,2560,0.08405973116556803
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,10240,2560,0.07414506276448568
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,10240,4096,0.07640853722890219
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,10240,2048,0.06986880302429199
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,10240,2048,0.06473813454310098
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,10240,2560,0.059418666362762454
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,10240,1536,0.053028265635172524
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,10240,1536,0.05128746827443441
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,10240,3072,0.06197653214136759
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,10240,1024,0.03731840054194133
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,10240,1024,0.045788800716400145
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,10240,2048,0.05344320138295492
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,10240,768,0.029069866736729937
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,10240,768,0.03905813296635945
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,10240,1024,0.0475381334622701
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,10240,512,0.02322346568107605
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,10240,512,0.03445226748784383
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,10240,1536,0.04899626572926839
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,10240,256,0.016777600844701132
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,10240,256,0.029341866572697956
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,10240,512,0.043652268250783284
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,10240,128,0.012873599926630655
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,10240,128,0.02606719930966695
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,10240,768,0.04462080001831055
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,10240,64,0.01206933359305064
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,10240,64,0.028592000404993694
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,10240,32,0.015277866522471109
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,10240,32,0.02725013295809428
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,10240,256,0.04363199869791667
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,10240,128,0.04400320053100586
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,8192,65536,1.385198974609375
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,8192,65536,1.8455914815266925
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,8192,51200,1.402410634358724
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,8192,51200,1.1080331166585287
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,8192,16384,0.4337034543355306
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,8192,16384,0.3807392120361328
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,10240,65536,0.7762378692626953
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,8192,12288,0.3243925412495931
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,8192,12288,0.26328852971394856
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,8192,16384,0.1826090653737386
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,8192,10240,0.2755039850870768
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,8192,10240,0.22198400497436524
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,8192,10240,0.1256650686264038
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,8192,8192,0.22136319478352867
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,8192,8192,0.17820693651835123
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,8192,12288,0.14012266794840494
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,8192,7168,0.19198400179545086
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,8192,7168,0.15138667424519855
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,8192,7168,0.09578986962636313
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,8192,6144,0.1660970687866211
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,8192,6144,0.13216959635416667
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,8192,8192,0.10254080295562744
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,8192,5120,0.1418922742207845
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,8192,5120,0.11567573547363282
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,8192,6144,0.08306986490885417
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,8192,4096,0.11219627062479656
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,8192,4096,0.08991040388743082
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,8192,5120,0.07587839762369791
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,8192,3584,0.09936853249867758
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,8192,3584,0.08488852977752685
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,8192,3584,0.05779306491216024
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,8192,3072,0.08539413611094157
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,8192,3072,0.0740330696105957
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,8192,4096,0.06466773351033529
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,8192,2560,0.07029866377512614
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,8192,2560,0.06444266637166342
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,8192,2560,0.048087465763092044
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,8192,2048,0.056510933240254725
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,8192,2048,0.05532480080922445
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,8192,3072,0.049865599473317465
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,8192,1536,0.04365866581598918
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,8192,1536,0.045235200723012285
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,8192,1536,0.0365066647529602
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,8192,1024,0.030050132671991987
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,8192,1024,0.03879040082295736
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,8192,2048,0.042005332310994466
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,8192,768,0.023282132546106973
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,8192,768,0.034007465839385985
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,8192,1024,0.03480319976806641
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,8192,512,0.0180565337340037
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,8192,512,0.03059733311335246
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,8192,768,0.030460800727208453
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,8192,256,0.013138133287429809
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,8192,256,0.02881386677424113
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,8192,512,0.030077866713205975
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,8192,128,0.013727999726931252
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,8192,128,0.026657066742579144
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,8192,256,0.03002133369445801
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,8192,64,0.013195733229319254
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,8192,64,0.024895999828974405
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,8192,32,0.013645866513252258
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,8192,128,0.030846933523813885
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,8192,32,0.026824533939361572
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,7168,65536,1.2755616505940757
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,7168,65536,1.5906378428141275
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,7168,51200,1.2191808064778646
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,7168,51200,1.0305461247762044
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,8192,51200,0.5337141036987305
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,7168,16384,0.35716800689697265
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,7168,16384,0.3365621248881022
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,7168,16384,0.16058773994445802
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,7168,12288,0.27061119079589846
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,7168,12288,0.24981226921081542
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,8192,65536,0.6747413635253906
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,7168,10240,0.24362773895263673
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,7168,10240,0.2007157325744629
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,7168,12288,0.12293226718902588
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,7168,8192,0.1929759979248047
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,7168,8192,0.15947945912679035
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,7168,10240,0.1112554629643758
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,7168,7168,0.17035412788391113
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,7168,7168,0.13949653307596843
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,7168,8192,0.09128106435139974
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,7168,6144,0.14338240623474122
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,7168,6144,0.12280853589375813
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,7168,7168,0.08568106492360433
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,7168,5120,0.11879146893819173
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,7168,5120,0.10676693121592204
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,7168,6144,0.07490986982981364
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,7168,4096,0.0941536029179891
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,7168,4096,0.08812053203582763
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,7168,5120,0.06758080323537191
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,7168,3584,0.08334399859110514
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,7168,3584,0.08046933015187582
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,7168,4096,0.05745706558227539
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,7168,3072,0.0713749329249064
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,7168,3072,0.07001706759134928
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,7168,3072,0.045823999245961505
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,7168,2560,0.06037653287251791
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,7168,2560,0.06080853144327799
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,7168,3584,0.05226773420969645
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,7168,2048,0.04956479867299397
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,7168,2048,0.051003734270731606
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,7168,51200,0.4663669268290202
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,7168,1536,0.03865493138631185
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,7168,1536,0.0411296010017395
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,7168,1536,0.03589973449707031
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,7168,1024,0.02624746759732564
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,7168,1024,0.03516800006230672
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,7168,2048,0.03999146620432536
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,7168,768,0.020403200387954713
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,7168,768,0.032239999373753866
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,7168,2560,0.044489598274230956
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,7168,512,0.015361066659291586
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,7168,512,0.028864000240961713
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,7168,1024,0.03415573438008626
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,7168,256,0.011615999539693196
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,7168,256,0.02400533358256022
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,7168,256,0.03163946668306987
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,7168,512,0.0316810667514801
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,7168,128,0.010429867108662923
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,7168,128,0.022338134050369263
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,7168,64,0.009518933296203614
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,7168,64,0.02274880011876424
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,7168,32,0.009595732887585957
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,7168,32,0.022987733284632363
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,7168,128,0.031404799222946166
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,7168,768,0.03170666694641113
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,6144,65536,1.2871381123860677
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,6144,65536,1.149831517537435
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,6144,51200,0.962283706665039
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,6144,51200,0.9052544275919596
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,6144,16384,0.30927680333455404
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,6144,16384,0.2942410786946615
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,6144,16384,0.14160213470458985
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,6144,12288,0.2357034683227539
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,6144,12288,0.22012480099995932
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,7168,65536,0.5859530766805012
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,6144,10240,0.202401065826416
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,6144,10240,0.1772864023844401
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,6144,10240,0.09527680079142252
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,6144,8192,0.16387200355529785
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,6144,8192,0.13731199900309246
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,6144,12288,0.10698453585306804
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,6144,7168,0.14477972984313964
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,6144,7168,0.1248586654663086
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,6144,7168,0.07438080310821533
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,6144,6144,0.12302719751993815
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,6144,6144,0.11031893094380696
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,6144,8192,0.07850026289621989
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,6144,5120,0.10002346833546956
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,6144,5120,0.09532480239868164
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,6144,5120,0.05923733313878378
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,6144,4096,0.07923413117726644
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,6144,4096,0.07620373566945395
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,6144,6144,0.06645866632461547
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,6144,3584,0.07081493536631266
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,6144,3584,0.07200640042622884
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,6144,3584,0.04523306687672933
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,6144,3072,0.06127039988835653
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,6144,3072,0.06226239999135336
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,6144,4096,0.05037013292312622
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,6144,2560,0.050680534044901526
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,6144,2560,0.053629867235819494
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,6144,2560,0.03852479855219523
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,6144,2048,0.041145598888397215
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,6144,2048,0.046404266357421876
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,6144,3072,0.04039146502812703
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,6144,1536,0.03229866623878479
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,6144,1536,0.03888426621754964
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,6144,1536,0.03135040005048116
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,6144,1024,0.022973867257436116
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,6144,1024,0.032778666416804
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,6144,2048,0.03465919891993205
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,6144,768,0.01797653237978617
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,6144,768,0.028998400767644244
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,6144,1024,0.030089600880940752
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,6144,512,0.013598933815956116
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,6144,512,0.026667733987172444
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,6144,512,0.027607466777165728
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,6144,256,0.009570133686065675
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,6144,256,0.023468800385793052
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,6144,768,0.028008532524108887
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,6144,128,0.009550933043162029
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,6144,128,0.020626133680343627
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,6144,256,0.027688533067703247
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,6144,64,0.00831573357184728
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,6144,128,0.027586134274800618
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,6144,64,0.020856533447901407
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,6144,32,0.008412800232569377
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,6144,32,0.020654932657877604
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,5120,65536,1.1414122263590494
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,5120,65536,1.0897674560546875
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,5120,51200,0.8755093256632487
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,5120,51200,0.8585866928100586
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,6144,51200,0.40111147562662763
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,5120,16384,0.27347199122111
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,5120,16384,0.28514134089152016
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,5120,16384,0.13821333249409992
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,5120,12288,0.2075157324473063
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,5120,12288,0.20490667025248208
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,6144,65536,0.5080010732014973
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,5120,10240,0.1841866652170817
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,5120,10240,0.16277333895365398
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,5120,12288,0.10487786928812663
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,5120,8192,0.14435839653015137
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,5120,8192,0.12820053100585938
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,5120,10240,0.09398720264434815
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,5120,7168,0.12642026742299398
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,5120,7168,0.11590186754862468
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,5120,8192,0.07814293702443441
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,5120,6144,0.10537919998168946
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,5120,6144,0.10253866513570149
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,5120,7168,0.07392746607462565
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,5120,5120,0.08784426848093668
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,5120,5120,0.08994666735331217
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,5120,6144,0.06505920092264811
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,5120,4096,0.06888426939646403
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,5120,4096,0.07248533566792806
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,5120,5120,0.05779733260472616
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,5120,3584,0.06094826857248942
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,5120,3584,0.0662058671315511
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,5120,4096,0.04968213240305583
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,5120,3072,0.05196693340937296
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,5120,3072,0.05771199862162272
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,5120,3584,0.04623573223749797
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,5120,2560,0.04462186495463054
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,5120,2560,0.050929065545399985
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,5120,3072,0.04155199925104777
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,5120,2048,0.035740800698598224
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,5120,2048,0.0435754656791687
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,5120,51200,0.4014698664347331
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,5120,1536,0.029029333591461183
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,5120,1536,0.03703253269195557
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,5120,2560,0.039928531646728514
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,5120,2048,0.03668479919433594
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,5120,1024,0.03298986752827962
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,5120,1024,0.020179200172424316
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,5120,768,0.016613333423932394
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,5120,1536,0.03246506651242574
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,5120,768,0.03057173291842143
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,5120,512,0.01330880026022593
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,5120,512,0.027772800127665205
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,5120,1024,0.03158186674118042
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,5120,512,0.02966826756795247
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,5120,256,0.010806399583816528
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,5120,256,0.025569067398707075
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,5120,128,0.007528533538182576
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,5120,768,0.029847466945648195
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,5120,128,0.02291946609814962
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,5120,64,0.010108799735705058
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,5120,64,0.022781866788864135
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,5120,32,0.01048533320426941
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,5120,256,0.02940373420715332
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,5120,32,0.022782933712005616
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,4096,65536,0.9140714645385742
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,4096,65536,0.9754208246866861
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,5120,128,0.029414399464925127
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,4096,51200,0.686187744140625
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,4096,51200,0.7755797068277995
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,4096,16384,0.2259925365447998
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,4096,16384,0.2527029355367025
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,4096,16384,0.11032426357269287
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,4096,12288,0.1720362663269043
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,4096,12288,0.18475200335184733
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,5120,65536,0.5049056053161621
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,4096,10240,0.14414933522542317
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,4096,10240,0.1452778657277425
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,4096,10240,0.07743253707885742
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,4096,12288,0.08577280044555664
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,4096,8192,0.1182517369588216
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,4096,8192,0.11606079737345379
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,4096,7168,0.10037120183308919
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,4096,7168,0.10452799797058106
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,4096,7168,0.05937706629435221
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,4096,6144,0.08533973693847656
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,4096,6144,0.09326826731363932
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,4096,8192,0.06424426635106405
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,4096,5120,0.07022826671600342
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,4096,5120,0.07979839642842611
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,4096,5120,0.04695253372192383
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,4096,4096,0.05645333528518677
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,4096,4096,0.06254613399505615
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,4096,6144,0.0521781325340271
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,4096,3584,0.05001706679662069
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,4096,3584,0.05736960172653198
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,4096,4096,0.040064001083374025
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,4096,3072,0.04260373512903849
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,4096,3584,0.03631680011749268
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,4096,3072,0.04969813426335652
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,4096,2560,0.03596373399098714
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,4096,2560,0.045389866828918456
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,4096,2560,0.03097813328107198
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,4096,2048,0.029179733991622925
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,4096,2048,0.03936320145924886
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,4096,3072,0.03224746584892273
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,4096,1536,0.023324799537658692
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,4096,1536,0.03377813498179118
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,4096,1536,0.023203200101852416
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,4096,1024,0.016421332955360413
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,4096,1024,0.030726399024327593
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,4096,2048,0.026444800694783527
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,4096,768,0.013303466637929282
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,4096,768,0.027159466346104937
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,4096,768,0.02291733423868815
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,4096,512,0.010980266332626342
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,4096,512,0.02477333347002665
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,4096,1024,0.022299732764561972
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,4096,256,0.009211732943852743
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,4096,256,0.022842667500178018
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,4096,256,0.02257919907569885
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,4096,128,0.009133866429328919
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,4096,128,0.020917334159215293
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,4096,512,0.022643200556437173
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,4096,64,0.008398933211962382
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,4096,64,0.020797866582870483
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,4096,32,0.007897600034872691
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,4096,128,0.021156267325083414
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,4096,32,0.021313067277272543
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,3584,65536,0.8525578816731771
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,3584,65536,0.9330645243326823
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,3584,51200,0.6261205037434896
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,3584,51200,0.74202454884847
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,4096,51200,0.31633920669555665
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,3584,16384,0.20187412897745766
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,3584,16384,0.23659200668334962
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,3584,16384,0.11087360382080078
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,3584,12288,0.1523146629333496
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,3584,12288,0.17162987391153972
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,4096,65536,0.3979082743326823
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,3584,10240,0.11745813687642415
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,3584,10240,0.1337941328684489
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,3584,10240,0.07898346583048502
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,3584,8192,0.09467946688334147
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,3584,8192,0.10694080193837482
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,3584,12288,0.08660693168640136
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,3584,7168,0.08389653364817301
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,3584,7168,0.09648213386535645
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,3584,7168,0.060976000626881924
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,3584,6144,0.07105706532796224
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,3584,6144,0.08611733118693034
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,3584,8192,0.06485866705576579
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,3584,5120,0.05838613510131836
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,3584,5120,0.07422719796498617
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,3584,5120,0.04846506516138713
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,3584,6144,0.05292373498280843
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,3584,4096,0.04681599934895833
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,3584,4096,0.05814079840977987
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,3584,3584,0.04195306698481242
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,3584,3584,0.05394239823023478
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,3584,3584,0.038680533568064376
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,3584,3072,0.03577386538187663
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,3584,3072,0.04644373257954915
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,3584,4096,0.04153813521067302
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,3584,2560,0.02979733347892761
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,3584,2560,0.043321601549784344
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,3584,3072,0.034332799911499026
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,3584,2048,0.024861866235733034
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,3584,2560,0.03288319905598958
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,3584,2048,0.037291733423868816
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,3584,1536,0.01967466672261556
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,3584,1536,0.03180586695671082
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,3584,2048,0.029772800207138062
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,3584,1024,0.013667200009028116
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,3584,1024,0.028344533840815228
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,3584,1536,0.026714666684468584
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,3584,768,0.01095360020796458
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,3584,768,0.025862399737040204
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,3584,1024,0.025614933172861738
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,3584,512,0.008754133184750875
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,3584,512,0.0228928009668986
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,3584,768,0.02395626703898112
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,3584,256,0.0067658667763074234
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,3584,51200,0.3198378562927246
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,3584,256,0.020920532941818237
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,3584,512,0.02677546739578247
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,3584,128,0.006713599960009258
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,3584,128,0.018974934021631876
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,3584,256,0.024101332823435465
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,3584,64,0.0058890665570894875
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,3584,128,0.02396799921989441
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,3584,64,0.01863893270492554
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,3584,32,0.005922133227189382
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,3584,32,0.0187882661819458
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,3072,65536,0.5872021357218424
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,3072,65536,0.8263466517130533
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,3072,51200,0.4590421358744304
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,3072,51200,0.6543338775634766
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,3072,16384,0.15199146270751954
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,3072,16384,0.21149759292602538
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,3072,16384,0.08835306962331137
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,3072,12288,0.11502292950948079
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,3072,12288,0.15428694089253742
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,3584,65536,0.401745065053304
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,3072,10240,0.10350826581319172
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,3072,10240,0.1288383960723877
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,3072,10240,0.0647541324297587
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,3072,8192,0.08348906834920247
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,3072,8192,0.10165226459503174
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,3072,12288,0.06997120380401611
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,3072,7168,0.07182613213857016
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,3072,7168,0.09077119827270508
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,3072,7168,0.049531733989715575
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,3072,6144,0.060969599088033045
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,3072,6144,0.0802069346110026
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,3072,8192,0.05301119883855184
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,3072,5120,0.051206401983896886
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,3072,5120,0.06842559973398844
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,3072,5120,0.04008426666259766
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,3072,4096,0.04116373459498088
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,3072,4096,0.053275732199350986
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,3072,6144,0.04350613355636597
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,3072,3584,0.03605120182037354
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,3072,3584,0.05002346833546957
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,3072,3584,0.03176106611887614
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,3072,3072,0.03133760094642639
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,3072,3072,0.044544001420338944
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,3072,4096,0.0348960002263387
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,3072,2560,0.026980266968409224
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,3072,2560,0.04127466678619385
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,3072,2560,0.027244800329208375
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,3072,2048,0.022345600525538127
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,3072,3072,0.028450133403142293
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,3072,2048,0.03573866685231526
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,3072,1536,0.017901867628097534
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,3072,1536,0.030851199229558306
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,3072,1536,0.021964800357818604
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,3072,1024,0.01304213305314382
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,3072,1024,0.027169066667556762
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,3072,2048,0.02393066684405009
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,3072,768,0.010145066181818645
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,3072,768,0.024753065903981526
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,3072,1024,0.02104640007019043
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,3072,512,0.008158933122952778
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,3072,512,0.022707200050354003
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,3072,768,0.019517866770426433
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,3072,256,0.006291200220584869
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,3072,256,0.02046826680501302
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,3072,256,0.019398399194081626
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,3072,512,0.01962453325589498
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,3072,128,0.005871999760468801
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,3072,128,0.018565332889556883
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,3072,64,0.005054933329423269
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,3072,64,0.01743146578470866
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,3072,32,0.005497600138187409
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,3072,128,0.01938986579577128
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,3072,32,0.017429333925247193
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,2560,65536,0.5379018783569336
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,2560,65536,0.7989482879638672
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,2560,51200,0.4220992088317871
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,2560,51200,0.6459519704182942
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,3072,51200,0.2417429288228353
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,2560,16384,0.13711466789245605
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,2560,16384,0.20544959704081217
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,2560,16384,0.08744640350341797
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,2560,12288,0.10407466888427734
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,2560,12288,0.15048106511433917
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,3072,65536,0.30137065251668294
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,2560,10240,0.09322453339894612
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,2560,10240,0.1209941307703654
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,2560,12288,0.0693674643834432
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,2560,8192,0.07403093179066976
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,2560,8192,0.09583360354105631
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,2560,10240,0.06341013511021933
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,2560,7168,0.06241600116093954
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,2560,7168,0.08595200379689535
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,2560,7168,0.047387735048929853
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,2560,6144,0.05265493392944336
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,2560,6144,0.07439466317494711
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,2560,8192,0.052496000130971276
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,2560,5120,0.0443338672320048
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,2560,5120,0.06368853251139323
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,2560,5120,0.039394132296244305
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,2560,4096,0.03567359844843547
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,2560,6144,0.04254613320032756
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,2560,4096,0.04984106620152791
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,2560,3584,0.03177599906921387
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,2560,3584,0.047650134563446044
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,2560,4096,0.03378133376439412
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,2560,3072,0.026923733949661254
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,2560,3072,0.04289919932683309
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,2560,3584,0.030618667602539062
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,2560,2560,0.02331519921620687
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,2560,2560,0.039137065410614014
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,2560,3072,0.027618134021759035
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,2560,2048,0.019157334168752035
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,2560,2048,0.0347765326499939
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,2560,2560,0.026029866933822633
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,2560,1536,0.015228799978892007
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,2560,1536,0.02987733284632365
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,2560,51200,0.24086079597473145
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,2560,1024,0.010962133606274922
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,2560,1024,0.026809600989023845
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,2560,1536,0.020680532852808634
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,2560,768,0.009495466947555542
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,2560,768,0.025144533316294355
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,2560,2048,0.023142399390538533
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,2560,512,0.008008533219496409
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,2560,512,0.025145600239435833
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,2560,1024,0.01964799960454305
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,2560,256,0.006742399930953979
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,2560,256,0.02021226684252421
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,2560,512,0.018172800540924072
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,2560,128,0.005522133409976959
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,2560,128,0.019773866732915243
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,2560,256,0.01814613342285156
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,2560,64,0.00625493327776591
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,2560,64,0.018910932540893554
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,2560,32,0.006347733239332835
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,2560,128,0.01790293256441752
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,2560,32,0.018787199258804323
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,2560,768,0.01821119983990987
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,2048,65536,0.42989012400309246
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,2048,51200,0.33737386067708336
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,2048,65536,0.7437365214029948
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,2048,51200,0.595898691813151
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,2048,16384,0.1100757360458374
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,2048,16384,0.1927893320719401
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,2048,16384,0.08559786478678386
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,2048,12288,0.08363733291625977
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,2048,12288,0.1415071964263916
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,2560,65536,0.30021759668986003
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,2048,10240,0.07276373704274496
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,2048,10240,0.1138207991917928
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,2048,10240,0.06145066817601522
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,2048,8192,0.0578325351079305
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,2048,8192,0.09003093242645263
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,2048,12288,0.0677781343460083
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,2048,7168,0.04945066769917806
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,2048,7168,0.0794485330581665
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,2048,7168,0.04630613327026367
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,2048,6144,0.0424511988957723
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,2048,6144,0.06788907051086426
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,2048,8192,0.049421866734822587
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,2048,5120,0.03606506586074829
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,2048,5120,0.05681066513061524
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,2048,6144,0.04065066576004028
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,2048,4096,0.028911999861399335
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,2048,4096,0.04718933502833049
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,2048,5120,0.037079465389251706
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,2048,3584,0.025769599278767902
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,2048,3584,0.04393920103708903
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,2048,4096,0.03205653429031372
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,2048,3072,0.022412800788879396
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,2048,3584,0.029099732637405396
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,2048,3072,0.040098134676615396
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,2048,2560,0.019419733683268228
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,2048,2560,0.03701759974161784
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,2048,2560,0.024851200977961223
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,2048,2048,0.016540799538294473
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,2048,2048,0.032917332649230954
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,2048,3072,0.025978666543960572
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,2048,1536,0.012875733772913614
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,2048,1536,0.028357332944869994
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,2048,1536,0.01862293283144633
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,2048,1024,0.00957546631495158
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,2048,1024,0.025100799401601155
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,2048,2048,0.021496532360712688
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,2048,768,0.00806933343410492
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,2048,768,0.024700800577799477
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,2048,768,0.016218666235605875
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,2048,1024,0.01776319940884908
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,2048,512,0.006750933329264323
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,2048,512,0.021306665738423665
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,2048,256,0.005858133236567179
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,2048,256,0.018986666202545167
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,2048,256,0.016105600198109946
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,2048,512,0.016131200393040977
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,2048,128,0.005039999882380167
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,2048,128,0.018730666240056357
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,2048,64,0.005514666438102722
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,2048,64,0.017588265736897788
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,2048,128,0.01569386621316274
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,2048,32,0.005936000247796377
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,2048,32,0.018081067005793254
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,1536,65536,0.3149322509765625
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,1536,65536,0.6783381144205729
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,1536,51200,0.24264000256856283
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,1536,51200,0.5348224004109701
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,2048,51200,0.2399903933207194
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,1536,16384,0.08167786598205566
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,1536,16384,0.17509652773539225
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,1536,16384,0.06556586821873983
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,1536,12288,0.0619978666305542
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,1536,12288,0.1278058687845866
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,2048,65536,0.30078293482462565
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,1536,10240,0.051305600007375086
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,1536,10240,0.10551679929097493
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,1536,12288,0.05333653291066488
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,1536,8192,0.041918933391571045
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,1536,8192,0.08070826530456543
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,1536,10240,0.04809813499450684
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,1536,7168,0.03630293210347493
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,1536,7168,0.07129706541697184
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,1536,8192,0.039185067017873124
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,1536,6144,0.03160746693611145
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,1536,6144,0.06041386524836222
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,1536,7168,0.035884801546732584
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,1536,5120,0.026280534267425538
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,1536,5120,0.05216853221257528
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,1536,6144,0.032891732454299924
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,1536,4096,0.022193066279093423
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,1536,4096,0.043253334363301595
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,1536,4096,0.02712000012397766
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,1536,3584,0.02003413240114848
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,1536,3584,0.04125653505325318
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,1536,5120,0.031177600224812824
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,1536,3072,0.017461333672205606
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,1536,3072,0.03712746699651082
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,1536,3584,0.02518186569213867
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,1536,2560,0.015542399883270264
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,1536,2560,0.033351465066274004
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,1536,2560,0.021556266148885093
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,1536,2048,0.012906666596730551
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,1536,2048,0.030780800183614093
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,1536,3072,0.02305813431739807
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,1536,1536,0.010408533612887065
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,1536,51200,0.16846079826354982
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,1536,1536,0.027509333690007527
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,1536,1024,0.00798826664686203
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,1536,1024,0.024807467063268026
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,1536,1536,0.017412267128626504
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,1536,768,0.006778666873772939
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,1536,768,0.022796799739201866
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,1536,2048,0.019425066312154134
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,1536,512,0.005842133363087972
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,1536,512,0.02068586746851603
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,1536,1024,0.016540799538294473
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,1536,256,0.004678399860858917
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,1536,768,0.015330132842063905
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,1536,256,0.018641066551208497
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,1536,128,0.0042346666256586705
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,1536,512,0.015362133582433065
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,1536,128,0.016636799772580466
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,1536,64,0.0038143999874591826
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,1536,256,0.015390933553377787
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,1536,32,0.004289066791534424
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,1536,64,0.017512534062067667
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,1536,128,0.01658560037612915
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,1536,32,0.016723199685414632
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,1024,65536,0.22420479456583658
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,1024,51200,0.1747381369272868
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,1024,65536,0.6552682876586914
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,1024,51200,0.4964159965515137
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,1024,16384,0.05807786782582601
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,1024,16384,0.16397652626037598
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,1024,16384,0.06297066609064737
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,1024,12288,0.04420373439788818
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,1024,12288,0.11833386421203614
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,1536,65536,0.2082752068837484
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,1024,10240,0.03696320056915283
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,1024,10240,0.09766080379486083
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,1024,10240,0.04425493478775024
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,1024,8192,0.029402667284011842
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,1024,8192,0.07245972951253256
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,1024,12288,0.050578133265177405
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,1024,7168,0.025966932376225788
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,1024,7168,0.06329066753387451
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,1024,7168,0.032979200283686325
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,1024,6144,0.022768000761667885
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,1024,8192,0.03538133303324382
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,1024,6144,0.054668800036112464
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,1024,5120,0.019630932807922365
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,1024,5120,0.04763199885686238
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,1024,6144,0.030497066179911297
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,1024,4096,0.016225066781044007
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,1024,4096,0.03929920196533203
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,1024,5120,0.02802773316701253
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,1024,3584,0.014717866977055868
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,1024,3584,0.03845653136571248
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,1024,4096,0.02436586618423462
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,1024,3072,0.012483200430870056
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,1024,3072,0.03354453245798747
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,1024,3584,0.02263360023498535
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,1024,3072,0.020252799987792967
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,1024,2560,0.011185066898663838
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,1024,2560,0.03167999982833862
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,1024,2048,0.010195199648539226
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,1024,2560,0.019351466496785482
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,1024,2048,0.03125866651535034
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,1024,1536,0.008468266328175862
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,1024,1536,0.02681279977162679
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,1024,2048,0.016922666629155477
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,1024,1024,0.006723199784755707
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,1024,1024,0.024510933955510458
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,1024,1536,0.01511360009511312
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,1024,768,0.006041599810123444
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,1024,768,0.02290239930152893
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,1024,51200,0.16524906158447267
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,1024,512,0.005103999873002371
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,1024,1024,0.014062933127085366
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,1024,512,0.02027946710586548
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,1024,256,0.0042346666256586705
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,1024,256,0.018759467204411826
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,1024,512,0.012873599926630655
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,1024,128,0.003819733361403147
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,1024,256,0.01288640002409617
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,1024,128,0.01700906753540039
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,1024,64,0.0035071998834609987
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,1024,768,0.013332266608874002
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,1024,64,0.014579199751218162
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,1024,128,0.013656533757845559
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,1024,32,0.003553066651026408
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,1024,32,0.015211733182271323
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,768,65536,0.17926613489786786
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,768,51200,0.14067519505818685
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,768,65536,0.6152170817057292
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,768,51200,0.49795840581258133
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,768,16384,0.04671359856923421
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,768,16384,0.1586293379465739
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,768,16384,0.06365866661071777
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,768,12288,0.03528853257497151
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,768,12288,0.11414933204650879
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,1024,65536,0.2046666622161865
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,768,10240,0.028549333413441975
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,768,10240,0.09191892941792806
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,768,10240,0.04470613400141398
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,768,8192,0.022258132696151733
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,768,12288,0.0509610652923584
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,768,8192,0.06727466583251954
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,768,7168,0.019645865758260092
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,768,7168,0.05914239883422852
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,768,7168,0.0338645339012146
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,768,6144,0.017180800437927246
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,768,6144,0.05074026584625244
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,768,8192,0.036008532842000326
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,768,5120,0.015237333377202353
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,768,5120,0.04485013484954834
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,768,5120,0.029277867078781127
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,768,4096,0.01256106694539388
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,768,6144,0.030904533465703328
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,768,4096,0.03715306520462036
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,768,3584,0.011213866869608562
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,768,3584,0.03518506685892741
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,768,3584,0.02358293334643046
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,768,3072,0.010473600029945374
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,768,4096,0.025446399052937822
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,768,3072,0.03204693396886189
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,768,2560,0.008876799543698629
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,768,2560,0.030384000142415362
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,768,2560,0.01994133392969767
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,768,3072,0.021473066012064616
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,768,2048,0.007667199770609538
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,768,2048,0.028785065809885664
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,768,1536,0.006346666812896728
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,768,1536,0.024693334102630617
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,768,1536,0.016114133596420287
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,768,2048,0.017745065689086913
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,768,1024,0.005141333242257436
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,768,1024,0.022756266593933105
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,768,768,0.004738133152325948
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,768,768,0.0208512008190155
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,768,1024,0.014917332927385965
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,768,512,0.004196266829967499
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,768,768,0.014085333546002707
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,768,512,0.019156267245610557
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,768,256,0.003453866640726725
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,768,256,0.01741546591122945
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,768,512,0.014267733693122864
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,768,128,0.003028266628583272
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,768,256,0.014468266566594442
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,768,128,0.01607039968172709
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,768,64,0.0029845332105954488
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,768,128,0.014547200004259745
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,768,64,0.01539413332939148
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,768,32,0.0030826665461063385
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,768,32,0.015229866902033488
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,512,65536,0.1435200055440267
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,768,51200,0.1657514731089274
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,512,65536,0.5862218856811523
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,512,51200,0.11450560092926025
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,512,51200,0.46611948013305665
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,512,16384,0.039342931906382245
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,512,16384,0.15549759864807128
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,768,65536,0.20495786666870117
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,512,12288,0.02916693290074666
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,512,12288,0.10988372961680096
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,512,16384,0.06201066573460897
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,512,10240,0.02422399918238322
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,512,10240,0.0884447971979777
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,512,12288,0.04935040076573689
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,512,8192,0.01883946657180786
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,512,8192,0.06361493269602457
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,512,10240,0.04216959873835246
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,512,7168,0.0165503998597463
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,512,7168,0.05378559827804565
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,512,8192,0.03381119966506958
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,512,7168,0.03223466674486796
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,512,6144,0.014672000209490457
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,512,6144,0.048241066932678225
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,512,5120,0.012850133577982583
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,512,5120,0.0436245322227478
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,512,5120,0.027192533016204834
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,512,6144,0.029242666562398274
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,512,4096,0.010710400342941285
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,512,4096,0.0361194650332133
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,512,3584,0.009873066345850627
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,512,51200,0.16374719937642415
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,512,3584,0.035021865367889406
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,512,3072,0.008796800176302593
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,512,3072,0.03089173237482707
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,512,4096,0.023986132939656575
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,512,2560,0.007947733501593272
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,512,3072,0.019828265905380248
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,512,2560,0.029025065898895263
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,512,2048,0.0070602665344874065
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,512,2048,0.026794666051864625
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,512,3584,0.021670399109522502
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,512,1536,0.006007466713587443
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,512,1536,0.024844799439112344
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,512,2560,0.018566399812698364
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,512,1024,0.005044266581535339
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,512,2048,0.016132266322771708
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,512,1024,0.022392533222834268
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,512,768,0.004649599889914194
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,512,768,0.020687999327977498
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,512,1024,0.013307733337084451
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,512,512,0.003828266759713491
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,512,768,0.012150399883588155
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,512,512,0.01872533361117045
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,512,256,0.0034143999218940735
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,512,256,0.017350399494171144
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,512,1536,0.014311466614405313
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,512,512,0.012455466389656066
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,512,128,0.016226133704185484
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,512,256,0.01243839959303538
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,512,64,0.0029919999341169994
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,512,128,0.012497066458066305
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,512,64,0.014587733149528503
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,512,32,0.0030762667457262674
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,512,32,0.014857600132624308
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,256,65536,0.10575786431630452
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,256,65536,0.5674805323282878
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,256,51200,0.08439253171284994
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,512,128,0.0030826665461063385
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,256,51200,0.44816106160481767
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,256,16384,0.03311893343925476
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,256,16384,0.15064746538798016
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,512,65536,0.20393706957499186
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,256,12288,0.024740266799926757
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,256,12288,0.10914026896158854
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,256,16384,0.061637334028879796
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,256,10240,0.020551466941833497
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,256,10240,0.08393493493398031
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,256,12288,0.04825173219045003
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,256,8192,0.017335466543833413
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,256,8192,0.060899198055267334
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,256,10240,0.04160106579462687
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,256,7168,0.014638933539390563
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,256,7168,0.04967146714528402
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,256,8192,0.03374293247858683
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,256,6144,0.013436800241470337
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,256,6144,0.04525119860967
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,256,7168,0.0317194660504659
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,256,5120,0.010872532924016316
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,256,5120,0.040958933035532635
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,256,6144,0.02881706754366557
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,256,4096,0.007579733431339264
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,256,4096,0.03397759993871053
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,256,5120,0.027236266930898027
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,256,3584,0.007112533350785573
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,256,3584,0.03284693360328674
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,256,4096,0.023204267024993896
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,256,3072,0.006318933268388112
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,256,3072,0.03076266646385193
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,256,51200,0.16368853251139323
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,256,2560,0.006165333092212677
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,256,3584,0.02148373325665792
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,256,2560,0.02887679934501648
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,256,2048,0.005356800059477488
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,256,3072,0.01950719952583313
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,256,2560,0.018590933084487914
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,256,2048,0.026586665709813433
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,256,1536,0.0046528001626332605
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,256,1536,0.023822933435440063
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,256,2048,0.015972266594568886
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,256,1024,0.003851733356714249
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,256,1024,0.0217738668123881
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,256,1024,0.013299199938774108
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,256,1536,0.01441493332386017
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,256,768,0.003419733295838038
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,256,768,0.020920532941818237
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,256,768,0.012417067090670269
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,256,512,0.0031658666829268134
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,256,512,0.018764799833297728
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,256,256,0.002885333448648453
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,256,65536,0.20407573382059732
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,256,256,0.01713493267695109
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,256,128,0.0028959999481836954
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,256,512,0.012611200412114462
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,256,128,0.01606186628341675
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,256,64,0.002665599932273229
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,256,64,0.015198933084805808
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,256,32,0.002696533252795537
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,256,128,0.012310399611790975
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,256,32,0.014565333724021912
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,128,65536,0.06897280216217042
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,128,65536,0.5611573537190755
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,128,51200,0.05552533467610678
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,256,256,0.01209493378798167
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,128,16384,0.021972266832987468
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,128,51200,0.4453472137451172
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,128,16384,0.14762667020161946
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,128,12288,0.012925866246223449
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,128,12288,0.10465706984202068
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,128,16384,0.059867731730143224
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,128,10240,0.011673600474993388
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,128,12288,0.046088532606760664
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,128,10240,0.08177066644032796
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,128,8192,0.012362666924794515
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,128,8192,0.05772906541824341
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,128,10240,0.03968426783879598
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,128,7168,0.010853333274523418
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,128,8192,0.032919466495513916
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,128,7168,0.04934293429056803
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,128,6144,0.009690666198730468
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,128,6144,0.043171199162801106
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,128,7168,0.03087573250134786
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,128,5120,0.008364799618721008
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,128,6144,0.028010666370391846
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,128,5120,0.039049601554870604
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,128,51200,0.16056426366170246
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,128,4096,0.007585066556930542
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,128,4096,0.0334666649500529
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,128,3584,0.006941866874694824
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,128,3584,0.03301546573638916
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,128,5120,0.025965867439905803
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,128,4096,0.02268799940745036
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,128,3072,0.006326400240262349
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,128,3072,0.030110933383305866
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,128,2560,0.0054613331953684485
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,128,65536,0.20094720522562662
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,128,3584,0.02062506675720215
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,128,2048,0.004659200211366018
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,128,3072,0.018551466862360637
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,128,2048,0.025766400496164958
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,128,1536,0.004242133100827535
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,128,1536,0.02289386590321859
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,128,2048,0.015238400300343832
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,128,1024,0.0034058667719364167
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,128,1024,0.02072426676750183
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,128,1536,0.013343999783198038
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,128,2560,0.017732266585032144
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,128,768,0.0033632000287373864
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,128,1024,0.012432000041007996
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,128,768,0.02071146567662557
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,128,512,0.003054933249950409
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,128,512,0.018667733669281004
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,128,256,0.0026538667579491934
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,128,512,0.011201066772143047
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,128,768,0.011725866794586181
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,128,256,0.01765759984652201
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,128,128,0.002771199991305669
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,128,256,0.011243733763694762
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,128,128,0.0151829332113266
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,128,2560,0.027540266513824463
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,128,32,0.002455466737349828
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,128,64,0.015050666530927024
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,128,32,0.014546133081118264
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,64,65536,0.06852693557739258
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,3072,128,128,0.011336533228556316
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,64,51200,0.0534495989481608
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,64,16384,0.01759679913520813
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,64,16384,0.14799572626749674
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,64,12288,0.01446399986743927
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,64,65536,0.5561333338419596
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,64,51200,0.4415285428365071
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,64,10240,0.012914133071899415
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,64,12288,0.10477973620096843
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,64,8192,0.010815999905268351
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,64,7168,0.00995733340581258
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,64,8192,0.05498453378677368
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,64,10240,0.07918933232625326
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,64,6144,0.009244799613952637
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,64,7168,0.04740053415298462
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,64,5120,0.008014933268229166
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,64,6144,0.0435263991355896
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,128,64,0.002194133400917053
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,64,4096,0.006817066669464111
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,64,3584,0.006442666550477346
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,64,4096,0.03290773431460063
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,64,3072,0.005514666438102722
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,64,3584,0.032902399698893234
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,64,3072,0.028856533765792846
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,64,2560,0.005157333115736643
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,64,2048,0.004660266637802124
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,64,2560,0.027124265829722088
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,64,1536,0.004060799876848856
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,64,1536,0.023231999079386393
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,64,2048,0.02537386616071065
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,64,1024,0.0034080001215140024
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,64,768,0.003036800026893616
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,64,768,0.019437867403030395
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,64,1024,0.021686400969823202
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,64,512,0.0026954665780067446
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,64,256,0.002570666621128718
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,64,512,0.018675200144449868
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,64,128,0.0023189333577950796
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,64,5120,0.03903786738713582
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,64,128,0.015289599696795145
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,64,64,0.002182399978240331
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,64,32,0.0022314667701721193
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,64,64,0.01456106702486674
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,64,32,0.014513066411018372
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,32,65536,0.06432000001271566
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,32,51200,0.05157013336817423
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,32,65536,0.554366938273112
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,32,16384,0.016292267044385276
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,32,51200,0.43970667521158857
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,32,12288,0.009737599889437358
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,32,16384,0.14436373710632325
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,64,256,0.016555733482042947
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,32,10240,0.00874133308728536
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,32,8192,0.007923200229803721
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,32,12288,0.10106453100840251
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,32,10240,0.07947093645731608
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,32,6144,0.006770133475462596
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,32,8192,0.055934933821360264
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,32,7168,0.04873600006103516
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,32,6144,0.04317546685536702
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,32,5120,0.006491733094056447
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,32,4096,0.008214400211970011
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,32,3584,0.007523199915885926
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,32,4096,0.03331200083096822
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,32,5120,0.040573867162068684
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,32,3584,0.03144853313763936
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,32,3072,0.0068693334857622785
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,32,2560,0.0061141331990559895
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,32,2048,0.004669866462548574
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,32,2560,0.027602134148279826
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,32,3072,0.02927253246307373
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,32,2048,0.02486720085144043
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,32,1536,0.003915733347336451
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,32,1024,0.0034101332227389016
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,32,768,0.0030026666820049284
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,32,1536,0.023117866118748984
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,32,1024,0.021347200870513915
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,32,512,0.0027306665976842242
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,32,768,0.019374932845433554
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,32,256,0.002570666621128718
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,32,512,0.018675200144449868
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,32,256,0.016778665781021117
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,32,128,0.002462933212518692
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,32,64,0.0021717332303524016
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,32,32,0.0021738665799299877
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,32,64,0.014730667074521383
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,32,32,0.014594133694966635
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,3072,32,7168,0.0071733335653940845
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,3072,32,128,0.015075199802716575
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,65536,16384,3.4134752909342447
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,65536,51200,5.905622355143229
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,65536,16384,1.8310591379801433
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,65536,12288,2.564488474527995
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,65536,51200,11.141142781575521
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,65536,12288,1.3682613372802734
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,65536,10240,2.2484405517578123
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,65536,10240,1.1540693918863931
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,65536,10240,0.6466943740844726
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,65536,8192,1.7151349385579426
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,65536,8192,0.8722133636474609
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,65536,12288,0.7296768188476562
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,65536,7168,1.4181514739990235
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,65536,7168,0.774066162109375
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,65536,16384,0.9760576248168945
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,65536,6144,1.2507092793782553
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,65536,6144,0.6694922765096029
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,65536,8192,0.5311957359313965
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,65536,5120,1.049062410990397
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,65536,5120,0.5617994944254557
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,65536,7168,0.48775466283162433
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,65536,4096,0.7928917566935222
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,65536,4096,0.45658133824666336
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,65536,6144,0.4187850634256999
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,65536,3584,0.708407465616862
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,65536,3584,0.40149758656819656
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,65536,5120,0.3818122545878092
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,65536,3072,0.5937706629435222
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,65536,3072,0.3446592013041178
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,65536,4096,0.33856531778971355
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,65536,2560,0.49956480662027997
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,65536,2560,0.2968319892883301
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,65536,3584,0.3075455983479818
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,65536,2048,0.40584853490193684
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,65536,2048,0.25124692916870117
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,65536,3072,0.2701727867126465
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,65536,1536,0.2985408147176107
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,65536,1536,0.19128959973653156
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,65536,2560,0.25586026509602866
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,65536,1024,0.21662507057189942
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,65536,1024,0.1427786668141683
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,65536,1024,0.18974612553914388
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,65536,768,0.16936426162719725
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,65536,768,0.12044266859690349
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,65536,1536,0.19383360544840494
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,65536,512,0.12129173278808594
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,65536,512,0.10071679751078289
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,65536,2048,0.22675414085388185
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,65536,256,0.08295146624247232
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,65536,256,0.08542079925537109
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,65536,768,0.18054186503092448
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,65536,128,0.0684554656346639
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,65536,128,0.07428267002105712
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,65536,512,0.1770378589630127
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,65536,64,0.059886932373046875
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,65536,64,0.0767626682917277
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,65536,32,0.06291520198186239
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,65536,32,0.07633066972096761
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,65536,256,0.17345280647277833
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,65536,128,0.18291093508402506
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,51200,65536,5.950948079427084
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,51200,65536,11.09131368001302
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,51200,51200,8.610335286458334
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,51200,51200,4.553416442871094
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,51200,16384,2.5640884399414063
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,51200,16384,1.3908096313476563
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,51200,16384,0.7677696228027344
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,51200,12288,1.9804074605305988
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,51200,12288,1.0537759780883789
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,51200,12288,0.5773429234822591
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,51200,10240,1.6218805948893231
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,51200,10240,0.8910346349080405
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,51200,10240,0.5120010693868001
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,51200,8192,1.2994752248128256
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,51200,8192,0.6961952209472656
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,65536,51200,3.1923594156901043
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,51200,7168,1.1135723114013671
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,51200,7168,0.6319007873535156
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,51200,8192,0.4175423940022786
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,51200,6144,0.9213471730550131
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,51200,6144,0.5174933433532715
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,51200,7168,0.38821118672688804
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,51200,5120,0.7823541641235352
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,51200,5120,0.4584757486979167
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,51200,6144,0.329637336730957
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,51200,4096,0.613922119140625
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,51200,4096,0.36735893885294596
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,51200,5120,0.29889920552571614
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,51200,3584,0.5507530848185221
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,51200,3584,0.32068265279134117
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,51200,51200,2.5491668701171877
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,51200,3072,0.4734058698018392
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,51200,3072,0.2793792088826498
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,51200,3584,0.24189653396606445
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,51200,2560,0.39157012303670247
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,51200,2560,0.237664000193278
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,51200,4096,0.26424214045206706
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,51200,2048,0.30587733586629234
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,51200,2048,0.1950485388437907
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,51200,2560,0.20351039568583168
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,51200,1536,0.24070080121358237
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,51200,1536,0.15543467203776043
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,51200,2048,0.17919573783874512
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,51200,1024,0.1666709264119466
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,51200,1024,0.1163103977839152
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,51200,3072,0.21222826639811196
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,51200,768,0.13376426696777344
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,51200,768,0.09800426959991455
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,51200,1024,0.15354453722635905
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,51200,512,0.09541760285695394
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,51200,512,0.08327999909718832
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,51200,768,0.14254506429036456
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,51200,256,0.06544106801350912
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,51200,256,0.07042773564656576
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,51200,256,0.1409429391225179
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,51200,128,0.05673386653264364
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,51200,128,0.06574079990386963
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,51200,1536,0.15717546145121258
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,51200,64,0.050979200998942056
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,51200,64,0.0682645320892334
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,51200,32,0.053865599632263186
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,51200,32,0.06798293590545654
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,51200,128,0.14343892733256022
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,51200,512,0.14192214012145996
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,16384,65536,3.56891835530599
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,16384,65536,2.1591957092285154
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,16384,51200,2.756956736246745
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,16384,51200,1.6754709879557292
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,16384,16384,0.820518430074056
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,16384,16384,0.5482698440551758
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,16384,16384,0.27852160135904946
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,16384,12288,0.5996810913085937
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,16384,12288,0.42746454874674483
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,16384,12288,0.2121386686960856
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,16384,10240,0.5529333114624023
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,16384,10240,0.3691295941670736
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,16384,10240,0.18863360087076825
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,16384,8192,0.43938347498575847
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,16384,8192,0.29276266098022463
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,16384,8192,0.15480213165283202
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,16384,7168,0.38343893686930336
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,16384,7168,0.24852906862894691
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,16384,7168,0.14250559806823732
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,16384,6144,0.3288842519124349
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,16384,6144,0.21179839769999184
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,16384,51200,0.8340106964111328
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,16384,5120,0.27128000259399415
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,16384,5120,0.19087893168131512
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,16384,5120,0.1134335994720459
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,16384,4096,0.21653119723002115
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,16384,4096,0.15032745997111002
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,16384,6144,0.12314773400624593
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,16384,3584,0.1878698666890462
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,16384,3584,0.12408426602681477
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,16384,3584,0.08879679838816325
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,16384,3072,0.16174933115641277
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,16384,3072,0.1060970703760783
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,16384,4096,0.09999039967854818
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,16384,2560,0.13626880645751954
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,16384,2560,0.0963690678278605
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,16384,3072,0.07837759653727214
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,16384,2048,0.10927146275838215
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,16384,2048,0.07847359975179037
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,16384,2560,0.07590506871541342
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,16384,1536,0.08042346636454265
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,16384,1536,0.06392319997151694
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,16384,1536,0.05538239876429239
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,16384,1024,0.05649493137995402
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,16384,1024,0.05347733497619629
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,16384,2048,0.06408960024515788
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,16384,768,0.04419306516647339
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,16384,768,0.04809600114822388
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,16384,1024,0.053284267584482826
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,16384,512,0.03309546709060669
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,16384,512,0.0418442686398824
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,16384,768,0.04586986700693767
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,16384,256,0.02278719941775004
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,16384,256,0.0348309318224589
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,16384,512,0.045613865057627365
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,16384,128,0.01955519914627075
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,16384,128,0.031362134218215945
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,16384,256,0.04444160064061482
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,16384,64,0.016547200083732606
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,16384,64,0.030532266696294146
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,16384,32,0.016956800222396852
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,16384,32,0.02972480058670044
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,16384,128,0.046139732996622725
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,12288,65536,2.576056416829427
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,12288,65536,1.6830059051513673
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,12288,51200,1.9655423482259113
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,12288,51200,1.3032341003417969
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,16384,65536,1.0525642395019532
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,12288,16384,0.5675285339355469
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,12288,16384,0.41949227650960286
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,51200,65536,3.1886367797851562
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,12288,12288,0.4333173433939616
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,12288,12288,0.31242345174153646
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,12288,12288,0.17545493443806964
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,12288,10240,0.36286932627360025
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,12288,10240,0.2558826605478922
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,12288,16384,0.22936639785766602
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,12288,8192,0.28814932505289714
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,12288,8192,0.2198474725087484
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,12288,8192,0.12778133551279705
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,12288,7168,0.2529685338338216
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,12288,7168,0.1860138734181722
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,12288,10240,0.15602134068806967
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,12288,6144,0.22152214050292968
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,12288,6144,0.16141440073649088
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,12288,7168,0.11887359619140625
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,12288,5120,0.1835349400838216
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,12288,5120,0.13834880193074545
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,12288,6144,0.10316906770070393
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,12288,4096,0.1508992036183675
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,12288,4096,0.11503146489461262
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,12288,5120,0.09447680314381918
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,12288,3584,0.13064746856689452
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,12288,3584,0.1017685333887736
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,12288,4096,0.08224106629689534
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,12288,3072,0.11417919794718426
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,12288,3072,0.08867093722025553
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,12288,3072,0.06513706843058267
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,12288,2560,0.09606080055236817
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,12288,2560,0.08019413153330485
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,12288,3584,0.07258133093516031
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,12288,2048,0.07738666534423828
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,12288,2048,0.06737813154856363
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,12288,2560,0.06335893472035727
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,12288,1536,0.0599616010983785
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,12288,1536,0.05436266660690307
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,12288,2048,0.05669333140055338
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,12288,1024,0.0415669322013855
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,12288,1024,0.04869759877522786
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,12288,1024,0.05058559974034628
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,12288,768,0.03269439935684204
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,12288,1536,0.051736533641815186
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,12288,768,0.042728531360626223
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,12288,512,0.027298132578531902
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,12288,512,0.03607573509216309
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,12288,512,0.046803200244903566
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,12288,256,0.018654932578404747
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,12288,256,0.030642133951187134
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,12288,768,0.04716053406397502
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,12288,128,0.01721280018488566
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,12288,128,0.029813333352406816
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,12288,256,0.046326398849487305
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,12288,64,0.01492586632569631
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,12288,64,0.028810666004816694
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,12288,32,0.015422933300336204
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,12288,32,0.029716267188390093
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,12288,128,0.04628373384475708
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,10240,65536,2.223901875813802
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,10240,65536,1.5235125223795571
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,10240,51200,1.6157802581787108
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,10240,51200,1.188385009765625
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,12288,51200,0.685254414876302
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,10240,16384,0.5060127894083659
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,10240,16384,0.38940693537394205
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,10240,16384,0.20399893124898277
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,10240,12288,0.38471145629882814
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,10240,12288,0.2873642603556315
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,12288,65536,0.8633941650390625
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,10240,10240,0.3201525370279948
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,10240,10240,0.22846080462137858
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,10240,12288,0.1562506675720215
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,10240,8192,0.2534997304280599
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,10240,8192,0.1860607941945394
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,10240,10240,0.13853972752888996
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,10240,7168,0.2194357395172119
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,10240,7168,0.162718931833903
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,10240,8192,0.11334293683369953
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,10240,6144,0.1888256072998047
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,10240,6144,0.14091839790344238
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,10240,7168,0.10498773256937663
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,10240,5120,0.15773226420084635
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,10240,5120,0.1187882661819458
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,10240,6144,0.09192960262298584
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,10240,4096,0.12677013079325358
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,10240,4096,0.1006229321161906
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,10240,5120,0.08351466655731202
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,10240,3584,0.11080959637959797
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,10240,3584,0.08816213607788086
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,10240,3584,0.06481493314107259
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,10240,3072,0.09549333254496256
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,10240,3072,0.0766538699467977
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,10240,4096,0.07243306636810302
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,10240,2560,0.08103253046671549
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,10240,2560,0.0707039992014567
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,10240,51200,0.6047391891479492
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,10240,2048,0.06523946523666382
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,10240,2048,0.05963413317998251
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,10240,2560,0.05646400054295858
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,10240,1536,0.05056426525115967
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,10240,1536,0.048120534420013426
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,10240,3072,0.05789866844813028
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,10240,1024,0.03507839838663737
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,10240,1024,0.04027200142542521
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,10240,2048,0.05075200001398722
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,10240,768,0.02733653386433919
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,10240,768,0.0359605352083842
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,10240,1024,0.04333653450012207
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,10240,512,0.019988266626993816
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,10240,512,0.030980267127354938
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,10240,1536,0.044707198937733963
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,10240,256,0.014892799655596414
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,10240,256,0.026820266246795656
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,10240,256,0.040223999818166094
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,10240,128,0.012056533495585125
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,10240,128,0.023285333315531412
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,10240,128,0.03999893267949422
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,10240,64,0.010792533556620281
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,10240,512,0.040232535203297934
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,10240,32,0.01076586643854777
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,10240,64,0.023256532351175942
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,10240,32,0.02394346594810486
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,10240,768,0.040445868174235025
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,8192,65536,1.3144373575846353
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,8192,65536,1.7708437601725262
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,8192,51200,1.3034912109375
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,8192,51200,1.0507498423258463
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,8192,16384,0.4211605389912923
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,8192,16384,0.33293546040852867
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,8192,16384,0.1589344024658203
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,8192,12288,0.29524161020914713
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,8192,12288,0.23969066937764488
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,10240,65536,0.7629610697428386
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,8192,10240,0.25764479637146
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,8192,10240,0.19775360425313313
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,8192,10240,0.10893332958221436
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,8192,8192,0.2042463938395182
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,8192,8192,0.1576181411743164
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,8192,12288,0.12198293209075928
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,8192,7168,0.18165866533915204
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,8192,7168,0.14494825998942057
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,8192,7168,0.08394026756286621
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,8192,6144,0.15527787208557128
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,8192,6144,0.12331946690877278
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,8192,8192,0.08944533665974935
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,8192,5120,0.1248960018157959
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,8192,5120,0.10523946285247802
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,8192,5120,0.06662826538085938
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,8192,4096,0.09994346300760905
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,8192,4096,0.08945706685384115
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,8192,6144,0.0730133295059204
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,8192,3584,0.09110079606374105
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,8192,3584,0.08035199642181397
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,8192,3584,0.05071893135706583
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,8192,3072,0.07897066275278727
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,8192,3072,0.07019626299540202
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,8192,4096,0.05603520075480143
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,8192,2560,0.06445120175679525
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,8192,2560,0.062495998541514074
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,8192,2560,0.04283413489659627
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,8192,2048,0.0522762656211853
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,8192,2048,0.052780799070994054
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,8192,3072,0.04456959962844849
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,8192,1536,0.04202026526133219
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,8192,1536,0.04301120042800903
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,8192,1536,0.03299946586290996
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,8192,1024,0.028959999481836956
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,8192,1024,0.03716266552607218
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,8192,2048,0.037862400213877365
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,8192,768,0.022294400135676067
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,8192,768,0.03309119939804077
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,8192,1024,0.03154559930165608
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,8192,512,0.017240534226099648
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,8192,512,0.029582933584849043
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,8192,768,0.028895999987920123
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,8192,256,0.012935466567675271
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,8192,256,0.026162133614222212
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,8192,512,0.028500266869862872
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,8192,128,0.010565333565076192
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,8192,128,0.022111999988555908
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,8192,256,0.028035199642181395
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,8192,64,0.009251200159390767
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,8192,64,0.022707200050354003
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,8192,32,0.00957973301410675
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,8192,32,0.02158400019009908
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,8192,128,0.02858346700668335
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,7168,65536,1.479431406656901
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,7168,65536,1.2038740793863931
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,7168,51200,1.1131434122721353
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,7168,51200,0.9605557123819987
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,8192,51200,0.4607978820800781
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,7168,16384,0.3433482805887858
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,7168,16384,0.30791893005371096
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,7168,16384,0.15781013170878094
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,7168,12288,0.25924800237019857
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,7168,12288,0.22787839571634927
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,8192,65536,0.579095458984375
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,7168,10240,0.22987306912740074
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,7168,10240,0.18117760022481283
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,7168,12288,0.12089707056681316
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,7168,8192,0.18384000460306804
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,7168,8192,0.14631253878275555
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,7168,10240,0.10909012953440349
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,7168,7168,0.15788693428039552
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,7168,7168,0.13025706609090168
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,7168,8192,0.08940587043762208
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,7168,6144,0.1350111961364746
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,7168,6144,0.11116906801859539
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,7168,7168,0.08356053034464518
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,7168,5120,0.11099413235982258
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,7168,5120,0.09841706752777099
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,7168,6144,0.07263466517130533
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,7168,4096,0.08935999870300293
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,7168,4096,0.08325013319651285
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,7168,5120,0.06555840174357096
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,7168,3584,0.07841386795043945
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,7168,3584,0.07542719841003417
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,7168,4096,0.055801598230997715
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,7168,3072,0.06833066940307617
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,7168,3072,0.06430399815241496
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,7168,3584,0.051006933053334556
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,7168,2560,0.05727893511454264
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,7168,2560,0.05786879857381185
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,7168,3072,0.045431466897328694
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,7168,2048,0.046728531519571945
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,7168,2048,0.04913066625595093
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,7168,2560,0.04410773515701294
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,7168,1536,0.03605653444925944
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,7168,1536,0.03936426639556885
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,7168,1536,0.03488746484120687
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,7168,1024,0.0252074658870697
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,7168,1024,0.03517119884490967
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,7168,2048,0.03981866836547852
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,7168,51200,0.46144854227701826
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,7168,768,0.020282665888468422
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,7168,768,0.03160213430722554
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,7168,512,0.016005333264668783
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,7168,512,0.028804266452789308
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,7168,768,0.031345067421595256
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,7168,256,0.01125759979089101
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,7168,256,0.023602133989334105
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,7168,1024,0.033768534660339355
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,7168,128,0.010455466310183207
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,7168,128,0.021875200668970744
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,7168,512,0.0311573326587677
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,7168,64,0.010058666268984478
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,7168,128,0.030983465909957885
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,7168,32,0.01011946698029836
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,7168,32,0.020744532346725464
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,7168,64,0.024218666553497314
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,7168,256,0.030828799804051715
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,6144,65536,1.1924821217854817
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,6144,65536,1.0544501622517903
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,6144,51200,0.8279136021931967
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,6144,51200,0.9477557500203451
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,6144,16384,0.3248096148173014
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,6144,16384,0.29136854807535806
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,6144,16384,0.13423892656962078
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,6144,12288,0.21803199450174965
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,6144,12288,0.19692586263020834
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,7168,65536,0.5792640050252278
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,6144,10240,0.1783786614735921
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,6144,10240,0.16244373321533204
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,6144,12288,0.10395306746164959
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,6144,8192,0.14737173716227214
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,6144,8192,0.13364373842875163
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,6144,10240,0.09390933513641357
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,6144,7168,0.12686506907145184
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,6144,7168,0.12016212940216064
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,6144,7168,0.07145280043284098
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,6144,6144,0.10964480241139729
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,6144,8192,0.07702293395996093
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,6144,6144,0.10437226295471191
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,6144,5120,0.09245759646097819
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,6144,5120,0.09167893727620444
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,6144,5120,0.05650346676508585
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,6144,4096,0.07496639887491861
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,6144,4096,0.07752319971720377
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,6144,6144,0.06256959835688272
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,6144,3584,0.0671189308166504
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,6144,3584,0.06786879698435465
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,6144,4096,0.04894506533940633
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,6144,3072,0.05761919816335043
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,6144,3072,0.05980053345362345
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,6144,3584,0.04410560131072998
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,6144,2560,0.04948800007502238
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,6144,2560,0.05218026638031006
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,6144,2560,0.03799253304799398
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,6144,2048,0.040345601240793866
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,6144,2048,0.045261867841084796
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,6144,3072,0.03964693148930867
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,6144,1536,0.03095039923985799
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,6144,1536,0.03716053167978923
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,6144,1536,0.030898133913675945
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,6144,1024,0.021950932343800862
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,6144,1024,0.034031999111175534
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,6144,2048,0.03416320085525513
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,6144,768,0.01758506695429484
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,6144,768,0.029475200176239013
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,6144,1024,0.02975146571795146
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,6144,512,0.014180266857147216
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,6144,512,0.02781760096549988
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,6144,768,0.02758293350537618
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,6144,256,0.010366933544476827
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,6144,256,0.023066665728886923
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,6144,256,0.027156267563501996
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,6144,128,0.009550933043162029
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,6144,128,0.022218666474024453
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,6144,512,0.02760533293088277
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,6144,64,0.008417066931724549
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,6144,64,0.021195733547210695
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,6144,32,0.008757332960764568
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,6144,128,0.02718399961789449
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,6144,32,0.020744532346725464
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,5120,65536,0.9838762919108073
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,5120,65536,0.9753994623819986
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,5120,51200,0.7524010976155598
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,5120,51200,0.7988330841064453
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,6144,51200,0.3892287890116374
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,5120,16384,0.24721813201904297
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,5120,16384,0.25010879834493005
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,5120,16384,0.13085760275522867
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,5120,12288,0.18670719464619953
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,5120,12288,0.18151572545369465
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,6144,65536,0.48946558634440107
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,5120,10240,0.16305599212646485
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,5120,10240,0.14720212618509929
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,5120,12288,0.09923840363820394
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,5120,8192,0.13208959897359213
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,5120,8192,0.1197856028874715
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,5120,10240,0.08841599623362223
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,5120,7168,0.11362240314483643
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,5120,7168,0.10770239830017089
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,5120,7168,0.06737600167592367
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,5120,8192,0.0738368034362793
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,5120,6144,0.09736320177714029
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,5120,6144,0.09400106271107991
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,5120,5120,0.08029546737670898
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,5120,5120,0.08171839714050293
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,5120,5120,0.051533865928649905
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,5120,4096,0.06375466585159302
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,5120,4096,0.06812907059987386
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,5120,6144,0.057706665992736814
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,5120,3584,0.05700266758600871
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,5120,3584,0.05875733296076456
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,5120,4096,0.04517333507537842
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,5120,3584,0.04089173475901286
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,5120,3072,0.05159786542256674
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,5120,3072,0.04845120112101237
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,5120,2560,0.04040426810582479
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,5120,2560,0.04695359865824382
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,5120,2560,0.03559039831161499
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,5120,2048,0.032857600847880045
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,5120,2048,0.041124268372853594
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,5120,3072,0.03707733154296875
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,5120,1536,0.02571093241373698
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,5120,1536,0.034168533484141034
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,5120,1536,0.028068266312281293
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,5120,1024,0.018237866957982383
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,5120,1024,0.03004693388938904
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,5120,2048,0.03156906763712565
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,5120,768,0.014266666769981385
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,5120,768,0.02775146762530009
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,5120,51200,0.38649174372355144
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,5120,512,0.010974933703740437
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,5120,512,0.024689066410064697
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,5120,1024,0.02727573315302531
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,5120,256,0.008308266599973042
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,5120,256,0.020782933632532755
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,5120,768,0.029296000798543293
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,5120,128,0.007132799923419952
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,5120,512,0.029313067595163982
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,5120,128,0.01972480018933614
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,5120,64,0.0063701331615448
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,5120,256,0.028968532880147297
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,5120,32,0.006718933085600535
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,5120,64,0.019921066363652547
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,5120,32,0.018598399559656777
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,5120,128,0.02849919994672139
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,4096,65536,0.7867210388183594
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,4096,65536,0.8898261388142904
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,4096,51200,0.6695498784383138
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,4096,51200,0.6984618504842122
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,4096,16384,0.201964807510376
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,4096,16384,0.2252906640370687
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,4096,16384,0.10889279842376709
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,4096,12288,0.15339093208312987
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,4096,12288,0.16252907117207843
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,5120,65536,0.4903552055358887
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,4096,10240,0.13444053332010905
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,4096,10240,0.13372586568196615
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,4096,12288,0.08409066994984946
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,4096,8192,0.10800746281941731
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,4096,8192,0.11014933586120605
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,4096,10240,0.07647039890289306
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,4096,7168,0.09302826722462973
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,4096,7168,0.09745173454284668
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,4096,7168,0.057690668106079104
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,4096,6144,0.0782975991566976
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,4096,6144,0.0842080036799113
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,4096,8192,0.0628383994102478
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,4096,5120,0.06613119840621948
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,4096,5120,0.07400320370992025
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,4096,5120,0.045577601591746016
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,4096,4096,0.05301760037740072
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,4096,4096,0.05984746615091959
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,4096,6144,0.04991146723429362
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,4096,3584,0.04602453311284383
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,4096,3584,0.053668268521626795
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,4096,3584,0.035571201642354326
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,4096,3072,0.03997759819030762
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,4096,3072,0.04733546574910481
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,4096,4096,0.039716267585754396
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,4096,2560,0.03419306675593058
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,4096,2560,0.04392000039418538
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,4096,2560,0.03054400086402893
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,4096,2048,0.028310400247573853
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,4096,2048,0.03858773310979207
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,4096,3072,0.03178880016009013
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,4096,1536,0.022231467564900718
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,4096,1536,0.031078400214513142
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,4096,1536,0.02269759972890218
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,4096,1024,0.01542080044746399
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,4096,1024,0.02863146662712097
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,4096,2048,0.026143999894460042
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,4096,768,0.012388267119725545
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,4096,768,0.025290666023890178
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,4096,1024,0.02188906669616699
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,4096,512,0.009612799684206644
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,4096,512,0.022789333264033
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,4096,768,0.02111999988555908
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,4096,256,0.007526400188604991
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,4096,512,0.02065920035044352
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,4096,256,0.0206112007300059
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,4096,256,0.021346133947372437
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,4096,128,0.007086933155854543
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,4096,128,0.019048533837000527
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,4096,128,0.019607466459274293
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,4096,64,0.006236800054709116
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,4096,64,0.01858560045560201
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,4096,32,0.006308266520500183
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,4096,32,0.02001706759134928
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,3584,65536,0.7102794647216797
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,3584,65536,0.8491551717122396
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,3584,51200,0.5537408192952473
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,3584,51200,0.6682463963826497
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,4096,51200,0.3144319852193197
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,3584,16384,0.17866133054097494
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,3584,16384,0.2108544031778971
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,3584,16384,0.1088597297668457
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,3584,12288,0.13456106185913086
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,3584,12288,0.15263039271036785
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,4096,65536,0.3962261199951172
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,3584,10240,0.11694080034891766
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,3584,10240,0.12902506987253826
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,3584,12288,0.08480640252431235
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,3584,8192,0.09595839977264405
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,3584,8192,0.10512746969858806
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,3584,10240,0.07726399898529053
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,3584,7168,0.08352213700612386
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,3584,7168,0.09446506500244141
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,3584,7168,0.05813013315200806
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,3584,6144,0.07127146720886231
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,3584,6144,0.0807925303777059
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,3584,8192,0.06297599871953329
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,3584,5120,0.05960533221562704
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,3584,5120,0.07005226612091064
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,3584,5120,0.0465173323949178
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,3584,4096,0.04732053279876709
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,3584,4096,0.05773653189341227
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,3584,6144,0.05081493457158407
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,3584,3584,0.041121065616607666
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,3584,3584,0.05153706471125284
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,3584,4096,0.04040640195210775
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,3584,3584,0.03628586530685425
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,3584,3072,0.03613226811091105
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,3584,3072,0.04790613253911336
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,3584,2560,0.030027733246485395
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,3584,2560,0.04303786754608154
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,3584,2560,0.03146026730537414
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,3584,2048,0.024819199244181314
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,3584,2048,0.03715733289718628
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,3584,3072,0.032561065753300984
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,3584,1536,0.02020053267478943
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,3584,1536,0.032686932881673174
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,3584,2048,0.02808000048001607
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,3584,1024,0.014938666423161825
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,3584,1024,0.029734400908152263
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,3584,51200,0.3154378573099772
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,3584,768,0.012379733721415202
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,3584,768,0.026089600721995038
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,3584,1024,0.0240447998046875
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,3584,512,0.00956053336461385
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,3584,512,0.02319999933242798
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,3584,768,0.022629332542419434
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,3584,256,0.007117866476376851
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,3584,256,0.022014933824539184
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,3584,1536,0.025332266092300416
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,3584,128,0.006306133170922597
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,3584,128,0.018592000007629395
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,3584,512,0.022408533096313476
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,3584,64,0.00547626664241155
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,3584,64,0.019349332650502524
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,3584,256,0.021925334135691324
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,3584,32,0.00559146652619044
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,3584,32,0.018924800554911296
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,3584,128,0.02226346731185913
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,3072,65536,0.5492874781290691
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,3072,65536,0.7699296315511067
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,3072,51200,0.44147520065307616
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,3072,51200,0.6070538838704427
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,3072,16384,0.143505064646403
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,3072,16384,0.1964853286743164
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,3072,16384,0.08764159679412842
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,3072,12288,0.10925013224283855
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,3072,12288,0.1430506706237793
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,3584,65536,0.39696321487426756
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,3072,10240,0.0982485294342041
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,3072,10240,0.11791786352793376
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,3072,10240,0.0632149338722229
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,3072,8192,0.07893760204315185
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,3072,8192,0.09670506318410238
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,3072,12288,0.06919679641723633
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,3072,7168,0.07020479838053385
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,3072,7168,0.08779626687367757
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,3072,8192,0.05100053151448568
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,3072,6144,0.05841066837310791
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,3072,6144,0.07353280385335287
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,3072,7168,0.04821759859720866
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,3072,5120,0.049004801114400226
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,3072,5120,0.06354346672693888
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,3072,5120,0.0391487995783488
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,3072,6144,0.04214293162027995
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,3072,4096,0.038933332761128744
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,3072,4096,0.05252159833908081
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,3072,3584,0.03509333531061808
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,3072,3584,0.04895253181457519
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,3072,3584,0.031221334139506025
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,3072,3072,0.030466133356094362
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,3072,3072,0.04304320017496745
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,3072,4096,0.03389546473821004
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,3072,2560,0.025310933589935303
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,3072,2560,0.0390559991200765
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,3072,2560,0.02690666715304057
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,3072,2048,0.021065600713094077
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,3072,2048,0.03501760164896647
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,3072,3072,0.027689599990844728
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,3072,1536,0.016945066054662068
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,3072,1536,0.030029867092768354
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,3072,1536,0.021735467513402305
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,3072,1024,0.012418133020401002
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,3072,1024,0.02687573234240214
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,3072,2048,0.02410773237546285
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,3072,768,0.009771733482678732
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,3072,768,0.024845866362253825
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,3072,1024,0.021028266350428263
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,3072,512,0.00763626645008723
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,3072,512,0.022436267137527464
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,3072,768,0.019381332397460937
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,3072,256,0.006292266647020976
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,3072,256,0.018986666202545167
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,3072,256,0.019761067628860474
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,3072,128,0.0058773333827654515
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,3072,512,0.019458132982254028
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,3072,128,0.019483733177185058
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,3072,64,0.005058133105436961
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,3072,64,0.01789439916610718
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,3072,32,0.005453866720199585
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,3072,32,0.020078933238983153
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,3072,128,0.019154133399327596
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,2560,65536,0.4828298568725586
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,2560,65536,0.7323989232381185
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,2560,51200,0.3810784022013346
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,2560,51200,0.6060949325561523
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,3072,51200,0.23838720321655274
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,2560,16384,0.12137813568115234
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,2560,16384,0.1855669339497884
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,2560,16384,0.0854314645131429
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,2560,12288,0.09219199816385905
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,2560,12288,0.1349247932434082
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,3072,65536,0.2972991943359375
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,2560,10240,0.0807861328125
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,2560,10240,0.10997653007507324
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,2560,12288,0.06744320392608642
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,2560,8192,0.06429546674092611
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,2560,8192,0.08954880237579346
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,2560,10240,0.06148266792297363
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,2560,7168,0.05661760171254476
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,2560,7168,0.08040213584899902
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,2560,8192,0.05021760066350302
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,2560,6144,0.047295999526977536
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,2560,6144,0.06677866776784261
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,2560,7168,0.046794664859771726
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,2560,5120,0.03952320019404094
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,2560,5120,0.057505067189534506
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,2560,6144,0.04200213352839152
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,2560,4096,0.031470932563145954
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,2560,4096,0.04799893299738566
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,2560,5120,0.03838506539662679
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,2560,3584,0.028310400247573853
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,2560,3584,0.04444906711578369
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,2560,4096,0.032604799667994185
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,2560,3072,0.024321067333221435
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,2560,3072,0.03986560106277466
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,2560,3584,0.030070400238037108
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,2560,2560,0.021303466955820718
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,2560,2560,0.037272532780965165
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,2560,3072,0.02688639958699544
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,2560,2048,0.01741866668065389
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,2560,2048,0.03298880060513814
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,2560,51200,0.23839573860168456
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,2560,2560,0.025230934222539265
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,2560,1536,0.014591999848683677
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,2560,1536,0.027661865949630736
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,2560,1024,0.010002133250236512
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,2560,1024,0.02555626630783081
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,2560,2048,0.022670932610829673
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,2560,768,0.00840106705824534
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,2560,768,0.023299199342727662
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,2560,1536,0.020334933201471964
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,2560,512,0.006734933455785115
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,2560,512,0.020906666914621987
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,2560,1024,0.019429334004720054
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,2560,256,0.00550186683734258
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,2560,256,0.018760534127553304
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,2560,768,0.01797013282775879
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,2560,128,0.005041066805521647
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,2560,128,0.01773759921391805
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,2560,256,0.017500799894332886
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,2560,64,0.004659200211366018
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,2560,64,0.01752426624298096
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,2560,32,0.004791466891765595
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,2560,512,0.01778986652692159
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,2560,32,0.016847999890645345
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,2560,128,0.017520000537236534
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,2048,65536,0.4022207895914714
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,2048,65536,0.6750794728597005
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,2048,51200,0.29606720606486003
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,2048,51200,0.533633041381836
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,2048,16384,0.09899093310038248
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,2048,16384,0.1718762715657552
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,2560,65536,0.297706667582194
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,2048,12288,0.0750272035598755
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,2048,12288,0.12525866826375326
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,2048,16384,0.08409919738769531
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,2048,10240,0.06769066651662191
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,2048,10240,0.10487679640452068
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,2048,10240,0.05889600118001302
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,2048,8192,0.05349440177281698
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,2048,8192,0.08440426985422769
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,2048,12288,0.06630719900131225
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,2048,7168,0.046011734008789065
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,2048,7168,0.07359146277109782
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,2048,7168,0.04452053308486938
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,2048,6144,0.03920853137969971
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,2048,6144,0.060577066739400234
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,2048,8192,0.04845013221104939
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,2048,5120,0.03285013238588969
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,2048,5120,0.05305600166320801
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,2048,5120,0.03606186707814534
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,2048,4096,0.026682666937510174
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,2048,4096,0.04591999848683675
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,2048,6144,0.039559467633565264
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,2048,3584,0.024055467049280802
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,2048,3584,0.041266131401062014
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,2048,3584,0.02834239999453227
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,2048,3072,0.02090346614519755
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,2048,4096,0.03115306695302327
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,2048,3072,0.037596801916758224
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,2048,2560,0.018145066499710084
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,2048,2560,0.03442560036977132
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,2048,2560,0.02391786575317383
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,2048,2048,0.015082666277885437
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,2048,2048,0.030876799424489336
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,2048,3072,0.02521173357963562
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,2048,1536,0.011609599987665812
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,2048,1536,0.027037866910298664
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,2048,2048,0.02068159977595011
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,2048,1024,0.008422399560610454
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,2048,1024,0.024768000841140746
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,2048,1024,0.01736853321393331
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,2048,768,0.007178666690985362
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,2048,768,0.02245546579360962
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,2048,1536,0.018155733744303383
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,2048,512,0.005986133217811584
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,2048,512,0.020911999543507895
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,2048,768,0.01565439999103546
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,2048,256,0.0050794666012128195
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,2048,256,0.018548266092936198
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,2048,512,0.015783466895421348
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,2048,256,0.015424000223477683
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,2048,128,0.004828799764315287
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,2048,128,0.017947733402252197
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,2048,64,0.004241066674391428
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,2048,64,0.016583466529846193
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,2048,32,0.004283733169237773
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,2048,128,0.01572160025437673
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,2048,32,0.017102932929992674
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,1536,65536,0.29601173400878905
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,1536,65536,0.6300522486368816
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,1536,51200,0.23416533470153808
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,1536,51200,0.49842131932576494
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,2048,51200,0.23654185930887855
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,1536,16384,0.07548373540242513
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,1536,16384,0.16184746424357097
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,1536,16384,0.06399146715799967
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,1536,12288,0.05703573226928711
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,1536,12288,0.11836586793263752
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,2048,65536,0.29673919677734373
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,1536,10240,0.04934293429056803
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,1536,10240,0.09584853649139405
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,1536,12288,0.051719466845194496
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,1536,8192,0.038794668515523274
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,1536,8192,0.07661333084106445
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,1536,10240,0.04552746613820394
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,1536,7168,0.0346997340520223
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,1536,7168,0.06592106819152832
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,1536,8192,0.03764479955037435
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,1536,6144,0.029458133379618327
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,1536,6144,0.056212266286214195
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,1536,7168,0.035936001936594644
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,1536,5120,0.025639466444651288
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,1536,5120,0.048630400498708086
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,1536,6144,0.032365866502126056
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,1536,4096,0.02102186679840088
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,1536,4096,0.04326826731363932
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,1536,5120,0.030437332391738892
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,1536,3584,0.019245866934458414
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,1536,3584,0.03928533395131429
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,1536,4096,0.02661120096842448
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,1536,3072,0.01660053332646688
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,1536,3072,0.035267198085784913
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,1536,3584,0.024382932980855306
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,1536,2560,0.014878933628400167
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,1536,2560,0.033123199144999185
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,1536,3072,0.022367999951044718
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,1536,2048,0.012332800030708312
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,1536,2048,0.029587199290593464
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,1536,51200,0.16488746007283528
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,1536,1536,0.009573333462079366
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,1536,1536,0.025041067600250246
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,1536,2048,0.01909546653429667
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,1536,1024,0.0076799998680750535
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,1536,1024,0.024244266748428344
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,1536,1536,0.016966400543848674
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,1536,768,0.006742399930953979
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,1536,768,0.022434133291244506
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,1536,1024,0.01618880033493042
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,1536,2560,0.021668267250061036
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,1536,768,0.015294933319091797
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,1536,512,0.00551146666208903
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,1536,256,0.004637866715590159
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,1536,512,0.021597866217295328
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,1536,256,0.016720000902811685
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,1536,128,0.003851733356714249
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,1536,128,0.018003199497858682
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,1536,256,0.01532266636689504
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,1536,64,0.003866666555404663
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,1536,512,0.015341867009798685
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,1536,32,0.003818666686614355
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,1536,64,0.01686720053354899
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,1536,32,0.015842133760452272
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,1024,65536,0.20251839955647788
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,1024,65536,0.5924906412760417
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,1024,51200,0.1618794600168864
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,1536,128,0.01617173353830973
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,1024,51200,0.468448003133138
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,1024,16384,0.052909866968790686
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,1024,16384,0.14975679715474446
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,1024,16384,0.061324799060821535
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,1024,12288,0.040453334649403885
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,1024,12288,0.10734933217366535
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,1536,65536,0.20433386166890463
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,1024,10240,0.03387413422266643
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,1024,10240,0.08703253269195557
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,1024,12288,0.04857600132624308
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,1024,8192,0.0269269327322642
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,1024,8192,0.06608213186264038
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,1024,10240,0.042056532700856526
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,1024,7168,0.02388906677563985
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,1024,7168,0.058088533083597814
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,1024,8192,0.034289065996805826
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,1024,6144,0.02065920035044352
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,1024,6144,0.04952853520711263
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,1024,7168,0.0321834663550059
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,1024,5120,0.01810879906018575
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,1024,5120,0.04502933422724406
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,1024,6144,0.029282132784525555
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,1024,4096,0.014965333541234336
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,1024,4096,0.03900800148646037
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,1024,5120,0.027296000719070436
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,1024,3584,0.013353600104649862
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,1024,3584,0.03566933472951253
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,1024,4096,0.02394346594810486
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,1024,3072,0.011620266238848369
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,1024,3072,0.03364799817403157
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,1024,3584,0.021612799167633055
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,1024,2560,0.01018346647421519
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,1024,2560,0.031023999055226643
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,1024,3072,0.020345600446065266
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,1024,2048,0.008808533350626629
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,1024,2048,0.028845866521199543
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,1024,2560,0.018675200144449868
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,1024,1536,0.0075882668296496075
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,1024,2048,0.016636799772580466
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,1024,1536,0.02476693391799927
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,1024,1024,0.005981866518656413
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,1024,1024,0.02289173404375712
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,1024,1536,0.014514133334159851
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,1024,768,0.005157333115736643
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,1024,1024,0.013741866747538248
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,1024,768,0.02061333258946737
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,1024,512,0.004565333326657614
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,1024,512,0.018755199511845906
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,1024,51200,0.1613653341929118
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,1024,256,0.003832533210515976
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,1024,256,0.016657066345214844
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,1024,512,0.012889599800109864
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,1024,128,0.003454933315515518
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,1024,256,0.012488533059755962
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,1024,128,0.015160533785820007
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,1024,64,0.0030165334542592366
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,1024,768,0.012910933295885722
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,1024,64,0.01602453291416168
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,1024,32,0.0031829332311948144
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,1024,128,0.0132533331712087
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,1024,32,0.017056000232696534
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,768,65536,0.1937546730041504
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,768,51200,0.15093226432800294
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,768,65536,0.5678581237792969
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,768,51200,0.4489450772603353
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,768,16384,0.051240531603495273
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,768,16384,0.14965333938598632
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,1024,65536,0.20102292696634927
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,768,12288,0.038780800501505536
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,768,12288,0.10739093621571857
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,768,16384,0.06189119815826416
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,768,10240,0.03237653374671936
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,768,10240,0.0858464002609253
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,768,12288,0.04891413450241089
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,768,8192,0.025540266434351606
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,768,8192,0.06425600051879883
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,768,10240,0.042721064885457356
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,768,7168,0.021689599752426146
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,768,7168,0.05649706522623697
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,768,8192,0.035131732622782394
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,768,6144,0.018936532735824584
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,768,7168,0.03295039931933085
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,768,6144,0.04780266682306926
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,768,5120,0.01617173353830973
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,768,5120,0.04380160172780355
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,768,5120,0.028576000531514482
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,768,4096,0.013558399677276612
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,768,4096,0.03876906633377075
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,768,6144,0.03050559957822164
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,768,3584,0.01202133297920227
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,768,3584,0.03511253197987874
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,768,4096,0.025220266977945965
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,768,3072,0.010441600282986959
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,768,3072,0.031113600730895995
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,768,3072,0.021115734179814657
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,768,3584,0.023091200987497965
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,768,2560,0.031049599250157673
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,768,51200,0.1621237277984619
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,768,2048,0.008142933249473572
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,768,2048,0.028222932418187456
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,768,2560,0.00925546685854594
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,768,1536,0.006859733164310456
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,768,1536,0.023641600211461385
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,768,2560,0.01986133257548014
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,768,1024,0.005538133283456167
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,768,1024,0.022822399934132896
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,768,2048,0.017783466974894205
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,768,768,0.0050687998533248905
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,768,1536,0.01569919983545939
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,768,768,0.021625600258509316
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,768,512,0.0042805333932240805
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,768,512,0.020101332664489747
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,768,768,0.01416213313738505
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,768,256,0.0038261334101359046
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,768,512,0.014057599504788718
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,768,256,0.017697066068649292
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,768,1024,0.014900267124176025
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,768,128,0.0033759998778502146
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,768,128,0.016037333011627197
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,768,256,0.01406826674938202
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,768,64,0.003005866706371307
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,768,128,0.014060800274213156
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,768,32,0.0030591999491055804
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,768,64,0.016657066345214844
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,768,32,0.01478506624698639
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,512,65536,0.12914880116780597
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,512,51200,0.10185386339823406
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,512,65536,0.5438367843627929
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,512,51200,0.43232107162475586
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,512,16384,0.03529493411382039
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,512,16384,0.14231146176656087
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,768,65536,0.20150933265686036
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,512,12288,0.025006934007008867
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,512,16384,0.060743467013041175
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,512,12288,0.09859733581542969
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,512,10240,0.020694400866826376
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,512,10240,0.07797119617462159
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,512,10240,0.04034880002339681
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,512,8192,0.017628800868988038
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,512,8192,0.05748480161031087
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,512,12288,0.04737173318862915
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,512,7168,0.015539200107256571
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,512,7168,0.05138453245162964
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,512,8192,0.03295786579449971
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,512,6144,0.01339413324991862
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,512,6144,0.04396266539891561
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,512,7168,0.031404799222946166
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,512,5120,0.011678933103879293
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,512,5120,0.04035626649856568
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,512,6144,0.028505599498748778
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,512,4096,0.01018346647421519
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,512,4096,0.03503679831822713
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,512,4096,0.023110399643580117
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,512,3584,0.009161600470542907
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,512,3584,0.032909866174062094
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,512,5120,0.02680319945017497
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,512,3072,0.008407466610272725
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,512,3072,0.031074132521947222
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,512,3584,0.021210666497548422
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,512,2560,0.007226666808128357
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,512,2560,0.02882453401883443
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,512,51200,0.16046613057454426
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,512,2048,0.006322133541107178
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,512,2048,0.02575040062268575
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,512,3072,0.01945599913597107
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,512,2560,0.01784320076306661
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,512,1536,0.005731200178464254
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,512,1536,0.02306879957516988
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,512,2048,0.016169599692026772
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,512,1024,0.004613333443800608
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,512,1024,0.021411200364430748
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,512,1536,0.014106667041778565
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,512,768,0.0043381333351135256
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,512,1024,0.013281066219011942
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,512,768,0.020782933632532755
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,512,512,0.00377813329299291
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,512,512,0.01951040029525757
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,512,512,0.012152533729871113
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,512,256,0.0031338666876157125
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,512,256,0.017100799083709716
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,512,256,0.012462932864824932
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,512,128,0.0029077333708604175
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,512,768,0.012034133076667786
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,512,128,0.016939733425776163
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,512,64,0.0026346666117509207
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,512,32,0.0026357332865397137
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,512,128,0.012660266955693564
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,512,64,0.01470186710357666
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,512,32,0.014727466305096946
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,256,65536,0.08480426470438639
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,256,51200,0.0673525333404541
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,256,65536,0.5217407862345378
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,256,51200,0.4128618558247884
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,256,16384,0.02702186703681946
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,256,16384,0.13844799995422363
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,512,65536,0.20041386286417642
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,256,12288,0.021916800737380983
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,256,12288,0.094922669728597
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,256,16384,0.06089813311894735
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,256,10240,0.018244266510009766
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,256,10240,0.0725162665049235
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,256,12288,0.044812798500061035
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,256,8192,0.01267306705315908
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,256,8192,0.05369173288345337
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,256,10240,0.03952000141143799
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,256,7168,0.011197866996129353
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,256,7168,0.04729386568069458
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,256,8192,0.03291413386662801
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,256,6144,0.010100266337394715
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,256,6144,0.0423253337542216
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,256,7168,0.03137493332227071
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,256,5120,0.008794666330019633
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,256,5120,0.037878398100535074
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,256,6144,0.02847786744435628
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,256,4096,0.0075765331586201985
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,256,4096,0.03388160069783529
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,256,5120,0.02681066592534383
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,256,3584,0.007188266515731812
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,256,4096,0.02315946618715922
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,256,3584,0.031027199824651082
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,256,3072,0.006287999947865804
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,256,3072,0.028870399792989093
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,256,51200,0.16020480791727704
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,256,2560,0.0060597335298856105
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,256,3584,0.0211136003335317
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,256,2560,0.027722666660944622
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,256,3072,0.019107200702031455
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,256,2048,0.005217066903909048
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,256,2048,0.025310933589935303
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,256,1536,0.0046741331617037455
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,256,1536,0.02164693276087443
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,256,2560,0.01785279909769694
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,256,1024,0.003815466662247976
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,256,1024,0.021552000443140665
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,256,1536,0.014043733477592468
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,256,768,0.003385599950949351
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,256,2048,0.01623679995536804
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,256,1024,0.012869333227475485
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,256,768,0.019099734226862588
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,256,512,0.003033600002527237
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,256,512,0.01834986607233683
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,256,256,0.00257493332028389
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,256,512,0.012059733271598816
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,256,768,0.012217600146929424
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,256,256,0.01206719974676768
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,256,128,0.0026421333352724713
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,256,128,0.014753066500027976
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,256,64,0.0022442666192849475
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,256,64,0.014595199624697366
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,256,128,0.012525866429011026
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,256,32,0.0025792000194390613
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,128,65536,0.06760426362355551
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,256,65536,0.19993707338968914
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,128,65536,0.5166783968607584
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,128,51200,0.051933864752451576
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,256,256,0.016642133394877114
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,128,51200,0.4076650619506836
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,128,16384,0.017282134294509886
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,256,32,0.015748266379038492
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,128,16384,0.13547306060791015
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,128,12288,0.01370133360226949
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,128,12288,0.093449600537618
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,128,12288,0.043202133973439534
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,128,10240,0.012065066893895467
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,128,10240,0.0710922638575236
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,128,16384,0.058310401439666745
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,128,10240,0.03825920025507609
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,128,8192,0.011666133006413778
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,128,8192,0.051826135317484534
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,128,51200,0.15702080726623535
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,128,7168,0.010328533252080281
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,128,8192,0.03207893371582031
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,128,7168,0.045311999320983884
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,128,6144,0.009191466371218364
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,128,6144,0.04090559879938761
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,128,7168,0.030133332808812457
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,128,5120,0.007977599898974102
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,128,65536,0.19607359568277996
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,128,6144,0.027616000175476073
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,128,4096,0.006694399813810985
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,128,4096,0.03431253433227539
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,128,3584,0.007021866738796234
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,128,5120,0.025856000185012818
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,128,4096,0.022000000874201456
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,128,3072,0.0058677335580190025
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,128,3072,0.028492800394694012
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,128,3584,0.031653332710266116
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,128,3584,0.020218666394551596
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,128,2560,0.006011733412742614
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,128,2560,0.026761599381764728
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,128,2048,0.004648533463478088
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,128,3072,0.018090667327245076
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,128,2048,0.024693334102630617
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,128,5120,0.037604268391927084
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,128,1536,0.00420906643072764
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,128,1536,0.022686932484308878
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,128,2560,0.017011199394861856
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,128,1024,0.003489066660404205
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,128,2048,0.014998400211334228
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,128,768,0.0033514666060606635
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,128,1536,0.013010133306185404
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,128,768,0.018776534001032512
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,128,1024,0.012183466553688049
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,128,512,0.0030250666042168934
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,128,1024,0.021010132630666097
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,128,768,0.011179733276367187
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,128,512,0.01846933364868164
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,128,256,0.0025941332181294756
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,128,512,0.01123199959595998
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,128,256,0.016201600432395935
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,128,128,0.0022890667120615644
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,128,256,0.011191466450691223
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,128,128,0.014615466197331747
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2817,128,128,0.011206400394439698
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,128,64,0.0155157337586085
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,128,32,0.014473600188891092
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,64,65536,0.5126282691955566
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,64,51200,0.04935786724090576
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,64,51200,0.4046634674072266
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,64,16384,0.012929067015647888
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,128,64,0.0022122666239738466
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,128,32,0.002201599876085917
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,64,12288,0.012436266740163167
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,64,16384,0.1328757365544637
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,64,12288,0.09192533493041992
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,64,8192,0.00914026697476705
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,64,10240,0.011165866255760193
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,64,8192,0.05050026575724283
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,64,7168,0.009153067072232565
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,64,7168,0.04550293286641439
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,64,6144,0.008356266220410665
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,64,6144,0.04069439967473348
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,64,5120,0.007915733257929485
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,64,5120,0.03701866865158081
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,64,4096,0.007578666508197785
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,64,4096,0.032986666758855185
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,64,3584,0.006648533542950948
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,64,3584,0.030784000953038532
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,64,3072,0.00626986672480901
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,64,3072,0.0286463995774587
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,64,2560,0.005392000079154968
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,64,2560,0.026744532585144042
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,64,2048,0.004942933221658071
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,64,1536,0.004195199906826019
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,64,2048,0.02474986712137858
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,64,65536,0.0641205350557963
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,64,1024,0.0034005333979924522
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,64,768,0.003033600002527237
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,64,1024,0.020886399348576865
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,64,10240,0.07170133590698242
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,64,512,0.0026975999275843303
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,64,768,0.019792000452677407
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,64,256,0.0026122666895389557
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,64,1536,0.022284799814224245
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,64,128,0.002233600119749705
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,64,512,0.01827413241068522
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,64,256,0.01658560037612915
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,64,64,0.002219733347495397
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,64,32,0.00222080002228419
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,64,128,0.015602133671442666
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,64,64,0.014382933576901754
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,64,32,0.01458026667435964
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,32,65536,0.061101865768432614
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,32,51200,0.04918613433837891
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,32,16384,0.012050132950146992
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,32,12288,0.009704533219337463
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,32,16384,0.13148266474405926
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,32,51200,0.40610774358113605
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,32,12288,0.09017919699350993
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,32,65536,0.5113141377766927
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,32,10240,0.00913279950618744
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,32,8192,0.008072533210118612
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,32,10240,0.07161813577016195
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,32,7168,0.00782719999551773
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,32,8192,0.05295466581980387
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,32,6144,0.0077909335494041445
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,32,7168,0.04729066689809163
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,32,5120,0.009531733393669129
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,32,4096,0.008920533458391826
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,32,5120,0.03849386771519979
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,32,3584,0.00805866668621699
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,32,3584,0.03086400032043457
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,32,4096,0.03303893407185872
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,32,3072,0.00710399995247523
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,32,2560,0.006267733375231425
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,32,2560,0.027036799987157183
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,32,3072,0.028907734155654907
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,32,2048,0.004666666686534882
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,32,1536,0.0038730666041374207
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,32,2048,0.025834665695826216
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,32,1536,0.020654932657877604
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,32,1024,0.004216533402601878
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,32,768,0.003110400090614955
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,32,1024,0.02059626579284668
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,32,512,0.0025727999707063037
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,32,512,0.017781333128611247
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,32,768,0.020785067478815714
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,32,256,0.0025685332715511323
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,32,6144,0.04259200096130371
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,32,128,0.014989866813023885
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,32,256,0.017046399911244712
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,32,64,0.0022143999735514323
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,32,32,0.002229333420594533
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,32,32,0.014333867033322654
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2817,32,64,0.014722133676211039
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2817,32,128,0.0022101332743962605
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,65536,16384,3.4022997538248694
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,65536,51200,5.781401570638021
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,65536,16384,1.756933339436849
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,65536,12288,2.5730239868164064
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,65536,51200,10.853782145182292
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,65536,12288,1.3452458699544272
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,65536,10240,2.3534304300944013
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,65536,10240,1.171893310546875
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,65536,10240,0.6366506576538086
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,65536,8192,1.8031786600748698
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,65536,8192,0.9257162729899088
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,65536,12288,0.7108704249064128
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,65536,7168,1.5819124857584634
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,65536,7168,0.8050069173177083
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,65536,16384,0.936306126912435
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,65536,6144,1.2439306894938151
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,65536,6144,0.676643180847168
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,65536,8192,0.5280810674031575
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,65536,5120,1.0121194839477539
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,65536,5120,0.5664149602254231
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,65536,7168,0.48587840398152665
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,65536,4096,0.8117663701375326
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,65536,4096,0.45235414505004884
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,65536,6144,0.423583984375
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,65536,3584,0.6899530410766601
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,65536,3584,0.402454408009847
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,65536,5120,0.3915103912353516
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,65536,3072,0.5996181488037109
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,65536,3072,0.3505887985229492
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,65536,4096,0.3447488149007162
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,65536,2560,0.4887584050496419
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,65536,2560,0.3084671974182129
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,65536,2560,0.262609068552653
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,65536,2048,0.3941194534301758
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,65536,2048,0.23693440755208334
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,65536,3072,0.27873706817626953
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,65536,1536,0.3102901458740234
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,65536,1536,0.18862080574035645
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,65536,3584,0.3139125188191732
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,65536,1024,0.2098837375640869
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,65536,1024,0.15813652674357098
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,65536,1024,0.19048426946004232
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,65536,768,0.164137601852417
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,65536,768,0.13700693448384602
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,65536,1536,0.19534719785054525
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,65536,512,0.13507307370503743
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,65536,512,0.12320640087127685
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,65536,768,0.18137386639912922
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,65536,256,0.1033898671468099
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,65536,256,0.10491519769032795
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,65536,512,0.17681280771891278
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,65536,128,0.06997546354929605
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,65536,128,0.07376320362091064
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,65536,2048,0.2313759962717692
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,65536,64,0.060244266192118326
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,65536,64,0.07616000175476074
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,65536,32,0.0607317328453064
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,65536,32,0.07565759817759196
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,65536,128,0.17811412811279298
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,65536,256,0.17976640065511068
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,51200,65536,5.873738606770834
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,51200,65536,10.937143961588543
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,51200,51200,4.649735514322916
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,51200,51200,8.516345723470051
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,51200,16384,2.6471626281738283
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,51200,16384,1.383173370361328
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,51200,16384,0.7477354685465495
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,51200,12288,1.9638186136881512
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,51200,12288,1.067852783203125
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,51200,12288,0.5674528121948242
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,51200,10240,1.6134293874104817
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,51200,10240,0.8356927871704102
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,51200,10240,0.5010496139526367
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,51200,8192,1.2988245646158854
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,51200,8192,0.7307135899861653
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,65536,51200,3.1249834696451826
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,51200,7168,1.0683060963948567
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,51200,7168,0.6505887985229493
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,51200,8192,0.41286719640096026
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,51200,6144,0.9146495819091797
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,51200,6144,0.5651754379272461
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,51200,7168,0.38172054290771484
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,51200,5120,0.720033073425293
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,51200,5120,0.43063999811808273
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,51200,5120,0.30117867787679037
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,51200,4096,0.5927018483479818
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,51200,4096,0.35492159525553385
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,51200,6144,0.3297792116800944
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,51200,3584,0.5138112068176269
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,51200,3584,0.31067094802856443
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,51200,3584,0.24278826713562013
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,51200,3072,0.44335892995198567
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,51200,3072,0.27807572682698567
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,51200,4096,0.2701621373494466
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,51200,2560,0.36568854649861654
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,51200,2560,0.23576639493306478
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,51200,51200,2.402939860026042
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,51200,2048,0.3046741485595703
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,51200,2048,0.18453332583109539
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,51200,2560,0.20577279726664224
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,51200,1536,0.22642134030659994
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,51200,1536,0.1457813262939453
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,51200,2048,0.1799946626027425
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,51200,3072,0.2171562671661377
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,51200,1024,0.15454932848612468
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,51200,1024,0.10987626711527507
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,51200,768,0.12274133364359538
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,51200,768,0.09263892968495688
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,51200,768,0.14241065979003906
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,51200,512,0.09069120089213054
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,51200,512,0.08048106829325358
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,51200,1024,0.15120959281921387
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,51200,256,0.06100266774495443
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,51200,256,0.06682133674621582
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,51200,512,0.14218346277872723
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,51200,128,0.05506879885991415
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,51200,128,0.060577066739400234
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,51200,256,0.14048105875651043
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,51200,64,0.046019200483957926
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,51200,64,0.0625055988629659
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,51200,32,0.04808533191680908
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,51200,32,0.06301653385162354
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,51200,128,0.14306772549947103
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,16384,65536,3.4448107401529944
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,16384,65536,2.1387252807617188
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,51200,1536,0.15569067001342773
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,16384,51200,1.6563701629638672
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,16384,51200,2.72466557820638
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,16384,16384,0.7774986902872721
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,16384,16384,0.537501843770345
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,16384,16384,0.28177172342936196
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,16384,12288,0.5856533050537109
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,16384,12288,0.4001525243123372
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,16384,12288,0.21485013961791993
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,16384,10240,0.5256202697753907
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,16384,10240,0.33650665283203124
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,16384,10240,0.19225172996520995
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,16384,8192,0.4190133412679036
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,16384,8192,0.28980159759521484
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,16384,8192,0.15838613510131835
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,16384,7168,0.3688213348388672
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,16384,7168,0.2370314598083496
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,16384,51200,0.8358399709065756
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,16384,6144,0.30908266703287757
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,16384,6144,0.2047797362009684
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,16384,7168,0.14666666984558105
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,16384,5120,0.25866026878356935
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,16384,5120,0.18451733589172364
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,16384,5120,0.11820586522420247
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,16384,4096,0.20511892636617027
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,16384,4096,0.1453770637512207
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,16384,6144,0.12602773507436116
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,16384,3584,0.17713066736857097
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,16384,3584,0.1234239975611369
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,16384,4096,0.10544746716817219
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,16384,3072,0.15356799761454265
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,16384,3072,0.10884266694386799
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,16384,3584,0.09225920041402182
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,16384,2560,0.12733653386433919
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,16384,2560,0.0948298692703247
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,16384,2560,0.07725760142008463
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,16384,2048,0.1033514658610026
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,16384,2048,0.07905920346577963
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,16384,3072,0.0839680035909017
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,16384,1536,0.08010559876759847
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,16384,1536,0.06409386793772379
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,16384,1536,0.056143999099731445
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,16384,1024,0.05629333257675171
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,16384,1024,0.054919465382893884
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,16384,2048,0.06564373175303141
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,16384,768,0.0436842679977417
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,16384,768,0.04743893146514892
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,16384,1024,0.053635199864705406
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,16384,512,0.033188267548878984
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,16384,512,0.04109760125478108
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,16384,768,0.045305601755778
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,16384,256,0.02304853399594625
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,16384,256,0.03915626605351766
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,16384,512,0.045458134015401205
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,16384,128,0.020886399348576865
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,16384,128,0.03745599985122681
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,16384,256,0.044044800599416095
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,16384,64,0.021963733434677123
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,16384,64,0.03518400192260742
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,16384,32,0.02227946718533834
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,16384,32,0.03498560190200806
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,16384,128,0.04573013385136922
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,12288,65536,2.662006378173828
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,12288,65536,1.7054730733235677
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,12288,51200,2.0506080627441405
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,12288,51200,1.3403146107991537
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,16384,65536,1.0552138646443685
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,12288,16384,0.6137184143066406
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,12288,16384,0.4277237256368001
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,51200,65536,3.0469014485677084
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,12288,12288,0.4481024106343587
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,12288,12288,0.31617921193440757
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,12288,12288,0.1615616003672282
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,12288,10240,0.36051092147827146
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,12288,10240,0.2704106648763021
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,12288,16384,0.21059093475341797
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,12288,8192,0.2839050610860189
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,12288,8192,0.19984854062398275
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,12288,8192,0.11886293093363445
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,12288,7168,0.2465973377227783
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,12288,7168,0.17913707097371417
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,12288,10240,0.14414292971293133
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,12288,6144,0.21348586082458496
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,12288,6144,0.15175573031107586
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,12288,6144,0.09705920219421386
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,12288,5120,0.1758090655008952
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,12288,5120,0.13014826774597169
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,12288,7168,0.10998613039652508
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,12288,4096,0.1383797327677409
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,12288,4096,0.10778026580810547
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,12288,4096,0.07672213713328044
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,12288,3584,0.12166933218638103
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,12288,3584,0.09528106848398844
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,12288,5120,0.08843839963277181
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,12288,3072,0.10620693365732829
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,12288,3072,0.08329813480377198
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,12288,3072,0.06041599909464518
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,12288,2560,0.08940587043762208
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,12288,2560,0.07596480051676432
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,12288,3584,0.06802453200022379
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,12288,2048,0.07192320028940836
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,12288,2048,0.06371306578318278
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,12288,2048,0.05236053466796875
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,12288,1536,0.05524266560872396
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,12288,1536,0.05108693440755209
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,12288,2560,0.0581216017405192
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,12288,1024,0.03924053510030111
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,12288,1024,0.04410986502965291
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,12288,1024,0.046214401721954346
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,12288,768,0.03073173364003499
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,12288,768,0.03920746644337972
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,12288,1536,0.04769706726074219
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,12288,512,0.02225173314412435
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,12288,512,0.03456213474273682
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,12288,768,0.042991999785105386
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,12288,256,0.01663253307342529
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,12288,256,0.029064534107844035
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,12288,512,0.04304639895757039
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,12288,128,0.013698133826255798
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,12288,128,0.02564799984296163
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,12288,128,0.04279146591822307
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,12288,64,0.01237440009911855
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,12288,64,0.025011199712753295
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,12288,32,0.01286186675230662
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,12288,32,0.02709546685218811
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,12288,256,0.04284053246180217
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,10240,65536,1.4476661682128906
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,10240,65536,2.1173749287923176
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,10240,51200,1.5286261240641275
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,10240,51200,1.1259146372477213
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,12288,51200,0.6166816075642904
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,10240,16384,0.47385813395182297
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,10240,16384,0.37551892598470055
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,10240,16384,0.18712639808654785
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,10240,12288,0.36000105539957683
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,10240,12288,0.2743274688720703
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,12288,65536,0.7762122472127279
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,10240,10240,0.3058549245198568
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,10240,10240,0.23928319613138832
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,10240,12288,0.1436789353688558
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,10240,8192,0.24781440099080404
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,10240,8192,0.18378772735595703
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,10240,10240,0.12824640274047852
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,10240,7168,0.2154538631439209
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,10240,7168,0.158627192179362
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,10240,8192,0.10608320236206055
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,10240,6144,0.18206186294555665
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,10240,6144,0.13789013226826985
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,10240,7168,0.09851946830749511
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,10240,5120,0.15081920623779296
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,10240,5120,0.11598292986551921
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,10240,6144,0.08637973467508951
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,10240,4096,0.120414932568868
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,10240,4096,0.09857280254364013
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,10240,5120,0.07867733637491861
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,10240,3584,0.10511360168457032
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,10240,3584,0.08617173035939535
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,10240,4096,0.06755733489990234
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,10240,3072,0.09146560033162435
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,10240,3072,0.07655466397603353
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,10240,3584,0.060071468353271484
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,10240,2560,0.0764031966527303
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,10240,2560,0.06928213437398276
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,10240,51200,0.537718391418457
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,10240,2560,0.05290453433990479
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,10240,2048,0.061028265953063966
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,10240,2048,0.05785066684087118
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,10240,1536,0.04771626790364583
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,10240,3072,0.05518399874369303
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,10240,1536,0.04574293295542399
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,10240,1024,0.03302506605784099
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,10240,1024,0.039153067270914714
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,10240,1024,0.04254719813664754
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,10240,768,0.025546665986378985
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,10240,768,0.035403732458750406
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,10240,1536,0.044099199771881106
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,10240,512,0.0195743997891744
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,10240,512,0.03151040077209473
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,10240,768,0.03960213263829549
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,10240,256,0.01441493332386017
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,10240,256,0.02608319918314616
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,10240,512,0.03968853155771891
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,10240,128,0.013285332918167114
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,10240,128,0.02493120034535726
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,10240,256,0.03947946627934774
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,10240,64,0.011230933666229247
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,10240,64,0.02479040026664734
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,10240,32,0.011244799693425496
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,10240,32,0.022717867294947305
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,10240,128,0.0395306666692098
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,10240,2048,0.04820266564687093
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,8192,65536,1.3431915283203124
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,8192,65536,1.830888493855794
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,8192,51200,1.05043093363444
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,8192,51200,1.3655595143636068
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,8192,16384,0.4394506772359212
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,8192,16384,0.36271893183390297
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,8192,16384,0.15982826550801593
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,8192,12288,0.3197632153828939
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,8192,12288,0.2564074675242106
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,10240,65536,0.6786474863688151
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,8192,10240,0.25797012646993
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,8192,10240,0.2016512076059977
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,8192,12288,0.12294399738311768
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,8192,8192,0.21278613408406577
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,8192,8192,0.16357653935750324
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,8192,10240,0.11081066926320393
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,8192,7168,0.1827583948771159
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,8192,7168,0.14168960253397625
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,8192,8192,0.09065492947896323
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,8192,6144,0.15993599891662597
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,8192,6144,0.12319680054982503
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,8192,7168,0.0846997340520223
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,8192,5120,0.12940159638722737
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,8192,5120,0.10549226601918538
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,8192,5120,0.06728959878285726
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,8192,4096,0.10214293003082275
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,8192,4096,0.09049386978149414
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,8192,6144,0.07382187048594156
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,8192,3584,0.09081172943115234
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,8192,3584,0.08065493106842041
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,8192,3584,0.05128746827443441
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,8192,3072,0.07812373638153076
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,8192,3072,0.07056000232696533
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,8192,4096,0.05682453314463297
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,8192,2560,0.06451093355814616
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,8192,2560,0.0641759991645813
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,8192,2560,0.043010131518046064
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,8192,2048,0.05274240175882975
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,8192,2048,0.05333439906438192
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,8192,3072,0.0446122686068217
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,8192,1536,0.04153813521067302
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,8192,1536,0.04354133208592732
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,8192,1536,0.033012266953786215
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,8192,1024,0.029183999697367353
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,8192,2048,0.03751893440882365
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,8192,1024,0.03733760118484497
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,8192,768,0.02341653307278951
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,8192,768,0.03309119939804077
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,8192,768,0.028463999430338543
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,8192,512,0.017204266786575318
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,8192,512,0.0288917342821757
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,8192,512,0.02800426681836446
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,8192,256,0.012923733393351237
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,8192,256,0.0271232008934021
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,8192,1024,0.03172159989674886
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,8192,128,0.010575999816258747
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,8192,128,0.02536746660868327
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,8192,256,0.028089600801467895
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,8192,64,0.011994666854540507
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,8192,64,0.024950400988260905
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,8192,32,0.012405332922935487
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,8192,32,0.024762666225433348
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,8192,128,0.028531199693679808
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,7168,65536,1.2458015441894532
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,7168,65536,1.6029098510742188
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,7168,51200,1.2391403198242188
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,7168,51200,0.9853994369506835
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,8192,51200,0.4629791895548503
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,7168,16384,0.38246612548828124
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,7168,16384,0.32091732025146485
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,7168,16384,0.1585312048594157
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,7168,12288,0.2890922546386719
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,7168,12288,0.23704746564229331
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,8192,65536,0.580732790629069
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,7168,10240,0.2358773390452067
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,7168,10240,0.18763413429260253
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,7168,12288,0.12169386545817058
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,7168,8192,0.19360639254252116
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,7168,8192,0.15077120463053387
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,7168,10240,0.1090879996617635
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,7168,7168,0.1653674602508545
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,7168,7168,0.13448959986368816
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,7168,8192,0.09014933109283448
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,7168,6144,0.14231252670288086
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,7168,6144,0.1152234633763631
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,7168,7168,0.08338239987691244
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,7168,5120,0.11489066282908122
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,7168,5120,0.1013493299484253
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,7168,5120,0.06571306784947714
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,7168,4096,0.09218346277872722
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,7168,4096,0.0863754669825236
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,7168,6144,0.07271146774291992
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,7168,3584,0.08203519980112711
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,7168,3584,0.07636480331420899
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,7168,4096,0.05621013243993124
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,7168,3584,0.05044800043106079
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,7168,3072,0.06981866359710694
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,7168,3072,0.06133013168970743
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,7168,2560,0.05954986810684204
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,7168,2560,0.055283200740814206
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,7168,2560,0.043644801775614424
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,7168,2048,0.04868693351745605
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,7168,2048,0.04662719964981079
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,7168,3072,0.04513173500696818
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,7168,1536,0.03400213321050008
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,7168,1536,0.040573867162068684
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,7168,1536,0.03459626833597819
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,7168,1024,0.02408746679623922
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,7168,1024,0.03668479919433594
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,7168,2048,0.03903893232345581
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,7168,768,0.019950934251149497
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,7168,768,0.03173973361651103
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,7168,768,0.03057066599527995
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,7168,512,0.0166485329469045
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,7168,512,0.02903253237406413
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,7168,1024,0.03306453426678975
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,7168,256,0.013318399588267008
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,7168,256,0.025564799706141155
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,7168,512,0.030753066142400105
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,7168,256,0.030460800727208453
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,7168,128,0.02379306753476461
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,7168,128,0.030529065926869707
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,7168,64,0.010379733641942342
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,7168,64,0.022858667373657226
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,7168,32,0.010915199915568035
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,7168,32,0.022758400440216063
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,7168,128,0.011198932925860088
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,6144,65536,1.2619274139404297
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,6144,65536,1.0774026234944662
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,6144,51200,0.9328522364298502
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,6144,51200,0.8502997080485025
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,7168,51200,0.4585610707600911
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,6144,16384,0.3003114700317383
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,6144,16384,0.27701972325642904
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,6144,16384,0.13693013191223144
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,6144,12288,0.22894506454467772
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,6144,12288,0.20447680155436196
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,7168,65536,0.5782335917154948
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,6144,10240,0.20157119433085122
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,6144,10240,0.16951893170674642
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,6144,12288,0.10528106689453125
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,6144,8192,0.16558720270792643
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,6144,8192,0.13780266443888348
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,6144,10240,0.09517013231913249
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,6144,7168,0.1417461395263672
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,6144,7168,0.12235199610392253
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,6144,8192,0.07909013430277506
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,6144,6144,0.12407999833424885
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,6144,6144,0.10523306528727214
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,6144,7168,0.0731658697128296
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,6144,5120,0.0998698631922404
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,6144,5120,0.09228479862213135
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,6144,6144,0.0636298656463623
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,6144,4096,0.07993813355763754
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,6144,4096,0.07842986583709717
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,6144,5120,0.05781439940134684
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,6144,3584,0.0689674695332845
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,6144,3584,0.07004053592681884
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,6144,4096,0.04983786741892497
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,6144,3072,0.06065813302993774
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,6144,3072,0.0599231998125712
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,6144,3584,0.04413119951883952
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,6144,2560,0.050460799535115564
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,6144,2560,0.05369813442230224
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,6144,3072,0.039614931742350264
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,6144,2048,0.04189013242721558
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,6144,2048,0.04576853513717651
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,6144,2560,0.03794986804326375
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,6144,1536,0.03189333279927571
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,6144,1536,0.037478399276733396
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,6144,2048,0.034216535091400144
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,6144,1024,0.023112533489863078
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,6144,1024,0.03328640063603719
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,6144,1536,0.03091413378715515
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,6144,768,0.01806400020917257
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,6144,768,0.029126399755477907
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,6144,1024,0.029547733068466187
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,6144,512,0.01359999974568685
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,6144,512,0.025250132878621417
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,6144,768,0.027345067262649535
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,6144,256,0.009664000074068705
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,6144,256,0.02331733306248983
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,6144,51200,0.3891317367553711
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,6144,512,0.027586134274800618
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,6144,128,0.008924800157546996
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,6144,128,0.020871466398239134
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,6144,64,0.007550933460394542
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,6144,64,0.020190932353337607
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,6144,32,0.00793280005455017
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,6144,32,0.021615999937057494
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,6144,128,0.02693866689999898
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,6144,256,0.02719893256823222
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,5120,65536,0.9999722798665365
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,5120,65536,0.9792885462443033
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,5120,51200,0.8161354700724284
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,5120,51200,0.7681055704752604
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,5120,16384,0.2433215936024984
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,5120,16384,0.2518389383951823
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,5120,16384,0.1139082670211792
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,5120,12288,0.18455146153767904
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,5120,12288,0.178110933303833
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,6144,65536,0.48809814453125
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,5120,10240,0.15016214052836102
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,5120,10240,0.14383467038472492
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,5120,12288,0.08811733722686768
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,5120,8192,0.12438613573710125
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,5120,8192,0.11708266735076904
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,5120,10240,0.07950507005055746
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,5120,7168,0.10560959974924725
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,5120,7168,0.10595839818318684
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,5120,8192,0.06651093165079752
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,5120,6144,0.0915712038675944
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,5120,6144,0.09047146638234457
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,5120,7168,0.06180479923884073
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,5120,5120,0.07377493381500244
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,5120,5120,0.07960320313771566
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,5120,6144,0.05437119801839193
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,5120,4096,0.060342399279276523
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,5120,4096,0.06625813245773315
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,5120,5120,0.04918293158213298
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,5120,3584,0.052349865436553955
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,5120,3584,0.05817386706670126
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,5120,4096,0.04257386525472005
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,5120,3072,0.04600106477737427
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,5120,3584,0.0388864000638326
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,5120,3072,0.04988693396250407
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,5120,2560,0.038891732692718506
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,5120,2560,0.04524159828821818
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,5120,2560,0.03339946667353312
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,5120,2048,0.0315285325050354
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,5120,2048,0.039274664719899495
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,5120,3072,0.03549866676330567
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,5120,1536,0.0244159996509552
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,5120,1536,0.03384960095087687
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,5120,1536,0.027930667002995808
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,5120,1024,0.017494400342305504
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,5120,1024,0.03025280038515727
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,5120,2048,0.0305184006690979
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,5120,768,0.013946666320164999
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,5120,768,0.02691199978192647
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,5120,768,0.029150933027267456
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,5120,512,0.010573866963386535
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,5120,512,0.02401706576347351
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,5120,1024,0.026652799050013228
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,5120,512,0.028893866141637164
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,5120,256,0.007968000074227651
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,5120,256,0.02118826707204183
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,5120,256,0.028961066404978437
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,5120,128,0.007100800176461537
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,5120,128,0.020615466435750327
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,5120,128,0.028933332363764448
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,5120,64,0.006320000191529592
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,5120,64,0.01962453325589498
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,5120,32,0.006692266464233399
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,5120,32,0.01916266679763794
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,4096,65536,0.874895985921224
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,4096,65536,0.9183498382568359
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,4096,51200,0.6725237528483073
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,4096,51200,0.7254645029703777
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,5120,51200,0.31982294718424475
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,4096,16384,0.22232319513956705
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,4096,16384,0.2374506632486979
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,4096,16384,0.10868586699167888
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,4096,12288,0.16755626996358236
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,4096,12288,0.17328853607177735
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,5120,65536,0.40145918528238933
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,4096,10240,0.13490559260050455
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,4096,10240,0.13313813209533693
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,4096,12288,0.08424746990203857
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,4096,8192,0.10780693689982097
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,4096,8192,0.11073386669158936
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,4096,10240,0.07585173447926839
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,4096,7168,0.0921781301498413
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,4096,7168,0.09840213457743327
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,4096,8192,0.06255679925282796
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,4096,6144,0.07915733655293783
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,4096,6144,0.08465173244476318
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,4096,7168,0.05794773499170939
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,4096,5120,0.06425493160883586
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,4096,5120,0.07418239911397298
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,4096,6144,0.05049386819203695
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,4096,4096,0.053097601731618246
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,4096,4096,0.060933331648508705
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,4096,5120,0.04585280021031697
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,4096,3584,0.0459935983022054
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,4096,3584,0.05384000142415365
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,4096,4096,0.03973333438237508
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,4096,3072,0.040480001767476397
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,4096,3072,0.04747519890467326
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,4096,3584,0.035062400499979655
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,4096,2560,0.03365439971288045
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,4096,2560,0.04340159893035889
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,4096,2560,0.030102399984995525
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,4096,2048,0.027738666534423827
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,4096,2048,0.03755626678466797
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,4096,3072,0.032092799743016556
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,4096,1536,0.02262079914410909
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,4096,1536,0.032467200358708696
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,4096,1536,0.023095466693242393
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,4096,1024,0.015793066223462424
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,4096,1024,0.029054933786392213
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,4096,2048,0.025995733340581258
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,4096,768,0.012077866991360982
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,4096,1024,0.02187733252843221
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,4096,768,0.026397866010665894
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,4096,512,0.009569066762924194
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,4096,512,0.023254400491714476
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,4096,51200,0.3161343892415365
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,4096,256,0.007534933090209961
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,4096,256,0.020803199211756388
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,4096,256,0.022317866484324135
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,4096,128,0.006740266581376393
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,4096,128,0.019768534104029338
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,4096,512,0.02060906688372294
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,4096,64,0.005914666752020518
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,4096,768,0.020205867290496827
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,4096,32,0.006291200220584869
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,4096,64,0.01874026656150818
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,4096,32,0.017334399620691936
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,4096,128,0.019441066185633342
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,3584,65536,0.8485717137654623
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,3584,65536,0.892689069112142
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,3584,51200,0.660638936360677
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,3584,51200,0.7020725250244141
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,3584,16384,0.20409919420878092
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,3584,16384,0.2232800006866455
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,3584,16384,0.1088917334874471
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,3584,12288,0.1509610652923584
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,3584,12288,0.16351787249247235
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,4096,65536,0.39778985977172854
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,3584,10240,0.12634027004241943
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,3584,10240,0.13153706391652426
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,3584,12288,0.0849514643351237
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,3584,8192,0.10195093154907227
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,3584,8192,0.10720213254292806
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,3584,10240,0.07678186893463135
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,3584,7168,0.08746986389160157
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,3584,7168,0.09599359830220541
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,3584,8192,0.06209280093510946
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,3584,6144,0.07511359850565592
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,3584,6144,0.08275520006815593
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,3584,7168,0.05742506583531698
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,3584,5120,0.06164586544036865
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,3584,5120,0.0700213352839152
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,3584,6144,0.050774399439493814
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,3584,4096,0.04980479876200358
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,3584,5120,0.04635733366012573
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,3584,4096,0.05796053409576416
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,3584,3584,0.043993600209554035
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,3584,3584,0.05226773420969645
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,3584,3584,0.03695253531138103
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,3584,3072,0.03773333231608073
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,3584,3072,0.04597226778666179
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,3584,4096,0.0399402658144633
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,3584,2560,0.03128106594085693
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,3584,2560,0.04292480150858562
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,3584,3072,0.032994133234024045
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,3584,2048,0.026296534140904743
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,3584,2048,0.038216535250345865
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,3584,2560,0.030963200330734252
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,3584,1536,0.020473599433898926
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,3584,1536,0.03128213286399841
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,3584,1536,0.025278933842976886
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,3584,1024,0.015244799852371215
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,3584,1024,0.027473066250483198
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,3584,2048,0.028283733129501342
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,3584,768,0.012361600001653036
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,3584,768,0.025010132789611818
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,3584,1024,0.023890133698781332
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,3584,512,0.009251200159390767
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,3584,512,0.02268480062484741
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,3584,768,0.022196267048517862
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,3584,256,0.007136000196139018
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,3584,256,0.02076586683591207
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,3584,512,0.022312533855438233
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,3584,128,0.005864533285299936
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,3584,256,0.022088533639907836
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,3584,128,0.01876693367958069
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,3584,64,0.005480533341566721
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,3584,64,0.017151999473571777
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,3584,32,0.005491200089454651
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,3584,32,0.01924906571706136
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,3584,128,0.02219840089480082
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,3072,65536,0.5824597040812175
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,3072,65536,0.7835360209147135
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,3072,51200,0.45645866394042967
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,3072,51200,0.6166421254475911
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,3584,51200,0.3103754679361979
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,3072,16384,0.14718079566955566
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,3072,16384,0.19882133801778157
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,3072,16384,0.08714880148569742
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,3072,12288,0.1119477351506551
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,3072,12288,0.1450549284617106
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,3584,65536,0.3916895866394043
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,3072,10240,0.09770987033843995
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,3072,10240,0.12105706532796223
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,3072,12288,0.06918293635050456
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,3072,8192,0.07973439693450927
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,3072,8192,0.0999338706334432
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,3072,10240,0.0632917324701945
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,3072,7168,0.06906346480051676
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,3072,7168,0.08911680380503337
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,3072,7168,0.04778346618016561
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,3072,6144,0.05922133525212606
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,3072,6144,0.07537493705749512
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,3072,8192,0.05176213184992472
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,3072,5120,0.05019520123799642
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,3072,5120,0.06412479877471924
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,3072,5120,0.03919999996821086
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,3072,4096,0.040742401281992593
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,3072,4096,0.053483732541402186
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,3072,6144,0.0424778660138448
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,3072,3584,0.03573973178863525
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,3072,3584,0.04882239898045858
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,3072,3584,0.03054400086402893
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,3072,3072,0.031233066320419313
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,3072,3072,0.043747198581695554
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,3072,4096,0.03413013219833374
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,3072,2560,0.026753065983454387
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,3072,2560,0.04038826624552409
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,3072,3072,0.02813226580619812
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,3072,2048,0.022297600905100502
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,3072,2048,0.036614398161570236
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,3072,2560,0.026315732796986895
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,3072,1536,0.018074667453765868
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,3072,1536,0.029772800207138062
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,3072,2048,0.02401919960975647
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,3072,1024,0.013452800114949546
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,3072,1024,0.02683626612027486
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,3072,1536,0.02170133392016093
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,3072,768,0.010403199990590414
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,3072,768,0.02462293306986491
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,3072,1024,0.02076693375905355
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,3072,512,0.007972266773382823
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,3072,512,0.02340266704559326
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,3072,768,0.019013333320617675
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,3072,512,0.019406932592391967
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,3072,256,0.006324266890684764
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,3072,256,0.020713599522908528
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,3072,128,0.00588266650835673
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,3072,128,0.018627200524012247
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,3072,128,0.01908906698226929
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,3072,64,0.005092266698678335
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,3072,64,0.01806079943974813
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,3072,51200,0.23869867324829103
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,3072,256,0.019126399358113607
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,3072,32,0.00514026681582133
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,3072,32,0.01704960068066915
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,2560,65536,0.47882239023844403
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,2560,65536,0.7184426625569661
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,2560,51200,0.3670506795247396
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,2560,51200,0.5653312047322591
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,2560,16384,0.11799893379211426
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,2560,16384,0.1825269381205241
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,2560,16384,0.08527039686838786
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,2560,12288,0.08912426630655924
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,2560,12288,0.13245440324147542
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,3072,65536,0.29728212356567385
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,2560,10240,0.08113813400268555
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,2560,10240,0.10998506546020508
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,2560,10240,0.06071466604868571
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,2560,8192,0.06553386847178141
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,2560,8192,0.0892906665802002
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,2560,12288,0.06787412961324056
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,2560,7168,0.055572267373402914
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,2560,7168,0.08029867013295491
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,2560,7168,0.04656000137329101
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,2560,6144,0.04718826611836751
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,2560,6144,0.06670506795247397
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,2560,8192,0.04928106864293416
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,2560,5120,0.03974826733271281
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,2560,5120,0.056289064884185794
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,2560,5120,0.0381930669148763
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,2560,4096,0.03208000063896179
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,2560,6144,0.04144853353500366
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,2560,4096,0.04771413405736287
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,2560,3584,0.028612265984217327
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,2560,3584,0.044050133228302
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,2560,3584,0.030023467540740967
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,2560,3072,0.02453119953473409
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,2560,3072,0.03927679856618245
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,2560,4096,0.033045333623886106
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,2560,2560,0.02095359961191813
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,2560,2560,0.03718613386154175
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,2560,3072,0.026848000288009644
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,2560,2048,0.017478400468826295
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,2560,2048,0.03165546655654907
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,2560,2560,0.02552853425343831
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,2560,1536,0.013736533125241599
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,2560,1536,0.02847786744435628
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,2560,1536,0.02060479919115702
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,2560,1024,0.009990400075912476
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,2560,1024,0.025140267610549927
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,2560,2048,0.022680532932281495
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,2560,768,0.008360532919565837
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,2560,768,0.022753065824508666
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,2560,1024,0.019693867365519205
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,2560,512,0.006723199784755707
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,2560,512,0.020898133516311646
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,2560,768,0.017753599087397258
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,2560,256,0.005459199845790863
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,2560,256,0.018178133169809978
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,2560,256,0.017476266622543334
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,2560,128,0.0051018665234247845
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,2560,128,0.01685439944267273
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,2560,128,0.01733013391494751
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,2560,64,0.00462719996770223
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,2560,64,0.018194133043289186
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,2560,32,0.004643199841181437
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,2560,32,0.016501333316167197
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,2048,65536,0.42475093205769854
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,2560,512,0.017679999272028603
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,2048,65536,0.6931519826253255
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,2048,51200,0.3328917185465495
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,2048,51200,0.5491754531860351
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,2560,51200,0.24015359878540038
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,2048,16384,0.1093173344930013
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,2048,16384,0.17950186729431153
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,2048,16384,0.08406186898549398
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,2048,12288,0.08293013572692871
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,2048,12288,0.1313258647918701
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,2560,65536,0.29976107279459635
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,2048,10240,0.06837973594665528
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,2048,10240,0.10439253648122152
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,2048,12288,0.06603413422902425
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,2048,8192,0.05342613458633423
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,2048,8192,0.08474240303039551
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,2048,10240,0.059740801652272545
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,2048,7168,0.04663893381754557
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,2048,7168,0.07383146286010742
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,2048,8192,0.048655998706817624
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,2048,6144,0.03897813161214193
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,2048,6144,0.06176533301671346
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,2048,7168,0.044819200038909913
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,2048,5120,0.03345280090967814
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,2048,5120,0.053838932514190675
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,2048,5120,0.036865067481994626
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,2048,4096,0.027243733406066895
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,2048,4096,0.045517865816752115
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,2048,6144,0.039494399229685465
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,2048,3584,0.024205867449442545
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,2048,3584,0.04237333138783773
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,2048,4096,0.03135679960250855
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,2048,3072,0.021499733130137123
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,2048,3584,0.028408533334732054
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,2048,3072,0.03762133518854777
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,2048,2560,0.018686934312184652
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,2048,2560,0.0352021336555481
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,2048,3072,0.02524799903233846
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,2048,2048,0.015210666259129844
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,2048,2048,0.031115732590357464
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,2048,2560,0.023995733261108397
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,2048,1536,0.01252906620502472
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,2048,1536,0.02683626612027486
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,2048,51200,0.2365546703338623
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,2048,1024,0.008663466572761536
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,2048,1024,0.024784000714619954
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,2048,1536,0.018321067094802856
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,2048,768,0.007416533430417378
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,2048,768,0.02285226583480835
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,2048,1024,0.017324799299240114
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,2048,512,0.005955199897289276
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,2048,2048,0.021067732572555543
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,2048,512,0.020667733748753865
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,2048,256,0.005087999999523163
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,2048,256,0.01747200091679891
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,2048,768,0.015735466281572977
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,2048,128,0.004283733169237773
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,2048,128,0.016809600591659545
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,2048,256,0.01567359964052836
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,2048,64,0.004221866528193155
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,2048,64,0.016103466351826988
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,2048,32,0.004216533402601878
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,2048,128,0.015433599551518759
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,2048,32,0.016703999042510985
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,2048,512,0.01574613352616628
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,1536,65536,0.31309439341227213
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,1536,51200,0.24303146998087563
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,1536,65536,0.6358357111612956
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,1536,51200,0.5306783994038899
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,1536,16384,0.07910613218943277
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,1536,16384,0.16310613950093586
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,1536,16384,0.06405333280563355
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,1536,12288,0.06042879819869995
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,1536,12288,0.11818772951761883
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,2048,65536,0.29831358591715496
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,1536,10240,0.0499669353167216
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,1536,10240,0.09736746946970622
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,1536,12288,0.0518613338470459
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,1536,8192,0.04080959955851237
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,1536,8192,0.07661120096842447
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,1536,10240,0.04518933296203613
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,1536,7168,0.03520533243815104
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,1536,7168,0.06634453137715658
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,1536,7168,0.03509013255437215
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,1536,6144,0.030705066521962483
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,1536,6144,0.056545066833496097
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,1536,8192,0.03741759856541951
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,1536,5120,0.026628265778223675
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,1536,5120,0.04918613433837891
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,1536,5120,0.030104533831278486
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,1536,4096,0.022150399287541707
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,1536,6144,0.03255680004755656
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,1536,4096,0.043274664878845216
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,1536,3584,0.020124799013137816
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,1536,3584,0.03958293199539185
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,1536,3584,0.024550400177637734
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,1536,3072,0.017411200205485027
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,1536,3072,0.035410133997599284
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,1536,4096,0.026745599508285523
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,1536,2560,0.015224533279736838
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,1536,2560,0.03324693242708842
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,1536,3072,0.022308266162872313
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,1536,2048,0.012600533167521157
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,1536,2048,0.029382399717966717
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,1536,2560,0.021154133478800456
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,1536,1536,0.010436266660690308
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,1536,1536,0.026612265904744463
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,1536,2048,0.01906026601791382
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,1536,1024,0.007964799801508587
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,1536,1024,0.02446720004081726
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,1536,1536,0.016902399063110352
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,1536,768,0.006797866523265838
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,1536,768,0.022696532805760703
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,1536,1024,0.01569706698258718
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,1536,512,0.005862399935722351
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,1536,512,0.020863999923070274
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,1536,768,0.015410133202870688
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,1536,256,0.005115733544031779
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,1536,256,0.018636800845464072
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,1536,512,0.015320533514022827
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,1536,128,0.0044832001129786175
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,1536,256,0.014913066228230795
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,1536,128,0.01660693287849426
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,1536,64,0.0038122666378815973
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,1536,64,0.016425599654515587
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,1536,128,0.016226133704185484
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,1536,32,0.004218666752179464
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,1536,32,0.01658560037612915
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,1024,65536,0.2112661361694336
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,1024,65536,0.5806346893310547
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,1024,51200,0.15875627199808756
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,1024,51200,0.45607147216796873
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,1536,51200,0.16550079981486004
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,1024,16384,0.05338773330052694
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,1024,16384,0.14919253985087078
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,1024,16384,0.061144534746805826
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,1024,12288,0.04049280087153117
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,1024,12288,0.10727893511454265
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,1536,65536,0.20398079554239906
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,1024,10240,0.033973332246144614
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,1024,10240,0.0871445337931315
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,1024,12288,0.04827626546223958
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,1024,8192,0.026814933617909747
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,1024,8192,0.06624639828999837
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,1024,10240,0.042005332310994466
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,1024,7168,0.023797333240509033
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,1024,7168,0.05783360004425049
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,1024,8192,0.034289065996805826
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,1024,6144,0.02069759964942932
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,1024,6144,0.04920426607131958
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,1024,7168,0.03213653365770976
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,1024,5120,0.01776319940884908
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,1024,5120,0.04493226607640584
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,1024,5120,0.027203200260798137
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,1024,4096,0.015054933230082192
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,1024,4096,0.0386954665184021
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,1024,6144,0.028883200883865357
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,1024,3584,0.013289599617322286
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,1024,3584,0.03502933184305827
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,1024,3584,0.02186346650123596
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,1024,3072,0.011753599842389424
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,1024,3072,0.03249813318252563
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,1024,4096,0.023974400758743287
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,1024,3072,0.0200053334236145
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,1024,51200,0.16224320729573566
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,1024,2560,0.01002346674601237
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,1024,2048,0.008840533097585042
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,1024,2560,0.030907734235127764
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,1024,2048,0.027965867519378663
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,1024,1536,0.007582933207352956
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,1024,1536,0.02480106751124064
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,1024,2048,0.016557866334915163
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,1024,1024,0.00589333325624466
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,1024,2560,0.018210132916768394
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,1024,1024,0.022909865776697794
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,1024,768,0.005121066669623057
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,1024,768,0.021129600207010903
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,1024,1536,0.014493866761525472
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,1024,512,0.004653866589069367
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,1024,1024,0.013716266552607218
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,1024,768,0.012828800082206725
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,1024,512,0.01900906761487325
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,1024,256,0.0038399999340375268
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,1024,256,0.018122667074203493
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,1024,128,0.0034453332424163817
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,1024,256,0.012488533059755962
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,1024,512,0.012560000022252401
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,1024,64,0.003033600002527237
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,1024,128,0.01594239970048269
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,1024,32,0.0034815999368826545
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,1024,64,0.016909867525100708
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,1024,128,0.013301333785057068
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,1024,32,0.014502400159835815
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,768,65536,0.16586666107177733
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,768,51200,0.13022720019022624
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,768,65536,0.5666986465454101
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,768,51200,0.4514922777811686
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,768,16384,0.044088534514109295
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,768,16384,0.14628267288208008
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,1024,65536,0.200873597462972
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,768,12288,0.03372693459192912
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,768,12288,0.1037706693013509
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,768,16384,0.062225067615509035
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,768,10240,0.027194666862487792
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,768,10240,0.08363839785257975
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,768,10240,0.04251840114593506
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,768,12288,0.04920320113499959
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,768,8192,0.021555199225743612
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,768,8192,0.0628330667813619
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,768,7168,0.019751467307408652
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,768,7168,0.053973333040873206
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,768,7168,0.03300266663233439
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,768,6144,0.017169066270192466
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,768,6144,0.04669866561889648
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,768,8192,0.035382401943206784
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,768,5120,0.014983466267585755
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,768,5120,0.04225920041402181
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,768,5120,0.02810773253440857
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,768,4096,0.012426666418711345
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,768,6144,0.030168533325195312
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,768,4096,0.0370250662167867
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,768,3584,0.011085866888364156
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,768,3584,0.03482133150100708
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,768,3584,0.02272746761639913
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,768,3072,0.009657599528630574
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,768,4096,0.025195733706156416
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,768,3072,0.03100159962972005
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,768,2560,0.008736000458399455
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,768,2560,0.02945706645647685
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,768,3072,0.020721065998077392
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,768,2048,0.007572266459465027
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,768,2048,0.02676266630490621
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,768,2048,0.017684266964594523
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,768,2560,0.019793067375818887
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,768,1536,0.006613333523273468
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,768,1536,0.02323733369509379
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,768,1536,0.01575040022532145
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,768,1024,0.005126399795214335
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,768,1024,0.022714666525522866
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,768,1024,0.014628266294797262
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,768,768,0.004636799792448679
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,768,768,0.020718934138615926
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,768,768,0.014108799894650779
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,768,512,0.003864533454179764
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,768,512,0.019371734062830607
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,768,51200,0.16263465881347655
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,768,256,0.003470933437347412
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,768,512,0.013710932930310568
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,768,256,0.016730666160583496
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,768,128,0.0030954666435718536
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,768,128,0.015290666619936624
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,768,256,0.01369599997997284
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,768,128,0.014060800274213156
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,768,64,0.0030954666435718536
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,768,32,0.0029663999875386557
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,768,64,0.01464959979057312
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,768,32,0.014667733510335287
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,512,65536,0.13001920382181803
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,512,51200,0.10237867037455242
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,512,65536,0.5420768102010091
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,512,51200,0.4309631983439128
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,512,16384,0.034884266058603924
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,512,16384,0.1404095967610677
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,512,16384,0.06087679862976074
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,512,12288,0.025118933121363325
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,512,12288,0.09843093554178874
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,768,65536,0.20213546752929687
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,512,10240,0.02092906634012858
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,512,10240,0.07834986845652261
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,512,12288,0.04668266773223877
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,512,8192,0.017778132359186807
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,512,8192,0.057272533575693764
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,512,10240,0.0400437315305074
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,512,7168,0.015449600418408713
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,512,7168,0.05169386863708496
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,512,8192,0.03304426670074463
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,512,6144,0.013408000270525614
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,512,6144,0.04422613382339478
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,512,6144,0.028515199820200603
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,512,7168,0.03136426607767741
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,512,5120,0.01172266701857249
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,512,5120,0.03998186588287354
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,512,4096,0.009771733482678732
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,512,4096,0.03690239985783895
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,512,5120,0.02645439902941386
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,512,3584,0.008884267012278239
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,512,3584,0.03321066697438558
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,512,4096,0.02314773400624593
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,512,3072,0.008071466783682505
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,512,3072,0.03024853269259135
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,512,3584,0.021143466234207153
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,512,2560,0.007387733459472657
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,512,51200,0.16054827372233074
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,512,2560,0.028835199276606244
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,512,2048,0.006316799918810527
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,512,2048,0.0253493328889211
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,512,2560,0.01779306729634603
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,512,3072,0.01941759983698527
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,512,1536,0.005534933507442474
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,512,2048,0.01567359964052836
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,512,1024,0.0046847999095916745
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,512,1536,0.022725333770116173
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,512,1024,0.02215999960899353
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,512,768,0.004178133110205332
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,512,768,0.02037973403930664
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,512,1536,0.014115200440088908
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,512,512,0.0037973334391911825
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,512,1024,0.012943999965985617
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,512,512,0.018797866503397622
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,512,768,0.011993599931399028
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,512,256,0.0030805334448814393
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,512,256,0.017042134205500284
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,512,128,0.002600533266862233
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,512,256,0.01246506671110789
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,512,128,0.015029333035151162
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,512,64,0.002573866645495097
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,512,64,0.016563199957211814
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,512,128,0.012436266740163167
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,512,32,0.0026677332818508146
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,512,32,0.014910933375358582
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,256,65536,0.08443840344746908
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,256,65536,0.5217226664225261
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,256,51200,0.06652373472849528
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,512,512,0.012074666221936543
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,256,51200,0.41373545328776046
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,256,16384,0.02657173275947571
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,256,16384,0.13982292811075847
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,512,65536,0.20030719439188638
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,256,12288,0.01946773330370585
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,256,12288,0.09595200220743814
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,256,16384,0.060502398014068606
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,256,10240,0.01955519914627075
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,256,10240,0.07545066674550374
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,256,12288,0.04521600008010864
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,256,8192,0.017040000359217326
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,256,10240,0.04032426675160726
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,256,8192,0.053858133157094326
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,256,7168,0.0143477330605189
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,256,7168,0.047915732860565184
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,256,8192,0.032869333028793336
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,256,51200,0.16044586499532063
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,256,6144,0.012451199690500896
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,256,7168,0.03089386622111003
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,256,6144,0.04182080030441284
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,256,5120,0.010104533036549885
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,256,5120,0.03906240065892537
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,256,4096,0.007640533149242401
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,256,4096,0.035768532752990724
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,256,6144,0.0284223993619283
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,256,5120,0.02632746696472168
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,256,3584,0.007096533477306366
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,256,3584,0.031768532594045
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,256,3072,0.0063285330931345625
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,256,3072,0.028904533386230467
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,256,3584,0.02101866602897644
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,256,2560,0.005876266459623972
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,256,3072,0.0194048007329305
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,256,4096,0.02350506583849589
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,256,2048,0.005182933310667674
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,256,2048,0.024980266888936363
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,256,65536,0.20007786750793458
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,256,2560,0.01772586703300476
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,256,1536,0.004726399978001913
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,256,1536,0.022383999824523926
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,256,1024,0.003806933263937632
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,256,1024,0.02137493292490641
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,256,1536,0.013848533233006796
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,256,768,0.0035071998834609987
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,256,2048,0.015901866555213928
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,256,1024,0.012899200121561686
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,256,768,0.019030400117238364
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,256,512,0.003035733352104823
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,256,2560,0.028125866254170732
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,256,512,0.018476800123850504
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,256,256,0.0026591998835404714
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,256,512,0.012026666601498922
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,256,256,0.016572800278663636
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,256,128,0.0025973332424958544
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,256,768,0.01199893355369568
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,256,256,0.011788800358772278
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,256,128,0.015032533804575601
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,256,64,0.0025941332181294756
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,256,32,0.00258240004380544
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,256,128,0.012058666348457337
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,256,64,0.014660267035166421
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,128,65536,0.06344853242238363
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,256,32,0.014622933665911355
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,128,51200,0.051881599426269534
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,128,65536,0.5165045420328777
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,128,51200,0.40903466542561845
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,128,16384,0.018663465976715088
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,128,16384,0.13598933219909667
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,128,12288,0.012691199779510498
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,128,12288,0.09475626945495605
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,128,16384,0.05854826768239339
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,128,10240,0.01128000020980835
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,128,10240,0.07227733135223388
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,128,12288,0.04427200158437093
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,128,8192,0.012019200126330058
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,128,8192,0.053133865197499595
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,128,10240,0.0382368008295695
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,128,7168,0.010549333691596986
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,128,7168,0.04737919966379801
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,128,7168,0.030078933636347456
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,128,6144,0.00958720048268636
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,128,8192,0.03209813237190247
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,128,51200,0.15765546162923177
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,128,6144,0.0413482666015625
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,128,5120,0.008348799745241801
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,128,5120,0.037690667311350505
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,128,4096,0.007551999886830647
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,128,4096,0.0330346663792928
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,128,5120,0.02563626567522685
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,128,3584,0.0066997334361076355
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,128,3584,0.031022934118906657
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,128,4096,0.02227946718533834
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,128,3072,0.006318933268388112
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,128,6144,0.02767146627108256
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,128,3072,0.028887466589609785
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,128,2560,0.005541333556175232
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,128,65536,0.1959114710489909
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,128,2560,0.02718293269475301
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,128,3584,0.020151466131210327
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,128,2048,0.0046741331617037455
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,128,2048,0.024717867374420166
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,128,1536,0.00420906643072764
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,128,2560,0.016899200280507405
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,128,2048,0.014831999937693277
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,128,3072,0.018288000424702962
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,128,1024,0.0034485332667827605
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,128,1024,0.021228800217310585
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,128,1024,0.012065066893895467
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,128,768,0.0034400001168251038
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,128,768,0.019782400131225585
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,128,768,0.011303466558456422
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,128,512,0.0029535998900731405
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,128,512,0.018582399686177573
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,128,512,0.010844799876213073
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,128,256,0.0026047999660174055
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,128,256,0.016615466276804606
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,128,1536,0.021042132377624513
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,128,128,0.002179199953873952
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,128,256,0.011203199625015259
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,128,128,0.014659200112024942
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,128,64,0.002162133405605952
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,128,128,0.011199999849001567
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2816,128,1536,0.012842667102813721
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,128,32,0.0021877333521842955
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,128,64,0.014457600315411887
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,64,65536,0.06410346825917562
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,128,32,0.015944533546765647
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,64,51200,0.04992213249206543
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,64,16384,0.014259200294812521
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,64,16384,0.13354132970174154
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,64,12288,0.010405332843462626
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,64,51200,0.4086154619852702
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,64,65536,0.5133301417032878
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,64,10240,0.009483733773231506
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,64,8192,0.00840106705824534
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,64,12288,0.09421226978302003
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,64,10240,0.06986880302429199
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,64,7168,0.01028053363164266
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,64,6144,0.009628799557685853
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,64,8192,0.05140373309453329
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,64,7168,0.04530986547470093
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,64,5120,0.008042666812737782
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,64,4096,0.006741333504517872
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,64,5120,0.03702933390935262
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,64,3584,0.006307200094064077
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,64,4096,0.033003733555475874
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,64,3072,0.0054666668176651
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,64,3584,0.030869332949320476
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,64,2560,0.00506986677646637
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,64,3072,0.028369067112604777
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,64,2048,0.0046015997727712
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,64,2048,0.02461013396581014
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,64,1536,0.003920000046491623
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,64,1536,0.02155733307202657
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,64,1024,0.003401600072781245
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,64,1024,0.021129600207010903
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,64,768,0.0029706666866938275
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,64,512,0.002587733417749405
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,64,512,0.01821333368619283
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,64,256,0.002551466723283132
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,64,6144,0.04118293523788452
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,64,256,0.016107733050982155
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,64,128,0.0024864000578721365
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,64,64,0.002370133250951767
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,64,128,0.014872533082962037
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,64,32,0.002170666555563609
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,64,2560,0.026809600989023845
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,64,32,0.014458666245142618
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,64,64,0.015046399831771851
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,32,65536,0.06108693281809489
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,32,51200,0.04805546601613363
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,32,16384,0.013318399588267008
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,32,51200,0.4046143849690755
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,32,65536,0.5108789443969727
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,32,12288,0.009471999605496724
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,32,16384,0.13161813418070475
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,32,10240,0.0086325337489446
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,32,8192,0.007650133470694225
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,32,12288,0.08992746671040854
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,32,10240,0.06921066443125407
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,32,7168,0.007157333195209503
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,32,6144,0.006664533416430156
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,32,8192,0.05074453353881836
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,64,768,0.020270933707555137
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,32,5120,0.006585599978764851
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,32,4096,0.008086400230725606
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,32,7168,0.046971734364827475
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,32,3584,0.007313066720962524
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,32,4096,0.032893866300582886
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,32,3072,0.006691200037797292
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,32,3584,0.031142399708429976
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,32,3072,0.0279968003431956
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,32,2560,0.006192000210285186
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,32,2048,0.004686933259169261
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,32,2560,0.02676266630490621
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,32,1536,0.0037994667887687682
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,32,2048,0.024922666947046916
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,32,1024,0.0033930666744709016
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,32,1536,0.020593067010243736
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,32,768,0.0029909332593282064
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,32,1024,0.020802134275436403
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,32,768,0.019368533293406168
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,32,512,0.0026335999369621276
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,32,256,0.0027114666998386385
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,32,512,0.017876267433166504
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,32,128,0.0021589333812395734
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,32,256,0.017056000232696534
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,32,128,0.014492799838383993
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,32,64,0.002288000037272771
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,32,64,0.014604799946149192
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,32,6144,0.04109866619110107
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2816,32,32,0.0022730665902296705
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,32,5120,0.03790719906489055
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2816,32,32,0.014270933469136557
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,65536,51200,5.430142720540365
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,65536,16384,3.161163838704427
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,65536,16384,1.616873550415039
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,65536,51200,10.03607889811198
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,65536,12288,2.36802978515625
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,65536,12288,1.1830186208089193
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,65536,10240,2.0668800354003904
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,65536,10240,1.0055466969807942
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,65536,10240,0.5920106887817382
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,65536,8192,1.483447519938151
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,65536,8192,0.8122688293457031
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,65536,12288,0.6679061253865559
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,65536,7168,1.2802635192871095
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,65536,7168,0.7063253402709961
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,65536,16384,0.88918825785319
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,65536,6144,1.1147413889567057
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,65536,6144,0.6022677103678385
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,65536,8192,0.4829226811726888
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,65536,5120,0.9655605316162109
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,65536,5120,0.5725311915079753
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,65536,7168,0.4426581382751465
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,65536,4096,0.7200298945109049
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,65536,4096,0.44348052342732747
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,65536,6144,0.38133974075317384
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,65536,3584,0.6358816146850585
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,65536,3584,0.4028629302978516
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,65536,5120,0.3495818773905436
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,65536,3072,0.5460778554280599
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,65536,3072,0.34941654205322265
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,65536,4096,0.3089482625325521
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,65536,2560,0.4553898811340332
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,65536,2560,0.2698911984761556
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,65536,3072,0.2450058619181315
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,65536,2048,0.36989545822143555
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,65536,2048,0.23599360783894857
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,65536,2560,0.2335829257965088
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,65536,1536,0.281164805094401
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,65536,1536,0.19593920707702636
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,65536,3584,0.27877225875854494
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,65536,1024,0.191101868947347
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,65536,1024,0.14571199417114258
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,65536,1024,0.1714431921641032
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,65536,768,0.1704330603281657
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,65536,768,0.12111999988555908
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,65536,1536,0.17779092788696288
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,65536,512,0.12576106389363606
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,65536,512,0.09949866930643717
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,65536,768,0.16127573649088542
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,65536,256,0.08187092940012614
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,65536,256,0.08513173262278238
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,65536,512,0.16358933448791504
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,65536,128,0.06079893509546915
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,65536,128,0.06936960220336914
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,65536,256,0.15837653477986652
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,65536,64,0.05474133491516113
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,65536,64,0.07195839881896973
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,65536,32,0.05841066837310791
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,65536,32,0.07336640357971191
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,65536,2048,0.20817492802937826
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,65536,128,0.1664394696553548
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,51200,65536,5.509794108072916
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,51200,65536,10.184383138020833
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,51200,51200,4.349759928385416
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,51200,51200,7.978765869140625
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,51200,16384,2.509668223063151
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,51200,16384,1.3396043141682943
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,51200,16384,0.7084341049194336
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,51200,12288,1.8382645924886067
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,51200,12288,0.9773824055989584
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,51200,12288,0.5337205251057943
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,51200,10240,1.5912928263346353
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,51200,10240,0.8134133021036783
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,51200,10240,0.47033173243204757
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,51200,8192,1.247048568725586
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,51200,8192,0.6427125295003255
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,65536,51200,2.9096522013346355
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,51200,7168,0.9901056289672852
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,51200,7168,0.6247093200683593
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,51200,8192,0.3836949348449707
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,51200,6144,0.8653909047444662
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,51200,6144,0.49607254664103195
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,51200,7168,0.35461546579996744
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,51200,5120,0.7028479894002279
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,51200,5120,0.40996694564819336
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,51200,6144,0.3024127960205078
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,51200,4096,0.5613418579101562
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,51200,4096,0.3492053349812826
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,51200,5120,0.27546345392862953
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,51200,3584,0.49817492167154953
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,51200,3584,0.29211839040120446
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,51200,3584,0.22169599533081055
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,51200,3072,0.4301130612691243
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,51200,3072,0.2615935961405436
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,51200,51200,2.2213343302408854
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,51200,2560,0.35015573501586916
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,51200,4096,0.24259093602498374
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,51200,2560,0.21517866452534995
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,51200,2048,0.282423464457194
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,51200,2048,0.17459413210550945
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,51200,2048,0.16567254066467285
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,51200,1536,0.21501332918802896
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,51200,1536,0.13594773610432942
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,51200,2560,0.18607786496480305
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,51200,1024,0.1514143943786621
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,51200,1024,0.10328000386555988
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,51200,3072,0.19571092923482258
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,51200,768,0.11843732992808025
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,51200,768,0.09327253500620523
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,51200,768,0.13086079756418864
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,51200,512,0.08707520167032877
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,51200,512,0.07924479643503825
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,51200,1024,0.14009920756022137
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,51200,256,0.06155733267466227
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,51200,256,0.06583893299102783
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,51200,1536,0.14347519874572753
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,51200,128,0.05273600021998087
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,51200,128,0.06180373430252075
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,51200,128,0.1307904005050659
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,51200,64,0.04727360010147095
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,51200,64,0.06376106739044189
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,51200,32,0.04934399922688802
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,51200,32,0.0642741322517395
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,51200,256,0.12890026569366456
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,16384,65536,1.9802122751871747
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,16384,65536,3.2130783081054686
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,16384,51200,2.4685897827148438
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,16384,51200,1.508037312825521
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,51200,512,0.1303008000055949
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,16384,16384,0.7445184071858724
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,16384,16384,0.4938026746114095
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,16384,16384,0.2527008056640625
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,16384,12288,0.5580928166707356
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,16384,12288,0.38122666676839195
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,16384,12288,0.19244693120320638
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,16384,10240,0.5079306602478028
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,16384,10240,0.3309194564819336
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,16384,10240,0.17165439923604328
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,16384,8192,0.4093514760335286
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,16384,8192,0.263754669825236
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,16384,8192,0.14077866872151693
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,16384,7168,0.355348269144694
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,16384,7168,0.22490347226460777
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,16384,51200,0.7482698440551758
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,16384,6144,0.29613014856974285
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,16384,6144,0.19429012934366863
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,16384,7168,0.13041173617045085
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,16384,5120,0.2486197312672933
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,16384,5120,0.16531413396199543
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,16384,5120,0.10400106906890869
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,16384,4096,0.19839466412862142
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,16384,4096,0.1373802661895752
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,16384,6144,0.112718931833903
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,16384,3584,0.16921493212382
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,16384,3584,0.12000426451365154
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,16384,3584,0.08137173652648926
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,16384,3072,0.1453759988149007
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,16384,3072,0.10497813224792481
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,16384,4096,0.09131093025207519
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,16384,2560,0.1249344031016032
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,16384,2560,0.09237759908040365
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,16384,2560,0.06824639638264975
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,16384,2048,0.10134720007578532
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,16384,2048,0.07688000202178955
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,16384,3072,0.07194773356119791
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,16384,1536,0.07694186369578043
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,16384,1536,0.06390186548233032
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,16384,1536,0.04991039832433065
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,16384,1024,0.05376000006993612
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,16384,1024,0.0536629319190979
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,16384,2048,0.058457601070404056
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,16384,768,0.04354133208592732
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,16384,768,0.047619199752807616
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,16384,1024,0.04857386747996013
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,16384,512,0.03219733238220215
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,16384,512,0.04013866583506266
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,16384,768,0.04172693490982056
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,16384,256,0.02307413419087728
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,16384,256,0.03293013374010722
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,16384,256,0.04114133516947428
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,16384,128,0.017754666010538735
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,16384,128,0.03113173246383667
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,16384,512,0.04190933307011922
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,16384,128,0.04196159839630127
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,16384,64,0.016562133034070333
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,16384,32,0.017298134167989095
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,16384,64,0.03060479958852132
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,16384,32,0.030830933650334673
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,12288,65536,1.5221088409423829
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,12288,65536,2.397078450520833
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,12288,51200,1.8123573303222655
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,12288,51200,1.2131669362386068
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,16384,65536,0.9492064158121745
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,12288,16384,0.5398250579833984
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,12288,16384,0.40732692082722977
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,51200,65536,2.892778778076172
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,12288,12288,0.4033727963765462
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,12288,12288,0.28292373021443684
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,12288,16384,0.20460799535115562
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,12288,10240,0.339242680867513
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,12288,10240,0.2535733381907145
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,12288,12288,0.15606719652811687
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,12288,8192,0.2716565450032552
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,12288,8192,0.1942325274149577
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,12288,10240,0.13961812655131023
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,12288,7168,0.24116586049397787
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,12288,7168,0.1718250592549642
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,12288,8192,0.11487893263498943
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,12288,6144,0.2058015982309977
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,12288,6144,0.14806826909383136
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,12288,7168,0.10639999707539875
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,12288,5120,0.17267519632975262
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,12288,5120,0.12339413166046143
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,12288,6144,0.0923413356145223
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,12288,4096,0.1355925401051839
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,12288,4096,0.1035434643427531
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,12288,5120,0.08395733038584391
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,12288,3584,0.1211850643157959
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,12288,3584,0.09103360176086425
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,12288,4096,0.07307199637095133
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,12288,3072,0.10237759749094646
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,12288,3072,0.08240319887797037
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,12288,3584,0.06554453372955323
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,12288,2560,0.08707306385040284
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,12288,2560,0.07266879876454671
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,12288,3072,0.05812906821568807
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,12288,2048,0.07033173243204752
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,12288,2048,0.061716266473134364
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,12288,2048,0.05142613252003988
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,12288,1536,0.05401599804560343
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,12288,1536,0.05130240122477213
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,12288,2560,0.056405333677927647
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,12288,1024,0.038199468453725176
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,12288,1024,0.04238506555557251
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,12288,1024,0.04580373366673787
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,12288,768,0.029366399844487506
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,12288,768,0.03762453397115072
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,12288,1536,0.04673279921213786
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,12288,512,0.02232426603635152
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,12288,512,0.03366506497065226
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,12288,512,0.04238080183664958
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,12288,768,0.042789332071940106
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,12288,256,0.016597333550453185
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,12288,256,0.02885226607322693
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,12288,128,0.014056533575057983
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,12288,128,0.024932267268498738
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,12288,128,0.04195839961369832
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,12288,64,0.012056533495585125
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,12288,64,0.02479040026664734
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,12288,32,0.012475732962290447
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,12288,32,0.02476373314857483
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,12288,256,0.04204479853312175
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,10240,65536,1.918115234375
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,10240,65536,1.3453984578450522
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,10240,51200,1.4118848164876303
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,10240,51200,1.0679370880126953
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,12288,51200,0.5993824005126953
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,10240,16384,0.44092585245768234
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,10240,16384,0.3433664004007975
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,10240,16384,0.1808512051900228
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,10240,12288,0.3390101432800293
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,10240,12288,0.2635658740997314
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,12288,65536,0.7540480295817058
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,10240,10240,0.2729098637898763
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,10240,10240,0.21647146542867027
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,10240,12288,0.1387466589609782
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,10240,8192,0.22667627334594725
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,10240,8192,0.17324585914611818
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,10240,10240,0.12282346884409587
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,10240,7168,0.1944266637166341
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,10240,7168,0.1504821300506592
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,10240,8192,0.10258986949920654
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,10240,6144,0.17570773760477704
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,10240,6144,0.13557333946228028
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,10240,7168,0.09422612984975179
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,10240,5120,0.14274880091349285
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,10240,5120,0.11027946472167968
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,10240,6144,0.08271466890970866
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,10240,4096,0.11658026377360027
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,10240,4096,0.09430399735768637
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,10240,5120,0.0748138666152954
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,10240,3584,0.1002079963684082
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,10240,3584,0.0831221342086792
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,10240,4096,0.06444906791051229
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,10240,3072,0.08854080041249593
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,10240,3072,0.07590080102284749
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,10240,3584,0.05913600126902262
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,10240,2560,0.07308373451232911
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,10240,2560,0.06571946541468301
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,10240,3072,0.05332906643549601
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,10240,2048,0.060670932133992515
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,10240,2048,0.05581119855244955
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,10240,2560,0.05185813506444296
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,10240,1536,0.04546026786168416
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,10240,1536,0.04557013511657715
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,10240,51200,0.5242858568827311
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,10240,2048,0.04772479931513469
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,10240,1024,0.03232426643371582
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,10240,1024,0.03911360104878743
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,10240,768,0.025044266382853193
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,10240,768,0.03464213212331136
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,10240,1536,0.04341119925181071
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,10240,512,0.01877546707789103
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,10240,512,0.031231999397277832
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,10240,768,0.03917653163274129
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,10240,512,0.039331201712290445
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,10240,256,0.015034666657447815
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,10240,256,0.027559467156728107
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,10240,128,0.01242453356583913
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,10240,128,0.025089067220687867
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,10240,128,0.03886293172836304
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,10240,64,0.010528000195821126
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,10240,64,0.023568000396092734
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,10240,32,0.010867200295130412
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,10240,32,0.022792534033457438
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,10240,256,0.03894720077514648
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,10240,1024,0.04217919905980428
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,8192,65536,1.240502421061198
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,8192,65536,1.5784842173258462
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,8192,51200,0.9637055714925131
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,8192,51200,1.2637098948160808
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,8192,16384,0.4343008041381836
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,8192,16384,0.3200768152872721
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,8192,16384,0.1544319947560628
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,8192,12288,0.3007871945699056
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,8192,12288,0.24475733439127603
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,10240,65536,0.6614069620768229
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,8192,10240,0.2518506685892741
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,8192,10240,0.18857280413309735
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,8192,12288,0.11835520267486573
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,8192,8192,0.20644159317016603
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,8192,8192,0.15362666447957357
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,8192,10240,0.10595839818318684
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,8192,7168,0.1786944071451823
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,8192,7168,0.13552640279134115
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,8192,7168,0.08139306704203288
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,8192,6144,0.1534997304280599
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,8192,6144,0.11911893685658773
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,8192,8192,0.08748799959818522
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,8192,5120,0.12733439604441327
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,8192,5120,0.10141440232594807
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,8192,5120,0.06354026794433594
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,8192,4096,0.10109120210011799
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,8192,4096,0.0865834633509318
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,8192,6144,0.07060906887054444
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,8192,3584,0.0868127981821696
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,8192,3584,0.07636906305948893
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,8192,3584,0.04875200192133586
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,8192,3072,0.07482986450195313
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,8192,3072,0.06972053050994872
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,8192,4096,0.05464106798171997
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,8192,2560,0.06397013266881307
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,8192,2560,0.059676798184712734
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,8192,2560,0.042139732837677
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,8192,2048,0.05226453145345052
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,8192,2048,0.048809599876403806
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,8192,3072,0.04322666724522908
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,8192,1536,0.03983999888102214
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,8192,1536,0.04055360158284505
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,8192,1536,0.03174613316853841
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,8192,1024,0.025568000475565594
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,8192,1024,0.0344320019086202
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,8192,2048,0.036611199378967285
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,8192,768,0.020855466524759926
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,8192,768,0.03140799999237061
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,8192,1024,0.03051946759223938
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,8192,512,0.015513599912325541
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,8192,512,0.027847466866175334
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,8192,768,0.027739733457565308
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,8192,256,0.011708799997965496
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,8192,256,0.022677334149678548
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,8192,256,0.02690560022989909
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,8192,128,0.010007466872533162
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,8192,128,0.022399999698003135
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,8192,512,0.02723520000775655
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,8192,64,0.009143466750780743
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,8192,64,0.021691733598709108
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,8192,32,0.00939626693725586
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,8192,32,0.02163413365681966
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,8192,128,0.027635200818379717
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,7168,65536,1.2867925008138021
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,7168,65536,1.096123758951823
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,7168,51200,0.9957632064819336
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,7168,51200,0.8640629450480143
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,8192,51200,0.4493247985839844
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,7168,16384,0.30924800237019856
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,7168,16384,0.2742197354634603
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,7168,16384,0.13133866786956788
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,7168,12288,0.23546346028645834
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,7168,12288,0.19956480662027995
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,8192,65536,0.5686943689982097
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,7168,10240,0.2224031925201416
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,7168,10240,0.17238826751708985
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,7168,12288,0.10176640351613361
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,7168,8192,0.17553067207336426
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,7168,8192,0.13503573735555013
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,7168,10240,0.09200106461842855
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,7168,7168,0.1564832051595052
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,7168,7168,0.12278613249460857
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,7168,8192,0.07602346738179525
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,7168,6144,0.1332693338394165
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,7168,6144,0.10836479663848878
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,7168,7168,0.07107626597086589
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,7168,5120,0.11186347007751465
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,7168,5120,0.09222826957702637
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,7168,5120,0.05564159949620565
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,7168,4096,0.08915200233459472
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,7168,6144,0.06203840176264445
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,7168,4096,0.07893013159434001
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,7168,3584,0.07616533438364664
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,7168,3584,0.06980053583780924
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,7168,3584,0.04362666606903076
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,7168,3072,0.06511253515879313
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,7168,3072,0.06244159936904907
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,7168,4096,0.048042666912078855
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,7168,2560,0.05436373154322306
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,7168,2560,0.055257598559061684
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,7168,3072,0.038846933841705324
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,7168,2048,0.04485546747843425
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,7168,2048,0.04605973164240519
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,7168,2560,0.03780906597773234
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,7168,1536,0.03424959977467855
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,7168,1536,0.04006826480229696
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,7168,1536,0.030583467086156207
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,7168,1024,0.023870933055877685
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,7168,1024,0.033506135145823165
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,7168,51200,0.37928107579549153
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,7168,768,0.01981226603190104
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,7168,768,0.031215999523798627
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,7168,2048,0.03428053458531698
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,7168,512,0.014709333578745524
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,7168,512,0.027124265829722088
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,7168,768,0.02762773235638936
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,7168,1024,0.029768532514572142
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,7168,256,0.011090133587519329
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,7168,256,0.02421120007832845
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,7168,128,0.009623466928799947
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,7168,128,0.0210698664188385
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,7168,512,0.02736213405927022
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,7168,64,0.00801386684179306
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,7168,128,0.027611732482910156
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,7168,64,0.02259520093599955
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,7168,32,0.008481066425641377
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,7168,256,0.027156267563501996
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,7168,32,0.020821332931518555
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,6144,65536,1.1025248209635417
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,6144,65536,0.9844277064005533
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,6144,51200,0.8430826822916666
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,6144,51200,0.7738698959350586
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,6144,16384,0.2719818751017252
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,6144,16384,0.25185599327087405
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,6144,16384,0.13090133666992188
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,6144,12288,0.20615785916646323
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,6144,12288,0.18370240529378254
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,7168,65536,0.47672001520792645
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,6144,10240,0.16949547131856282
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,6144,10240,0.14952534039815266
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,6144,12288,0.10107839902242025
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,6144,8192,0.13936746915181478
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,6144,8192,0.12359999815622966
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,6144,10240,0.09108053048451742
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,6144,7168,0.1206933339436849
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,6144,7168,0.11135359605153401
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,6144,7168,0.06993813514709472
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,6144,6144,0.10408533414204915
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,6144,6144,0.09972266356150308
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,6144,8192,0.0756053368250529
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,6144,5120,0.08905920187632242
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,6144,5120,0.08475946585337321
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,6144,5120,0.054843731721242274
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,6144,6144,0.060533332824707034
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,6144,4096,0.07212479909261069
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,6144,4096,0.0719551960627238
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,6144,3584,0.06248746713002523
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,6144,3584,0.0638154665629069
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,6144,3584,0.04307093222935994
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,6144,3072,0.05405439933141073
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,6144,3072,0.057750399907430014
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,6144,4096,0.04800533453623454
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,6144,2560,0.0468725323677063
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,6144,2560,0.05149759848912557
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,6144,3072,0.038770135243733725
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,6144,2048,0.03829760154088338
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,6144,2048,0.044810668627421064
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,6144,2560,0.03785706758499145
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,6144,1536,0.030954666932423908
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,6144,1536,0.0372597336769104
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,6144,1536,0.030195200443267824
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,6144,1024,0.023367466529210408
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,6144,1024,0.031089067459106445
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,6144,2048,0.033986131350199386
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,6144,768,0.016657066345214844
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,6144,768,0.02762133280436198
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,6144,1024,0.02938026587168376
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,6144,512,0.012147200107574464
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,6144,512,0.027461334069569902
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,6144,768,0.027270400524139406
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,6144,512,0.026867200930913288
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,6144,256,0.011613866686820984
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,6144,256,0.02363626758257548
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,6144,128,0.008142933249473572
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,6144,128,0.020273067553838096
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,6144,128,0.026958932479222614
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,6144,256,0.026770132780075073
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,6144,64,0.007095466554164887
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,6144,32,0.007387733459472657
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,6144,64,0.02020373344421387
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,6144,32,0.022269866863886514
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,5120,65536,0.9521898905436197
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,5120,65536,0.9132415771484375
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,5120,51200,0.7208682378133138
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,5120,51200,0.7311242421468098
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,6144,51200,0.37676372528076174
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,5120,16384,0.2260298728942871
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,5120,16384,0.22702399889628092
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,5120,16384,0.10908160209655762
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,5120,12288,0.1709503968556722
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,5120,12288,0.16497599283854167
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,6144,65536,0.47317012151082355
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,5120,10240,0.14846720695495605
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,5120,10240,0.13685866991678874
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,5120,12288,0.08479893207550049
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,5120,8192,0.11753919919331868
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,5120,8192,0.11044373512268066
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,5120,10240,0.07711359659830729
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,5120,7168,0.10346773465474446
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,5120,7168,0.1012447992960612
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,5120,8192,0.06463573376337686
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,5120,6144,0.08681920369466146
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,5120,6144,0.09003520011901855
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,5120,7168,0.05959999958674113
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,5120,5120,0.07291519641876221
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,5120,5120,0.07531946500142415
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,5120,6144,0.0528767983118693
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,5120,4096,0.05805973211924235
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,5120,4096,0.06261866490046183
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,5120,5120,0.04740159908930461
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,5120,3584,0.051520001888275144
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,5120,3584,0.05569066603978475
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,5120,3584,0.03834240039189656
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,5120,3072,0.044582398732503255
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,5120,3072,0.05114133358001709
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,5120,4096,0.04200960000356038
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,5120,2560,0.037826132774353025
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,5120,2560,0.04574933449427287
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,5120,2560,0.03309973279635112
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,5120,2048,0.030686932802200317
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,5120,2048,0.039212799072265624
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,5120,3072,0.0343669335047404
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,5120,1536,0.024296534061431885
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,5120,1536,0.03457066615422567
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,5120,51200,0.3094485282897949
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,5120,1024,0.01767359972000122
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,5120,1024,0.028997333844502766
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,5120,1536,0.027446399132410686
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,5120,768,0.0139957328637441
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,5120,768,0.027049599091211955
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,5120,1024,0.026409600178400678
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,5120,512,0.010614400108655293
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,5120,512,0.024754132827123007
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,5120,2048,0.03012266755104065
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,5120,256,0.007948799928029378
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,5120,256,0.023126399517059325
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,5120,256,0.02858026623725891
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,5120,128,0.006683733562628429
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,5120,512,0.02873813311258952
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,5120,128,0.01876373291015625
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,5120,64,0.006258133550484974
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,5120,64,0.019066667556762694
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,5120,32,0.006347733239332835
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,5120,32,0.018631466229756675
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,5120,128,0.02799573342005412
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,4096,65536,0.7008277257283528
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,4096,65536,0.8130399703979492
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,5120,768,0.0287882665793101
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,4096,51200,0.5702943801879883
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,4096,51200,0.6320565541585286
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,4096,16384,0.18108372688293456
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,4096,16384,0.19868799845377605
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,4096,16384,0.10496426423390706
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,4096,12288,0.13705600102742513
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,4096,12288,0.14443306922912597
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,5120,65536,0.38692054748535154
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,4096,10240,0.12050879796346028
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,4096,10240,0.12134293715159099
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,4096,12288,0.08108479976654052
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,4096,8192,0.09749759833017985
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,4096,8192,0.09977706273396811
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,4096,10240,0.07339626948038737
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,4096,7168,0.0862506628036499
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,4096,7168,0.09053119818369547
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,4096,8192,0.06099626620610556
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,4096,6144,0.07315093676249186
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,4096,6144,0.08201493422190348
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,4096,7168,0.05553386608759562
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,4096,5120,0.06117013295491537
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,4096,5120,0.06654400030771891
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,4096,6144,0.04882880051930745
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,4096,4096,0.04841173489888509
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,4096,4096,0.055859200159708654
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,4096,5120,0.044148266315460205
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,4096,3584,0.042744533220926924
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,4096,3584,0.04949546655019124
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,4096,3584,0.034509865442911784
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,4096,3072,0.03689173460006714
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,4096,3072,0.04532586733500163
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,4096,4096,0.038660267988840744
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,4096,2560,0.031191466252009074
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,4096,2560,0.04164906740188599
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,4096,2560,0.02937600016593933
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,4096,2048,0.025486934185028075
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,4096,3072,0.030503465731938677
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,4096,2048,0.037275731563568115
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,4096,1536,0.02036799987157186
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,4096,1536,0.03243199984232585
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,4096,1536,0.022290132443110146
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,4096,1024,0.015154133240381876
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,4096,1024,0.028620799382527668
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,4096,2048,0.025421865781148273
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,4096,768,0.012197333574295043
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,4096,768,0.025187200307846068
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,4096,1024,0.02153600056966146
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,4096,768,0.01925440033276876
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,4096,512,0.009494400024414063
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,4096,512,0.022872533400853476
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,4096,256,0.007157333195209503
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,4096,256,0.020357332626978555
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,4096,256,0.01901866594950358
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,4096,128,0.006274133423964183
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,4096,128,0.018802134195963542
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,4096,128,0.019025067488352455
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,4096,64,0.005436799923578898
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,4096,64,0.01825493375460307
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,4096,32,0.00553599993387858
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,4096,32,0.017262933651606242
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,4096,512,0.019052799542744955
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,3584,65536,0.6688938776652018
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,3584,65536,0.7783893585205078
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,3584,51200,0.5202624003092449
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,3584,51200,0.6619914372762044
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,4096,51200,0.30670401255289714
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,3584,16384,0.17299200693766276
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,3584,16384,0.1974474589029948
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,3584,16384,0.08449493249257406
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,3584,12288,0.13069120248158772
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,3584,12288,0.14538666407267253
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,4096,65536,0.38440532684326173
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,3584,10240,0.1111957311630249
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,3584,10240,0.11915520032246907
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,3584,12288,0.06726079781850179
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,3584,8192,0.09117226600646973
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,3584,8192,0.0976149320602417
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,3584,10240,0.061782399813334145
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,3584,7168,0.07855786482493082
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,3584,7168,0.08801493644714356
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,3584,8192,0.050271999835968015
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,3584,6144,0.0680021365483602
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,3584,6144,0.07926293214162192
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,3584,7168,0.047387735048929853
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,3584,5120,0.05503253142038981
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,3584,5120,0.06382506688435872
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,3584,5120,0.03878399928410848
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,3584,6144,0.04222826560338338
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,3584,4096,0.04445759852727254
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,3584,4096,0.05441173315048218
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,3584,3584,0.040508798758188885
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,3584,3584,0.04928106864293416
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,3584,3584,0.03054186701774597
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,3584,3072,0.035306668281555174
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,3584,3072,0.045526401201883955
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,3584,4096,0.03412479956944783
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,3584,2560,0.02935466567675273
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,3584,2560,0.04002559979756673
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,3584,3072,0.027617067098617554
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,3584,2048,0.024784000714619954
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,3584,2048,0.03509013255437215
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,3584,2560,0.026891734202702838
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,3584,1536,0.01973653237024943
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,3584,1536,0.031658667325973514
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,3584,1536,0.021861332654953002
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,3584,1024,0.013371733824412027
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,3584,1024,0.026946133375167845
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,3584,2048,0.023859200874964397
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,3584,768,0.011020800471305848
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,3584,768,0.024580266078313193
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,3584,1024,0.02108906706174215
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,3584,512,0.008475733796755473
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,3584,512,0.022326399882634483
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,3584,768,0.01944533387819926
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,3584,256,0.006030933558940887
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,3584,256,0.02002453406651815
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,3584,512,0.019436800479888917
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,3584,128,0.005976533393065134
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,3584,256,0.019028266270955406
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,3584,128,0.020666666825612388
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,3584,64,0.005474133292833964
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,3584,128,0.019040000438690186
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,3584,64,0.018336000045140584
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,3584,32,0.005854933460553487
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,3584,32,0.017181867361068727
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,3072,65536,0.5336693445841472
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,3072,65536,0.7107103983561198
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,3072,51200,0.40076052347819013
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,3072,51200,0.581381352742513
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,3584,51200,0.2288287957509359
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,3072,16384,0.13200960159301758
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,3072,16384,0.17711466153462727
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,3072,16384,0.08353919982910156
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,3072,12288,0.09997440179189046
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,3072,12288,0.12998080253601074
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,3584,65536,0.2859615961710612
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,3072,10240,0.08453760147094727
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,3072,10240,0.10819413661956787
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,3072,10240,0.060178132851918545
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,3072,8192,0.06841493447621663
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,3072,8192,0.08815253575642903
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,3072,12288,0.06614613135655721
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,3072,7168,0.060558935006459556
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,3072,7168,0.0806325356165568
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,3072,8192,0.04878720045089722
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,3072,7168,0.0460810661315918
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,3072,6144,0.05215573310852051
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,3072,6144,0.07115306854248046
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,3072,5120,0.044556800524393717
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,3072,5120,0.05885226726531982
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,3072,5120,0.03790719906489055
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,3072,4096,0.03616426785786946
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,3072,4096,0.04970133304595947
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,3072,6144,0.04087786674499512
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,3072,3584,0.032013867298762
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,3072,3584,0.04524053335189819
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,3072,3584,0.03015786608060201
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,3072,3072,0.02800426681836446
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,3072,3072,0.04214719931284587
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,3072,4096,0.03333760102589925
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,3072,2560,0.024487467606862386
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,3072,2560,0.0391050656636556
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,3072,3072,0.026850134134292603
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,3072,2560,0.026054400205612182
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,3072,2048,0.02040106654167175
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,3072,2048,0.034340266386667886
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,3072,1536,0.016591999928156534
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,3072,1536,0.030858665704727173
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,3072,1536,0.021423999468485513
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,3072,1024,0.012247467041015625
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,3072,1024,0.02677759925524394
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,3072,2048,0.023516800006230673
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,3072,768,0.01046399970849355
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,3072,768,0.022858667373657226
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,3072,768,0.01944213310877482
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,3072,512,0.007940266529719036
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,3072,512,0.021950932343800862
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,3072,1024,0.020269866784413657
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,3072,256,0.006355200211207073
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,3072,256,0.021163733800252278
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,3072,512,0.018959999084472656
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,3072,128,0.0054058666030565895
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,3072,128,0.018074667453765868
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,3072,128,0.018984532356262206
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,3072,64,0.004643199841181437
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,3072,64,0.016555733482042947
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,3072,256,0.018703999121983846
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,3072,32,0.005115733544031779
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,3072,32,0.017448532581329345
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,2560,65536,0.44369707107543943
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,2560,65536,0.6646154403686524
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,2560,51200,0.3448543866475423
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,2560,51200,0.5247658729553223
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,3072,51200,0.2270261287689209
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,2560,16384,0.1123744010925293
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,2560,16384,0.16866025924682618
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,2560,16384,0.08155413468678793
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,2560,12288,0.08535253206888835
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,2560,12288,0.1232757329940796
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,3072,65536,0.2844085375467936
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,2560,10240,0.07804693380991617
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,2560,10240,0.10321599642435711
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,2560,10240,0.05932693481445313
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,2560,8192,0.06170346736907959
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,2560,12288,0.0649621327718099
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,2560,8192,0.08272639910380045
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,2560,7168,0.05374293327331543
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,2560,7168,0.07421867052714029
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,2560,7168,0.04537599881490072
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,2560,6144,0.045902931690216066
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,2560,6144,0.06487573385238647
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,2560,8192,0.04789653221766154
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,2560,5120,0.037611734867095944
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,2560,5120,0.052892800172170004
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,2560,6144,0.040727468331654866
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,2560,4096,0.03145813345909119
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,2560,5120,0.03656426668167114
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,2560,4096,0.04654506842295329
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,2560,3584,0.02777600089708964
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,2560,3584,0.04224746624628703
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,2560,3584,0.029389866193135578
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,2560,4096,0.03214933276176453
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,2560,3072,0.02394346594810486
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,2560,3072,0.03922666708628337
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,2560,2560,0.02136853337287903
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,2560,2560,0.03618026574452718
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,2560,3072,0.026340266068776447
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,2560,2048,0.01695893406867981
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,2560,2048,0.031144533554712934
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,2560,2560,0.024936532974243163
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,2560,1536,0.013974400361378989
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,2560,1536,0.028758400678634645
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,2560,1536,0.01983573238054911
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,2560,1024,0.010036266843477885
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,2560,1024,0.024793599049250284
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,2560,2048,0.022318933407465616
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,2560,768,0.008389332890510559
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,2560,768,0.022740266720453896
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,2560,51200,0.22962133089701334
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,2560,512,0.006788266698519389
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,2560,512,0.02071146567662557
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,2560,768,0.017463467518488564
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,2560,256,0.005499733487764994
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,2560,256,0.01840959986050924
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,2560,256,0.017369600137074788
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,2560,128,0.004709333181381226
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,2560,128,0.016722132762273155
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,2560,1024,0.018984532356262206
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,2560,128,0.017406932512919106
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,2560,64,0.004257066547870636
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,2560,32,0.004546133180459341
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,2560,64,0.016642133394877114
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,2560,32,0.016501333316167197
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,2048,65536,0.34744106928507484
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,2560,512,0.017143466075261436
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,2048,65536,0.6199498494466146
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,2048,51200,0.27459306716918946
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,2048,51200,0.4865536053975423
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,2048,16384,0.09141440391540527
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,2048,16384,0.15509653091430664
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,2048,16384,0.08059093157450357
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,2048,12288,0.06952640215555826
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,2048,12288,0.11384320259094238
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,2560,65536,0.28788798650105796
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,2048,10240,0.06062293450037638
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,2048,10240,0.09389119942982992
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,2048,10240,0.056517334779103603
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,2048,8192,0.04882560173670451
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,2048,8192,0.07399679819742838
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,2048,12288,0.06411626736323038
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,2048,7168,0.042796798547108966
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,2048,7168,0.06673173109690347
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,2048,7168,0.0432426651318868
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,2048,6144,0.036246399084726974
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,2048,6144,0.05745173295338949
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,2048,8192,0.046088532606760664
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,2048,5120,0.029502934217453
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,2048,5120,0.04817066589991252
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,2048,5120,0.034414935111999514
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,2048,4096,0.023912533124287923
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,2048,4096,0.04284160137176514
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,2048,6144,0.038389333089192706
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,2048,3584,0.0217685341835022
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,2048,3584,0.039293865362803146
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,2048,3584,0.02776319980621338
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,2048,3072,0.019330133994420372
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,2048,3072,0.03660906553268432
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,2048,4096,0.03025599916776021
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,2048,2560,0.016286933422088624
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,2048,2560,0.03294826745986938
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,2048,2560,0.02318506638209025
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,2048,2048,0.013677866260210673
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,2048,2048,0.029100799560546876
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,2048,3072,0.024179200331370033
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,2048,1536,0.011194666226704914
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,2048,1536,0.026951466004053754
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,2048,1536,0.017485866943995156
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,2048,1024,0.00865280032157898
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,2048,1024,0.024605866273244223
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,2048,2048,0.020232532421747842
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,2048,768,0.007233066856861115
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,2048,768,0.022101332743962608
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,2048,1024,0.017002665996551515
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,2048,512,0.006004266440868378
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,2048,512,0.020821332931518555
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,2048,768,0.01527253290017446
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,2048,256,0.005051733553409576
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,2048,256,0.016993065675099693
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,2048,256,0.01493013302485148
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,2048,128,0.004238933324813843
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,2048,512,0.01541759967803955
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,2048,128,0.01617280046145121
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,2048,64,0.003789866715669632
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,2048,64,0.017272533973058064
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,2048,32,0.003841066608826319
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,2048,32,0.01562879979610443
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,1536,65536,0.2737173398335775
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,1536,65536,0.599777094523112
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,2048,128,0.015178666512171427
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,1536,51200,0.2203082720438639
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,1536,51200,0.45393921534220383
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,2048,51200,0.22619199752807617
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,1536,16384,0.07189973195393881
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,1536,16384,0.14681280453999837
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,1536,16384,0.06040853261947632
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,1536,12288,0.054202667872111
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,1536,12288,0.10618240038553875
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,2048,65536,0.2861301422119141
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,1536,10240,0.044336001078287765
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,1536,10240,0.0869429349899292
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,1536,12288,0.04911359945933024
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,1536,8192,0.036467198530832926
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,1536,8192,0.06750079790751139
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,1536,10240,0.04337919950485229
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,1536,7168,0.031363199154535934
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,1536,7168,0.06025280157725016
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,1536,8192,0.03608106772104899
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,1536,6144,0.02780906756718953
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,1536,6144,0.053282133738199865
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,1536,7168,0.03392746845881144
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,1536,5120,0.023386667172114052
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,1536,5120,0.04541866779327393
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,1536,6144,0.030935466289520264
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,1536,4096,0.019900800784428914
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,1536,4096,0.03941333293914795
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,1536,5120,0.0285589337348938
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,1536,3584,0.01762239933013916
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,1536,3584,0.037088000774383546
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,1536,4096,0.025446399052937822
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,1536,3072,0.015899733702341715
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,1536,3584,0.02353066603342692
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,1536,3072,0.03477546771367391
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,1536,2560,0.01393066644668579
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,1536,2560,0.03179413278897603
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,1536,2560,0.02023786703745524
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,1536,2048,0.011236266295115153
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,1536,2048,0.028620799382527668
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,1536,3072,0.0214303990205129
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,1536,1536,0.009194667140642803
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,1536,1536,0.025623466571172076
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,1536,1536,0.016246400276819863
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,1536,1024,0.007500799993673961
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,1536,1024,0.02377706567446391
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,1536,2048,0.018286933501561485
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,1536,768,0.006284800171852112
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,1536,768,0.02182506720225016
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,1536,1024,0.016004266341527303
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,1536,512,0.0055285334587097164
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,1536,768,0.014932266871134438
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,1536,512,0.020924800634384157
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,1536,256,0.004321066538492838
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,1536,256,0.01845226685206095
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,1536,51200,0.1533418655395508
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,1536,128,0.003930666546026865
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,1536,512,0.01458026667435964
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,1536,256,0.014552533626556396
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,1536,64,0.0034613333642482757
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,1536,64,0.01705066760381063
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,1536,32,0.0037759999434153237
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,1536,128,0.015820800264676412
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,1536,128,0.01685973405838013
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,1536,32,0.01633280018965403
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,1024,65536,0.19363199869791667
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,1024,51200,0.15164373715718588
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,1024,65536,0.5489877065022786
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,1024,51200,0.418286927541097
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,1024,16384,0.04951146841049194
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,1024,16384,0.13516586621602375
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,1024,16384,0.058101332187652587
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,1024,12288,0.03749973376592
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,1024,12288,0.09649813175201416
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,1536,65536,0.19059839248657226
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,1024,10240,0.030755199988683063
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,1024,10240,0.07659839789072673
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,1024,12288,0.0448853333791097
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,1024,8192,0.024999467531840007
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,1024,8192,0.05816320180892944
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,1024,10240,0.039666132132212324
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,1024,7168,0.021498666206995646
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,1024,7168,0.0517408013343811
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,1024,8192,0.033163734277089435
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,1024,6144,0.019374932845433554
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,1024,6144,0.04751360019048055
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,1024,7168,0.030495999256769817
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,1024,5120,0.01634986698627472
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,1024,5120,0.04109546740849813
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,1024,5120,0.025634133815765382
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,1024,4096,0.0138154665629069
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,1024,6144,0.027774933973948163
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,1024,4096,0.03575786749521891
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,1024,3584,0.012242133418718975
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,1024,3584,0.03297386765480041
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,1024,4096,0.022669865687688192
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,1024,3072,0.011062399546305338
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,1024,3072,0.030867199103037517
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,1024,3584,0.021042132377624513
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,1024,2560,0.00942186713218689
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,1024,2560,0.02914560039838155
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,1024,3072,0.018716800212860107
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,1024,2048,0.008065066734949748
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,1024,2048,0.0262773334980011
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,1024,2560,0.018217599391937254
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,1024,2048,0.016115199526151022
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,1024,1536,0.007125333448251088
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,1024,1536,0.02424959937731425
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,1024,51200,0.15119786262512208
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,1024,1024,0.0059008002281188965
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,1024,1536,0.014033066232999167
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,1024,1024,0.022402133544286093
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,1024,768,0.005015466610590617
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,1024,768,0.021901865800221763
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,1024,512,0.004507733384768168
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,1024,1024,0.013363200426101684
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,1024,768,0.012273066242535909
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,1024,512,0.020824533700942994
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,1024,256,0.0037632000943024955
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,1024,256,0.017242666085561117
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,1024,256,0.012449066837628682
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,1024,512,0.012433066964149475
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,1024,128,0.0035391998787721
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,1024,128,0.014629333217938741
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,1024,64,0.003193599979082743
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,1024,64,0.014471466342608133
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,1024,128,0.012827733159065246
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,1024,32,0.014578133821487427
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,768,65536,0.15693333943684895
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,768,65536,0.5233248074849446
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,768,51200,0.12117973168690999
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,1024,32,0.003272533416748047
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,768,51200,0.4256725311279297
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,768,16384,0.04058560132980347
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,768,16384,0.13021120230356853
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,768,16384,0.05851093530654907
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,768,12288,0.030321067571640013
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,768,12288,0.09238293170928955
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,1024,65536,0.18777920405069987
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,768,10240,0.024796799818674723
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,768,10240,0.07317226727803548
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,768,12288,0.046462933222452804
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,768,8192,0.02030613422393799
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,768,10240,0.040839465459187825
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,768,8192,0.055221335093180335
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,768,7168,0.017829332749048868
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,768,7168,0.049659732977549234
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,768,7168,0.031751465797424314
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,768,6144,0.01572160025437673
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,768,6144,0.04502826531728109
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,768,8192,0.0333621342976888
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,768,5120,0.013386666774749756
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,768,5120,0.03930986722310384
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,768,5120,0.02726400097211202
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,768,4096,0.011251200238863628
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,768,6144,0.028917332490285236
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,768,4096,0.03498986562093099
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,768,3584,0.010039466619491576
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,768,3584,0.033086933692296344
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,768,3584,0.022380799055099487
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,768,3072,0.009175466497739156
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,768,3072,0.03089066743850708
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,768,4096,0.02392959992090861
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,768,2560,0.008029866715272267
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,768,2560,0.02876799901326497
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,768,3072,0.02021546761194865
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,768,2048,0.007155199845631917
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,768,2048,0.025778132677078246
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,768,2560,0.01901866594950358
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,768,2048,0.017044266064961754
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,768,1536,0.006357333560784657
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,768,1536,0.023538132508595787
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,768,1024,0.005108266572157542
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,768,1024,0.021913599967956544
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,768,1536,0.014939733346303306
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,768,768,0.004645333190759023
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,768,1024,0.014528000354766845
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,768,768,0.019827200969060262
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,768,512,0.0038143999874591826
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,768,512,0.018577067057291667
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,768,768,0.013772799571355184
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,768,256,0.0034122665723164878
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,768,512,0.013651200135548911
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,768,256,0.01766293247540792
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,768,128,0.002998399982849757
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,768,51200,0.15210132598876952
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,768,128,0.015619200468063355
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,768,256,0.013708800077438354
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,768,64,0.003047466774781545
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,768,32,0.0029909332593282064
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,768,64,0.014499200383822122
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,768,32,0.016401066382726034
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,512,65536,0.14343679745992025
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,768,128,0.013795199990272521
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,512,51200,0.11388053099314373
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,512,51200,0.39801174799601235
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,512,65536,0.49869651794433595
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,512,16384,0.038786133130391434
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,512,16384,0.12666560014088948
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,512,16384,0.05758080085118612
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,512,12288,0.02919680078824361
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,512,12288,0.08984533150990805
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,768,65536,0.18845973014831544
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,512,10240,0.02309760053952535
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,512,10240,0.06847786903381348
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,512,12288,0.04443519910176595
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,512,8192,0.01819733381271362
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,512,8192,0.05263253450393677
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,512,10240,0.037945600350697835
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,512,8192,0.03169599970181783
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,512,7168,0.016375466187795003
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,512,7168,0.048205868403116865
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,512,6144,0.014518400033315023
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,512,6144,0.04314773480097453
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,512,7168,0.029988267024358112
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,512,5120,0.012777599692344665
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,512,5120,0.03722773392995198
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,512,6144,0.027313067515691118
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,512,4096,0.01034773290157318
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,512,4096,0.03471146821975708
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,512,5120,0.025123200813929242
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,512,3584,0.009221333265304565
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,512,3584,0.03102826674779256
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,512,4096,0.022359466552734374
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,512,3072,0.008390399813652038
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,512,51200,0.14983146985371906
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,512,3072,0.029717334111531574
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,512,3584,0.020439465840657554
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,512,2560,0.007612800101439159
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,512,2560,0.02728640039761861
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,512,2048,0.006772266825040181
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,512,2048,0.025043199459711712
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,512,2560,0.01738986571629842
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,512,1536,0.005801600217819214
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,512,2048,0.015414399902025857
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,512,1536,0.022665599981943764
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,512,1024,0.0047189335028330484
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,512,1024,0.021486934026082358
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,512,1536,0.013666133085886637
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,512,768,0.004296533266703288
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,512,1024,0.01284160017967224
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,512,3072,0.01876693367958069
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,512,768,0.019180800517400107
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,512,512,0.0039381332695484165
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,512,512,0.019504000743230186
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,512,256,0.0030218665798505146
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,512,768,0.01244586706161499
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,512,256,0.01636693378289541
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,512,128,0.002598399917284648
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,512,128,0.016017066438992818
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,512,512,0.012160000205039979
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,512,64,0.0028245332340399425
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,512,128,0.011973333358764649
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,512,32,0.002810666710138321
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,512,32,0.014475733041763306
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,256,65536,0.07335786819458008
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,256,65536,0.4776000022888184
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,512,256,0.012011733651161195
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,256,51200,0.058728531996409095
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,512,64,0.01480959951877594
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,256,51200,0.38016319274902344
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,256,16384,0.022966400782267252
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,256,16384,0.1265994628270467
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,512,65536,0.1869226614634196
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,256,12288,0.018756266434987387
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,256,12288,0.08409173488616943
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,256,12288,0.04197653134663899
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,256,16384,0.05708373387654623
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,256,10240,0.017061332861582436
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,256,10240,0.06351039807001749
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,256,8192,0.014533332983652749
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,256,8192,0.04836053450902303
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,256,10240,0.03768959840138753
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,256,7168,0.012437333663304646
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,256,8192,0.03173653284708659
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,256,7168,0.045075198014577225
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,256,6144,0.011175466577212016
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,256,6144,0.04044373432795207
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,256,7168,0.030114134152730305
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,256,5120,0.009738666812578838
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,256,6144,0.02714879910151164
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,256,5120,0.0356768012046814
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,256,4096,0.007503999769687653
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,256,4096,0.032935466368993124
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,256,5120,0.025518933931986492
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,256,3584,0.0070816000302632645
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,256,4096,0.02221119999885559
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,256,3584,0.030002133051554365
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,256,3072,0.006276266773541768
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,256,3072,0.028990934292475384
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,256,3584,0.020606933037439983
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,256,2560,0.005769599974155426
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,256,3072,0.018230400482813516
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,256,2560,0.026997333765029906
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,256,51200,0.14942186673482258
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,256,2048,0.005129600067933401
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,256,2048,0.02513706684112549
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,256,1536,0.004653866589069367
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,256,2560,0.017362133661905924
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,256,2048,0.01575146714846293
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,256,1536,0.023336533705393472
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,256,1024,0.02110613385836283
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,256,1536,0.013706666231155396
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,256,768,0.0034261333445707956
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,256,1024,0.013032533725102744
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,256,768,0.020694400866826376
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,256,1024,0.003769599894682566
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,256,512,0.003089066594839096
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,256,768,0.012106666962305706
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,256,512,0.018926932414372762
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,256,256,0.002937600016593933
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,256,256,0.016531200210253397
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,256,512,0.012170666456222534
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,256,65536,0.18627413113911945
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,256,256,0.012006400028864543
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,256,128,0.016407466928164163
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,256,64,0.002219733347495397
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,256,64,0.01444906691710154
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,256,32,0.0026208000878492994
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,256,32,0.014860799908638
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,128,65536,0.06709226767222086
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,128,65536,0.47153174082438154
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,256,128,0.0027232001225153605
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,128,51200,0.05149333477020264
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,256,128,0.01202133297920227
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,128,16384,0.013363200426101684
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,128,51200,0.37474985122680665
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,128,16384,0.12126719951629639
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,128,12288,0.012026666601498922
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,128,12288,0.08522026538848877
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,128,12288,0.04032320181528727
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,128,10240,0.011193600296974183
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,128,10240,0.06401386658350626
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,128,10240,0.036534400780995686
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,128,8192,0.01123306651910146
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,128,16384,0.05587519804636637
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,128,8192,0.04754453500111898
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,128,7168,0.010759466886520385
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,128,7168,0.04526079893112182
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,128,7168,0.028857600688934327
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,128,51200,0.14613547325134277
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,128,6144,0.010032000144322713
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,128,8192,0.03054400086402893
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,128,6144,0.04030079841613769
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,128,5120,0.009217066566149394
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,128,5120,0.035004798571268717
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,128,4096,0.007577600081761678
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,128,4096,0.0314517339070638
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,128,6144,0.02600319981575012
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,128,3584,0.00684799998998642
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,128,3584,0.030001066128412884
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,128,4096,0.02105813423792521
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,128,65536,0.18258879979451498
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,128,3072,0.006297599772612255
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,128,3072,0.028178133567174274
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,128,5120,0.024385066827138265
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,128,2560,0.005058133105436961
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,128,3584,0.019800533850987755
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,128,2048,0.004636799792448679
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,128,2560,0.027589333057403565
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,128,3072,0.01774186690648397
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,128,2048,0.024119466543197632
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,128,1536,0.0042303999265035
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,128,2560,0.016565333803494772
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,128,1536,0.023107200860977173
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,128,1024,0.0034677334129810332
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,128,1024,0.020770132541656494
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,128,1536,0.012534399827321371
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,128,768,0.0033759998778502146
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,128,1024,0.01202239990234375
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,128,768,0.01872746745745341
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,128,512,0.0030080000559488933
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,128,512,0.017816533644994102
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,128,768,0.011178666353225708
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,128,256,0.0025792000194390613
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,128,512,0.011172266801198323
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,128,256,0.01634880006313324
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,128,128,0.002201599876085917
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,128,256,0.010781866312026978
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,128,128,0.014839466412862143
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,128,64,0.0022079999248186748
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,128,32,0.002178133279085159
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,128,128,0.0107722669839859
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,128,64,0.014556800325711569
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,128,32,0.01407360037167867
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,64,65536,0.05678613185882568
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,64,51200,0.045203200976053876
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,64,16384,0.01226026713848114
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,64,16384,0.11866239706675212
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,64,12288,0.010120532910029094
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,64,51200,0.36880534489949546
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,64,10240,0.010699733098347982
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,64,12288,0.08063466548919677
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,64,8192,0.009178666273752849
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,64,10240,0.061211733023325596
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,64,7168,0.008752000331878663
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,64,8192,0.04582933187484741
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2561,128,2048,0.014582399527231851
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,64,6144,0.007945600152015685
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,64,7168,0.04315199851989746
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,64,5120,0.007897600034872691
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,64,4096,0.006720000008742015
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,64,5120,0.03510719935099284
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,64,6144,0.039450665314992264
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,64,4096,0.031128533681233722
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,64,3584,0.006297599772612255
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,64,3072,0.005481599768002828
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,64,3584,0.029021867116292316
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,64,2560,0.005102933446566264
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,64,3072,0.02757333318392436
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,64,2048,0.004640000065167745
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,64,2560,0.02654293378194173
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,64,1536,0.0038751999537150065
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,64,2048,0.02330986658732096
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,64,1024,0.003398400048414866
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,64,1536,0.021270400285720824
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,64,768,0.0029685333371162414
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,64,1024,0.021038933595021566
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,64,512,0.0025941332181294756
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,64,768,0.018632533152898152
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,64,65536,0.46791893641153975
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,64,256,0.002595199892918269
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,64,512,0.01718399922053019
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,64,128,0.0021685334543387097
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,64,64,0.0021685334543387097
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,64,256,0.016478932897249856
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,64,128,0.01492693324883779
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,64,32,0.002221866697072983
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,64,64,0.014453333616256715
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,32,65536,0.05464959939320883
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,32,51200,0.04406826496124268
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,32,65536,0.46605974833170577
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,32,51200,0.3703872044881185
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,32,16384,0.011513599753379821
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,32,12288,0.00956053336461385
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,32,16384,0.11844800313313801
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,32,12288,0.07941013177235921
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,32,10240,0.00885759989420573
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,32,8192,0.007982933521270752
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,32,10240,0.0607968012491862
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,32,7168,0.0073749333620071415
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,32,8192,0.04746666749318441
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,32,6144,0.007160533467928569
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,32,7168,0.04384746551513672
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,32,5120,0.009976533055305482
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,32,6144,0.04083733161290486
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,32,4096,0.008593066533406576
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,32,5120,0.03501439889272054
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,32,3584,0.007924266656239827
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,32,4096,0.032374399900436404
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,64,32,0.014379733800888061
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,32,3072,0.007124266525109608
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,32,3584,0.02898240089416504
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,32,2560,0.006298666695753734
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,32,3072,0.027805866797765096
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,32,2560,0.026858667532602947
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,32,1536,0.0038133333126703895
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,32,2048,0.022924800713857017
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,32,1024,0.003436800092458725
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,32,1536,0.02145493427912394
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,32,768,0.0029696000119050344
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,32,1024,0.020707199970881142
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,32,512,0.0026367999613285064
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,32,768,0.018568533658981323
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,32,256,0.0025727999707063037
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,32,512,0.017534933487574258
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,32,256,0.015146666765213012
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,32,128,0.01511253317197164
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,32,64,0.0021898667017618816
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,32,64,0.014670933286348978
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,32,32,0.002194133400917053
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2561,32,32,0.014044800400733947
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,32,2048,0.0046304002404212955
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2561,32,128,0.0021759999295075734
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,65536,16384,3.1463648478190107
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,65536,51200,5.2511444091796875
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,65536,16384,1.6406421661376953
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,65536,12288,2.2914708455403643
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,65536,51200,9.827396647135417
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,65536,12288,1.2059338887532554
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,65536,10240,2.0307071685791014
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,65536,10240,0.9680458704630533
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,65536,10240,0.5830495834350586
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,65536,8192,1.476902389526367
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,65536,8192,0.8384298960367838
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,65536,12288,0.6537866592407227
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,65536,7168,1.2907039642333984
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,65536,7168,0.6686869303385417
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,65536,16384,0.8588789621988931
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,65536,6144,1.0740331013997397
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,65536,6144,0.5742197036743164
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,65536,8192,0.48395201365152996
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,65536,5120,0.8721482594807943
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,65536,5120,0.49983892440795896
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,65536,7168,0.4476778666178386
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,65536,4096,0.6711125055948893
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,65536,4096,0.3968778610229492
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,65536,6144,0.38611199061075846
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,65536,3584,0.5862037022908528
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,65536,3584,0.3772885322570801
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,65536,5120,0.35772692362467445
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,65536,3072,0.5127519925435384
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,65536,3072,0.30498879750569663
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,65536,4096,0.31903254191080727
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,65536,2560,0.4286304155985515
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,65536,2560,0.25854080518086753
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,65536,2560,0.23992427190144858
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,65536,2048,0.3467925389607748
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,65536,2048,0.22172907193501792
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,65536,3072,0.2559754689534505
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,65536,1536,0.2675989468892416
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,65536,1536,0.16980160077412923
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,65536,3584,0.28599039713541663
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,65536,1024,0.18274346987406415
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,65536,1024,0.12381866772969563
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,65536,1024,0.17420266469319662
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,65536,768,0.14213333129882813
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,65536,768,0.10330986976623535
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,65536,1536,0.17702399889628093
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,65536,512,0.10571306546529133
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,65536,512,0.0900544007619222
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,65536,2048,0.2126538594563802
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,65536,256,0.07063360214233398
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,65536,256,0.07482346693674723
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,65536,512,0.1597546736399333
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,65536,128,0.06578239997227987
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,65536,128,0.06799999872843424
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,65536,768,0.1650613307952881
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,65536,64,0.05496213436126709
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,65536,64,0.07003626823425294
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,65536,32,0.05659519831339518
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,65536,32,0.0697760025660197
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,65536,256,0.1597333272298177
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,65536,128,0.1616437276204427
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,51200,65536,5.330341084798177
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,51200,65536,9.652220662434896
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,51200,51200,4.355149841308593
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,51200,51200,7.7586817423502605
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,51200,16384,2.4247690836588545
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,51200,16384,1.2764661153157553
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,51200,16384,0.689735476175944
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,51200,12288,1.7974154154459636
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,51200,12288,0.9528703689575195
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,51200,12288,0.5220575968424479
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,51200,10240,1.489249038696289
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,51200,10240,0.7758837381998698
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,51200,10240,0.46304747263590496
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,51200,8192,1.1342379252115884
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,51200,8192,0.6215936024983724
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,65536,51200,2.7495318094889325
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,51200,7168,0.9564128239949545
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,51200,7168,0.5312063852945964
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,51200,7168,0.3523189226786295
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,51200,6144,0.7941312154134115
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,51200,6144,0.48197011947631835
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,51200,8192,0.3826346715291341
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,51200,5120,0.6735850652058919
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,51200,5120,0.3923850695292155
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,51200,5120,0.2799861272176107
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,51200,4096,0.5423744201660157
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,51200,4096,0.3237087885538737
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,51200,6144,0.30414292017618816
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,51200,3584,0.46395734151204426
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,51200,3584,0.27490240732828775
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,51200,3584,0.22376532554626466
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,51200,3072,0.39935146967569984
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,51200,3072,0.25263360341389973
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,51200,4096,0.25101332664489745
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,51200,2560,0.3406325340270996
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,51200,2560,0.20872106552124023
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,51200,51200,2.164777628580729
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,51200,2048,0.27429866790771484
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,51200,2048,0.16881173451741535
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,51200,2560,0.1895594596862793
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,51200,1536,0.21198719342549643
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,51200,1536,0.13338665962219237
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,51200,3072,0.1989759922027588
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,51200,1024,0.14358506202697754
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,51200,1024,0.10161386330922444
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,51200,2048,0.1679797331492106
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,51200,768,0.11291626294453938
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,51200,768,0.08555200099945068
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,51200,1024,0.13839359283447267
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,51200,512,0.08168213367462158
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,51200,512,0.07539093494415283
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,51200,1536,0.14130345980326336
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,51200,256,0.05618346532185873
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,51200,256,0.06170346736907959
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,51200,256,0.12902506987253826
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,51200,128,0.045081599553426104
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,51200,128,0.05773226817448934
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,51200,768,0.13125332991282146
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,51200,64,0.04344853162765503
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,51200,64,0.05606933434804281
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,51200,32,0.04302613337834676
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,51200,32,0.05770453214645386
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,51200,128,0.13139306704203288
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,16384,65536,3.1006144205729167
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,51200,512,0.1293002684911092
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,16384,65536,1.9279637654622395
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,16384,51200,1.4804213205973307
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,16384,51200,2.4326324462890625
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,16384,16384,0.69989439646403
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,16384,16384,0.47422507603963215
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,16384,16384,0.25236053466796876
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,16384,12288,0.5171466509501139
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,16384,12288,0.34965651830037436
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,16384,12288,0.1931381384531657
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,16384,10240,0.4351893424987793
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,16384,10240,0.28953278859456383
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,16384,10240,0.17264746030171713
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,16384,8192,0.34852587381998695
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,16384,8192,0.23750294049580895
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,16384,8192,0.14235092798868815
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,16384,7168,0.3046016057332357
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,16384,7168,0.20953173637390138
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,16384,51200,0.7518880208333333
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,16384,7168,0.13213333288828533
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,16384,6144,0.18000213305155438
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,16384,6144,0.2649610678354899
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,16384,5120,0.22131627400716147
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,16384,5120,0.16000426610310872
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,16384,5120,0.10439146359761556
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,16384,4096,0.1752639929453532
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,16384,4096,0.12396586736043294
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,16384,6144,0.1147157351175944
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,16384,3584,0.1512053330739339
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,16384,3584,0.10828053156534831
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,16384,4096,0.0926367998123169
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,16384,3072,0.13276159763336182
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,16384,3072,0.0964682658513387
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,16384,3584,0.08358720143636068
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,16384,2560,0.1096000035603841
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,16384,2560,0.08431466420491537
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,16384,2560,0.07029120127360025
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,16384,2048,0.089137069384257
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,16384,2048,0.0705013354619344
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,16384,3072,0.07267093658447266
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,16384,1536,0.0696949323018392
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,16384,1536,0.05768746534983317
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,16384,1536,0.04986879825592041
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,16384,1024,0.0485482652982076
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,51200,65536,2.7916704813639326
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,16384,1024,0.049132800102233885
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,16384,768,0.0381440003712972
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,16384,768,0.042455466588338216
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,16384,2048,0.05945493380228678
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,16384,512,0.02797546585400899
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,16384,512,0.037862400213877365
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,16384,768,0.04184533357620239
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,16384,256,0.020012799898783365
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,16384,256,0.03755413293838501
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,16384,512,0.04166186650594075
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,16384,128,0.01696959932645162
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,16384,256,0.04086720148722331
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,16384,128,0.03541546662648519
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,16384,64,0.020696532726287842
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,16384,64,0.036184533437093096
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,16384,32,0.021915733814239502
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,16384,32,0.0350165327390035
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,16384,1024,0.04893866777420044
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,16384,128,0.04177920023600261
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,12288,65536,1.6128885904947918
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,12288,65536,2.530718994140625
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,12288,51200,1.9313524881998698
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,12288,51200,1.2598475138346354
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,12288,16384,0.5763541539510091
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,12288,16384,0.4076554616292317
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,16384,65536,0.9471296310424805
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,12288,12288,0.4341429392496745
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,12288,12288,0.308242130279541
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,12288,16384,0.20459839502970376
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,12288,10240,0.33651841481526695
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,12288,10240,0.24246613184611002
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,12288,12288,0.1572437286376953
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,12288,8192,0.2733525276184082
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,12288,8192,0.19379307428995768
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,12288,10240,0.1401631991068522
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,12288,7168,0.23489920298258463
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,12288,7168,0.16720959345499675
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,12288,8192,0.11592426300048828
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,12288,6144,0.2025013287862142
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,12288,6144,0.14845439592997234
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,12288,7168,0.10741439660390217
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,12288,5120,0.16570773124694824
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,12288,5120,0.12239680290222169
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,12288,6144,0.09372693697611491
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,12288,4096,0.13337705930074056
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,12288,4096,0.10308266480763753
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,12288,5120,0.08557439645131429
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,12288,3584,0.11613547007242839
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,12288,3584,0.09092373053232829
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,12288,3584,0.06583466529846191
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,12288,3072,0.10036479632059733
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,12288,3072,0.08228586514790853
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,12288,4096,0.07439146836598715
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,12288,2560,0.0847658634185791
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,12288,2560,0.07299839655558268
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,12288,2560,0.056834133466084805
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,12288,2048,0.06887359619140625
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,12288,2048,0.0612394650777181
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,12288,3072,0.05806933244069418
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,12288,1536,0.05299093325932821
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,12288,1536,0.051193598906199136
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,12288,2048,0.051173333326975504
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,12288,1024,0.03743786811828613
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,12288,1024,0.043886931737264
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,12288,1536,0.04651519854863485
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,12288,768,0.029011199871699016
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,12288,768,0.03907519976298014
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,12288,768,0.042414931456247966
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,12288,512,0.023481599489847817
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,12288,512,0.03396159807840983
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,12288,1024,0.045799465974171956
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,12288,256,0.016588800152142844
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,12288,256,0.02908586661020915
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,12288,512,0.041891201337178545
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,12288,128,0.013218133648236593
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,12288,128,0.030158933003743487
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,12288,256,0.04163306554158529
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,12288,64,0.016681599617004394
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,12288,64,0.030778666337331135
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,12288,32,0.016977065801620485
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,12288,128,0.041883734862009685
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,12288,32,0.02904319961865743
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,10240,65536,1.3431391398111978
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,10240,65536,1.934823481241862
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,10240,51200,1.4650975545247396
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,10240,51200,1.054468282063802
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,12288,51200,0.6020213445027669
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,10240,16384,0.4529866536458333
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,10240,16384,0.3435157457987467
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,10240,16384,0.1808234691619873
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,10240,12288,0.34301331837972004
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,10240,12288,0.2548415978749593
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,12288,65536,0.7589333216349284
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,10240,10240,0.29082988103230795
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,10240,10240,0.21156373023986816
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,10240,12288,0.13953493436177572
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,10240,8192,0.23443093299865722
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,10240,8192,0.177731196085612
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,10240,10240,0.12430079778035481
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,10240,7168,0.20292906761169432
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,10240,7168,0.15023147265116374
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,10240,8192,0.10313173135121663
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,10240,6144,0.17549333572387696
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,10240,6144,0.1342741330464681
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,10240,7168,0.09571946461995443
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,10240,5120,0.14316266377766926
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,10240,5120,0.10963199933369953
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,10240,6144,0.0840831995010376
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,10240,4096,0.11494932969411213
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,10240,4096,0.0937119960784912
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,10240,5120,0.07613973617553711
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,10240,3584,0.09994986852010092
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,10240,3584,0.0830410639444987
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,10240,4096,0.06573653221130371
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,10240,3072,0.0868117332458496
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,10240,3072,0.0756661335627238
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,10240,51200,0.5231370608011882
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,10240,2560,0.07152000268300375
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,10240,2560,0.06604373455047607
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,10240,3584,0.05919466813405355
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,10240,2048,0.058813865979512533
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,10240,2048,0.05543680191040039
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,10240,2560,0.052019198735555015
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,10240,1536,0.045127467314402266
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,10240,1536,0.04570560057957967
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,10240,3072,0.052668801943461095
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,10240,1024,0.031941332419713336
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,10240,1024,0.03915946483612061
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,10240,2048,0.047429335117340085
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,10240,768,0.02469546596209208
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,10240,768,0.03342399994532268
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,10240,768,0.03893440167109172
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,10240,512,0.018857600291570027
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,10240,512,0.032953600088755294
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,10240,1024,0.04216426610946655
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,10240,256,0.01698453426361084
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,10240,256,0.028946133454640706
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,10240,1536,0.0429535984992981
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,10240,128,0.012423466642697651
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,10240,128,0.023961599667867026
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,10240,128,0.03884266614913941
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,10240,64,0.010429867108662923
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,10240,64,0.024153600136439003
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,10240,32,0.010877866546312969
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,10240,256,0.03882346550623576
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,10240,32,0.022850133975346885
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,10240,512,0.03885333140691121
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,8192,65536,1.1531115214029948
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,8192,65536,1.500055440266927
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,8192,51200,1.142185592651367
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,8192,51200,0.9825610478719076
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,8192,16384,0.36005865732828773
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,8192,16384,0.30176852544148763
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,10240,65536,0.661572265625
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,8192,12288,0.2638805389404297
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,8192,12288,0.21540373166402182
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,8192,16384,0.155077330271403
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,8192,10240,0.23997653325398766
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,8192,10240,0.18818559646606445
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,8192,10240,0.10644266605377198
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,8192,8192,0.1929130713144938
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,8192,8192,0.15123945871988934
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,8192,12288,0.11941013336181641
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,8192,7168,0.1665013313293457
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,8192,7168,0.13334933916727704
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,8192,8192,0.08839253584543863
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,8192,6144,0.14348799387613934
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,8192,6144,0.11920320192972819
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,8192,7168,0.08241813182830811
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,8192,5120,0.1163530667622884
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,8192,5120,0.09855360190073649
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,8192,5120,0.06478826602300009
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,8192,4096,0.09317333698272705
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,8192,4096,0.0842688004175822
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,8192,6144,0.07164586385091146
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,8192,3584,0.08289493719736735
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,8192,3584,0.07478720347086588
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,8192,3584,0.04972906510035197
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,8192,3072,0.07142720222473145
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,8192,3072,0.06760213375091553
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,8192,4096,0.055544535319010414
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,8192,2560,0.059628800551096595
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,8192,2560,0.05843733151753744
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,8192,2560,0.042121601104736325
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,8192,2048,0.04887253443400065
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,8192,2048,0.04838933149973552
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,8192,3072,0.04345813194910685
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,8192,1536,0.03782080014546712
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,8192,1536,0.04113813241322835
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,8192,1536,0.03158933321634928
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,8192,1024,0.02617173393567403
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,8192,1024,0.03453973531723022
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,8192,2048,0.03659520149230957
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,8192,768,0.020520534118016562
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,8192,768,0.031178667147954302
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,8192,768,0.02709866762161255
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,8192,512,0.015478400389353433
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,8192,512,0.027999999125798543
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,8192,1024,0.030859732627868654
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,8192,256,0.011783466736475626
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,8192,256,0.024861866235733034
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,8192,256,0.026577067375183106
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,8192,128,0.00915839970111847
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,8192,128,0.02218773365020752
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,8192,512,0.027101866404215497
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,8192,64,0.008523733417193095
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,8192,64,0.02265920042991638
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,8192,32,0.008774399757385254
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,8192,32,0.02070080041885376
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,8192,128,0.02719786763191223
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,7168,65536,1.198185602823893
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,7168,65536,1.0619989395141602
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,7168,51200,0.9540704091389974
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,7168,51200,0.8358410517374674
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,8192,51200,0.4484256108601888
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,7168,16384,0.2974272092183431
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,7168,16384,0.2651637395222982
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,7168,16384,0.13183786869049072
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,7168,12288,0.22588159243265787
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,7168,12288,0.19612800280253093
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,8192,65536,0.5667818705240886
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,7168,10240,0.1993941307067871
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,7168,10240,0.16084373792012532
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,7168,12288,0.10221760272979737
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,7168,8192,0.15785706837972005
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,7168,8192,0.12888426780700685
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,7168,10240,0.09263253211975098
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,7168,7168,0.14107093811035157
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,7168,7168,0.11571093400319417
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,7168,8192,0.07642239729563395
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,7168,6144,0.11717973550160725
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,7168,6144,0.10414079825083415
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,7168,7168,0.07135466734568277
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,7168,5120,0.09894506931304932
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,7168,5120,0.08705066839853923
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,7168,5120,0.05623466571172079
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,7168,4096,0.07864639759063721
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,7168,4096,0.07463146845499674
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,7168,6144,0.062157865365346274
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,7168,3584,0.06780800024668375
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,7168,3584,0.06583893299102783
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,7168,3584,0.04336853424708049
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,7168,3072,0.058829867839813234
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,7168,3072,0.059667201836903896
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,7168,4096,0.04866346518198649
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,7168,2560,0.05034026702245077
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,7168,2560,0.05147626797358194
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,7168,2560,0.03790080149968465
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,7168,2048,0.04081600109736125
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,7168,2048,0.043435732523600265
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,7168,3072,0.039112532138824464
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,7168,1536,0.031102933486302692
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,7168,1536,0.03824959993362427
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,7168,51200,0.37962026596069337
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,7168,1024,0.022898133595784506
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,7168,1536,0.030247465769449873
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,7168,1024,0.03254826664924622
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,7168,768,0.01807360053062439
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,7168,768,0.02879679997762044
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,7168,2048,0.03393280108769735
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,7168,512,0.012934399644533792
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,7168,768,0.027619200944900512
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,7168,512,0.026349866390228273
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,7168,1024,0.029767467578252153
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,7168,256,0.009619200229644775
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,7168,256,0.022549333175023396
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,7168,128,0.00969493289788564
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,7168,128,0.02066453297932943
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,7168,128,0.0271616001923879
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,7168,64,0.007930666704972585
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,7168,512,0.027204267183939618
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,7168,64,0.020997333526611327
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,7168,256,0.027130667368570966
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,7168,32,0.02007253368695577
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,6144,65536,1.009999974568685
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,6144,65536,1.1872117360432943
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,7168,32,0.008340266346931458
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,6144,51200,0.8716010411580404
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,6144,51200,0.8032490412394205
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,6144,16384,0.2861109415690104
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,6144,16384,0.2751253445943197
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,6144,16384,0.13147733211517335
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,6144,12288,0.2172309398651123
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,6144,12288,0.19223252932230633
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,7168,65536,0.47645333607991536
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,6144,10240,0.18233493169148762
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,6144,10240,0.14750506083170573
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,6144,12288,0.10165653228759766
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,6144,8192,0.14798186620076498
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,6144,8192,0.12137920061747234
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,6144,10240,0.09160746733347574
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,6144,7168,0.12509546279907227
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,6144,7168,0.10904320081075032
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,6144,8192,0.076472536722819
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,6144,6144,0.1085045337677002
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,6144,6144,0.09837226867675782
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,6144,7168,0.07106026808420816
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,6144,5120,0.08724586963653565
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,6144,5120,0.08250666459401448
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,6144,5120,0.05530986785888672
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,6144,4096,0.06987733046213786
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,6144,4096,0.07015679677327474
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,6144,6144,0.06162773370742798
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,6144,3584,0.06273386478424073
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,6144,3584,0.06077866554260254
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,6144,3584,0.04332799911499023
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,6144,3072,0.05408960183461507
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,6144,3072,0.05491733153661092
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,6144,4096,0.048129065831502275
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,6144,2560,0.04491413434346517
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,6144,2560,0.04901013374328613
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,6144,3072,0.038551465670267744
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,6144,2048,0.0372383991877238
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,6144,2048,0.04205226500829061
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,6144,2560,0.03758293390274048
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,6144,1536,0.028017065922419232
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,6144,1536,0.03720106681187947
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,6144,2048,0.03327893416086833
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,6144,1024,0.020422399044036865
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,6144,1024,0.03277759949366252
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,6144,1536,0.03009386658668518
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,6144,768,0.016492799917856852
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,6144,768,0.029342933495839434
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,6144,1024,0.02895680069923401
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,6144,512,0.013678933183352152
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,6144,512,0.026804266373316447
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,6144,768,0.02685333291689555
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,6144,256,0.010894933342933654
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,6144,256,0.025126399596532185
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,6144,512,0.02667413353919983
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,6144,128,0.007506133119265239
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,6144,128,0.023040000597635904
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,6144,256,0.02674773335456848
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,6144,128,0.026342399915059406
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,6144,64,0.010125866532325745
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,6144,64,0.022760534286499025
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,6144,32,0.022299732764561972
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,5120,65536,0.9413536071777344
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,6144,32,0.010428800185521444
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,5120,65536,0.9430933634440104
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,5120,51200,0.7171829223632813
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,5120,51200,0.725706672668457
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,6144,51200,0.3751733462015788
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,5120,16384,0.22570452690124512
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,5120,16384,0.23374932607014975
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,5120,16384,0.10938560167948405
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,5120,12288,0.1710015932718913
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,5120,12288,0.1646154721577962
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,6144,65536,0.47345279057820633
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,5120,10240,0.14600319862365724
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,5120,10240,0.13527040481567382
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,5120,12288,0.08421226342519125
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,5120,8192,0.11535999774932862
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,5120,8192,0.10998186270395915
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,5120,10240,0.07591893672943115
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,5120,7168,0.10166079998016357
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,5120,7168,0.10018880367279052
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,5120,8192,0.06405333280563355
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,5120,6144,0.08547413349151611
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,5120,7168,0.05976213216781616
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,5120,6144,0.09017919699350993
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,5120,5120,0.07134613196055094
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,5120,5120,0.07471040089925131
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,5120,5120,0.04721706708272298
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,5120,4096,0.05745919942855835
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,5120,4096,0.0630517323811849
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,5120,6144,0.05247040192286173
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,5120,3584,0.05125653346379598
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,5120,3584,0.05453866720199585
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,5120,3584,0.037853864828745525
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,5120,3072,0.04423146645228068
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,5120,3072,0.049457065264383954
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,5120,4096,0.041303467750549314
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,5120,2560,0.03731093406677246
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,5120,2560,0.04533546765645345
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,5120,2560,0.033214932680130003
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,5120,2048,0.030409600337346392
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,5120,2048,0.039246932665507
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,5120,3072,0.03462186654408773
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,5120,1536,0.024011733134587605
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,5120,1536,0.0343285322189331
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,5120,1536,0.027407999833424884
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,5120,1024,0.017553067207336424
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,5120,1024,0.029040000836054486
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,5120,2048,0.03004586696624756
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,5120,768,0.013794133067131042
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,5120,768,0.027914667129516603
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,5120,1024,0.02626986702283223
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,5120,512,0.010390399893124899
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,5120,512,0.024769065777460735
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,5120,768,0.029340799649556475
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,5120,256,0.009236266215642292
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,5120,51200,0.3086144129435221
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,5120,256,0.02287893295288086
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,5120,128,0.008308266599973042
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,5120,128,0.020949333906173706
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,5120,128,0.025997867186864216
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,5120,512,0.028751999139785767
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,5120,64,0.0075530668099721265
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,5120,32,0.007890133559703827
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,5120,64,0.02183253367741903
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,5120,256,0.02834026614824931
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,5120,32,0.020895999670028687
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,4096,65536,0.8575679779052734
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,4096,65536,0.8644255956013998
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,4096,51200,0.6692778905232747
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,4096,51200,0.6832053502400716
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,4096,16384,0.20509546597798667
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,4096,16384,0.22332800229390465
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,4096,16384,0.10473386446634929
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,4096,12288,0.15502400398254396
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,4096,12288,0.15912853876749675
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,5120,65536,0.38685439427693685
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,4096,10240,0.1206976016362508
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,4096,10240,0.12132906913757324
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,4096,12288,0.08155306975046793
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,4096,8192,0.09750933647155761
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,4096,8192,0.09906880060831705
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,4096,10240,0.07376533349355062
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,4096,7168,0.08393386999766031
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,4096,7168,0.09045120080312094
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,4096,8192,0.06025280157725016
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,4096,6144,0.0716480016708374
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,4096,6144,0.08143893082936605
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,4096,7168,0.05577280124028524
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,4096,5120,0.05857386589050293
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,4096,5120,0.0663424015045166
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,4096,5120,0.04417813221613566
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,4096,4096,0.04793599843978882
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,4096,4096,0.05542613267898559
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,4096,6144,0.04817066589991252
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,4096,3584,0.041833599408467606
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,4096,3584,0.049982933203379314
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,4096,4096,0.038711468378702804
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,4096,3584,0.034626134236653644
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,4096,3072,0.035988267262776694
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,4096,3072,0.04514453411102295
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,4096,2560,0.030758400758107502
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,4096,2560,0.04116053183873494
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,4096,2560,0.02932800054550171
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,4096,2048,0.024990934133529662
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,4096,2048,0.035069866975148516
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,4096,3072,0.03081386685371399
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,4096,1536,0.019618133703867592
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,4096,1536,0.031326933701833086
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,4096,2048,0.02547520001729329
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,4096,1024,0.01418880025545756
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,4096,1024,0.02746880054473877
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,4096,1024,0.02137706677118937
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,4096,768,0.011547733346621196
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,4096,768,0.02474666635195414
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,4096,768,0.018739199638366698
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,4096,512,0.008690133690834045
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,4096,1536,0.02214399973551432
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,4096,512,0.022600533564885457
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,4096,256,0.006649599969387054
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,4096,256,0.02027626633644104
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,4096,256,0.018653867642084758
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,4096,128,0.005932799975077311
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,4096,128,0.01836479902267456
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,4096,128,0.018988800048828126
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,4096,512,0.018718934059143065
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,4096,64,0.005500799914201101
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,4096,64,0.018696532646814982
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,4096,32,0.005514666438102722
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,4096,32,0.01768853267033895
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,3584,65536,0.6234677632649739
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,3584,65536,0.7356682459513346
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,3584,51200,0.46857598622639973
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,3584,51200,0.5858826955159505
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,4096,51200,0.3060511906941732
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,3584,16384,0.1504319985707601
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,3584,16384,0.18554879824320475
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,3584,16384,0.08412906328837076
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,3584,12288,0.11374399662017823
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,3584,12288,0.13576745986938477
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,4096,65536,0.38458026250203453
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,3584,10240,0.09815359910329183
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,3584,10240,0.11291733582814534
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,3584,12288,0.06632533470789591
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,3584,8192,0.08078186511993408
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,3584,8192,0.09230720202128093
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,3584,10240,0.0616096019744873
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,3584,7168,0.06903253396352133
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,3584,7168,0.0840288003285726
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,3584,7168,0.047022934754689535
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,3584,6144,0.058924798170725504
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,3584,6144,0.07486080328623454
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,3584,8192,0.04994986852010091
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,3584,5120,0.05014933347702026
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,3584,5120,0.06055573225021362
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,3584,5120,0.038228265444437665
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,3584,4096,0.04018239974975586
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,3584,4096,0.05146986643473307
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,3584,6144,0.041788800557454424
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,3584,3584,0.035087998708089194
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,3584,3584,0.04719680150349935
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,3584,3584,0.03014613389968872
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,3584,3072,0.030713599920272828
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,3584,3072,0.043253334363301595
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,3584,4096,0.033080534140268965
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,3584,2560,0.02622186740239461
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,3584,2560,0.03914133310317993
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,3584,2560,0.02648426691691081
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,3584,2048,0.021874133745829263
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,3584,2048,0.03397013346354167
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,3584,3072,0.028056534131368
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,3584,1536,0.017417599757512413
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,3584,1536,0.031086933612823487
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,3584,1536,0.02180373271306356
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,3584,1024,0.012686933080355326
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,3584,1024,0.02685439984003703
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,3584,2048,0.023753599325815836
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,3584,768,0.010261332988739014
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,3584,768,0.02360639969507853
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,3584,768,0.019368533293406168
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,3584,512,0.007993599772453308
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,3584,512,0.021286400159200032
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,3584,1024,0.02101759910583496
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,3584,256,0.006311466793219249
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,3584,256,0.01914773384730021
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,3584,256,0.019377066691716512
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,3584,128,0.00592853327592214
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,3584,128,0.018149334192276
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,3584,512,0.019001599152882895
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,3584,64,0.005030400057633718
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,3584,64,0.016927999258041383
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,3584,32,0.005044266581535339
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,3584,32,0.016693333784739174
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,3072,65536,0.546066157023112
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,3072,65536,0.718611208597819
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,3584,128,0.01938986579577128
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,3072,51200,0.42163626352945965
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,3072,51200,0.5787978490193685
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,3584,51200,0.22828906377156577
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,3072,16384,0.14156160354614258
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,3072,16384,0.18234346707661947
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,3072,16384,0.0827999989191691
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,3072,12288,0.10687039693196614
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,3072,12288,0.1337557315826416
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,3584,65536,0.2844309488932292
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,3072,10240,0.08874560197194417
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,3072,10240,0.11302719910939533
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,3072,12288,0.06600746711095175
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,3072,8192,0.07759786446889241
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,3072,8192,0.0917248010635376
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,3072,10240,0.06015040079752604
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,3072,7168,0.06662186781565348
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,3072,7168,0.08342080116271973
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,3072,7168,0.045966935157775876
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,3072,6144,0.057462398211161295
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,3072,6144,0.07287466526031494
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,3072,8192,0.04934933185577393
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,3072,5120,0.04890880187352498
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,3072,5120,0.059539198875427246
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,3072,5120,0.03766826788584392
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,3072,4096,0.039554135004679365
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,3072,4096,0.05135039885838827
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,3072,6144,0.04092479944229126
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,3072,3584,0.035181868076324466
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,3072,3584,0.046817068258921304
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,3072,4096,0.03233493367830913
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,3072,3072,0.03078506588935852
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,3072,3072,0.043186132113138834
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,3072,3072,0.02704426646232605
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,3072,2560,0.026331732670466106
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,3072,2560,0.03910613457361857
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,3072,3584,0.02974826693534851
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,3072,2048,0.02222613294919332
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,3072,2048,0.03464213212331136
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,3072,2560,0.025961599747339886
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,3072,1536,0.017861332496007284
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,3072,1536,0.03101973334948222
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,3072,2048,0.023574399948120116
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,3072,1024,0.012567466497421265
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,3072,1024,0.02699413299560547
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,3072,1536,0.021170133352279664
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,3072,768,0.010041600465774536
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,3072,768,0.024283732970555624
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,3072,51200,0.22676053047180175
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,3072,512,0.008513066172599792
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,3072,512,0.022682666778564453
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,3072,1024,0.020269866784413657
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,3072,256,0.00636053333679835
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,3072,768,0.019067732493082683
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,3072,256,0.020644267400105797
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,3072,128,0.00592853327592214
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,3072,512,0.018961066007614137
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,3072,128,0.018988800048828126
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,3072,64,0.0055871998270352686
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,3072,256,0.01834986607233683
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,3072,64,0.02006933291753133
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,3072,32,0.0056202664971351625
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,3072,128,0.018590933084487914
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,3072,32,0.01856000026067098
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,2560,65536,0.4537856101989746
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,2560,65536,0.6633962631225586
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,2560,51200,0.37257601420084635
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,2560,51200,0.5220032056172689
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,2560,16384,0.1121781349182129
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,2560,16384,0.16770025889078777
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,2560,16384,0.08127573331197104
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,2560,12288,0.08530240058898926
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,2560,12288,0.12330453395843506
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,3072,65536,0.2847381273905436
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,2560,10240,0.07552320162455241
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,2560,10240,0.09921173254648843
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,2560,10240,0.05853439966837565
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,2560,8192,0.0604586680730184
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,2560,8192,0.08060586452484131
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,2560,12288,0.06444693406422933
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,2560,7168,0.050665601094563806
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,2560,7168,0.07223146756490072
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,2560,7168,0.045025066534678145
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,2560,6144,0.04330986738204956
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,2560,6144,0.0620959997177124
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,2560,8192,0.0474453330039978
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,2560,5120,0.036954665184021
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,2560,5120,0.05229546626408895
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,2560,5120,0.03630079825719197
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,2560,6144,0.039442133903503415
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,2560,4096,0.029982932408650714
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,2560,4096,0.04562346537907918
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,2560,3584,0.026199465990066527
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,2560,3584,0.0420693318049113
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,2560,3584,0.02911893328030904
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,2560,3072,0.023025067647298177
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,2560,3072,0.03914560079574585
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,2560,4096,0.0313920001188914
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,2560,2560,0.0200000007947286
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,2560,2560,0.03501546780268351
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,2560,3072,0.025976532697677614
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,2560,2048,0.016557866334915163
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,2560,2560,0.024800000588099162
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,2560,2048,0.031124265988667806
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,2560,1536,0.0131221334139506
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,2560,1536,0.02986239989598592
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,2560,1536,0.019834667444229126
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,2560,1024,0.009708799918492635
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,2560,1024,0.02516053318977356
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,2560,2048,0.021851734320322672
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,2560,768,0.008032000064849854
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,2560,768,0.022772266467412313
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,2560,1024,0.019035732746124266
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,2560,512,0.006749866902828217
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,2560,512,0.022320000330607097
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,2560,768,0.01736853321393331
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,2560,256,0.005868799984455109
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,2560,256,0.018710400660832724
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,2560,256,0.01695573329925537
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,2560,128,0.005031466484069824
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,2560,512,0.017366399367650352
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,2560,128,0.016516266266504924
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,2560,64,0.004702933132648468
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,2560,64,0.01802133321762085
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,2560,32,0.005053866902987162
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,2560,128,0.01731520096460978
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,2560,32,0.016671999295552572
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,2048,65536,0.4132394790649414
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,2048,65536,0.6454314549763998
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,2048,51200,0.3245024045308431
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,2048,51200,0.5096255938212078
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,2560,51200,0.22778560320536295
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,2048,16384,0.10798933506011962
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,2048,16384,0.16525972684224446
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,2048,16384,0.08041600386301676
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,2048,12288,0.08235733509063721
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,2048,12288,0.12548906803131105
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,2560,65536,0.2858517328898112
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,2048,10240,0.06406186819076538
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,2048,10240,0.09657706419626871
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,2048,12288,0.06360960006713867
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,2048,8192,0.05339093208312988
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,2048,8192,0.07592639923095704
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,2048,10240,0.05706666707992554
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,2048,7168,0.0435370683670044
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,2048,7168,0.06683306694030762
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,2048,8192,0.0465173323949178
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,2048,6144,0.03760853211085002
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,2048,6144,0.057385599613189696
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,2048,7168,0.04327573378880818
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,2048,5120,0.031378134091695147
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,2048,5120,0.049848532676696776
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,2048,6144,0.03831466833750407
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,2048,4096,0.026254934072494508
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,2048,4096,0.04331839879353841
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,2048,5120,0.035131732622782394
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,2048,3584,0.022709333896636964
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,2048,3584,0.040814932187398276
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,2048,4096,0.030564266443252563
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,2048,3072,0.020563199122746786
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,2048,3072,0.03712533315022786
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,2048,3584,0.027646932999293012
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,2048,2560,0.017682133118311565
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,2048,2560,0.03471573193868001
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,2048,3072,0.02438933253288269
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,2048,2048,0.014845866958300272
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,2048,2048,0.02962239980697632
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,2048,2048,0.020247467358907065
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,2048,1536,0.011281067132949829
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,2048,1536,0.026973867416381837
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,2048,2560,0.023449599742889404
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,2048,1024,0.008391466736793519
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,2048,1024,0.025521065791447955
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,2048,1536,0.017448532581329345
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,2048,51200,0.22536320686340333
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,2048,768,0.0074080000321070354
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,2048,768,0.022366933027903237
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,2048,512,0.006343466540177663
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,2048,512,0.020779732863108316
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,2048,768,0.01530880033969879
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,2048,256,0.005151999990145365
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,2048,512,0.014891733725865683
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,2048,256,0.018785067399342856
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,2048,128,0.004678399860858917
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,2048,128,0.016598400473594666
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,2048,256,0.015293866395950317
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,2048,1024,0.017064533631006875
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,2048,64,0.004260266820589701
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,2048,32,0.004590933521588644
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,2048,128,0.015422933300336204
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,2048,64,0.01717653274536133
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,2048,32,0.01653439998626709
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,1536,65536,0.2964106559753418
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,1536,51200,0.23016853332519532
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,1536,65536,0.5989119847615559
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,1536,51200,0.49753281275431316
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,1536,16384,0.07546453475952149
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,1536,16384,0.15219626426696778
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,1536,16384,0.06022186676661173
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,1536,12288,0.05787626504898071
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,1536,12288,0.11051733493804931
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,2048,65536,0.2846783955891927
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,1536,10240,0.04741866588592529
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,1536,10240,0.09008853435516358
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,1536,10240,0.042632532119750974
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,1536,8192,0.03821333249409993
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,1536,12288,0.04863679806391398
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,1536,8192,0.0697930653889974
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,1536,7168,0.033846398194630936
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,1536,7168,0.061205331484476724
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,1536,7168,0.03343786795934041
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,1536,6144,0.029522132873535157
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,1536,6144,0.0538101315498352
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,1536,8192,0.035735468069712326
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,1536,5120,0.02590293288230896
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,1536,5120,0.04565013249715169
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,1536,6144,0.030564266443252563
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,1536,5120,0.028518400589625043
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,1536,4096,0.0211136003335317
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,1536,4096,0.041007999579111734
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,1536,3584,0.019288533926010133
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,1536,3584,0.03731306791305542
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,1536,3584,0.023135999838511147
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,1536,3072,0.017143466075261436
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,1536,3072,0.0347434679667155
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,1536,4096,0.025260800123214723
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,1536,2560,0.014807466665903726
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,1536,2560,0.032230399052302044
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,1536,3072,0.021095466613769532
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,1536,2048,0.012530133128166199
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,1536,2560,0.020320000251134236
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,1536,2048,0.028860799471537274
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,1536,2048,0.018590933084487914
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,1536,1536,0.009668266773223877
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,1536,1536,0.026500266790390015
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,1536,1024,0.007706666489442189
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,1536,1024,0.02437439958254496
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,1536,1536,0.01636799971262614
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,1536,768,0.006738133231798808
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,1536,768,0.022682666778564453
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,1536,1024,0.015756799777348836
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,1536,512,0.005475200215975444
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,1536,512,0.020282665888468422
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,1536,768,0.014940800269444785
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,1536,512,0.014963199694951376
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,1536,256,0.004673066735267639
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,1536,256,0.017607466379801432
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,1536,128,0.004199466605981191
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,1536,128,0.016229333480199178
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,1536,256,0.014478933811187745
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,1536,64,0.0038143999874591826
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,1536,128,0.015288533767064414
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,1536,64,0.01681386629740397
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,1536,32,0.003819733361403147
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,1536,32,0.01694933374722799
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,1024,65536,0.1976821263631185
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,1024,65536,0.5495018641153971
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,1024,51200,0.14967039426167805
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,1024,51200,0.43794879913330076
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,1536,51200,0.15368960698445638
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,1024,16384,0.050508801142374674
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,1024,16384,0.13598507245381672
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,1024,16384,0.0576533317565918
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,1024,12288,0.038179198900858566
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,1024,12288,0.09767573674519857
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,1536,65536,0.18928853670756024
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,1024,10240,0.03197759985923767
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,1024,10240,0.07765973409016927
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,1024,12288,0.04616959889729817
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,1024,8192,0.026180267333984375
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,1024,8192,0.059470931688944496
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,1024,10240,0.03917866547902425
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,1024,7168,0.02322453260421753
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,1024,7168,0.052602668603261314
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,1024,8192,0.0326581339041392
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,1024,6144,0.020491733153661092
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,1024,6144,0.048078934351603195
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,1024,7168,0.03012053370475769
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,1024,5120,0.017417599757512413
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,1024,5120,0.04117439985275269
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,1024,6144,0.02764799992243449
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,1024,4096,0.014828800161679586
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,1024,4096,0.03653759956359863
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,1024,5120,0.02565760016441345
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,1024,3584,0.01318186620871226
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,1024,3584,0.034439468383789064
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,1024,3584,0.02096959948539734
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,1024,3072,0.011370666821797689
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,1024,3072,0.030985599756240843
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,1024,4096,0.022666666905085245
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,1024,2560,0.01001706620057424
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,1024,2560,0.029709867636362713
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,1024,3072,0.018668800592422485
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,1024,2048,0.008430932958920796
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,1024,2048,0.026318933566411334
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,1024,51200,0.15037226676940918
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,1024,1536,0.007259733478228251
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,1024,1536,0.025013333559036253
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,1024,2560,0.018225065867106118
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,1024,1024,0.005885866781075796
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,1024,1024,0.022858667373657226
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,1024,2048,0.01572160025437673
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,1024,768,0.005049600203831991
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,1024,768,0.020814933379491172
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,1024,1536,0.01447466711203257
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,1024,512,0.004364799956480662
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,1024,512,0.019221333662668864
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,1024,768,0.012033067146937053
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,1024,256,0.003755733370780945
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,1024,256,0.016692266861597697
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,1024,512,0.012277332941691081
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,1024,128,0.0034517332911491393
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,1024,256,0.012076800068219502
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,1024,128,0.01684266726175944
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,1024,64,0.003869866579771042
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,1024,64,0.01651093363761902
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,1024,128,0.012444800138473511
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,1024,32,0.0038474666575590765
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,1024,32,0.01588159998257955
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,768,65536,0.1672320048014323
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,768,65536,0.520794677734375
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,768,51200,0.13313493728637696
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,1024,1024,0.013143466909726462
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,768,51200,0.4144298553466797
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,768,16384,0.044846932093302413
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,768,16384,0.1332202672958374
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,1024,65536,0.18669333457946777
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,768,12288,0.03440426588058472
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,768,12288,0.09515626430511474
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,768,16384,0.058465067545572916
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,768,10240,0.02834026614824931
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,768,10240,0.07468159993489584
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,768,12288,0.04655893246332805
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,768,8192,0.02299413283665975
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,768,8192,0.05775680144627889
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,768,10240,0.0403498649597168
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,768,7168,0.019900800784428914
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,768,7168,0.0497439980506897
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,768,8192,0.03376746575037638
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,768,6144,0.01761066714922587
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,768,6144,0.04673173427581787
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,768,7168,0.031724800666173295
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,768,5120,0.015013333161671957
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,768,5120,0.03915733496348063
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,768,6144,0.028563199440638225
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,768,4096,0.013209600249926248
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,768,4096,0.035087998708089194
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,768,5120,0.027215999364852906
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,768,3584,0.011223466197649638
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,768,3584,0.0328874667485555
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,768,4096,0.023907200495402018
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,768,3072,0.010290132959683736
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,768,3072,0.031040000915527343
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,768,3584,0.021898667017618813
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,768,2560,0.008757332960764568
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,768,2560,0.028883200883865357
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,768,3072,0.01983893314997355
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,768,2048,0.007913599908351897
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,768,2048,0.0268010675907135
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,768,2560,0.01935466726620992
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,768,1536,0.006723199784755707
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,768,1536,0.023116799195607503
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,768,2048,0.01697280009587606
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,768,1024,0.005523199836413065
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,768,1024,0.021252266565958657
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,768,1536,0.015280000368754067
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,768,768,0.005041066805521647
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,768,768,0.02032426595687866
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,768,1024,0.014502400159835815
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,768,512,0.004316799839337667
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,768,512,0.01879253387451172
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,768,768,0.013680000106493631
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,768,256,0.0034730667869249977
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,768,512,0.013225600123405457
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,768,256,0.016797866423924765
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,768,128,0.0030303999781608583
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,768,128,0.016290133198102318
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,768,51200,0.15149760246276855
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,768,256,0.013658666610717773
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,768,64,0.015735466281572977
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,768,128,0.013757866621017457
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,768,64,0.0035946667194366455
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,768,32,0.002997333308060964
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,768,32,0.014542933305104574
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,512,65536,0.12433280150095623
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,512,51200,0.09812266826629638
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,512,65536,0.4979850769042969
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,512,51200,0.3978069305419922
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,512,16384,0.03418346643447876
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,512,16384,0.12686186631520588
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,512,16384,0.0572703997294108
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,512,12288,0.024338134129842124
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,512,12288,0.08841066360473633
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,768,65536,0.18741226196289062
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,512,10240,0.020071466763814293
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,512,10240,0.06871999899546305
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,512,12288,0.043272535006205246
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,512,8192,0.017191465695699057
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,512,8192,0.051736533641815186
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,512,10240,0.03750613530476888
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,512,7168,0.014935466647148132
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,512,7168,0.047340798377990725
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,512,8192,0.03175999919573466
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,512,6144,0.013150933384895324
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,512,6144,0.043329068024953205
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,512,7168,0.029655466477076214
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,512,5120,0.011411199967066448
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,512,5120,0.037356801827748615
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,512,5120,0.02553066611289978
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,512,4096,0.009572266538937887
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,512,4096,0.033056000868479415
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,512,6144,0.026845866441726686
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,512,3584,0.008885332942008972
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,512,3584,0.03125333388646444
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,512,4096,0.022332799434661866
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,512,3072,0.008005333443482716
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,512,3072,0.029538132747014362
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,512,51200,0.14893120129903156
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,512,2560,0.007100800176461537
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,512,2560,0.026843732595443724
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,512,3584,0.020350933074951172
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,512,2048,0.006326400240262349
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,512,3072,0.01867413322130839
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,512,2048,0.025099732478459674
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,512,1536,0.005497600138187409
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,512,2560,0.017833600441614784
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,512,1536,0.02299626668294271
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,512,1024,0.004709333181381226
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,512,1024,0.021314134200414024
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,512,2048,0.01536853313446045
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,512,1536,0.013680000106493631
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,512,768,0.0038261334101359046
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,512,768,0.02062293291091919
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,512,1024,0.01288640002409617
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,512,512,0.003491200009981791
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,512,768,0.011985066533088683
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,512,512,0.019792000452677407
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,512,256,0.003065599997838338
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,512,256,0.01686826745669047
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,512,256,0.012053333719571431
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,512,128,0.002914133419593175
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,512,128,0.01558080017566681
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,512,128,0.012012799580891926
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,512,64,0.0026367999613285064
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,512,64,0.01557866632938385
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,512,32,0.0027615999182065325
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,512,512,0.01162986656030019
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,512,32,0.014512000481287637
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,256,65536,0.07441386381785074
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,256,65536,0.47882239023844403
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,256,51200,0.060429867108662924
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,256,51200,0.37992852528889975
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,256,16384,0.02400319973627726
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,256,16384,0.12574613094329834
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,512,65536,0.18588800430297853
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,256,12288,0.018338133891423546
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,256,12288,0.08533973693847656
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,256,12288,0.042005332310994466
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,256,10240,0.018385066588719686
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,256,10240,0.06381866534550985
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,256,16384,0.05706026554107666
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,256,8192,0.015187199910481772
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,256,8192,0.04851200183232625
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,256,10240,0.037062398592631024
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,256,7168,0.013986133535703025
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,256,7168,0.04320853153864543
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,256,8192,0.031300266583760575
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,256,6144,0.012851199507713318
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,256,6144,0.04030933380126953
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,256,7168,0.029734400908152263
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,256,5120,0.010450133681297302
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,256,5120,0.036614398161570236
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,256,5120,0.02515520056088765
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,256,6144,0.02690560022989909
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,256,4096,0.007309866448243459
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,256,4096,0.0331007997194926
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,256,3584,0.006905599931875865
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,256,3584,0.030474666754404706
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,256,51200,0.14840639432271321
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,256,3072,0.006404266754786174
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,256,3072,0.02895786762237549
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,256,4096,0.021859200795491536
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,256,2560,0.005508266886075338
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,256,2560,0.026754132906595868
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,256,3584,0.02027946710586548
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,256,3072,0.01829119920730591
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,256,2048,0.004714666803677877
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,256,2048,0.024997333685557045
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,256,1536,0.0044064000248909
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,256,1536,0.02263360023498535
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,256,1536,0.013276799519856771
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,256,1024,0.0038058665891488397
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,256,2048,0.015507200360298156
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,256,1024,0.020725333690643312
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,256,768,0.003453866640726725
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,256,768,0.01902079979578654
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,256,1024,0.012893866499265036
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,256,512,0.002994133283694585
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,256,768,0.011620266238848369
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,256,512,0.018705066045125326
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,256,2560,0.01735360026359558
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,256,256,0.0026335999369621276
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,256,512,0.0116565336783727
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,256,128,0.0025941332181294756
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,256,65536,0.18527679443359374
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,256,128,0.01585599978764852
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,256,256,0.011678933103879293
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,256,128,0.011587199568748475
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,256,32,0.002402133246262868
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,256,64,0.014696533481280008
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,256,32,0.014320000012715658
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,128,65536,0.05954346656799316
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,128,65536,0.4722751935323079
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,128,51200,0.048995200792948404
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,256,256,0.016309332847595216
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,128,51200,0.37473812103271487
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,128,16384,0.014757333199183145
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,256,64,0.0022986667851607008
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,128,16384,0.12064747015635173
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,128,12288,0.011325866977373759
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,128,12288,0.08317546844482422
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,128,16384,0.05474239985148112
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,128,12288,0.04034133354822795
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,128,10240,0.010410666465759277
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,128,10240,0.06238506635030111
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,128,8192,0.009583999713261921
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,128,8192,0.047387735048929853
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,128,10240,0.036278398831685384
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,128,7168,0.011494400103886922
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,128,8192,0.030444800853729248
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,128,7168,0.04411946535110474
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,128,7168,0.028475733598073323
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,128,6144,0.010442666212717692
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,128,6144,0.040660266081492105
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,128,51200,0.1453397274017334
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,128,5120,0.008356266220410665
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,128,5120,0.034991999467213944
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,128,65536,0.18188907305399576
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,128,4096,0.007531733314196269
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,128,5120,0.023995733261108397
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,128,4096,0.0326474666595459
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,128,3584,0.006937600175539653
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,128,3584,0.030178133646647138
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,128,4096,0.02104426622390747
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,128,3072,0.006052266558011373
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,128,3072,0.02726293404897054
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,128,2560,0.005433600147565206
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,128,6144,0.0259552001953125
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,128,2560,0.026616533597310383
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,128,2048,0.004610133171081543
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,128,2048,0.023571199178695677
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,128,3072,0.01745599905649821
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,128,3584,0.01941759983698527
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,128,2560,0.01660160024960836
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,128,1536,0.004300799965858459
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,128,1024,0.0035157332817713416
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,128,1536,0.022734934091567995
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,128,2048,0.014570666352907815
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,128,1536,0.012880000472068786
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,128,768,0.018722132841746012
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,128,1024,0.012116266290346782
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,128,512,0.0029845332105954488
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,128,768,0.01088533302148183
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,128,512,0.018631466229756675
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,128,256,0.002701866626739502
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,128,256,0.01660693287849426
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,128,256,0.010840533177057902
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,128,128,0.002502399931351344
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,128,768,0.003335466732581457
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,128,128,0.015818666418393454
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,128,64,0.0022272000710169475
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,128,128,0.010819199681282043
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,128,32,0.002181333303451538
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,128,32,0.014605866869290671
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,64,65536,0.05772906541824341
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2560,128,512,0.011160533626874287
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,64,51200,0.04627093474070231
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,64,65536,0.46921599706013994
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,64,16384,0.016898133357365928
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,64,51200,0.3709664026896159
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,64,12288,0.014037332932154336
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,64,16384,0.11998933156331379
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,64,10240,0.012442666292190551
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,64,12288,0.08241279919942221
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,128,1024,0.021569067239761354
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,64,8192,0.010774399836858113
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,64,7168,0.009967999656995137
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,64,10240,0.059614932537078856
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,64,8192,0.04593173265457153
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,64,6144,0.009147733449935913
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,64,7168,0.04326933224995931
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,64,5120,0.007914666831493378
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,64,6144,0.039084800084431964
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,64,4096,0.006737066805362702
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,128,64,0.015095466375350952
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,64,5120,0.0365066647529602
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,64,4096,0.032807467381159465
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,64,3072,0.005627733469009399
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,64,2560,0.005053866902987162
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,64,3072,0.027543467283248902
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,64,3584,0.029718399047851562
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,64,2560,0.027138133843739826
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,64,2048,0.004645333190759023
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,64,1536,0.0038965334494908653
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,64,1024,0.0033930666744709016
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,64,1536,0.02190720041592916
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,64,2048,0.02445866664250692
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,64,768,0.0029834667841593427
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,64,1024,0.021266132593154907
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,64,512,0.002630399912595749
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,64,256,0.0025439999997615816
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,64,512,0.01858773430188497
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,64,768,0.019272534052530925
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,64,256,0.015761066476504007
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,64,128,0.0022272000710169475
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,64,64,0.002169599880774816
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,64,32,0.002196266750494639
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,64,64,0.015003732840220132
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,64,128,0.015145599842071533
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,64,32,0.01397119959195455
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,32,65536,0.0559061328570048
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,32,51200,0.045262932777404785
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,32,16384,0.011825066804885865
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,32,16384,0.11651840209960937
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,32,12288,0.009337600072224934
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,32,51200,0.3700373331705729
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,64,3584,0.0062826668222745255
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,32,10240,0.008679466446240743
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,32,8192,0.007921066880226136
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,32,12288,0.07825493017832438
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,32,10240,0.05922773281733194
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,32,8192,0.04680213530858358
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,32,7168,0.04323840141296387
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,32,6144,0.006778666873772939
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,32,6144,0.039868799845377605
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,32,5120,0.0063733334342638654
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,32,5120,0.035233068466186526
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,32,4096,0.007356800138950348
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,32,4096,0.03172373374303182
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,32,3584,0.006664533416430156
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,32,3584,0.029602134227752687
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,32,3072,0.006163200239340464
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,32,65536,0.4675466537475586
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,32,3072,0.0290613333384196
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,32,2560,0.0060127998391787205
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,32,2560,0.02672213315963745
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,32,2048,0.00476800004641215
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,32,7168,0.00703359991312027
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,32,1536,0.003934933245182038
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,32,2048,0.0236842672030131
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,32,1536,0.021290665864944457
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,32,1024,0.003446399917205175
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,32,768,0.0029813334345817565
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,32,512,0.002616533388694127
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,32,1024,0.020716800292332967
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,32,256,0.002525866776704788
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,32,768,0.019827200969060262
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,32,512,0.016764799753824867
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,32,128,0.0023306667804718017
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,32,256,0.016263467073440552
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,32,64,0.0022837333381175995
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,32,128,0.014630400141080222
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2560,32,32,0.002234666546185811
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,32,32,0.014173866311709086
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2560,32,64,0.01456106702486674
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,65536,16384,2.893548838297526
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,65536,51200,4.898640950520834
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,65536,16384,1.462826665242513
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,65536,12288,2.161412302652995
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,65536,51200,9.023195393880208
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,65536,12288,1.0850772857666016
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,65536,10240,1.8734133402506512
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,65536,10240,0.8966837565104167
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,65536,10240,0.5543402353922526
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,65536,8192,1.290229288736979
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,65536,8192,0.7325589497884114
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,65536,12288,0.6279594421386718
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,65536,7168,1.2006272633870443
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,65536,7168,0.6531914393107097
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,65536,16384,0.8347295761108399
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,65536,6144,0.9901578903198243
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,65536,6144,0.5807146708170573
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,65536,8192,0.4562090555826823
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,65536,5120,0.8455135981241861
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,65536,5120,0.4620917320251465
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,65536,7168,0.4214357376098633
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,65536,4096,0.6588234583536784
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,65536,4096,0.37685972849527993
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,65536,6144,0.36427733103434246
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,65536,3584,0.5916757583618164
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,65536,3584,0.3389034589131673
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,65536,5120,0.32785174051920574
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,65536,3072,0.49797760645548506
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,65536,3072,0.3280202547709147
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,65536,4096,0.2933504104614258
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,65536,2560,0.4171157201131185
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,65536,2560,0.27970027923583984
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,65536,3584,0.26362454096476234
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,65536,2048,0.3311253229777018
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,65536,2048,0.23475947380065917
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,65536,2560,0.2172544002532959
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,65536,1536,0.29156373341878256
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,65536,1536,0.18361066182454427
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,65536,3072,0.22720746994018554
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,65536,1024,0.20007786750793458
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,65536,1024,0.13457706769307454
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,65536,1024,0.15342613855997722
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,65536,768,0.15738879839579265
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,65536,768,0.11346133550008138
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,65536,2048,0.19393919308980306
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,65536,512,0.11575893561045329
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,65536,512,0.09180160363515219
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,65536,1536,0.1632522741953532
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,65536,256,0.07578453222910563
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,65536,256,0.07556373278299967
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,65536,256,0.14685333569844564
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,65536,128,0.05765973329544068
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,65536,128,0.06552533308664957
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,65536,768,0.14408426284790038
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,65536,64,0.05013546546300253
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,65536,64,0.06590933402379354
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,65536,32,0.05261546770731608
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,65536,32,0.06629653374354044
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,65536,128,0.15027093887329102
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,65536,512,0.14539732933044433
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,51200,65536,4.972685750325521
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,51200,65536,9.08433837890625
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,51200,51200,7.341928609212239
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,51200,51200,3.9691594441731772
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,51200,16384,2.075881576538086
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,51200,16384,1.1688597361246744
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,51200,16384,0.654253896077474
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,51200,12288,1.6426026662190754
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,51200,12288,0.8707295735677084
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,51200,12288,0.49364693959554035
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,51200,10240,1.265064493815104
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,51200,10240,0.7200927734375
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,51200,10240,0.43675947189331055
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,51200,8192,1.0678549448649088
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,51200,8192,0.5779178619384766
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,65536,51200,2.6146420796712238
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,51200,7168,0.8949749628702799
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,51200,7168,0.5284192085266113
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,51200,7168,0.33250134785970054
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,51200,6144,0.7768277486165365
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,51200,6144,0.4627488136291504
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,51200,8192,0.35865599314371743
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,51200,5120,0.6313589096069336
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,51200,5120,0.3799253463745117
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,51200,5120,0.25790506998697915
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,51200,4096,0.5115392049153645
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,51200,4096,0.34000107447306316
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,51200,6144,0.28616107304890953
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,51200,3584,0.46208426157633464
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,51200,3584,0.3117504119873047
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,51200,3584,0.20276907285054527
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,51200,3072,0.39673067728678385
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,51200,3072,0.2629322687784831
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,51200,4096,0.22829759915669762
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,51200,2560,0.32354345321655276
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,51200,2560,0.22034239768981934
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,51200,51200,1.9941802978515626
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,51200,2048,0.29735679626464845
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,51200,2048,0.18327040672302247
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,51200,2560,0.17028053601582843
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,51200,1536,0.22596799532572426
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,51200,1536,0.1463136037190755
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,51200,3072,0.17994027137756347
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,51200,1024,0.16005013783772787
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,51200,1024,0.11101653575897216
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,51200,1024,0.1283189296722412
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,51200,768,0.1252063989639282
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,51200,768,0.09360960324605307
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,51200,1536,0.13059093157450358
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,51200,512,0.08991466363271078
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,51200,512,0.07714880307515462
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,51200,2048,0.15084800720214844
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,51200,256,0.05977173248926798
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,51200,256,0.06256106694539389
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,51200,256,0.11784746646881103
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,51200,128,0.0472981333732605
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,51200,128,0.057107198238372806
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,51200,128,0.12000640233357747
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,51200,64,0.04199359814325969
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,51200,64,0.058341332276662196
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,51200,32,0.044369065761566163
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,51200,32,0.05799253384272257
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,51200,512,0.11805120309193928
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,16384,65536,1.7309546152750652
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,16384,65536,2.917041015625
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,51200,768,0.1185909350713094
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,16384,51200,1.390994135538737
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,16384,51200,2.2173375447591144
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,16384,16384,0.758888562520345
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,16384,16384,0.44159892400105794
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,16384,16384,0.23242452939351402
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,16384,12288,0.48644374211629227
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,16384,12288,0.32907091776529945
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,16384,12288,0.1778058687845866
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,16384,10240,0.42709652582804364
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,16384,10240,0.27696641286214196
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,16384,10240,0.15852799415588378
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,16384,8192,0.3441258748372396
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,16384,8192,0.22493972778320312
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,16384,8192,0.13001920382181803
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,16384,7168,0.2973589261372884
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,16384,7168,0.19825812975565593
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,16384,51200,0.6727807998657227
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,16384,6144,0.25625599225362145
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,16384,6144,0.17564479509989422
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,16384,7168,0.12072319984436035
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,16384,5120,0.2097760041554769
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,16384,5120,0.15239466031392415
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,16384,6144,0.10490880012512208
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,16384,4096,0.1724511941274007
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,16384,4096,0.12399040063222248
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,16384,5120,0.0956821362177531
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,16384,3584,0.14790080388387045
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,16384,3584,0.11042239665985107
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,16384,3584,0.0747925360997518
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,16384,3072,0.12743146419525148
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,16384,3072,0.09768426418304443
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,16384,4096,0.08465173244476318
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,16384,2560,0.11623679796854655
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,16384,2560,0.08376533190409342
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,16384,3072,0.065174400806427
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,16384,2048,0.09226880073547364
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,16384,2048,0.07038613160451254
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,16384,2560,0.061470933755238856
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,16384,1536,0.07131733099619547
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,16384,1536,0.06040000120798746
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,51200,65536,2.5771530151367186
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,16384,1024,0.05036906798680624
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,16384,1024,0.04867626825968425
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,16384,1536,0.0457535982131958
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,16384,768,0.039613866806030275
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,16384,768,0.043597865104675296
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,16384,1024,0.04452053308486938
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,16384,512,0.02956800063451131
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,16384,512,0.03691413402557373
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,16384,2048,0.05198186635971069
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,16384,256,0.020489599307378134
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,16384,256,0.031224532922108965
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,16384,512,0.038070400555928544
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,16384,128,0.01775040030479431
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,16384,128,0.028753066062927247
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,16384,256,0.03741759856541951
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,16384,64,0.015288533767064414
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,16384,64,0.028587732712427778
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,16384,32,0.015438933173815408
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,16384,32,0.028972800572713214
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,16384,768,0.038574934005737305
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,16384,128,0.03840106725692749
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,12288,65536,1.396567408243815
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,12288,65536,2.012211227416992
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,12288,51200,1.578231430053711
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,12288,51200,1.0830997467041015
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,12288,16384,0.48166828155517577
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,12288,16384,0.34249067306518555
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,16384,65536,0.8483968098958334
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,12288,12288,0.35618454615275064
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,12288,12288,0.2775349299112956
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,12288,16384,0.18255893389383954
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,12288,10240,0.3386602719624837
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,12288,10240,0.23598400751749674
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,12288,12288,0.1388927936553955
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,12288,8192,0.27131732304890954
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,12288,8192,0.19222720464070636
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,12288,10240,0.1250058650970459
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,12288,7168,0.24036800066630043
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,12288,7168,0.164247465133667
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,12288,8192,0.10247360070546467
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,12288,6144,0.20693227450052895
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,12288,6144,0.1433333396911621
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,12288,7168,0.09557653268178304
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,12288,5120,0.17134720484415691
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,12288,5120,0.11956480344136555
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,12288,6144,0.08256213665008545
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,12288,4096,0.13668373425801594
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,12288,4096,0.09810026486714682
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,12288,5120,0.07517973581949869
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,12288,3584,0.11866772969563802
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,12288,3584,0.09061760107676188
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,12288,4096,0.06411840120951334
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,12288,3072,0.1035744031270345
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,12288,3072,0.0809216022491455
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,12288,3584,0.057707734902699793
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,12288,2560,0.08655146757761636
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,12288,2560,0.06955946286519368
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,12288,3072,0.05162239869435629
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,12288,2048,0.07094720204671225
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,12288,2048,0.05965226491292318
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,12288,2560,0.05003093481063843
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,12288,1536,0.052970667680104576
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,12288,1536,0.05130453507105509
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,12288,2048,0.04585280021031697
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,12288,1024,0.03777920007705689
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,12288,1024,0.04121706485748291
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,12288,1024,0.04098453521728516
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,12288,768,0.029651200771331786
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,12288,768,0.03708159923553467
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,12288,1536,0.04234026670455933
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,12288,512,0.021754666169484457
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,12288,512,0.03201280037562053
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,12288,512,0.03789333502451579
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,12288,256,0.015591466426849365
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,12288,256,0.027011199792226152
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,12288,768,0.03755093415578206
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,12288,128,0.01370133360226949
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,12288,128,0.02656746705373128
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,12288,128,0.037435734272003175
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,12288,64,0.011636267105738323
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,12288,64,0.024639999866485594
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,12288,32,0.011762133240699768
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,12288,32,0.02477759917577108
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,12288,256,0.037419732411702475
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,10240,65536,1.6999615987141925
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,10240,65536,1.2459551493326821
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,10240,51200,1.351905059814453
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,10240,51200,0.9799285252888998
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,12288,51200,0.5254677454630534
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,10240,16384,0.423961607615153
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,10240,16384,0.3203573226928711
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,10240,16384,0.15815359751383465
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,10240,12288,0.3212394714355469
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,10240,12288,0.23751680056254068
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,12288,65536,0.6621088027954102
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,10240,10240,0.2831541379292806
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,10240,10240,0.21517440478006997
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,10240,12288,0.12163946628570557
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,10240,8192,0.2353226661682129
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,10240,8192,0.16458133061726887
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,10240,10240,0.10818346341451009
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,10240,7168,0.19929493268330892
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,10240,7168,0.14247679710388184
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,10240,8192,0.08929813702901204
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,10240,6144,0.17280534108479817
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,10240,6144,0.1276138703028361
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,10240,7168,0.08359253406524658
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,10240,5120,0.14421119689941406
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,10240,5120,0.1052575985590617
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,10240,6144,0.07243306636810302
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,10240,4096,0.11419306596120198
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,10240,4096,0.08732799688975015
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,10240,51200,0.4491200129191081
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,10240,3584,0.09872213204701742
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,10240,3584,0.07987413406372071
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,10240,5120,0.06505920092264811
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,10240,3072,0.08462826410929361
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,10240,3072,0.07182613213857016
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,10240,3072,0.046452267964680986
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,10240,4096,0.05613226493199667
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,10240,2560,0.07389012972513834
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,10240,2560,0.06285973389943442
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,10240,2048,0.05825599829355875
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,10240,2048,0.052891735235850015
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,10240,3584,0.05195200045903524
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,10240,1536,0.044546135266621906
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,10240,1536,0.04541973272959392
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,10240,1536,0.03811200062433879
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,10240,1024,0.03155413269996643
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,10240,1024,0.037221332391103104
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,10240,2048,0.041553068161010745
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,10240,768,0.02476159930229187
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,10240,768,0.033139199018478394
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,10240,2560,0.0456000010172526
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,10240,512,0.018467199802398682
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,10240,512,0.028868265946706134
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,10240,1024,0.03717439969380697
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,10240,256,0.013442132870356241
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,10240,256,0.025256532430648803
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,10240,768,0.03460906744003296
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,10240,128,0.011640533804893494
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,10240,128,0.02446933388710022
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,10240,512,0.034689064820607504
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,10240,64,0.010409599542617798
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,10240,64,0.02289386590321859
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,10240,32,0.010788266857465107
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,10240,32,0.02271040081977844
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,10240,128,0.034371201197306314
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,10240,256,0.03444799979527791
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,8192,65536,1.4201620737711589
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,8192,65536,1.082118352254232
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,8192,51200,1.108907699584961
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,8192,51200,0.8497653325398764
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,8192,16384,0.33190825780232747
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,8192,16384,0.2671797434488932
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,10240,65536,0.5610282897949219
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,8192,12288,0.24409173329671224
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,8192,12288,0.19862933158874513
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,8192,16384,0.1320693333943685
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,8192,10240,0.22604907353719078
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,8192,10240,0.17214612960815429
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,8192,10240,0.09155519803365073
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,8192,8192,0.18690667152404786
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,8192,8192,0.13894400596618653
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,8192,12288,0.10094613234202068
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,8192,7168,0.15753599802652996
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,8192,7168,0.11899733543395996
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,8192,7168,0.06942719618479411
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,8192,6144,0.1354442596435547
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,8192,6144,0.10624639987945557
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,8192,8192,0.07504639625549317
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,8192,5120,0.11525546709696452
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,8192,5120,0.09492906729380289
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,8192,5120,0.05479679902394613
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,8192,4096,0.09282240072886148
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,8192,4096,0.07706773281097412
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,8192,6144,0.060350934664408364
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,8192,3584,0.07805973688761393
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,8192,3584,0.06819946765899658
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,8192,3584,0.042531200249989826
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,8192,3072,0.06649493376413981
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,8192,3072,0.06406826575597127
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,8192,4096,0.047161599000295
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,8192,2560,0.057752533753712976
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,8192,2560,0.054231464862823486
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,8192,2560,0.03615573247273763
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,8192,2048,0.04689813454945882
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,8192,2048,0.04604053497314453
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,8192,3072,0.037523198127746585
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,8192,1536,0.035974399248758955
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,8192,1536,0.040040532747904464
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,8192,1536,0.028038400411605834
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,8192,1024,0.02472533384958903
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,8192,1024,0.03347626527150472
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,8192,2048,0.032151466608047484
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,8192,768,0.020202666521072388
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,8192,768,0.031048534313837688
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,8192,768,0.02430079976717631
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,8192,1024,0.02683839996655782
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,8192,512,0.014936533570289613
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,8192,512,0.026841600735982258
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,8192,256,0.01125333309173584
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,8192,256,0.022796799739201866
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,8192,512,0.02408000032107035
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,8192,128,0.00902826686700185
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,8192,128,0.020730666319529214
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,8192,128,0.023989333709081014
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,8192,256,0.02392639915148417
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,8192,64,0.008081066608428954
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,8192,64,0.020826667547225952
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,8192,32,0.008379733562469483
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,8192,32,0.020233599344889323
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,7168,65536,0.9800682703653971
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,7168,65536,1.1996799468994142
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,7168,51200,0.9346314748128256
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,7168,51200,0.755785624186198
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,8192,51200,0.37514772415161135
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,7168,16384,0.2837098757425944
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,7168,16384,0.24290453592936195
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,7168,16384,0.13037652969360353
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,7168,12288,0.21537493069966634
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,7168,12288,0.18135040601094563
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,8192,65536,0.4716821352640788
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,7168,10240,0.17889173825581867
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,7168,10240,0.1516640027364095
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,7168,12288,0.10070186456044514
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,7168,8192,0.14488639831542968
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,7168,8192,0.12530666987101238
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,7168,10240,0.09110079606374105
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,7168,7168,0.12655146916707355
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,7168,7168,0.11200106938680013
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,7168,8192,0.07455039819081624
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,7168,6144,0.10996479988098144
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,7168,6144,0.09895040194193522
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,7168,7168,0.06917973359425864
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,7168,5120,0.09222079912821451
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,7168,5120,0.0881493330001831
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,7168,6144,0.05976959864298502
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,7168,4096,0.07492053508758545
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,7168,4096,0.07111360232035319
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,7168,5120,0.053770665327707926
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,7168,3584,0.06634453137715658
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,7168,3584,0.06633706490198771
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,7168,3584,0.04260053237279256
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,7168,3072,0.057918934027353916
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,7168,3072,0.05815360148747763
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,7168,4096,0.046834135055541994
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,7168,2560,0.04887359937032064
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,7168,2560,0.049277865886688234
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,7168,2560,0.03662079970041911
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,7168,2048,0.04241173267364502
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,7168,2048,0.04276373386383057
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,7168,3072,0.03850346803665161
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,7168,1536,0.031022934118906657
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,7168,1536,0.037564798196156816
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,7168,1536,0.03015039960543315
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,7168,1024,0.022028799851735434
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,7168,1024,0.031275733311971025
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,7168,2048,0.03329919974009196
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,7168,768,0.01783039967219035
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,7168,768,0.027010132869084675
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,7168,1024,0.029468800624211627
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,7168,512,0.013121066490809121
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,7168,512,0.025617067019144697
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,7168,768,0.027427200476328534
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,7168,256,0.00960533320903778
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,7168,51200,0.36953280766805013
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,7168,256,0.027114667495091754
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,7168,256,0.02156053384145101
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,7168,512,0.027292799949645997
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,7168,128,0.008379733562469483
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,7168,64,0.007561600208282471
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,7168,128,0.020217599471410115
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,7168,32,0.007529599964618683
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,7168,64,0.020574933290481566
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,7168,32,0.019080533583958944
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,7168,128,0.026859732468922932
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,6144,65536,1.0078229268391927
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,6144,65536,0.8984458923339844
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,6144,51200,0.7061887741088867
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,6144,51200,0.811359977722168
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,6144,16384,0.26564480463663737
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,6144,16384,0.22779839833577475
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,6144,16384,0.10705706278483074
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,6144,12288,0.1882080078125
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,6144,12288,0.16842133204142254
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,7168,65536,0.4669994672139485
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,6144,10240,0.1702773412068685
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,6144,10240,0.14550719261169434
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,6144,12288,0.08334826628367106
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,6144,8192,0.13365012804667156
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,6144,8192,0.11369493007659912
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,6144,10240,0.0757578690846761
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,6144,7168,0.12047146956125895
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,6144,7168,0.10216106573740642
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,6144,7168,0.058101332187652587
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,6144,6144,0.10308053493499755
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,6144,6144,0.09348693688710531
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,6144,8192,0.06265279849370321
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,6144,5120,0.0841269334157308
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,6144,5120,0.08012693723042806
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,6144,5120,0.045716265837351486
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,6144,4096,0.06628586848576865
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,6144,4096,0.06368746757507324
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,6144,6144,0.04962666829427083
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,6144,3584,0.060082133611043295
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,6144,3584,0.057384534676869714
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,6144,3584,0.03584213256835937
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,6144,3072,0.05089600086212158
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,6144,3072,0.05348266760508219
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,6144,4096,0.040085331598917646
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,6144,2560,0.041945600509643556
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,6144,2560,0.04581226507822673
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,6144,2560,0.03136746684710185
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,6144,2048,0.03512746493021647
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,6144,2048,0.040199466546376544
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,6144,3072,0.03261013428370158
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,6144,1536,0.027137066920598345
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,6144,1536,0.036421334743499754
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,6144,1536,0.025518933931986492
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,6144,1024,0.020165334145228066
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,6144,1024,0.029689600070317585
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,6144,2048,0.02807253400484721
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,6144,768,0.015196800231933594
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,6144,768,0.027010132869084675
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,6144,1024,0.025073067347208662
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,6144,512,0.011797333757082622
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,6144,512,0.023998934030532836
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,6144,768,0.026732800404230754
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,6144,256,0.008846933643023174
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,6144,256,0.020915200312932335
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,6144,256,0.026395734151204425
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,6144,128,0.007192533214886982
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,6144,128,0.01888213356335958
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,6144,512,0.026834134260813398
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,6144,64,0.006331733365853627
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,6144,64,0.019447465737660728
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,6144,32,0.006670933465162914
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,6144,32,0.018568533658981323
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,6144,128,0.02648426691691081
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,5120,65536,0.878773307800293
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,5120,65536,0.8228042602539063
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,5120,51200,0.6524415969848633
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,5120,51200,0.7248586654663086
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,6144,51200,0.2995039939880371
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,5120,16384,0.21410133043924967
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,5120,16384,0.20432000160217284
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,5120,16384,0.1067797342936198
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,5120,12288,0.16303466161092123
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,5120,12288,0.15270613034566244
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,6144,65536,0.3750965436299642
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,5120,10240,0.1416063944498698
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,5120,10240,0.13259627024332682
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,5120,12288,0.08381226857503256
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,5120,8192,0.1125823974609375
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,5120,8192,0.10815146764119465
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,5120,10240,0.07585919698079427
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,5120,7168,0.09762880007425943
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,5120,7168,0.09561173121134439
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,5120,8192,0.06213119824727377
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,5120,6144,0.08367253144582112
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,5120,6144,0.08451306819915771
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,5120,7168,0.05779733260472616
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,5120,5120,0.07013653119405111
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,5120,5120,0.07383999824523926
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,5120,6144,0.05019199848175049
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,5120,4096,0.0572927991549174
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,5120,4096,0.055490132172902426
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,5120,5120,0.04594133297602336
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,5120,3584,0.050348798433939614
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,5120,3584,0.05171413421630859
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,5120,3584,0.037281068166097005
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,5120,3072,0.043731200695037845
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,5120,3072,0.047729067007700604
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,5120,4096,0.04006186723709106
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,5120,2560,0.03691306511561076
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,5120,2560,0.04226453304290771
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,5120,3072,0.03362239996592204
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,5120,2048,0.028974932432174683
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,5120,2048,0.037062398592631024
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,5120,2560,0.03204053243001302
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,5120,1536,0.022356265783309938
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,5120,1536,0.03253759940465291
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,5120,1536,0.02681066592534383
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,5120,1024,0.01588373382886251
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,5120,1024,0.027564799785614012
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,5120,2048,0.029395200808842975
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,5120,768,0.012894933422406515
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,5120,768,0.025330134232838947
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,5120,1024,0.02605866591135661
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,5120,512,0.010011733571688334
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,5120,512,0.02318399945894877
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,5120,51200,0.2973600069681803
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,5120,256,0.007522133489449819
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,5120,256,0.020592000087102255
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,5120,512,0.02868586579958598
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,5120,128,0.00631466656923294
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,5120,256,0.028171734015146895
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,5120,768,0.02885546684265137
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,5120,128,0.0261354664961497
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,5120,64,0.005885866781075796
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,5120,64,0.01858666737874349
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,5120,32,0.005970133344332377
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,5120,128,0.020111999909083047
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,5120,32,0.018756266434987387
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,4096,65536,0.8214176177978516
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,4096,65536,0.8030730565388998
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,4096,51200,0.6423967997233073
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,4096,51200,0.6336117426554362
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,4096,16384,0.21000852584838867
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,4096,16384,0.20209919611612953
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,4096,16384,0.081931734085083
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,4096,12288,0.15859626134236654
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,4096,12288,0.1525290648142497
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,5120,65536,0.3727861404418945
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,4096,10240,0.11547520160675048
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,4096,10240,0.11810346444447835
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,4096,12288,0.06477440198262532
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,4096,8192,0.09299093087514242
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,4096,8192,0.09818452994028727
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,4096,10240,0.05878293514251709
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,4096,7168,0.08206506570180258
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,4096,7168,0.08683306376139323
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,4096,8192,0.04765546719233195
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,4096,6144,0.06998826662699381
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,4096,6144,0.07279040018717448
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,4096,7168,0.04389760096867879
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,4096,5120,0.05882026751836141
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,4096,5120,0.06191039880116781
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,4096,5120,0.03553280035654704
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,4096,4096,0.04708906809488932
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,4096,4096,0.050443732738494874
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,4096,6144,0.03898986577987671
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,4096,3584,0.041324798266092935
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,4096,3584,0.04680853287378947
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,4096,3584,0.02792106668154399
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,4096,3072,0.03325120011965434
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,4096,3072,0.04341333309809367
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,4096,4096,0.03065173427263896
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,4096,2560,0.028717867533365887
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,4096,2560,0.03909866809844971
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,4096,2560,0.02361066738764445
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,4096,2048,0.0238154669602712
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,4096,2048,0.034924801190694174
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,4096,3072,0.02473919987678528
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,4096,1536,0.018759467204411826
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,4096,1536,0.030887466669082642
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,4096,2048,0.021040000518163047
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,4096,1024,0.014013866583506266
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,4096,1024,0.026242132981618243
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,4096,1536,0.018592000007629395
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,4096,768,0.010640000303586323
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,4096,768,0.02482666571935018
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,4096,1024,0.017784533898035686
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,4096,512,0.008141866823037466
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,4096,512,0.021116799116134642
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,4096,512,0.01579093337059021
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,4096,256,0.005957333246866862
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,4096,256,0.01871253252029419
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,4096,768,0.01583253343900045
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,4096,128,0.005952000121275584
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,4096,128,0.01949333349863688
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,4096,256,0.015715199708938598
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,4096,64,0.005099733173847198
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,4096,128,0.015386666854222616
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,4096,32,0.0055402666330337524
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,4096,64,0.01864746610323588
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,4096,32,0.018669867515563966
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,3584,65536,0.5630218505859375
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,4096,51200,0.22601280212402344
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,3584,65536,0.6793952306111654
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,3584,51200,0.4269685427347819
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,3584,51200,0.5327935854593913
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,3584,16384,0.13960639635721844
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,3584,16384,0.1663285255432129
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,3584,16384,0.08250880241394043
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,3584,12288,0.10585599740346271
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,3584,12288,0.1261013348897298
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,4096,65536,0.28158613840738933
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,3584,10240,0.0892906665802002
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,3584,10240,0.11071253617604573
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,3584,12288,0.06552319924036662
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,3584,8192,0.08542826970418295
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,3584,8192,0.09141973654429117
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,3584,10240,0.05952426592508951
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,3584,7168,0.07397119998931885
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,3584,7168,0.07999893029530844
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,3584,7168,0.045805867513020834
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,3584,6144,0.06273279984792074
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,3584,8192,0.049406933784484866
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,3584,6144,0.06973439852396647
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,3584,5120,0.051378134886423746
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,3584,5120,0.05800319910049438
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,3584,5120,0.037191466490427656
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,3584,4096,0.040863998730977374
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,3584,4096,0.04896533489227295
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,3584,6144,0.0403711994489034
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,3584,3584,0.0361685315767924
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,3584,3584,0.045714131991068524
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,3584,4096,0.032570666074752806
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,3584,3072,0.031285333633422854
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,3584,3072,0.042572800318400064
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,3584,3584,0.029523199796676634
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,3584,2560,0.026471465826034546
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,3584,2560,0.03813013235727946
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,3584,3072,0.026601600646972656
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,3584,2048,0.02151040037473043
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,3584,2048,0.03290346662203471
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,3584,2560,0.025965867439905803
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,3584,1536,0.01738133430480957
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,3584,1536,0.029139200846354168
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,3584,2048,0.023423999547958374
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,3584,1024,0.012035199999809265
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,3584,1024,0.02577386697133382
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,3584,1536,0.02153173287709554
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,3584,768,0.0100490669409434
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,3584,768,0.025101866324742633
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,3584,1024,0.020601600408554077
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,3584,512,0.007905066510041555
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,3584,512,0.02092693249384562
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,3584,768,0.01914026737213135
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,3584,256,0.006362666686375936
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,3584,256,0.018939733505249023
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,3584,512,0.01899306575457255
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,3584,128,0.0054624001185099285
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,3584,128,0.01651093363761902
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,3584,128,0.01867199937502543
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,3584,64,0.005047466854254405
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,3584,64,0.017921066284179686
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,3584,32,0.005129600067933401
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,3584,256,0.01885546644528707
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,3584,32,0.016583466529846193
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,3072,65536,0.48094507853190105
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,3072,65536,0.6422037124633789
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,3072,51200,0.3785802523295084
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,3072,51200,0.505406920115153
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,3584,51200,0.22506346702575683
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,3072,16384,0.12370986938476562
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,3072,16384,0.162117338180542
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,3072,16384,0.0809343973795573
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,3072,12288,0.09462080001831055
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,3072,12288,0.1206485350926717
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,3584,65536,0.28021440505981443
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,3072,10240,0.08082133134206136
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,3072,10240,0.10237973531087238
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,3072,12288,0.06466986735661825
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,3072,8192,0.06469120184580485
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,3072,8192,0.08411839803059896
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,3072,10240,0.05796906550725302
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,3072,7168,0.057860267162322995
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,3072,7168,0.07345066865285238
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,3072,7168,0.044818135102589925
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,3072,6144,0.04964906771977742
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,3072,6144,0.06403839985529582
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,3072,8192,0.04805759986241658
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,3072,5120,0.04093546470006307
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,3072,5120,0.05356266498565674
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,3072,5120,0.03624853293100993
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,3072,4096,0.03328533371289571
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,3072,6144,0.03999786774317424
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,3072,4096,0.0457696000734965
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,3072,3584,0.029653332630793255
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,3072,3584,0.04311466614405314
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,3072,3584,0.02894933422406514
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,3072,3072,0.02583786646525065
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,3072,3072,0.04071040153503418
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,3072,4096,0.031879466772079465
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,3072,2560,0.02300800085067749
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,3072,2560,0.03588373263676961
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,3072,3072,0.026658133665720625
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,3072,2048,0.01908586621284485
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,3072,2048,0.03237760066986084
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,3072,2560,0.025593600670496625
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,3072,1536,0.015558399756749473
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,3072,1536,0.02871253291765849
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,3072,1536,0.020708266894022623
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,3072,1024,0.011705600221951802
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,3072,1024,0.025126399596532185
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,3072,2048,0.023161600033442177
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,3072,768,0.009666132926940917
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,3072,768,0.022827732563018798
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,3072,1024,0.020615466435750327
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,3072,512,0.007097599903742473
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,3072,512,0.02059733271598816
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,3072,768,0.018547199169794717
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,3072,256,0.005445333321889242
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,3072,256,0.019489065806070963
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,3072,256,0.01843520005544027
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,3072,512,0.01861013372739156
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,3072,128,0.005061333378156027
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,3072,128,0.01767146587371826
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,3072,64,0.004589866598447164
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,3072,64,0.01684160033861796
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,3072,32,0.004676266511281332
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,3072,128,0.018620800971984864
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,3072,32,0.016537599762280784
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,2560,65536,0.4413290659586589
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,2560,65536,0.6033962885538737
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,2560,51200,0.3507199923197428
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,2560,51200,0.47612905502319336
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,3072,51200,0.22620479265848795
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,2560,16384,0.10203200181325275
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,2560,16384,0.1446239948272705
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,2560,16384,0.07977279822031656
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,2560,12288,0.07696533203125
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,2560,12288,0.10883519649505616
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,3072,65536,0.28245652516682945
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,2560,10240,0.06996479829152426
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,2560,10240,0.09486186504364014
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,2560,12288,0.0631231983502706
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,2560,8192,0.05688746770222982
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,2560,8192,0.07627306779225668
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,2560,10240,0.055864532788594566
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,2560,7168,0.04930026531219482
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,2560,7168,0.06633173227310181
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,2560,8192,0.04630506833394368
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,2560,6144,0.04250986576080322
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,2560,6144,0.055895467599232994
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,2560,7168,0.043622398376464845
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,2560,5120,0.03445226748784383
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,2560,5120,0.049586133162180586
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,2560,5120,0.03518506685892741
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,2560,4096,0.028911999861399335
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,2560,4096,0.04297813177108765
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,2560,6144,0.0379744013150533
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,2560,3584,0.024870399634043375
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,2560,3584,0.03917333285013835
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,2560,3584,0.028149332602818804
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,2560,3072,0.021539199352264404
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,2560,3072,0.03705493211746216
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,2560,4096,0.030482133229573567
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,2560,2560,0.01902079979578654
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,2560,2560,0.033165866136550905
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,2560,3072,0.025153066714604693
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,2560,2048,0.015639467040697734
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,2560,2048,0.029125332832336426
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,2560,2560,0.024306132396062215
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,2560,1536,0.01218239963054657
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,2560,1536,0.0268885334332784
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,2560,2048,0.02137493292490641
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,2560,1024,0.008847999572753906
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,2560,1024,0.023564799626668295
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,2560,1536,0.019423999389012656
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,2560,768,0.007335466643174489
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,2560,768,0.0228383998076121
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,2560,1024,0.018605866034825645
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,2560,512,0.005884799857934316
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,2560,512,0.020115200678507486
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,2560,512,0.01698453426361084
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,2560,256,0.005106133222579956
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,2560,256,0.01756160060564677
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,2560,768,0.016987733046213784
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,2560,128,0.0046293333172798155
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,2560,128,0.01694933374722799
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,2560,256,0.016733866930007935
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,2560,64,0.0042015999555587765
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,2560,64,0.016568533579508462
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,2560,32,0.004306133091449738
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,2560,128,0.016935465733210246
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,2560,32,0.016151466965675355
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,2048,65536,0.34819841384887695
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,2560,51200,0.22363306681315104
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,2048,65536,0.5760032018025716
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,2048,51200,0.2785567919413249
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,2048,51200,0.4529130617777507
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,2048,16384,0.09023893674214681
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,2048,16384,0.14205759366353352
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,2048,16384,0.05709120035171509
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,2048,12288,0.06900160312652588
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,2048,12288,0.10593706766764324
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,2560,65536,0.27974507013956706
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,2048,10240,0.05743786493937174
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,2048,10240,0.08852586746215821
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,2048,10240,0.040408531824747726
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,2048,8192,0.04634453455607097
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,2048,8192,0.07024959723154703
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,2048,12288,0.04513920148213704
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,2048,7168,0.04002559979756673
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,2048,7168,0.05925120115280151
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,2048,7168,0.031065599123636885
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,2048,6144,0.03403626680374146
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,2048,6144,0.05145279963811239
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,2048,8192,0.033614933490753174
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,2048,5120,0.02826346755027771
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,2048,5120,0.04660586516062419
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,2048,5120,0.02645333409309387
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,2048,4096,0.023399466276168825
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,2048,4096,0.04041813214619954
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,2048,6144,0.02845226724942525
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,2048,3584,0.0210805336634318
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,2048,3584,0.03693759838740031
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,2048,3584,0.021449599663416544
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,2048,4096,0.023187200228373207
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,2048,3072,0.01747093399365743
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,2048,3072,0.03557013273239136
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,2048,2560,0.015315199891726175
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,2048,2560,0.03193813363711039
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,2048,2560,0.018595200777053834
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,2048,3072,0.01946880022684733
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,2048,2048,0.012882133324941
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,2048,2048,0.028442666927973433
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,2048,1536,0.010636799534161885
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,2048,1536,0.026978133122126262
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,2048,2048,0.016537599762280784
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,2048,1024,0.008030933141708375
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,2048,1024,0.022899200518925987
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,2048,1024,0.014107732971509298
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,2048,768,0.006867200136184692
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,2048,768,0.02183039983113607
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,2048,768,0.012851199507713318
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,2048,512,0.00591893345117569
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,2048,1536,0.014458666245142618
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,2048,512,0.0191103994846344
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,2048,256,0.004624000191688538
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,2048,256,0.01739733417828878
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,2048,256,0.012782933314641318
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,2048,512,0.012898133198420206
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,2048,128,0.003824000060558319
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,2048,128,0.016591999928156534
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,2048,64,0.0038250667353471124
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,2048,64,0.01688746611277262
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,2048,128,0.012807466586430869
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,2048,32,0.0038773333032925926
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,2048,32,0.015020799636840821
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,1536,65536,0.2684447924296061
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,1536,65536,0.5483658472696941
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,1536,51200,0.21309760411580406
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,2048,51200,0.1499381383260091
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,1536,51200,0.42017065684000654
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,1536,16384,0.07101120154062907
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,1536,16384,0.1321120023727417
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,1536,16384,0.05831360022226969
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,1536,12288,0.05438719987869263
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,1536,12288,0.09803093274434407
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,2048,65536,0.18502079645792643
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,1536,10240,0.04587733348210653
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,1536,10240,0.08127146561940511
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,1536,12288,0.04620159864425659
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,1536,8192,0.03586560090382894
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,1536,8192,0.06233919858932495
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,1536,10240,0.03994133472442627
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,1536,7168,0.03172373374303182
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,1536,7168,0.05531626542409261
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,1536,8192,0.03393386602401734
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,1536,6144,0.02722346584002177
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,1536,6144,0.04924906492233276
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,1536,7168,0.031498666604359946
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,1536,5120,0.02346986730893453
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,1536,5120,0.04464319944381714
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,1536,6144,0.028974932432174683
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,1536,4096,0.019406932592391967
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,1536,4096,0.03723733425140381
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,1536,5120,0.027323732773462932
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,1536,3584,0.01801919937133789
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,1536,3584,0.03509226640065511
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,1536,4096,0.02400533358256022
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,1536,3072,0.015658666690190635
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,1536,3072,0.03289706707000732
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,1536,3584,0.022317866484324135
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,1536,2560,0.013784533739089966
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,1536,2560,0.03062613407770793
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,1536,3072,0.02060479919115702
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,1536,2048,0.011179733276367187
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,1536,2048,0.02704533338546753
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,1536,2048,0.017774933576583864
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,1536,1536,0.008408533533414204
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,1536,2560,0.01962453325589498
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,1536,1536,0.025550933678944905
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,1536,1024,0.006725333134333293
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,1536,1024,0.02296853264172872
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,1536,1024,0.015366400281588236
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,1536,768,0.005885866781075796
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,1536,1536,0.015742933750152587
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,1536,768,0.02065066695213318
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,1536,512,0.005051733553409576
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,1536,512,0.019192532698313395
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,1536,51200,0.14875946044921876
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,1536,256,0.003929600119590759
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,1536,512,0.01418239971001943
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,1536,768,0.014533332983652749
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,1536,128,0.0038730666041374207
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,1536,128,0.016535466909408568
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,1536,128,0.014924800395965577
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,1536,64,0.003470933437347412
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,1536,64,0.017195733388264973
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,1536,256,0.017016534010569254
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,1536,32,0.003487999985615412
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,1536,32,0.0146506667137146
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,1024,65536,0.20111145973205566
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,1536,256,0.014057599504788718
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,1024,51200,0.1592586676279704
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,1024,65536,0.5010506629943847
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,1024,51200,0.3958911895751953
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,1024,16384,0.046615465482076006
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,1024,16384,0.11934613386789958
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,1024,16384,0.056157867113749184
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,1024,12288,0.0353333314259847
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,1024,12288,0.08725546995798747
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,1536,65536,0.18440747261047363
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,1024,10240,0.02844799955685933
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,1024,10240,0.07017813523610433
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,1024,10240,0.03750400145848592
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,1024,8192,0.022503467400868733
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,1024,8192,0.053641601403554284
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,1024,12288,0.042520534992218015
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,1024,7168,0.01995840072631836
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,1024,7168,0.04789653221766154
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,1024,7168,0.02922559976577759
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,1024,6144,0.017544533809026083
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,1024,6144,0.044309333960215254
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,1024,8192,0.031354665756225586
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,1024,5120,0.015271466970443726
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,1024,5120,0.03921706676483154
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,1024,5120,0.025227733453114826
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,1024,4096,0.01249066690603892
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,1024,4096,0.03405226469039917
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,1024,6144,0.026422399282455444
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,1024,3584,0.010840533177057902
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,1024,3584,0.030867199103037517
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,1024,3584,0.01997973322868347
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,1024,3072,0.009815466403961182
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,1024,4096,0.02188373406728109
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,1024,3072,0.030900265773137408
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,1024,2560,0.008598400155703227
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,1024,2560,0.028802132606506346
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,1024,2560,0.01733120083808899
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,1024,2048,0.007548800110816956
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,1024,3072,0.01825066606203715
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,1024,2048,0.025407999753952026
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,1024,1536,0.0063391998410224915
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,1024,1536,0.022746666272481283
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,1024,51200,0.14647572835286457
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,1024,1536,0.013267200191815695
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,1024,1024,0.005497600138187409
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,1024,1024,0.020794665813446044
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,1024,2048,0.015689599514007568
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,1024,768,0.004665599763393402
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,1024,768,0.020730666319529214
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,1024,512,0.004297600189844767
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,1024,512,0.018946133057276406
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,1024,768,0.012051199873288471
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,1024,1024,0.013125333189964294
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,1024,256,0.0034048000971476236
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,1024,512,0.012060800194740295
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,1024,256,0.017729065815607705
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,1024,128,0.0030389333764712016
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,1024,128,0.014590932925542196
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,1024,64,0.003032533327738444
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,1024,256,0.012030933300654094
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,1024,32,0.0029834667841593427
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,1024,64,0.014579199751218162
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,1024,32,0.014507733782132468
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,768,65536,0.1593557357788086
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,768,65536,0.4789461453755696
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,768,51200,0.12205013434092205
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,768,51200,0.3802069346110026
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,1024,128,0.012831999858220419
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,768,16384,0.03944960037867228
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,768,16384,0.11620266437530517
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,768,16384,0.05719253222147623
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,768,12288,0.028742400805155437
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,768,12288,0.08224319616953532
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,1024,65536,0.1820853392283122
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,768,10240,0.022924800713857017
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,768,10240,0.06567573149998983
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,768,12288,0.04387840032577515
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,768,8192,0.01800533334414164
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,768,8192,0.05067413250605265
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,768,10240,0.038173866271972653
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,768,7168,0.016115199526151022
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,768,7168,0.0460426648457845
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,768,8192,0.03221333424250285
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,768,6144,0.01404800017674764
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,768,6144,0.041230932871500654
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,768,7168,0.030247465769449873
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,768,5120,0.012274133165677388
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,768,5120,0.037272532780965165
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,768,5120,0.02622186740239461
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,768,6144,0.02805333336194356
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,768,4096,0.010471466183662414
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,768,4096,0.03348373174667359
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,768,3584,0.009237333138783773
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,768,3584,0.030829866727193195
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,768,3584,0.02148266633351644
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,768,3072,0.008255999783674877
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,768,3072,0.029373866319656373
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,768,4096,0.023156267404556275
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,768,2560,0.007521066566308339
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,768,2560,0.02717973391215007
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,768,3072,0.019452800353368126
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,768,2048,0.006389333307743073
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,768,2048,0.02487466732660929
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,768,2560,0.018525866667429607
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,768,1536,0.005909333129723867
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,768,1536,0.023062400023142495
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,768,2048,0.01681386629740397
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,768,1024,0.004609066744645437
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,768,1024,0.020835200945536293
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,768,51200,0.14669119517008464
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,768,1536,0.015014400084813436
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,768,1024,0.014503467082977294
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,768,768,0.004249600072701773
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,768,768,0.019330133994420372
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,768,512,0.003435733417669932
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,768,512,0.018857600291570027
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,768,512,0.013245866696039835
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,768,256,0.003028266628583272
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,768,768,0.013350400328636169
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,768,256,0.01641386648019155
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,768,128,0.002994133283694585
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,768,128,0.014760532975196838
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,768,256,0.013643733660380044
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,768,64,0.002771199991305669
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,768,128,0.013434666395187377
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,768,32,0.002644266684850057
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,768,64,0.016219733158747356
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,768,32,0.014490666985511779
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,512,65536,0.13523839314778646
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,512,51200,0.09523839950561523
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,512,65536,0.4524074554443359
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,512,51200,0.3615509351094564
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,512,16384,0.030929066737492877
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,512,16384,0.1172714630762736
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,768,65536,0.18302292823791505
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,512,12288,0.023895466327667238
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,512,16384,0.05552639961242676
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,512,12288,0.08324693044026693
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,512,10240,0.01927466591199239
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,512,10240,0.06657173236211142
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,512,10240,0.03671893278757731
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,512,8192,0.015964800119400026
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,512,8192,0.051710931460062656
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,512,12288,0.041203200817108154
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,512,7168,0.015312000115712484
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,512,7168,0.04332693417867024
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,512,7168,0.028492800394694012
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,512,6144,0.013401599725087485
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,512,8192,0.03095573385556539
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,512,6144,0.03917866547902425
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,512,5120,0.00988266666730245
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,512,5120,0.036252800623575845
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,512,5120,0.024321067333221435
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,512,4096,0.01016960044701894
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,512,4096,0.031292800108591715
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,512,6144,0.025994666417439777
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,512,3584,0.00809386670589447
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,512,3584,0.028794666131337483
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,512,4096,0.021473066012064616
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,512,3072,0.0070816000302632645
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,512,3072,0.028907734155654907
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,512,3584,0.019466666380564372
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,512,2560,0.006669866542021434
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,512,3072,0.01814080079396566
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,512,2560,0.0260917325814565
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,512,2048,0.005846400062243144
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,512,2048,0.023041067520777385
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,512,51200,0.14506667455037434
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,512,2560,0.01693333387374878
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,512,1536,0.004931200047334035
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,512,1536,0.023118933041890465
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,512,2048,0.015270400047302245
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,512,1024,0.004119466741879781
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,512,1024,0.021899733940760294
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,512,768,0.003819733361403147
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,512,768,0.018655999501546224
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,512,1024,0.012809600432713827
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,512,512,0.00344106654326121
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,512,512,0.018492799997329713
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,512,768,0.01197760005791982
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,512,256,0.002974933385848999
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,512,256,0.015429332852363586
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,512,512,0.011690666278203327
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,512,128,0.0025994665920734406
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,512,256,0.011658666531244914
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,512,128,0.014851199587186179
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,512,64,0.0026335999369621276
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,512,128,0.01164906620979309
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,512,64,0.014637866616249084
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,512,32,0.0025888000925381976
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,512,32,0.014843733112017313
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,256,65536,0.06772800286610922
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,512,1536,0.013260799646377563
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,256,65536,0.4297439893086751
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,256,51200,0.05505173206329346
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,256,51200,0.3424352010091146
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,256,16384,0.019717333714167277
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,256,16384,0.11157013575236004
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,256,16384,0.054421333471933994
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,256,12288,0.0193066676457723
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,512,65536,0.18148053487141927
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,256,12288,0.07315626939137777
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,256,10240,0.014327466487884521
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,256,10240,0.05726079940795899
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,256,10240,0.03543786605199178
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,256,8192,0.011823999881744384
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,256,8192,0.045645864804585774
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,256,12288,0.03932906786600749
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,256,7168,0.010622933506965637
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,256,8192,0.030453334252039593
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,256,7168,0.04132266839345296
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,256,7168,0.028487465778986615
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,256,6144,0.0092031995455424
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,256,6144,0.03992319901784261
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,256,51200,0.1447861353556315
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,256,5120,0.007934933404127757
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,256,5120,0.03462400039037068
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,256,4096,0.007066666583220164
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,256,4096,0.030804266532262165
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,256,6144,0.026372265815734864
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,256,3584,0.006307200094064077
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,256,3584,0.028696533044179278
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,256,4096,0.02143893241882324
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,256,3072,0.00586346685886383
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,256,5120,0.024650667111078897
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,256,3072,0.028724267085393267
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,256,2560,0.005443199972311656
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,256,2560,0.02481386661529541
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,256,3584,0.01981546680132548
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,256,2048,0.0046528001626332605
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,256,2048,0.02296533385912577
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,256,2560,0.016862932840983072
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,256,1536,0.004218666752179464
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,256,65536,0.18037439982096354
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,256,1536,0.021505065759023032
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,256,1024,0.0034474665919939675
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,256,1024,0.02030186653137207
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,256,3072,0.017917867501576742
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,256,2048,0.015267200271288552
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,256,768,0.0035829332967599234
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,256,768,0.018705066045125326
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,256,1536,0.013242666920026144
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,256,512,0.0029813334345817565
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,256,1024,0.01279253363609314
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,256,512,0.018268799781799315
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,256,256,0.002571733295917511
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,256,512,0.01164906620979309
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,256,256,0.016453333695729575
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,256,128,0.0023445333043734234
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,256,256,0.011614933609962463
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,256,128,0.015940266847610473
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,256,64,0.0022698665658632917
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,256,32,0.002233600119749705
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,256,128,0.011664000153541566
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,256,64,0.014735999703407287
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,256,32,0.014465066790580749
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,128,65536,0.05757546822230021
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,128,51200,0.046868268648783365
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,128,65536,0.42489919662475584
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,128,51200,0.3366549491882324
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,128,16384,0.01658560037612915
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,128,16384,0.1051754633585612
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,256,768,0.01176106631755829
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,128,12288,0.012528000275293985
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,128,12288,0.07207893530527751
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,128,16384,0.052441600958506265
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,128,10240,0.010844799876213073
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,128,10240,0.05516586701075236
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,128,12288,0.03848106861114502
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,128,8192,0.01138879954814911
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,128,8192,0.04556266864140828
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,128,10240,0.03488213221232096
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,128,7168,0.01000853379567464
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,128,7168,0.04114880164464314
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,128,8192,0.02932586669921875
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,128,6144,0.009367466966311137
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,128,7168,0.027654399474461872
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,128,6144,0.03760106563568115
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,128,51200,0.14193387031555177
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,128,5120,0.007933866480986278
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,128,5120,0.034270934263865155
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,128,4096,0.006727466483910878
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,128,4096,0.02986133297284444
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,128,6144,0.025016532341639204
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,128,3584,0.0062943999965985615
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,128,3584,0.029820799827575684
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,128,5120,0.023642667134602866
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,128,3072,0.005655466516812643
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,128,65536,0.17707093556722003
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,128,4096,0.02025066614151001
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,128,2560,0.0050346667567888895
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,128,3072,0.027635200818379717
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,128,3584,0.018703999121983846
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,128,2560,0.024741333723068238
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,128,2048,0.004666666686534882
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,128,2048,0.02264960010846456
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,128,3072,0.016979199647903443
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,128,1536,0.00383146678407987
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,128,1536,0.020757333437601725
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,128,2048,0.014061866203943887
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,128,1024,0.0033791999022165934
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,128,1536,0.012432000041007996
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,128,1024,0.020717867215474448
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,128,768,0.0030730667213598887
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,128,1024,0.01162559986114502
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,128,512,0.0026176000634829206
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,128,768,0.01076479951540629
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,128,512,0.017950934171676636
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,128,256,0.0025792000194390613
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,128,512,0.010804266730944315
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,128,256,0.015127467115720114
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,128,256,0.010434133807818095
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,128,128,0.00223786657055219
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,128,128,0.014526933431625366
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,128,128,0.010431999961535137
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,128,64,0.002183466653029124
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,128,64,0.014326399564743042
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,128,32,0.0025813333690166474
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2305,128,2560,0.016242133577664693
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,128,32,0.014500266313552857
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,64,65536,0.050551466147104894
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,64,51200,0.041051733493804934
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,64,65536,0.42093760172526046
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,64,16384,0.01286186675230662
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,64,51200,0.334445858001709
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,64,16384,0.10389440059661866
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,64,12288,0.013902933398882548
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,64,10240,0.011457066734631855
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,64,12288,0.06927039623260497
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,64,8192,0.010250666737556457
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,64,10240,0.05400853157043457
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,64,8192,0.04373546838760376
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,128,768,0.01885973413785299
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,64,7168,0.04104426701863607
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,64,6144,0.008729599912961324
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,64,5120,0.008377599716186523
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,64,6144,0.03712000052134196
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,64,4096,0.0071050668756167095
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,64,5120,0.034886399904886886
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,64,3584,0.0067071999112765
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,64,4096,0.029330132404963176
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,64,3072,0.005996799965699514
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,64,3584,0.02908266584078471
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,64,2560,0.005412266651789347
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,64,3072,0.02701759934425354
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,64,2048,0.0042357335488001505
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,64,2560,0.0249674657980601
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,64,1536,0.0038421332836151125
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,64,2048,0.02335253357887268
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,64,1024,0.0034261333445707956
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,64,1536,0.023013333479563393
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,64,768,0.0030271999537944795
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,64,1024,0.019606399536132812
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,64,512,0.0027306665976842242
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,64,768,0.01916586756706238
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,64,256,0.0022367998957633973
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,64,512,0.016709333658218382
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,64,128,0.002164266755183538
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,64,64,0.0021802666286627453
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,64,128,0.014667733510335287
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,64,32,0.0021717332303524016
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,64,64,0.013527466853459676
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,32,65536,0.052501332759857175
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,64,7168,0.009733333190282186
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,32,51200,0.039077333609263104
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,32,65536,0.4197322527567546
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,32,16384,0.00981760025024414
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,32,51200,0.33294827143351235
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,32,16384,0.100764799118042
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,32,12288,0.008875733613967896
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,32,12288,0.06752959887186685
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,32,10240,0.007949866851170858
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,32,8192,0.007624533275763195
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,32,10240,0.06162026723225912
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,32,7168,0.0066890666882197065
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,64,256,0.01668693423271179
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,32,6144,0.006331733365853627
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,32,7168,0.0464842677116394
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,32,8192,0.051644798119862875
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,32,5120,0.016963199774424235
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,32,6144,0.04311360120773315
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,64,32,0.01434879998366038
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,32,3584,0.01243839959303538
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,32,5120,0.039060266812642415
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,32,4096,0.030903466542561847
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,32,3584,0.02712000012397766
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,32,3072,0.01114026705423991
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,32,3072,0.02754986683527629
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,32,2560,0.00960533320903778
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,32,2048,0.004301866888999939
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,32,1536,0.003884800026814143
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,32,2048,0.0226090669631958
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,32,2560,0.025358933210372924
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,32,1024,0.0031221332649389905
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,32,1536,0.022474666436513267
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,32,768,0.002974933385848999
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,32,1024,0.020713599522908528
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,32,512,0.002583466718594233
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,32,768,0.01884160041809082
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,32,256,0.0026026666164398193
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,32,512,0.01797119975090027
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,32,128,0.002160000056028366
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,32,256,0.016411733627319337
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,32,64,0.002161066730817159
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,32,32,0.002170666555563609
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,32,128,0.014739200472831726
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,32,32,0.013985066612561544
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2305,32,4096,0.01411733329296112
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2305,32,64,0.014538666605949402
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,65536,51200,4.755431620279948
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,65536,16384,1.4215775807698567
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,65536,16384,2.7191541035970053
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,65536,51200,8.659871419270832
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,65536,12288,2.0532405853271483
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,65536,12288,1.061126391092936
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,65536,10240,1.867470932006836
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,65536,10240,0.9114026387532552
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,65536,10240,0.5282538731892903
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,65536,8192,1.4199583689371744
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,65536,8192,0.7185802459716797
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,65536,12288,0.5929898579915365
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,65536,7168,1.2348138173421224
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,65536,7168,0.6196981430053711
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,65536,16384,0.7912160237630208
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,65536,6144,0.985319455464681
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,65536,6144,0.603497568766276
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,65536,8192,0.4385205268859863
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,65536,5120,0.8121983846028646
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,65536,5120,0.5070165316263835
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,65536,7168,0.4063125292460124
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,65536,4096,0.6280789057413737
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,65536,4096,0.3696127891540527
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,65536,6144,0.3525941212972005
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,65536,3584,0.5730143864949544
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,65536,3584,0.33408212661743164
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,65536,5120,0.3225013415018717
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,65536,3072,0.4793226559956868
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,65536,3072,0.3142815907796224
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,65536,4096,0.28960320154825847
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,65536,2560,0.3977450688680013
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,65536,2560,0.24032533963521324
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,65536,3584,0.2606730620066325
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,65536,2048,0.3347071965535482
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,65536,2048,0.20173439979553223
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,65536,2560,0.2172032038370768
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,65536,1536,0.24837546348571776
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,65536,1536,0.16118826866149902
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,65536,3072,0.2312074661254883
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,65536,1024,0.175273593266805
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,65536,1024,0.11385599772135417
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,65536,1024,0.1582357406616211
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,65536,768,0.1383562723795573
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,65536,768,0.09502293268839518
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,65536,1536,0.1600213368733724
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,65536,512,0.09351572990417481
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,65536,512,0.08256320158640543
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,65536,512,0.14391679763793946
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,65536,256,0.06373653411865235
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,65536,256,0.06919679641723633
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,65536,768,0.14901332855224608
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,65536,128,0.05532906850179037
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,65536,128,0.0643285314242045
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,65536,2048,0.19044052759806315
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,65536,64,0.05020693143208822
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,65536,64,0.06484906673431397
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,65536,32,0.05069653193155924
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,65536,32,0.06423786481221518
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,65536,256,0.14219093322753906
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,65536,128,0.15048106511433917
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,51200,65536,4.67188466389974
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,51200,65536,8.54221700032552
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,51200,51200,3.7323318481445313
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,51200,51200,6.9749704996744795
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,51200,16384,2.0504778544108073
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,51200,16384,1.1460309346516928
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,51200,16384,0.6156714757283528
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,51200,12288,1.5426954905192056
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,51200,12288,0.806492805480957
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,51200,12288,0.4667946815490723
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,51200,10240,1.441381327311198
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,51200,10240,0.7048714955647786
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,51200,10240,0.41530345280965164
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,51200,8192,1.0421599706013998
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,51200,8192,0.5755936304728191
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,65536,51200,2.5146517435709637
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,51200,7168,0.9005962371826172
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,51200,7168,0.5396640141805013
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,51200,8192,0.34227307637532556
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,51200,6144,0.7622741063435872
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,51200,6144,0.4340799967447917
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,51200,7168,0.31832853953043616
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,51200,5120,0.6299434661865234
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,51200,5120,0.37491413752237956
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,51200,5120,0.2519584019978841
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,51200,4096,0.5014656066894532
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,51200,4096,0.3173247973124186
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,51200,6144,0.27463998794555666
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,51200,3584,0.4312661488850911
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,51200,3584,0.259389861424764
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,51200,3584,0.20190186500549318
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,51200,3072,0.37203839619954426
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,51200,3072,0.2270847956339518
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,51200,4096,0.2258783976236979
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,51200,2560,0.3203637440999349
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,51200,2560,0.21133227348327638
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,51200,51200,1.9735893249511718
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,51200,2048,0.2658133347829183
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,51200,2048,0.16531306902567547
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,51200,2048,0.15338346163431804
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,51200,1536,0.20058879852294922
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,51200,1536,0.12205440203348797
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,51200,2560,0.1717354615529378
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,51200,1024,0.1377994696299235
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,51200,1024,0.09219199816385905
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,51200,3072,0.18091626167297364
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,51200,768,0.09935359954833985
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,51200,768,0.07846079667409261
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,51200,768,0.1184127966562907
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,51200,512,0.0734549363454183
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,51200,512,0.06876373291015625
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,51200,1024,0.1258570671081543
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,51200,256,0.050819198290506996
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,51200,256,0.05733973185221354
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,51200,1536,0.13058026631673175
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,51200,128,0.04600746631622314
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,51200,128,0.05335359970728556
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,51200,256,0.11669119993845622
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,51200,64,0.03880319992701213
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,51200,64,0.05286080042521159
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,51200,32,0.0399125337600708
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,51200,32,0.053505067030588785
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,51200,128,0.11886080106099446
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,16384,65536,2.753778076171875
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,51200,512,0.1179861307144165
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,16384,65536,1.7089227040608723
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,16384,51200,1.3309429168701172
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,16384,51200,2.1342623392740885
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,16384,16384,0.679639498392741
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,16384,16384,0.42041600545247393
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,16384,16384,0.22776640256245934
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,16384,12288,0.4603583971659343
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,16384,12288,0.3294282595316569
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,16384,12288,0.17494826316833495
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,16384,10240,0.4094058672587077
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,16384,10240,0.2853482564290365
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,16384,10240,0.15605759620666504
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,16384,8192,0.3284885406494141
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,16384,8192,0.21325119336446127
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,16384,8192,0.12882880369822186
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,16384,7168,0.2898357391357422
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,16384,7168,0.19828906059265136
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,16384,7168,0.11964480082194011
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,16384,6144,0.24657707214355468
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,16384,6144,0.16885013580322267
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,16384,51200,0.6738858540852865
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,16384,5120,0.20926399230957032
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,16384,5120,0.1421877384185791
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,16384,5120,0.09601386388142905
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,16384,4096,0.16715200742085773
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,16384,4096,0.11527040004730224
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,16384,6144,0.10407573382059734
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,16384,3584,0.14162240028381348
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,16384,3584,0.10230613549550374
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,16384,4096,0.0852181355158488
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,16384,3072,0.1256767988204956
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,16384,3072,0.09366506735483805
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,16384,3584,0.07550826867421469
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,16384,2560,0.10205120245615643
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,16384,2560,0.0805951992670695
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,16384,2560,0.06255786816279094
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,16384,2048,0.08289600213368734
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,16384,2048,0.0679093360900879
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,16384,3072,0.06674986680348714
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,16384,1536,0.06600746711095175
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,16384,1536,0.0587936004002889
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,16384,1536,0.045347201824188235
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,16384,1024,0.04644373257954915
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,16384,1024,0.04737919966379801
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,16384,2048,0.052756265799204505
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,16384,768,0.036321067810058595
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,16384,768,0.041679998238881424
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,16384,1024,0.04410453240076701
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,16384,512,0.027543467283248902
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,16384,512,0.037305601437886554
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,16384,768,0.038318932056427
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,16384,256,0.019037866592407228
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,16384,256,0.030937600135803222
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,16384,256,0.03758399883906047
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,16384,128,0.017786665757497152
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,16384,128,0.029131732384363812
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,16384,128,0.038355199495951335
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,16384,64,0.015057067076365152
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,16384,64,0.027638399600982667
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,16384,32,0.01534293293952942
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,16384,512,0.03758080005645752
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,16384,32,0.02874133388201396
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,12288,65536,1.33786989847819
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,12288,65536,2.0416491190592447
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,12288,51200,1.5359626770019532
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,12288,51200,1.0719317118326823
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,51200,65536,2.5209386189778646
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,12288,16384,0.4707029342651367
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,12288,16384,0.3398976008097331
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,16384,65536,0.8487562815348306
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,12288,12288,0.35208638509114587
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,12288,12288,0.2544821262359619
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,12288,12288,0.1393664042154948
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,12288,10240,0.31615146001180017
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,12288,10240,0.22105813026428223
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,12288,16384,0.18121493657430013
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,12288,8192,0.2526581287384033
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,12288,8192,0.18312427202860515
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,12288,10240,0.12472853660583497
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,12288,7168,0.22210559844970704
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,12288,7168,0.1531775951385498
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,12288,8192,0.10312639872233073
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,12288,6144,0.18829760551452637
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,12288,6144,0.13265706698099772
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,12288,7168,0.09595733483632406
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,12288,5120,0.15749653180440265
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,12288,5120,0.11299413045247395
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,12288,6144,0.08394986788431803
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,12288,4096,0.1231882651646932
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,12288,4096,0.09396053155263265
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,12288,5120,0.07637226581573486
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,12288,3584,0.10859839916229248
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,12288,3584,0.08379200299580893
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,12288,4096,0.0657973329226176
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,12288,3072,0.09278079668680826
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,12288,3072,0.07822186946868896
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,12288,3584,0.05782933235168457
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,12288,2560,0.0787285327911377
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,12288,2560,0.06753066380818686
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,12288,3072,0.05177173217137655
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,12288,2048,0.06381760040918985
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,12288,2048,0.057716266314188636
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,12288,2048,0.04603519837061564
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,12288,1536,0.04955093463261922
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,12288,1536,0.04981013139088948
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,12288,2560,0.050058666865030924
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,12288,1024,0.03460479974746704
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,12288,1024,0.04001813332239787
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,12288,1024,0.040797865390777587
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,12288,768,0.027313067515691118
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,12288,768,0.03485759894053141
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,12288,1536,0.042003198464711504
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,12288,512,0.02030293345451355
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,12288,512,0.03153493404388428
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,12288,768,0.037588266531626384
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,12288,256,0.014929067095120749
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,12288,256,0.026716800530751546
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,12288,512,0.037378132343292236
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,12288,128,0.01328000028928121
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,12288,128,0.02579733331998189
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,12288,256,0.037367467085520426
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,12288,64,0.011202133695284526
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,12288,64,0.02291946609814962
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,12288,32,0.011615999539693196
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,12288,32,0.02274880011876424
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,12288,128,0.0371018648147583
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,10240,65536,1.8351647694905597
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,10240,65536,1.265166982014974
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,10240,51200,1.3944586435953776
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,10240,51200,0.9938720067342123
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,12288,51200,0.5245535850524903
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,10240,16384,0.4297173182169597
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,10240,16384,0.3281920115152995
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,10240,16384,0.15813546180725097
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,10240,12288,0.32642240524291993
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,10240,12288,0.24249280293782555
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,12288,65536,0.6569557189941406
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,10240,10240,0.2583701292673747
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,10240,10240,0.18829867045084636
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,10240,12288,0.12186346848805744
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,10240,8192,0.21560959815979003
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,10240,8192,0.15466987291971843
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,10240,10240,0.10869759718577068
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,10240,7168,0.1832362651824951
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,10240,7168,0.13239146868387858
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,10240,8192,0.09023893674214681
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,10240,6144,0.16252907117207843
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,10240,6144,0.11849599679311115
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,10240,7168,0.08369386990865071
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,10240,5120,0.12965973218282062
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,10240,5120,0.10107093652089436
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,10240,6144,0.07352213064829508
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,10240,4096,0.10571520328521729
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,10240,4096,0.08455466429392497
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,10240,5120,0.06632639964421591
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,10240,3584,0.08930880228678385
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,10240,3584,0.07597973346710205
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,10240,4096,0.05776853164037069
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,10240,3072,0.0799722671508789
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,10240,3072,0.07037546634674072
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,10240,3584,0.051481600602467856
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,10240,2560,0.06468693415323892
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,10240,2560,0.06006933450698852
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,10240,2560,0.045653335253397626
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,10240,2048,0.052909866968790686
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,10240,2048,0.051121068000793454
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,10240,51200,0.45029439926147463
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,10240,1536,0.04147839943567912
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,10240,3072,0.04710719982783
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,10240,1536,0.04428906838099162
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,10240,1024,0.029227733612060547
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,10240,1024,0.036722131570180255
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,10240,1024,0.036908801396687826
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,10240,768,0.023312000433603923
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,10240,768,0.0328821341196696
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,10240,1536,0.0380181352297465
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,10240,512,0.017573332786560057
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,10240,512,0.030022400617599487
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,10240,2048,0.041572264830271405
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,10240,256,0.012969600160916648
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,10240,256,0.026897066831588747
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,10240,256,0.03420159816741943
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,10240,128,0.01055999994277954
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,10240,128,0.02509760061899821
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,10240,512,0.034423466523488364
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,10240,64,0.012035199999809265
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,10240,64,0.025044266382853193
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,10240,32,0.012538666526476542
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,10240,128,0.034295467535654704
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,10240,768,0.03434453407923381
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,10240,32,0.024884267648061117
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,8192,65536,1.0352170944213868
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,8192,65536,1.24716796875
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,8192,51200,0.8084021250406901
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,8192,51200,0.9658154805501302
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,8192,16384,0.356606928507487
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,8192,16384,0.2851744016011556
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,8192,16384,0.13209066390991211
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,8192,12288,0.2358784039815267
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,8192,12288,0.19564159711201984
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,10240,65536,0.568232536315918
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,8192,10240,0.21274347305297853
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,8192,10240,0.16153705914815267
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,8192,12288,0.1014367977778117
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,8192,8192,0.1755562623341878
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,8192,8192,0.13330773512522381
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,8192,10240,0.09098986784617105
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,8192,7168,0.1508778731028239
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,8192,7168,0.11494932969411213
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,8192,7168,0.07060159842173258
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,8192,6144,0.12921173572540284
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,8192,6144,0.10201600392659504
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,8192,8192,0.07602346738179525
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,8192,5120,0.10899199644724529
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,8192,5120,0.09015786647796631
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,8192,5120,0.0543232003847758
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,8192,4096,0.0859071969985962
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,8192,4096,0.07378880182902017
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,8192,6144,0.06122133334477743
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,8192,3584,0.07352320353190103
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,8192,3584,0.0665120005607605
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,8192,3584,0.04248213370641073
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,8192,3072,0.06329600016276041
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,8192,3072,0.06106453339258829
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,8192,4096,0.04740373293558757
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,8192,2560,0.05424106518427531
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,8192,2560,0.05334399938583374
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,8192,3072,0.037511467933654785
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,8192,2560,0.03597653309504191
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,8192,2048,0.04386773506800334
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,8192,2048,0.04476586580276489
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,8192,1536,0.03409706751505534
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,8192,1536,0.03904213507970174
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,8192,1536,0.027961599826812743
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,8192,1024,0.024647466341654458
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,8192,1024,0.031946667035420734
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,8192,2048,0.03194239934285482
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,8192,768,0.019297067324320474
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,8192,768,0.028820266326268513
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,8192,1024,0.026975999275843304
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,8192,512,0.014495999614397685
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,8192,512,0.02589226762453715
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,8192,768,0.024598399798075356
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,8192,256,0.010409599542617798
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,8192,256,0.022997333606084188
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,8192,512,0.02426346739133199
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,8192,128,0.009573333462079366
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,8192,128,0.02060799996058146
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,8192,256,0.02392959992090861
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,8192,64,0.008508800466855367
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,8192,64,0.021091200908025107
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,8192,32,0.008687999844551087
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,8192,32,0.020639999707539877
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,8192,128,0.02435413400332133
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,7168,65536,1.1552874247233071
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,7168,65536,1.0025279998779297
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,7168,51200,0.9304416020711264
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,7168,51200,0.7855818430582683
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,8192,51200,0.3724085489908854
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,7168,16384,0.2850549379984538
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,7168,16384,0.25154239336649575
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,7168,16384,0.13045866489410402
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,7168,12288,0.21566294034322103
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,7168,12288,0.18142612775166828
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,8192,65536,0.47121601104736327
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,7168,10240,0.18391359647115071
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,7168,10240,0.15210453669230145
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,7168,12288,0.10090773105621338
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,7168,8192,0.14570453961690266
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,7168,8192,0.12519893646240235
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,7168,10240,0.0911733309427897
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,7168,7168,0.12934719721476237
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,7168,7168,0.10875519911448162
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,7168,8192,0.07533226807912191
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,7168,6144,0.11064319610595703
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,7168,6144,0.09652693271636963
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,7168,7168,0.07008319695790609
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,7168,5120,0.09400106271107991
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,7168,5120,0.08458879788716635
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,7168,6144,0.060939733187357584
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,7168,4096,0.075382399559021
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,7168,4096,0.06957226594289144
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,7168,5120,0.05480959812800089
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,7168,3584,0.06451946496963501
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,7168,3584,0.0620746652285258
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,7168,4096,0.047041066487630206
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,7168,3072,0.055371733506520596
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,7168,3072,0.05531946818033854
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,7168,3584,0.04263893365859985
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,7168,2560,0.046982399622599286
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,7168,2560,0.04772479931513469
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,7168,2560,0.03715733289718628
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,7168,2048,0.0384607990582784
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,7168,2048,0.04122560024261475
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,7168,3072,0.037970133622487384
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,7168,1536,0.028479999303817748
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,7168,1536,0.0371946652730306
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,7168,51200,0.3720255851745605
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,7168,1024,0.021286400159200032
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,7168,1024,0.03270933429400126
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,7168,1536,0.029960532983144123
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,7168,768,0.01714026729265849
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,7168,1024,0.02905600070953369
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,7168,768,0.029303467273712157
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,7168,512,0.014072533448537192
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,7168,512,0.027033599217732747
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,7168,512,0.026892799139022826
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,7168,256,0.01083626647790273
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,7168,256,0.022948267062505086
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,7168,768,0.02694293260574341
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,7168,128,0.008398933211962382
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,7168,128,0.02066133419672648
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,7168,256,0.02653119961420695
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,7168,64,0.007906133433183034
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,7168,64,0.020753065745035805
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,7168,32,0.007933866480986278
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,7168,128,0.026846933364868163
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,7168,32,0.01944533387819926
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,6144,65536,0.9574698766072591
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,6144,65536,0.8669216156005859
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,6144,51200,0.7324330647786458
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,6144,51200,0.6828128178914388
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,7168,2048,0.03359786669413249
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,6144,16384,0.2360095977783203
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,6144,16384,0.2149738629659017
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,6144,16384,0.10728533267974853
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,6144,12288,0.17876267433166504
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,6144,12288,0.15886613527933757
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,7168,65536,0.46755307515462236
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,6144,10240,0.16030186017354328
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,6144,10240,0.13462079366048177
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,6144,10240,0.07620906829833984
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,6144,8192,0.1291103998819987
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,6144,8192,0.10988907019297282
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,6144,12288,0.08411093552907309
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,6144,7168,0.11240959962209066
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,6144,7168,0.09880853494008382
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,6144,7168,0.05834240118662516
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,6144,6144,0.09636800289154053
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,6144,6144,0.08864959875742594
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,6144,8192,0.06286293268203735
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,6144,5120,0.08017280101776122
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,6144,5120,0.07642239729563395
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,6144,6144,0.050392532348632814
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,6144,4096,0.06293973525365194
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,6144,5120,0.04581013520558675
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,6144,4096,0.0627893328666687
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,6144,3584,0.05546559890111288
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,6144,3584,0.0560426672299703
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,6144,3584,0.036448001861572266
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,6144,3072,0.04777599970499675
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,6144,3072,0.05138560136159261
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,6144,4096,0.040006399154663086
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,6144,2560,0.040408531824747726
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,6144,2560,0.04449280103047688
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,6144,3072,0.03269973397254944
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,6144,2048,0.03330986698468526
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,6144,2048,0.03929813305536906
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,6144,2560,0.03111039996147156
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,6144,1536,0.025415466229120894
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,6144,1536,0.03491520086924235
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,6144,1536,0.025135999917984007
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,6144,1024,0.01872426668802897
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,6144,1024,0.02916053334871928
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,6144,2048,0.028076799710591634
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,6144,768,0.015397333105405173
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,6144,768,0.02682773272196452
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,6144,1024,0.024387200673421226
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,6144,512,0.01086293359597524
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,6144,512,0.0228928009668986
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,6144,768,0.026769065856933595
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,6144,256,0.008551466464996337
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,6144,256,0.021910399198532104
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,6144,512,0.026391466458638508
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,6144,128,0.0071050668756167095
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,6144,256,0.026088533798853557
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,6144,128,0.02063039938608805
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,6144,128,0.026358399788538617
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,6144,64,0.00673280010620753
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,6144,64,0.019010132551193236
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,6144,32,0.006775466601053874
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,6144,32,0.018713599443435668
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,5120,65536,0.8493770599365235
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,5120,65536,0.8135818481445313
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,5120,51200,0.6375989278157552
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,5120,51200,0.6568042755126953
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,6144,51200,0.30175679524739585
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,5120,16384,0.20475093523661295
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,5120,16384,0.2011349360148112
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,5120,16384,0.10681706269582111
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,5120,12288,0.1551370620727539
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,5120,12288,0.14851733843485515
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,6144,65536,0.3811242739359538
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,5120,10240,0.1416309356689453
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,5120,10240,0.13120746612548828
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,5120,10240,0.07665066719055176
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,5120,8192,0.11363519827524822
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,5120,8192,0.10357013543446858
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,5120,12288,0.08399786949157714
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,5120,7168,0.09747947057088216
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,5120,7168,0.09113066991170247
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,5120,7168,0.05756053527196249
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,5120,6144,0.08316373030344645
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,5120,6144,0.08203413486480712
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,5120,8192,0.062370133399963376
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,5120,5120,0.07083093325297038
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,5120,5120,0.07183466752370199
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,5120,5120,0.046002133687337236
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,5120,4096,0.05480639934539795
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,5120,4096,0.05728746652603149
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,5120,6144,0.050197335084279385
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,5120,3584,0.04692586660385132
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,5120,3584,0.05163946549097696
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,5120,3584,0.037140266100565596
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,5120,3072,0.04089813232421875
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,5120,3072,0.047882668177286786
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,5120,4096,0.04053759972254435
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,5120,2560,0.035792001088460285
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,5120,2560,0.043133866786956784
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,5120,2560,0.032305065790812174
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,5120,2048,0.029608533779780073
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,5120,2048,0.03685973485310872
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,5120,3072,0.03360746701558431
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,5120,1536,0.02276373306910197
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,5120,1536,0.033124266068140666
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,5120,1536,0.026741333802541095
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,5120,1024,0.016126933693885803
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,5120,1024,0.028835199276606244
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,5120,2048,0.029577600955963134
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,5120,768,0.013132799665133157
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,5120,768,0.02507200042406718
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,5120,1024,0.026014933983484905
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,5120,512,0.009655466675758362
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,5120,512,0.022665599981943764
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,5120,768,0.02741760015487671
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,5120,256,0.007190399865309398
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,5120,256,0.02026240030924479
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,5120,256,0.02754453420639038
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,5120,128,0.006682666639486949
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,5120,128,0.01916266679763794
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,5120,512,0.02781013250350952
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,5120,64,0.005818666517734527
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,5120,64,0.01886506676673889
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,5120,32,0.00589333325624466
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,5120,128,0.02442986567815145
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,5120,32,0.01848213275273641
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,4096,65536,0.6085311889648437
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,4096,65536,0.705234146118164
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,4096,51200,0.46034987767537433
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,4096,51200,0.5744618733723958
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,5120,51200,0.3031872113545736
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,4096,16384,0.1515135924021403
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,4096,16384,0.1742304007212321
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,4096,16384,0.08178133169809977
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,4096,12288,0.11467946370442708
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,4096,12288,0.13089066346486408
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,5120,65536,0.3780874570210775
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,4096,10240,0.10474666754404705
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,4096,10240,0.11031786600748698
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,4096,12288,0.06420586506525675
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,4096,8192,0.08423253695170084
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,4096,8192,0.09257919788360595
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,4096,10240,0.0583893338839213
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,4096,7168,0.07420693238576254
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,4096,7168,0.08234773476918539
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,4096,8192,0.04865386486053467
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,4096,6144,0.06289706627527872
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,4096,6144,0.07253546714782715
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,4096,7168,0.0440224011739095
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,4096,5120,0.052561068534851076
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,4096,5120,0.06141226689020792
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,4096,5120,0.03557440042495728
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,4096,4096,0.042139732837677
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,4096,4096,0.04938986698786418
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,4096,6144,0.03908906777699788
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,4096,3584,0.03788906733194987
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,4096,3584,0.047237332661946616
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,4096,4096,0.03095253308614095
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,4096,3072,0.032433066765467325
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,4096,3072,0.043057068188985186
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,4096,3584,0.028255999088287354
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,4096,2560,0.027291733026504516
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,4096,2560,0.03922239939371745
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,4096,3072,0.024887466430664064
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,4096,2048,0.022875734170277915
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,4096,2048,0.03336426814397176
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,4096,2560,0.023166932662328086
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,4096,1536,0.018230400482813516
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,4096,1536,0.031074132521947222
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,4096,2048,0.021057067314783733
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,4096,1024,0.01374613344669342
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,4096,1536,0.018227199713389076
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,4096,1024,0.02583786646525065
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,4096,768,0.0108106662829717
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,4096,768,0.024870399634043375
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,4096,768,0.01537493367989858
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,4096,512,0.008152533570925396
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,4096,512,0.020718934138615926
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,4096,51200,0.22387839953104655
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,4096,256,0.006332799792289734
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,4096,256,0.01885226567586263
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,4096,1024,0.017729065815607705
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,4096,128,0.005904000004132589
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,4096,512,0.01577600042025248
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,4096,128,0.01918826699256897
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,4096,64,0.005182933310667674
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,4096,64,0.01761066714922587
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,4096,128,0.015427199999491372
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,4096,32,0.0054837331175804135
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,4096,32,0.0165503998597463
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,4096,256,0.015281066298484802
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,3584,65536,0.5518954594930012
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,3584,65536,0.6861183802286784
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,3584,51200,0.4710645357767741
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,3584,51200,0.5367925643920899
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,3584,16384,0.14257920583089193
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,3584,16384,0.1677248001098633
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,3584,16384,0.0818613370259603
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,3584,12288,0.10839680035909016
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,3584,12288,0.12631253401438397
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,4096,65536,0.2804682731628418
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,3584,10240,0.09156053066253662
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,3584,10240,0.10770560105641683
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,3584,10240,0.05901120106379191
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,3584,8192,0.07449173132578532
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,3584,8192,0.08971520264943442
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,3584,12288,0.06492160161336263
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,3584,7168,0.06498560110727945
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,3584,7168,0.07875413099924723
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,3584,7168,0.045290664831797285
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,3584,6144,0.056109865506490074
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,3584,6144,0.06990826924641927
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,3584,8192,0.048349865277608234
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,3584,5120,0.047739732265472415
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,3584,5120,0.059197866916656496
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,3584,5120,0.03717120091120402
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,3584,4096,0.03924800157546997
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,3584,4096,0.0496394673983256
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,3584,6144,0.04003413518269856
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,3584,3584,0.03417173226674398
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,3584,3584,0.045575467745463054
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,3584,3584,0.02975146571795146
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,3584,4096,0.03295146624247233
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,3584,3072,0.043030401070912674
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,3584,3072,0.03123093247413635
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,3584,2560,0.026476800441741943
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,3584,2560,0.037817601362864176
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,3584,2560,0.02611733277638753
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,3584,2048,0.022336000204086305
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,3584,3072,0.02677333354949951
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,3584,2048,0.03292479912439982
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,3584,1536,0.017720532417297364
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,3584,1536,0.029285333553949994
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,3584,1536,0.021036799748738608
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,3584,1024,0.012478933731714884
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,3584,1024,0.026445867617925008
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,3584,2048,0.023543467124303184
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,3584,768,0.010046933094660442
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,3584,768,0.024718934297561647
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,3584,1024,0.020653865734736123
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,3584,512,0.008370133241017659
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,3584,512,0.02193173368771871
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,3584,768,0.019006933768590292
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,3584,256,0.006434133152167003
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,3584,256,0.01884373426437378
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,3584,512,0.018845866123835243
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,3584,128,0.005460266768932342
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,3584,256,0.018756266434987387
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,3584,128,0.018312533696492515
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,3584,64,0.00505920002857844
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,3584,64,0.01684160033861796
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,3584,32,0.0050570666790008545
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,3584,128,0.01873706579208374
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,3584,32,0.01663040022055308
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,3072,65536,0.5163317362467448
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,3072,65536,0.630398941040039
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,3072,51200,0.3674858729044596
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,3072,51200,0.51157652537028
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,3584,51200,0.22482453982035316
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,3072,16384,0.1164415995279948
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,3072,16384,0.1537343978881836
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,3072,16384,0.08109440008799235
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,3072,12288,0.08846933046976725
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,3072,12288,0.11663466294606525
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,3584,65536,0.28040959040323893
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,3072,10240,0.08140906492869059
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,3072,10240,0.10075946648915608
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,3072,12288,0.06491520007451376
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,3072,8192,0.06258133252461752
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,3072,8192,0.081223464012146
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,3072,10240,0.05837440093358358
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,3072,7168,0.05664746761322022
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,3072,7168,0.07032960255940755
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,3072,7168,0.044742401440938315
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,3072,6144,0.0457856019337972
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,3072,6144,0.060779734452565515
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,3072,8192,0.047917866706848146
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,3072,5120,0.03893760045369466
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,3072,5120,0.05237866640090942
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,3072,5120,0.03631893396377563
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,3072,4096,0.03163093328475952
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,3072,4096,0.04488319953282674
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,3072,6144,0.039572266737620036
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,3072,3584,0.028861866394678755
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,3072,3584,0.041282133261362715
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,3072,3584,0.02895893255869548
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,3072,3072,0.02413866718610128
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,3072,3072,0.03914026816685994
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,3072,4096,0.03170880079269409
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,3072,2560,0.02135573426882426
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,3072,2560,0.03521813154220581
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,3072,3072,0.02643839915593465
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,3072,2048,0.017436800400416057
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,3072,2048,0.03146986762682597
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,3072,51200,0.22668159802754723
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,3072,2560,0.02553386688232422
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,3072,1536,0.014062933127085366
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,3072,1536,0.0273632009824117
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,3072,1024,0.00979200005531311
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,3072,1024,0.025244800249735515
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,3072,2048,0.022757333517074586
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,3072,1024,0.020042665799458823
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,3072,768,0.022798933585484824
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,3072,1536,0.020623999834060668
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,3072,512,0.006555733581384023
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,3072,512,0.02067199945449829
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,3072,768,0.008775466680526733
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,3072,256,0.0055178667108217875
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,3072,256,0.01867306629816691
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,3072,512,0.018246400356292724
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,3072,128,0.005070933202902476
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,3072,128,0.016641066471735636
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,3072,256,0.018252799908320107
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,3072,768,0.018689066171646118
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,3072,128,0.018550399939219156
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,3072,64,0.017223467429478966
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,3072,32,0.016825600465138753
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,2560,65536,0.3986154556274414
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,3072,64,0.0046858668327331545
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,2560,65536,0.6153919855753581
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,3072,32,0.004659200211366018
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,2560,51200,0.3107733408610026
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,2560,51200,0.4743349393208821
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,2560,16384,0.10456000169118244
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,2560,16384,0.14768106142679852
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,2560,16384,0.07986773649851481
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,2560,12288,0.07904000282287597
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,2560,12288,0.1105994701385498
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,3072,65536,0.2828949292500814
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,2560,10240,0.07203199863433837
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,2560,10240,0.09413546721140544
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,2560,10240,0.056284801165262854
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,2560,8192,0.05599679946899414
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,2560,8192,0.07640960216522216
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,2560,12288,0.06347519954045613
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,2560,7168,0.048956799507141116
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,2560,7168,0.06637760003407797
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,2560,7168,0.0430400013923645
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,2560,6144,0.042507731914520265
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,2560,6144,0.05559680064519247
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,2560,8192,0.04634666840235392
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,2560,5120,0.03615573247273763
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,2560,5120,0.049355733394622806
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,2560,5120,0.03532693386077881
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,2560,4096,0.028124799331029255
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,2560,4096,0.041943466663360594
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,2560,6144,0.038072534402211505
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,2560,3584,0.025196800629297893
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,2560,3584,0.03938239812850952
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,2560,3584,0.028061866760253906
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,2560,3072,0.022127999862035116
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,2560,3072,0.03764586846033732
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,2560,4096,0.030573866764704388
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,2560,2560,0.018441599607467652
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,2560,2560,0.03293226758639018
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,2560,3072,0.025363200902938844
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,2560,2048,0.016167466839154564
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,2560,2048,0.029167999823888142
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,2560,2560,0.023571199178695677
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,2560,1536,0.01169706682364146
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,2560,1536,0.027164800961812334
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,2560,2048,0.0215445339679718
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,2560,1024,0.008805333574612936
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,2560,1024,0.023164800802866616
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,2560,1536,0.019436800479888917
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,2560,768,0.007154133419195812
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,2560,768,0.021577600638071695
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,2560,1024,0.018592000007629395
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,2560,512,0.005894400179386139
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,2560,512,0.020300799608230592
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,2560,512,0.017010132471720375
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,2560,768,0.016953599452972413
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,2560,256,0.005070933202902476
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,2560,256,0.017143466075261436
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,2560,128,0.004626133541266123
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,2560,128,0.01651946703592936
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,2560,256,0.016545066237449647
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,2560,64,0.004199466605981191
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,2560,128,0.01694719990094503
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,2560,32,0.004287999868392944
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,2560,64,0.01632213294506073
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,2560,32,0.01553600033124288
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,2048,65536,0.3293951988220215
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,2048,65536,0.5575103759765625
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,2048,51200,0.24917759895324706
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,2048,51200,0.4428906758626302
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,2560,51200,0.2227989355723063
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,2048,16384,0.08192426363627116
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,2048,16384,0.13626880645751954
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,2048,16384,0.05726720094680786
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,2048,12288,0.062283734480539955
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,2048,12288,0.10368746916453045
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,2560,65536,0.2812821388244629
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,2048,10240,0.051690665880839023
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,2048,10240,0.08626346588134766
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,2048,12288,0.045159467061360675
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,2048,8192,0.04213759899139404
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,2048,8192,0.0682581345240275
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,2048,10240,0.040567465623219806
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,2048,7168,0.03629333178202311
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,2048,7168,0.05911573171615601
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,2048,8192,0.033037867148717245
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,2048,6144,0.03185813426971436
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,2048,6144,0.05151253143946329
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,2048,7168,0.03141866723696391
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,2048,5120,0.026596266031265258
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,2048,5120,0.046366933981577554
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,2048,5120,0.02666986584663391
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,2048,4096,0.022235733270645142
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,2048,4096,0.04002133210500081
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,2048,6144,0.028009599447250365
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,2048,3584,0.0202239990234375
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,2048,3584,0.03700480063756307
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,2048,3584,0.021042132377624513
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,2048,3072,0.017761067549387614
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,2048,3072,0.03511573473612468
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,2048,4096,0.02288533250490824
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,2048,2560,0.015481600165367126
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,2048,2560,0.03135573267936707
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,2048,2560,0.018177066246668497
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,2048,2048,0.012968533237775168
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,2048,2048,0.02877226670583089
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,2048,3072,0.019382399320602418
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,2048,1536,0.010379733641942342
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,2048,1536,0.026443733771642046
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,2048,51200,0.14927466710408527
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,2048,1024,0.007963733375072479
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,2048,1024,0.023075199127197264
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,2048,2048,0.016545066237449647
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,2048,768,0.006754133105278015
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,2048,768,0.022001065810521445
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,2048,1024,0.014082133769989014
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,2048,512,0.005860266586144766
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,2048,512,0.018793600797653198
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,2048,768,0.012497066458066305
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,2048,256,0.004721066852410635
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,2048,256,0.017884800831476845
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,2048,512,0.012934399644533792
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,2048,1536,0.014565333724021912
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,2048,128,0.004147200038035711
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,2048,128,0.016578132907549538
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,2048,256,0.01284160017967224
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,2048,64,0.003952000041802724
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,2048,32,0.003956266740957896
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,2048,32,0.01623679995536804
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,1536,65536,0.24120960235595704
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,1536,65536,0.5281397183736165
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,2048,128,0.012852266430854797
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,1536,51200,0.1900373299916585
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,1536,51200,0.41900053024291994
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,2048,64,0.016537599762280784
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,1536,16384,0.061212801933288576
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,1536,16384,0.12720533212025958
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,2048,65536,0.1843733310699463
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,1536,12288,0.04825280110041301
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,1536,16384,0.05842133363087972
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,1536,12288,0.09517119725545248
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,1536,10240,0.04152746597925822
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,1536,10240,0.07881813049316407
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,1536,10240,0.0408458669980367
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,1536,8192,0.03241920073827108
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,1536,8192,0.0617248018582662
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,1536,12288,0.04577920039494832
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,1536,7168,0.02852800091107686
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,1536,7168,0.05339200099309286
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,1536,7168,0.0321941335995992
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,1536,6144,0.023963733514149984
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,1536,6144,0.047574400901794434
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,1536,8192,0.03394240140914917
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,1536,5120,0.02135573426882426
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,1536,5120,0.04327893257141113
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,1536,5120,0.027897600332895917
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,1536,4096,0.017560533682505288
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,1536,4096,0.03723520040512085
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,1536,6144,0.029253333806991577
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,1536,3584,0.016081066926320393
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,1536,3584,0.03410666783650716
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,1536,3584,0.022731733322143555
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,1536,3072,0.013894400000572205
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,1536,3072,0.03289600014686585
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,1536,4096,0.024112000068028768
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,1536,2560,0.012058666348457337
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,1536,2560,0.029848533868789672
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,1536,2560,0.01945599913597107
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,1536,2048,0.010154666503270467
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,1536,3072,0.020642133553822835
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,1536,2048,0.027082665761311846
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,1536,1536,0.008459732929865519
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,1536,1536,0.024618667364120484
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,1536,1536,0.015705600380897522
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,1536,1024,0.0067093332608540845
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,1536,1024,0.022473599513371786
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,1536,2048,0.01773973306020101
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,1536,768,0.005907199780146281
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,1536,768,0.021555199225743612
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,1536,1024,0.014937600493431092
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,1536,512,0.004866133133570353
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,1536,768,0.014454399545987448
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,1536,512,0.0191648006439209
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,1536,256,0.004228266576925913
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,1536,256,0.01741866668065389
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,1536,512,0.014125866691271463
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,1536,128,0.0035573333501815797
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,1536,128,0.016562133034070333
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,1536,128,0.014897066354751586
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,1536,64,0.0034005333979924522
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,1536,64,0.0161461333433787
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,1536,32,0.0034826666116714476
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,1536,32,0.017117865880330405
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,1024,65536,0.17173120180765789
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,1536,256,0.014144000411033631
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,1024,65536,0.48720318476359054
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,1024,51200,0.13497172991434733
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,1024,51200,0.38767147064208984
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,1536,51200,0.14922773043314616
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,1024,16384,0.04524266719818115
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,1024,16384,0.1184874693552653
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,1024,16384,0.05522773265838623
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,1024,12288,0.03402773141860962
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,1024,12288,0.08597226937611899
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,1024,12288,0.04241173267364502
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,1024,10240,0.027721599737803145
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,1024,10240,0.06916266282399496
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,1536,65536,0.1842005411783854
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,1024,8192,0.022449066241582237
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,1024,8192,0.05458773374557495
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,1024,8192,0.03138346672058105
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,1024,7168,0.01997013290723165
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,1024,10240,0.03755093415578206
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,1024,7168,0.04888000090916951
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,1024,6144,0.0173962672551473
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,1024,6144,0.04357866843541463
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,1024,7168,0.028945066531499225
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,1024,5120,0.014834133783976236
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,1024,5120,0.03860053221384684
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,1024,6144,0.02606933315594991
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,1024,4096,0.012555733323097229
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,1024,4096,0.03325653274854024
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,1024,5120,0.024779733022054037
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,1024,3584,0.010856533050537109
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,1024,3584,0.03096533417701721
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,1024,51200,0.14653120040893555
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,1024,3072,0.009698133667310078
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,1024,3072,0.0310208002726237
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,1024,3584,0.020272000630696615
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,1024,2560,0.008746666709582011
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,1024,2560,0.02695573369661967
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,1024,4096,0.02156053384145101
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,1024,2048,0.007512533167997997
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,1024,2048,0.025169066588083905
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,1024,3072,0.018311466773351034
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,1024,1536,0.0066655998428662615
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,1024,1536,0.023415466149648033
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,1024,2560,0.016942934195200602
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,1024,1024,0.005095466474692027
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,1024,2048,0.015416533748308817
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,1024,1024,0.021385600169499717
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,1024,1536,0.013705600301424661
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,1024,768,0.004689066608746847
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,1024,768,0.02034453352292379
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,1024,1024,0.012906666596730551
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,1024,768,0.012360533078511555
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,1024,512,0.004197333256403605
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,1024,512,0.0205567995707194
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,1024,256,0.0034111998975276947
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,1024,256,0.01723626653353373
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,1024,512,0.012078932921091716
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,1024,128,0.003032533327738444
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,1024,256,0.011684266726175944
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,1024,128,0.01612053314844767
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,1024,64,0.0031040000418821974
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,1024,128,0.012523733576138816
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,1024,32,0.0029919999341169994
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,1024,64,0.01663253307342529
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,1024,32,0.015711999932924905
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,768,65536,0.14426026344299317
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,768,51200,0.11298986275990804
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,768,65536,0.46411625544230145
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,768,51200,0.3665866533915202
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,768,16384,0.0384117325146993
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,768,16384,0.11502719720204671
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,1024,65536,0.18175679842631023
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,768,12288,0.028334933519363403
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,768,12288,0.08245440324147543
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,768,16384,0.056987734635670986
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,768,10240,0.022363734245300294
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,768,10240,0.06449280182520548
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,768,12288,0.043662933508555095
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,768,8192,0.01892906626065572
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,768,8192,0.051096534729003905
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,768,10240,0.03855786720911662
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,768,7168,0.016169599692026772
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,768,7168,0.04574506680170695
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,768,8192,0.03234773278236389
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,768,6144,0.014707199732462563
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,768,6144,0.04118826786677043
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,768,7168,0.030282666285832722
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,768,5120,0.012217600146929424
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,768,5120,0.037033601601918535
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,768,6144,0.02760639985402425
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,768,4096,0.010443733135859171
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,768,4096,0.03269973397254944
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,768,5120,0.025941334168116253
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,768,3584,0.0092031995455424
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,768,3584,0.030719999472300214
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,768,4096,0.023121066888173423
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,768,3072,0.008379733562469483
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,768,3072,0.028911999861399335
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,768,3584,0.02152106761932373
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,768,2560,0.007551999886830647
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,768,2560,0.026834134260813398
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,768,3072,0.019413334131240845
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,768,2048,0.006723199784755707
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,768,2048,0.024660267432530723
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,768,2560,0.01818986733754476
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,768,1536,0.005975466469923655
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,768,1536,0.022715733448664347
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,768,1536,0.014919466773668923
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,768,2048,0.016530133287111917
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,768,1024,0.004670933385690053
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,768,1024,0.02065066695213318
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,768,768,0.0042357335488001505
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,768,768,0.019435733556747437
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,768,1024,0.014434132973353067
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,768,512,0.0038730666041374207
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,768,768,0.013240533073743186
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,768,512,0.01879146695137024
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,768,256,0.0030432000756263735
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,768,256,0.015995732943216958
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,768,512,0.013297067085901896
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,768,128,0.002976000060637792
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,768,256,0.01323946714401245
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,768,128,0.01447466711203257
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,768,64,0.0025962665677070618
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,768,128,0.013261866569519044
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,768,64,0.014469333489735923
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,768,32,0.0026464000344276427
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,768,32,0.014487466216087342
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,512,65536,0.12069439888000488
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,512,65536,0.4503775914510091
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,512,51200,0.095524263381958
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,512,51200,0.35855468114217126
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,768,51200,0.14703253110249836
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,512,16384,0.03268373409907023
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,512,16384,0.11314880053202311
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,512,16384,0.05560959974924723
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,512,12288,0.022809600830078124
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,512,12288,0.07861013412475586
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,768,65536,0.1825834592183431
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,512,10240,0.01955839991569519
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,512,12288,0.04063680171966553
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,512,10240,0.062059732278188076
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,512,8192,0.017362133661905924
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,512,8192,0.04959786732991536
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,512,10240,0.03588053385416667
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,512,8192,0.030165332555770873
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,512,7168,0.014273066322008768
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,512,7168,0.04434666633605957
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,512,7168,0.028475733598073323
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,512,6144,0.012355200449625651
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,512,6144,0.04122879902521769
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,512,5120,0.010903466741243999
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,512,5120,0.037058134873708085
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,512,51200,0.1451541264851888
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,512,4096,0.00925439993540446
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,512,4096,0.031178667147954302
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,512,6144,0.02605760097503662
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,512,3584,0.008361599842707316
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,512,3584,0.030745599667231244
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,512,5120,0.024369066953659056
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,512,3072,0.007658666869004567
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,512,3072,0.029007999102274577
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,512,4096,0.021448532740275063
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,512,2560,0.0070602665344874065
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,512,2560,0.026473599672317504
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,512,3584,0.01949440042177836
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,512,2048,0.005929600199063619
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,512,2048,0.024701867500940958
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,512,3072,0.01789439916610718
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,512,1536,0.005508266886075338
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,512,2560,0.016607999801635742
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,512,2048,0.014919466773668923
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,512,1024,0.004266666869322458
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,512,1024,0.02072319984436035
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,512,1536,0.0129120002190272
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,512,768,0.003870933254559835
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,512,768,0.01880000034968058
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,512,1024,0.01286720037460327
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,512,512,0.0034400001168251038
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,512,512,0.01837013363838196
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,512,768,0.011552000045776367
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,512,256,0.002994133283694585
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,512,1536,0.022870399554570518
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,512,256,0.01639359990755717
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,512,512,0.011362133423487346
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,512,128,0.0026496000587940215
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,512,128,0.016340266664822897
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,512,64,0.0025898667673269907
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,512,128,0.011633066336313884
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,512,256,0.011750400066375732
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,512,32,0.0026357332865397137
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,512,64,0.014786133170127868
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,256,65536,0.06582933266957601
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,512,32,0.014625066518783569
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,256,51200,0.05271253188451132
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,256,65536,0.4313717206319173
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,256,51200,0.3405482610066732
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,256,16384,0.021384533246358237
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,256,16384,0.10676693121592204
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,512,65536,0.1813663959503174
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,256,12288,0.01652479966481527
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,256,12288,0.07292586962381999
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,256,12288,0.03915093342463176
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,256,10240,0.014114133516947427
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,256,10240,0.05603306690851847
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,256,16384,0.053496531645456945
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,256,8192,0.012070399522781373
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,256,8192,0.046133331457773846
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,256,10240,0.0353770653406779
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,256,7168,0.010794666409492493
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,256,7168,0.041152000427246094
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,256,7168,0.028479999303817748
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,256,6144,0.00958293378353119
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,256,8192,0.03009066581726074
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,256,6144,0.03721813360850017
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,256,5120,0.00837546686331431
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,256,5120,0.034916265805562334
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,256,5120,0.023976532618204753
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,256,4096,0.0071946665644645694
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,256,4096,0.030668799082438154
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,256,6144,0.025573333104451496
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,256,3584,0.006795733173688252
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,256,51200,0.14504106839497882
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,256,3584,0.029123200972874956
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,256,4096,0.02103466590245565
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,256,3072,0.02762453357378642
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,256,3584,0.019869865973790486
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,256,2560,0.005504000186920166
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,256,3072,0.017746132612228394
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,256,2560,0.02555733323097229
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,256,2048,0.004667733112970988
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,256,2048,0.025143466393152875
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,256,2560,0.016520532965660095
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,256,3072,0.006312533219655354
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,256,2048,0.01493119994799296
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,256,1536,0.022046933571497597
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,256,1024,0.003803733239571253
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,256,1024,0.021660800774892172
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,256,1536,0.01330880026022593
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,256,768,0.003399466723203659
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,256,768,0.01904426614443461
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,256,1024,0.012487467130025227
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,256,512,0.0030858665704727173
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,256,768,0.011622400085131327
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,256,512,0.018576000134150186
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,256,65536,0.18097920417785646
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,256,512,0.011566932996114094
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,256,256,0.0027082666754722597
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,256,1536,0.004257066547870636
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,256,256,0.016547200083732606
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,256,128,0.002586666742960612
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,256,128,0.014652799566586813
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,256,256,0.011240532994270325
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,256,64,0.002195200075705846
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,256,128,0.01163093348344167
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,256,32,0.00257493332028389
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,256,64,0.01575040022532145
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,256,32,0.01404159963130951
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,128,65536,0.054413866996765134
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,128,51200,0.04440426826477051
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,128,65536,0.42482452392578124
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,128,51200,0.33558505376180015
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,128,16384,0.011672533551851908
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,128,16384,0.10510400136311848
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,128,16384,0.052154668172200526
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,128,12288,0.012016000350316365
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,128,12288,0.07067306836446127
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,128,10240,0.010863999525705974
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,128,10240,0.05549120108286539
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,128,10240,0.03511679967244466
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,128,8192,0.011642666657765706
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,128,12288,0.0396938681602478
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,128,8192,0.0452128012975057
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,128,7168,0.010403199990590414
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,128,7168,0.041196799278259276
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,128,51200,0.141866668065389
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,128,6144,0.009176533420880635
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,128,7168,0.027923200527826948
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,128,8192,0.029318400224049884
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,128,6144,0.037360000610351565
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,128,5120,0.008438400427500407
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,128,5120,0.03367253144582112
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,128,4096,0.006717866659164429
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,128,4096,0.029498666524887085
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,128,6144,0.02518613338470459
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,128,65536,0.17691413561503094
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,128,3584,0.006325333317120869
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,128,3584,0.02940373420715332
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,128,3072,0.005927466849486033
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,128,4096,0.020310399929682414
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,128,3072,0.026858667532602947
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,128,2560,0.005469866593678792
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,128,5120,0.023504000902175904
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,128,2560,0.024784000714619954
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,128,2048,0.004634666442871094
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,128,3072,0.01693120002746582
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,128,2048,0.02295253276824951
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,128,1536,0.0042357335488001505
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,128,1536,0.021090133984883627
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,128,2048,0.014014933506647745
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,128,1024,0.003399466723203659
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,128,2560,0.01590079963207245
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,128,1024,0.021090133984883627
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,128,1536,0.012497066458066305
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,128,768,0.003049599876006444
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,128,768,0.018802134195963542
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,128,512,0.0025792000194390613
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,128,768,0.010431999961535137
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,128,512,0.016936532656351724
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,128,1024,0.012124799688657125
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,128,512,0.010413866241772969
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,128,3584,0.018956800301869713
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,128,256,0.017177599668502807
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,128,256,0.010771200060844421
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,128,128,0.002221866697072983
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,128,64,0.002203733225663503
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2304,128,128,0.010422399640083313
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,128,64,0.014499200383822122
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,128,32,0.00258240004380544
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,128,32,0.015275733669598899
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,64,65536,0.4222400029500325
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,64,51200,0.04147306680679321
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,128,256,0.0025888000925381976
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,64,16384,0.01402773360411326
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,64,51200,0.3339146614074707
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,64,16384,0.10353279908498128
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,64,12288,0.011610666910807293
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,64,10240,0.010289067029953003
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,64,12288,0.0697312037150065
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,64,8192,0.008733866612116496
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,64,10240,0.05458879868189494
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,128,128,0.015537066260973611
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,64,7168,0.008379733562469483
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,64,8192,0.04431999921798706
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,64,6144,0.007560533285140991
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,64,65536,0.05366400082906088
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,64,5120,0.0071168000499407455
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,64,4096,0.006683733562628429
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,64,6144,0.03702293237050374
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,64,5120,0.03377813498179118
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,64,7168,0.04132800102233887
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,64,3584,0.00625600020090739
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,64,4096,0.030869332949320476
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,64,3072,0.005709866682688395
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,64,2560,0.005075199902057648
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,64,3072,0.02676266630490621
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,64,2048,0.004622933268547058
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,64,2048,0.02308373252550761
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,64,2560,0.02522346576054891
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,64,1536,0.003824000060558319
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,64,1024,0.0033887999753157297
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,64,1024,0.02026559909184774
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,64,1536,0.022537599007288613
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,64,768,0.002997333308060964
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,64,512,0.0026730666557947796
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,64,512,0.0170741339524587
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,64,768,0.01917120019594828
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,64,256,0.0026079999903837843
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,64,128,0.002182399978240331
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,64,256,0.017207467555999757
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,64,128,0.014643200238545737
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,64,64,0.00236160010099411
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,64,64,0.014479999740918478
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,64,32,0.0022869333624839783
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,64,32,0.014105600118637086
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,32,65536,0.052009598414103186
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,32,51200,0.039094400405883786
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,64,3584,0.027147734165191652
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,32,16384,0.00957546631495158
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,32,65536,0.4224960009256999
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,32,12288,0.00801813354094823
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,32,51200,0.3327669461568197
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,32,10240,0.007587199906508128
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,32,12288,0.07089813550313315
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,32,8192,0.011627733707427979
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,32,10240,0.05586773157119751
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,32,8192,0.04538133144378662
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,32,7168,0.01064639985561371
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,32,6144,0.009630933403968811
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,32,7168,0.04036159912745158
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,32,5120,0.009139200051625569
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,32,5120,0.03322346607844035
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,32,6144,0.037332268555959065
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,32,4096,0.008291199803352356
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,32,3584,0.006372266511122386
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,32,4096,0.03024959961573283
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,32,3584,0.02758079965909322
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,32,3072,0.005886933207511902
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,32,2560,0.005133866767088572
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,32,3072,0.026795732975006103
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,32,16384,0.10557546615600585
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,32,2048,0.00487360010544459
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,32,1536,0.0038272000849246977
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,32,2560,0.02487466732660929
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,32,1024,0.003373866776625315
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,32,768,0.0030154667794704436
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,32,1536,0.02211946646372477
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,32,1024,0.020595200856526694
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,32,512,0.0027104000250498454
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,32,256,0.002595199892918269
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,32,768,0.01857066750526428
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,32,512,0.017029333114624023
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,32,128,0.002204799900452296
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,32,64,0.002164266755183538
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,32,128,0.014819199840227762
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2304,32,32,0.002163200080394745
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,32,64,0.014906666676203408
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,32,32,0.016141866644223533
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,32,2048,0.022833067178726196
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2304,32,256,0.01625599960486094
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,65536,16384,2.4978401184082033
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,65536,51200,4.320553588867187
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,65536,16384,1.3021279652913411
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,65536,12288,1.818508783976237
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,65536,51200,8.116837565104166
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,65536,12288,0.975593630472819
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,65536,10240,1.535045369466146
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,65536,10240,0.8206282933553061
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,65536,10240,0.48537387847900393
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,65536,8192,1.1610591888427735
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,65536,8192,0.6427616119384766
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,65536,12288,0.5516223907470703
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,65536,7168,1.0680842081705728
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,65536,7168,0.5742752075195312
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,65536,16384,0.7274186452229817
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,65536,6144,0.8553130467732748
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,65536,6144,0.5093269348144531
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,65536,8192,0.3985109329223633
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,65536,5120,0.7232618967692057
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,65536,5120,0.486190923055013
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,65536,7168,0.36799147923787434
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,65536,4096,0.5832533518473307
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,65536,4096,0.3639466603597005
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,65536,6144,0.31293439865112305
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,65536,3584,0.5016096115112305
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,65536,3584,0.31640745798746744
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,65536,4096,0.25839786529541015
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,65536,3072,0.47044798533121746
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,65536,3072,0.2909930547078451
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,65536,5120,0.2853834788004557
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,65536,2560,0.3880352020263672
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,65536,2560,0.2410912036895752
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,65536,3584,0.22474452654520669
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,65536,2048,0.3104501406351725
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,65536,2048,0.19562239646911622
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,65536,2560,0.19139946301778157
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,65536,1536,0.24124266306559244
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,65536,1536,0.152566401163737
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,65536,3072,0.20309012730916343
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,65536,1024,0.17512106895446777
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,65536,1024,0.11346986293792724
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,65536,1024,0.14024213155110676
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,65536,768,0.13144853115081787
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,65536,768,0.0964085340499878
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,65536,1536,0.14586666425069172
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,65536,512,0.0950762669245402
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,65536,512,0.08245866298675537
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,65536,2048,0.16786773999532062
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,65536,256,0.0644874652226766
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,65536,256,0.07026133537292481
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,65536,512,0.1316810687383016
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,65536,128,0.050614400704701745
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,65536,128,0.057962667942047116
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,65536,768,0.13082986672719318
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,65536,64,0.044072532653808595
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,65536,64,0.05814079840977987
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,65536,32,0.04530773162841797
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,65536,32,0.05980053345362345
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,65536,128,0.13269440333048504
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,65536,256,0.12897173563639325
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,51200,65536,8.096431986490886
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,51200,65536,4.382437133789063
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,51200,51200,3.478131103515625
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,51200,51200,6.587405904134114
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,51200,16384,1.9328319549560546
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,51200,16384,1.1064362843831381
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,51200,16384,0.5875733057657878
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,51200,12288,1.4313578287760416
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,51200,12288,0.7831605275472004
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,51200,12288,0.4431167920430501
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,51200,10240,1.378664525349935
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,51200,10240,0.6517653147379557
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,51200,10240,0.39121920267740884
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,51200,8192,1.0639839808146159
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,51200,8192,0.5154058774312337
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,65536,51200,2.2977760314941404
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,51200,7168,0.8742687861124674
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,51200,7168,0.449565855662028
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,51200,8192,0.3211626688639323
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,51200,6144,0.6724138895670573
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,51200,6144,0.4214912096659343
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,51200,7168,0.29447040557861326
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,51200,5120,0.558676274617513
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,51200,5120,0.3357173283894857
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,51200,6144,0.252455472946167
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,51200,4096,0.44988587697347004
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,51200,4096,0.2792330741882324
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,51200,5120,0.22794987360636393
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,51200,3584,0.3927562713623047
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,51200,3584,0.24951146443684896
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,51200,3584,0.1815231959025065
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,51200,3072,0.33876800537109375
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,51200,3072,0.2046602725982666
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,51200,4096,0.2038304011027018
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,51200,2560,0.28719679514567054
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,51200,2560,0.1753546714782715
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,51200,51200,1.8141354878743488
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,51200,2048,0.23014399210611977
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,51200,2048,0.14180053075154622
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,51200,2560,0.15398292541503905
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,51200,1536,0.1781354745229085
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,51200,1536,0.11726613044738769
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,51200,3072,0.16099947293599445
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,51200,1024,0.1222645362218221
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,51200,1024,0.08638719717661539
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,51200,1024,0.11525119940439861
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,51200,768,0.09615999857584635
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,51200,768,0.07485013008117676
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,51200,1536,0.11800533135732014
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,51200,512,0.07106666564941407
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,51200,512,0.06499199867248535
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,51200,2048,0.13552746772766114
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,51200,256,0.04827839930852254
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,51200,256,0.05654186805089315
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,51200,256,0.10589120388031006
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,51200,128,0.04250133434931437
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,51200,128,0.05069653193155924
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,51200,768,0.10644693374633789
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,51200,64,0.03662613232930501
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,51200,64,0.052140800158182773
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,51200,32,0.038268800576527914
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,51200,32,0.05264426469802856
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,51200,128,0.10709333419799805
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,16384,65536,2.5333567301432294
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,16384,65536,1.565985107421875
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,51200,512,0.10663999716440838
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,16384,51200,1.2413162231445312
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,16384,51200,1.978071467081706
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,16384,16384,0.6388074874877929
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,16384,16384,0.3828960100809733
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,16384,16384,0.2222058614095052
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,16384,12288,0.4432064056396484
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,16384,12288,0.28466453552246096
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,16384,12288,0.17034239768981935
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,16384,10240,0.459717337290446
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,16384,10240,0.28131732940673826
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,16384,10240,0.1512106736501058
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,16384,8192,0.3683231989542643
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,16384,8192,0.22594985961914063
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,16384,8192,0.12470080057779949
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,16384,7168,0.32437121073404945
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,16384,7168,0.2026517391204834
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,16384,51200,0.6558879852294922
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,16384,6144,0.2821077346801758
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,16384,6144,0.17738134066263836
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,16384,7168,0.11545813083648682
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,16384,5120,0.22463040351867675
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,16384,5120,0.13923947016398114
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,16384,5120,0.09054826895395915
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,16384,4096,0.17940692901611327
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,16384,4096,0.10383040110270184
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,16384,6144,0.09874239762624105
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,16384,3584,0.15857812563578289
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,16384,3584,0.09785172939300538
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,16384,3584,0.06982186635335287
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,16384,3072,0.14200533231099446
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,16384,3072,0.08510186672210693
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,16384,4096,0.0791605313618978
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,16384,2560,0.1033354679743449
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,16384,2560,0.07373546759287516
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,51200,65536,2.316221872965495
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,16384,2048,0.07821333408355713
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,16384,2048,0.06166719992955526
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,16384,2560,0.05822933514912924
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,16384,3072,0.061324799060821535
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,16384,1536,0.06029119888941446
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,16384,1536,0.05530773401260376
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,16384,1024,0.04305493434270223
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,16384,1024,0.044300798575083414
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,16384,2048,0.04928106864293416
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,16384,768,0.03314346671104431
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,16384,768,0.03919893503189087
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,16384,768,0.03527359962463379
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,16384,512,0.024465066194534302
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,16384,512,0.035835735003153485
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,16384,1024,0.041620266437530515
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,16384,256,0.017935999234517417
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,16384,256,0.029741867383321123
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,16384,512,0.034969600041707356
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,16384,256,0.03442026774088542
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,16384,128,0.016620799899101257
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,16384,128,0.026898133754730224
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,16384,64,0.013273599743843078
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,16384,64,0.025649066766103106
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,16384,32,0.014106667041778565
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,16384,32,0.027101866404215497
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,16384,128,0.0352565328280131
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,16384,1536,0.043005867799123125
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,12288,65536,1.9547147115071613
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,12288,65536,1.250534439086914
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,12288,51200,1.4696479797363282
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,12288,51200,0.974293327331543
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,12288,16384,0.43115625381469724
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,12288,16384,0.31829973856608074
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,16384,65536,0.8293792088826498
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,12288,12288,0.32928854624430337
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,12288,12288,0.23607145945231117
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,12288,16384,0.17566506067911786
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,12288,10240,0.27287467320760095
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,12288,10240,0.1984597365061442
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,12288,12288,0.13449813524881998
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,12288,8192,0.21983787218729653
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,12288,8192,0.16429972648620605
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,12288,10240,0.12026666800181071
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,12288,7168,0.20947413444519042
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,12288,7168,0.13417919476826984
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,12288,8192,0.09917653401692708
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,12288,6144,0.1832149346669515
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,12288,6144,0.1151039997736613
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,12288,7168,0.09206613699595133
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,12288,5120,0.148251740137736
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,12288,5120,0.10084799925486247
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,12288,6144,0.07980053424835205
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,12288,4096,0.11979093551635742
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,12288,4096,0.08202986717224121
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,12288,5120,0.07215999762217204
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,12288,3584,0.09395840167999267
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,12288,3584,0.07783040205637613
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,12288,4096,0.0624181350072225
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,12288,3072,0.08255360126495362
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,12288,3072,0.06788907051086426
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,12288,3584,0.056194134553273524
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,12288,2560,0.06861013571421305
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,12288,2560,0.0605568011601766
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,12288,3072,0.050632532437642416
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,12288,2048,0.05686080058415731
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,12288,2048,0.05003519852956136
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,12288,2560,0.049269334475199385
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,12288,1536,0.04282453457514445
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,12288,1536,0.04391253391901652
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,12288,1536,0.04119466543197632
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,12288,1024,0.029846400022506714
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,12288,1024,0.03632746537526448
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,12288,2048,0.04495999813079834
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,12288,768,0.023639466365178427
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,12288,768,0.032868266105651855
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,12288,1024,0.04044266541798909
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,12288,512,0.017478400468826295
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,12288,512,0.029715200265248615
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,12288,768,0.037316266695658365
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,12288,256,0.013247999548912048
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,12288,256,0.024668800830841064
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,12288,256,0.03624853293100993
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,12288,128,0.01323946714401245
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,12288,128,0.023798400163650514
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,12288,512,0.036830933888753255
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,12288,64,0.01207360029220581
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,12288,64,0.02490026752154032
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,12288,32,0.012801067034403483
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,12288,128,0.03658560117085775
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,12288,32,0.02294720013936361
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,10240,65536,1.5540181477864583
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,10240,65536,1.094259262084961
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,10240,51200,1.194050089518229
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,10240,51200,0.8526218414306641
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,12288,51200,0.5109578768412273
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,10240,16384,0.3716522534688314
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,10240,16384,0.2725386619567871
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,10240,16384,0.1527114709218343
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,10240,12288,0.281881586710612
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,10240,12288,0.2092949390411377
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,12288,65536,0.643662961324056
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,10240,10240,0.2522400061289469
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,10240,10240,0.17537706693013508
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,10240,12288,0.11768960158030192
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,10240,8192,0.20767146746317544
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,10240,8192,0.14738240242004394
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,10240,10240,0.10527466932932536
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,10240,7168,0.17711253166198732
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,10240,7168,0.1285930633544922
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,10240,8192,0.08713280359903972
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,10240,6144,0.1555605411529541
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,10240,6144,0.11128640174865723
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,10240,7168,0.08075946966807047
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,10240,5120,0.1254250685373942
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,10240,5120,0.09777173201243082
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,10240,6144,0.07002879778544108
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,10240,4096,0.10131093660990398
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,10240,4096,0.08035733699798583
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,10240,5120,0.06316266854604086
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,10240,51200,0.4364789326985677
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,10240,3584,0.08729386329650879
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,10240,3584,0.07556053002675375
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,10240,3072,0.07417600154876709
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,10240,3072,0.0664789319038391
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,10240,4096,0.05542933146158854
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,10240,2560,0.06454826593399048
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,10240,2560,0.058208000659942624
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,10240,3072,0.04612586498260498
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,10240,2048,0.051268267631530764
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,10240,2048,0.04564906756083171
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,10240,3584,0.05025279919306437
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,10240,1536,0.04066239992777507
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,10240,1536,0.0401962677637736
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,10240,1536,0.03716906706492106
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,10240,1024,0.025436800718307496
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,10240,1024,0.03330346743265788
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,10240,2560,0.04471466541290283
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,10240,768,0.020670932531356812
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,10240,768,0.03044266700744629
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,10240,2048,0.04149973392486572
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,10240,512,0.015337600310643514
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,10240,512,0.02772586743036906
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,10240,768,0.03373333215713501
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,10240,256,0.011902933319409687
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,10240,512,0.034371201197306314
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,10240,256,0.024081067244211832
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,10240,128,0.00921066701412201
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,10240,128,0.020965333779652914
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,10240,128,0.0334389328956604
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,10240,64,0.008379733562469483
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,10240,64,0.02109439969062805
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,10240,32,0.008387200037638346
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,10240,32,0.02086720069249471
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,10240,256,0.03417386611302693
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,8192,65536,0.941107177734375
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,8192,65536,1.2285247802734376
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,10240,1024,0.03629120190938313
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,8192,51200,0.9744874954223632
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,8192,51200,0.7811733245849609
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,8192,16384,0.29035838445027673
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,8192,16384,0.24559253056844077
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,10240,65536,0.5499776204427083
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,8192,12288,0.21899092992146812
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,8192,12288,0.1747872034708659
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,8192,16384,0.1262933333714803
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,8192,10240,0.18562666575113934
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,8192,10240,0.16262613932291667
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,8192,12288,0.09719680150349935
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,8192,8192,0.14888639450073243
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,8192,8192,0.13501866658528644
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,8192,10240,0.08798293272654215
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,8192,7168,0.13173866271972656
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,8192,7168,0.12307413419087727
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,8192,7168,0.06699199676513672
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,8192,6144,0.11415359973907471
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,8192,6144,0.10609493255615235
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,8192,8192,0.07189013163248698
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,8192,5120,0.11652586460113526
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,8192,5120,0.09383467038472494
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,8192,5120,0.052426668008168545
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,8192,4096,0.0980511983235677
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,8192,4096,0.0698794682820638
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,8192,6144,0.05801920096079508
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,8192,3584,0.08534293174743653
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,8192,3584,0.06640106836954753
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,8192,3584,0.0412224014600118
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,8192,3072,0.06994880040486653
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,8192,3072,0.05859306653340658
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,8192,4096,0.045059200127919516
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,8192,2560,0.05835946798324585
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,8192,2560,0.05027733246485392
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,8192,2560,0.035190399487813315
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,8192,2048,0.04245013395945231
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,8192,2048,0.04230826695760091
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,8192,3072,0.03649386564890544
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,8192,1536,0.03333973487218221
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,8192,1536,0.03888426621754964
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,8192,1536,0.02757973273595174
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,8192,1024,0.023851732412974037
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,8192,1024,0.03241706689198812
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,8192,2048,0.031146667400995892
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,8192,768,0.018722132841746012
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,8192,768,0.029049599170684816
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,8192,1024,0.026826665798823042
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,8192,512,0.014452266693115234
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,8192,512,0.025704532861709595
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,8192,768,0.02348053256670634
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,8192,256,0.01037013332049052
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,8192,256,0.022681599855422972
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,8192,256,0.02316373387972514
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,8192,512,0.02333546678225199
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,8192,128,0.00874133308728536
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,8192,128,0.02084266742070516
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,8192,64,0.007518933216730754
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,8192,64,0.02062826752662659
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,8192,32,0.008006399869918824
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,8192,32,0.021157334248224892
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,8192,128,0.023504000902175904
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,7168,65536,0.9987765630086264
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,7168,65536,0.859715207417806
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,7168,51200,0.7799360275268554
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,7168,51200,0.6752127965291341
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,8192,51200,0.361625607808431
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,7168,16384,0.24989546140034996
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,7168,16384,0.20855040550231935
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,7168,16384,0.12564906279246013
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,7168,12288,0.18891305923461915
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,7168,12288,0.15575253168741862
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,8192,65536,0.4549258550008138
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,7168,10240,0.15534933408101398
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,7168,10240,0.1316106637318929
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,7168,12288,0.097380264600118
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,7168,8192,0.12741440137227375
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,7168,8192,0.11025706926981609
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,7168,10240,0.08784639835357666
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,7168,7168,0.110425599416097
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,7168,7168,0.1007317304611206
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,7168,8192,0.07220160166422526
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,7168,6144,0.09631893634796143
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,7168,6144,0.0878175973892212
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,7168,7168,0.06662506659825643
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,7168,5120,0.0801141341527303
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,7168,5120,0.07807573477427164
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,7168,5120,0.05214826663335165
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,7168,4096,0.06491520007451376
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,7168,4096,0.06366613308588663
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,7168,6144,0.057988266150156655
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,7168,3584,0.05786240100860596
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,7168,3584,0.06141226689020792
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,7168,4096,0.04578773180643718
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,7168,3072,0.050018131732940674
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,7168,3072,0.05339733362197876
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,7168,3584,0.04164479970932007
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,7168,2560,0.04314560095469157
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,7168,2560,0.0452021320660909
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,7168,51200,0.35878934860229494
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,7168,2048,0.03553493420283
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,7168,2048,0.03811200062433879
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,7168,2560,0.03644373416900635
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,7168,1536,0.0286954661210378
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,7168,1536,0.03535360097885132
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,7168,2048,0.03343466520309448
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,7168,1024,0.01894506613413493
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,7168,3072,0.03768853346506755
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,7168,1024,0.029638399680455525
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,7168,768,0.015457066893577575
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,7168,768,0.02676266630490621
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,7168,768,0.02648640076319377
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,7168,512,0.012115200360616047
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,7168,512,0.025304534037907916
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,7168,1536,0.029364265998204547
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,7168,256,0.008827733000119527
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,7168,256,0.021230934063593547
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,7168,256,0.026131200790405273
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,7168,1024,0.028787199656168622
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,7168,512,0.02638826568921407
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,7168,128,0.007950933277606964
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,7168,128,0.018887466192245482
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,7168,64,0.0071274667978286745
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,7168,64,0.02024959921836853
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,7168,32,0.007099733253320058
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,7168,32,0.020080000162124634
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,7168,128,0.026060799757639568
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,6144,65536,0.9003658930460612
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,6144,65536,0.7872693379720052
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,6144,51200,0.6626741409301757
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,6144,51200,0.6183893203735351
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,6144,16384,0.21878399848937988
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,6144,16384,0.1919381300608317
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,7168,65536,0.455515734354655
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,6144,12288,0.16718079249064127
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,6144,12288,0.1434655984242757
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,6144,16384,0.10273706912994385
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,6144,10240,0.1529930591583252
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,6144,10240,0.12597866853078205
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,6144,10240,0.07323199907938639
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,6144,8192,0.12376853624979656
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,6144,8192,0.10338239669799805
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,6144,12288,0.0806816021601359
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,6144,7168,0.1116703987121582
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,6144,7168,0.09634772936503092
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,6144,7168,0.055676798025767006
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,6144,6144,0.09514559904734293
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,6144,8192,0.06035733222961426
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,6144,6144,0.08296960194905599
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,6144,5120,0.07836159865061441
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,6144,5120,0.07453760306040445
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,6144,5120,0.044225064913431804
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,6144,4096,0.06187413136164347
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,6144,4096,0.05841919978459677
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,6144,6144,0.04925440152486165
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,6144,3584,0.05375359853108724
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,6144,3584,0.05624106725056967
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,6144,4096,0.03901760180791219
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,6144,3584,0.03537813425064087
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,6144,3072,0.04711360136667887
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,6144,3072,0.049377067883809404
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,6144,2560,0.03999360005060832
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,6144,2560,0.043567999203999834
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,6144,2560,0.030776532491048177
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,6144,2048,0.03290560046831767
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,6144,2048,0.03796586592992147
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,6144,3072,0.03181546727816264
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,6144,1536,0.025229867299397784
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,6144,1536,0.03482346534729004
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,6144,1536,0.02523733377456665
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,6144,1024,0.018894932667414346
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,6144,1024,0.0279423991839091
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,6144,2048,0.027829333146413164
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,6144,768,0.01530026694138845
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,6144,768,0.02648746569951375
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,6144,1024,0.024363734324773154
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,6144,512,0.011078400413195293
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,6144,512,0.023026132583618165
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,6144,768,0.02632213234901428
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,6144,256,0.00841813286145528
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,6144,256,0.022936532894770302
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,6144,256,0.025854933261871337
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,6144,512,0.025995733340581258
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,6144,128,0.00674773355325063
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,6144,128,0.018914133310317993
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,6144,64,0.006274133423964183
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,6144,64,0.018565332889556883
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,6144,32,0.00631466656923294
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,6144,32,0.01859626571337382
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,6144,128,0.025991467634836833
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,5120,65536,0.7685002644856771
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,5120,65536,0.7360917409261067
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,5120,51200,0.5785834630330403
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,5120,51200,0.5761333465576172
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,6144,51200,0.29009065628051756
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,5120,16384,0.18754879633585614
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,5120,16384,0.17793599764506024
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,5120,16384,0.10268266995747884
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,5120,12288,0.1415434678395589
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,5120,12288,0.13130026658376057
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,6144,65536,0.3646464029947917
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,5120,10240,0.11741013526916504
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,5120,10240,0.10775466759999593
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,5120,12288,0.080840531984965
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,5120,8192,0.09214933713277182
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,5120,8192,0.08912106355031332
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,5120,10240,0.07269653479258219
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,5120,7168,0.08136320114135742
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,5120,7168,0.08188587029774984
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,5120,8192,0.059640534718831385
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,5120,6144,0.06754879951477051
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,5120,6144,0.07019626299540202
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,5120,7168,0.05531839927037557
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,5120,5120,0.05736426512400309
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,5120,5120,0.06112213134765625
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,5120,5120,0.043987198670705156
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,5120,4096,0.04656320015589396
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,5120,4096,0.04944640000661214
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,5120,6144,0.0485429326693217
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,5120,3584,0.04126826524734497
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,5120,3584,0.04702506860097249
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,5120,3584,0.03625493446985881
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,5120,3072,0.03558613459269206
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,5120,3072,0.0426528016726176
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,5120,4096,0.039043200016021726
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,5120,2560,0.02959146698315938
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,5120,2560,0.03864106734593709
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,5120,3072,0.03263253370920817
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,5120,2048,0.02453546722730001
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,5120,2048,0.033199999729792276
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,5120,51200,0.28918933868408203
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,5120,2560,0.03179840048154195
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,5120,1536,0.019542400042215982
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,5120,1536,0.030997333923975627
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,5120,1024,0.013895466923713684
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,5120,1024,0.02690560022989909
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,5120,1024,0.025697066386540728
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,5120,768,0.011080533266067505
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,5120,768,0.024682666858037314
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,5120,1536,0.02605653405189514
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,5120,512,0.008805333574612936
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,5120,512,0.02260479927062988
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,5120,768,0.02746986746788025
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,5120,256,0.006797866523265838
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,5120,256,0.02008426586786906
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,5120,2048,0.02878933350245158
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,5120,128,0.005895466605822245
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,5120,128,0.017835734287897746
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,5120,512,0.0244159996509552
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,5120,64,0.005578666428724925
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,5120,128,0.023363200823465984
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,5120,64,0.018415999412536622
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,5120,32,0.005485866467158
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,5120,256,0.026625066995620728
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,5120,32,0.018477867046991982
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,4096,65536,0.5845546722412109
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,4096,65536,0.6384021123250325
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,4096,51200,0.4639295895894368
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,4096,51200,0.5020693461100261
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,4096,16384,0.14024853706359863
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,4096,16384,0.15562987327575684
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,4096,16384,0.07764373620351156
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,4096,12288,0.10638613700866699
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,4096,12288,0.11611093680063884
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,5120,65536,0.36523520151774086
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,4096,10240,0.10692479610443115
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,4096,10240,0.10369919935862223
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,4096,10240,0.05583360195159912
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,4096,8192,0.08606506983439127
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,4096,8192,0.08587733109792074
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,4096,12288,0.06185919841130575
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,4096,7168,0.07434986432393392
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,4096,7168,0.07816853523254394
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,4096,7168,0.04263039827346802
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,4096,6144,0.06305919885635376
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,4096,6144,0.06633813381195068
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,4096,8192,0.04633493423461914
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,4096,5120,0.05328640143076578
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,4096,5120,0.05725013415018717
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,4096,6144,0.03836479981740316
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,4096,4096,0.04226346810658772
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,4096,4096,0.04783466657002767
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,4096,5120,0.03457066615422567
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,4096,3584,0.037267200152079266
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,4096,3584,0.04513066609700521
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,4096,3584,0.027536000808080035
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,4096,3072,0.031699200471242264
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,4096,3072,0.04086720148722331
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,4096,4096,0.03023146589597066
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,4096,2560,0.026594134171803792
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,4096,2560,0.03708266814549764
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,4096,2560,0.02325013279914856
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,4096,2048,0.021834667523701987
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,4096,2048,0.03309440016746521
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,4096,3072,0.024449066321055094
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,4096,1536,0.017691733439763387
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,4096,1536,0.02895786762237549
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,4096,1536,0.017951999107996622
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,4096,1024,0.013004799683888754
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,4096,1024,0.02485760052998861
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,4096,2048,0.020304000377655028
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,4096,768,0.010030933221181234
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,4096,768,0.024130133787790935
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,4096,1024,0.01734293301900228
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,4096,512,0.007940266529719036
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,4096,512,0.02078826626141866
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,4096,768,0.015314132968584696
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,4096,256,0.0063296000162760425
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,4096,256,0.018685867389043175
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,4096,256,0.015340800086657206
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,4096,128,0.00551146666208903
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,4096,512,0.015284267067909241
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,4096,128,0.017805866400400796
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,4096,64,0.005093333125114441
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,4096,64,0.017385600010553996
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,4096,32,0.005154133339722951
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,4096,128,0.015331199765205384
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,4096,32,0.017373865842819212
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,3584,65536,0.4885845184326172
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,3584,65536,0.6263573328653972
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,3584,51200,0.37233492533365886
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,3584,51200,0.4717983881632487
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,4096,51200,0.21298133532206215
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,3584,16384,0.12380906740824382
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,3584,16384,0.14403947194417316
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,3584,16384,0.07768959999084472
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,3584,12288,0.09370986620585123
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,3584,12288,0.1085653305053711
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,4096,65536,0.2658261299133301
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,3584,10240,0.07838186422983805
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,3584,10240,0.09391146500905355
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,3584,12288,0.06246293385823568
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,3584,8192,0.0637941320737203
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,3584,8192,0.07702826658884684
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,3584,10240,0.05629760026931763
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,3584,7168,0.05596053202946981
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,3584,7168,0.07083413600921631
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,3584,8192,0.046861867109934494
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,3584,6144,0.0487008015314738
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,3584,6144,0.05947519938151041
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,3584,7168,0.04386773506800334
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,3584,5120,0.04132159948348999
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,3584,5120,0.05202133258183798
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,3584,6144,0.03885973294576009
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,3584,4096,0.03376746575037638
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,3584,4096,0.04525546630223592
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,3584,5120,0.03552533388137817
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,3584,3584,0.03025919993718465
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,3584,3584,0.04374826749165853
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,3584,4096,0.031033599376678468
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,3584,3072,0.02638079921404521
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,3584,3072,0.038957866032918294
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,3584,3584,0.02849493424097697
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,3584,2560,0.023051732778549196
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,3584,2560,0.03521066506703695
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,3584,3072,0.026238934199015303
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,3584,2048,0.019550933440526327
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,3584,2048,0.030861866474151612
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,3584,2560,0.0251093327999115
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,3584,1536,0.016030933459599814
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,3584,1536,0.028952533006668092
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,3584,2048,0.02275200088818868
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,3584,1024,0.011683199803034465
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,3584,1024,0.026029866933822633
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,3584,51200,0.21442133585611978
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,3584,1024,0.020148267348607383
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,3584,768,0.009727999567985535
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,3584,768,0.023782400290171306
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,3584,512,0.008030933141708375
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,3584,512,0.021597866217295328
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,3584,1536,0.02061333258946737
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,3584,256,0.0066997334361076355
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,3584,256,0.01958400011062622
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,3584,256,0.018320000171661376
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,3584,128,0.005502933263778686
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,3584,512,0.01864853302637736
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,3584,128,0.01857173244158427
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,3584,768,0.01884160041809082
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,3584,64,0.004695466657479604
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,3584,32,0.005067733426888784
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,3584,64,0.016614400347073875
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,3584,128,0.018590933084487914
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,3584,32,0.017708800236384072
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,3072,65536,0.4138784090677897
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,3072,65536,0.5681280136108399
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,3072,51200,0.327782408396403
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,3072,51200,0.44626560211181643
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,3072,16384,0.10812479654947918
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,3072,16384,0.13481067021687826
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,3072,16384,0.07653013070424398
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,3072,12288,0.08178346951802572
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,3072,12288,0.10106773376464843
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,3584,65536,0.2685173352559408
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,3072,10240,0.06975786685943604
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,3072,10240,0.08880320390065512
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,3072,10240,0.05475946664810181
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,3072,8192,0.0564736008644104
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,3072,8192,0.07243733406066895
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,3072,12288,0.061622401078542076
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,3072,7168,0.048964265982309976
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,3072,7168,0.06430613199869792
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,3072,7168,0.04249279896418254
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,3072,6144,0.04211306571960449
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,3072,6144,0.05450239976247152
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,3072,8192,0.0457258661588033
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,3072,5120,0.03604053258895874
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,3072,5120,0.049420801798502605
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,3072,5120,0.03546666701634725
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,3072,4096,0.029726932446161907
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,3072,4096,0.04206186532974243
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,3072,6144,0.03784000078837077
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,3072,3584,0.026106667518615723
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,3072,3584,0.04030400117238363
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,3072,4096,0.030806400378545123
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,3072,3072,0.023575466871261597
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,3072,3072,0.03707520167032878
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,3072,3584,0.02765546639760335
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,3072,2560,0.019949867328008016
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,3072,2560,0.033662935098012284
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,3072,3072,0.026107732454935712
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,3072,2048,0.017965867122014364
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,3072,2048,0.030247465769449873
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,3072,2560,0.02480319937070211
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,3072,1536,0.013722667098045349
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,3072,1536,0.0294869323571523
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,3072,2048,0.022285866737365722
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,3072,1024,0.010472533106803895
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,3072,1024,0.023940267165501912
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,3072,1536,0.020308266083399452
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,3072,768,0.00890773336092631
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,3072,768,0.022722133000691733
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,3072,1024,0.019847466548283895
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,3072,512,0.006984533369541168
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,3072,768,0.018152532974878947
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,3072,512,0.021415466070175172
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,3072,256,0.0054517333706219995
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,3072,256,0.01789120038350423
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,3072,256,0.017808000246683754
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,3072,128,0.004716800153255462
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,3072,512,0.018152532974878947
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,3072,128,0.016925867398579916
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,3072,64,0.004220800101757049
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,3072,64,0.016846932967503867
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,3072,32,0.004628266890843709
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,3072,128,0.018179200092951455
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,3072,32,0.017450666427612303
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,2560,65536,0.37586774826049807
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,2560,65536,0.5472362518310547
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,2560,51200,0.2946122805277506
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,2560,51200,0.4291263898213704
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,3072,51200,0.21432639757792155
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,2560,16384,0.09449066321055094
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,2560,16384,0.13164160251617432
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,2560,16384,0.07599360148111979
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,2560,12288,0.07160533269246419
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,2560,12288,0.098471466700236
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,3072,65536,0.2705984115600586
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,2560,10240,0.06390186548233032
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,2560,10240,0.0844544013341268
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,2560,12288,0.0606058677037557
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,2560,8192,0.049686400095621745
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,2560,8192,0.06718506813049316
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,2560,10240,0.0542794664700826
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,2560,7168,0.04439146518707275
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,2560,7168,0.060836267471313474
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,2560,8192,0.04434773524602254
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,2560,6144,0.03746133248011271
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,2560,6144,0.05147733290990194
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,2560,7168,0.04125226736068725
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,2560,5120,0.03224746584892273
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,2560,5120,0.047253334522247316
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,2560,6144,0.03701119820276896
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,2560,4096,0.026151466369628906
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,2560,4096,0.0398357351620992
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,2560,5120,0.03382506767908732
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,2560,3584,0.023800534009933472
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,2560,3584,0.03911999861399333
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,2560,4096,0.030136533578236896
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,2560,3072,0.020733867088953653
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,2560,3072,0.03572160005569458
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,2560,3584,0.026858667532602947
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,2560,2560,0.017937066157658894
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,2560,3072,0.024727465709050496
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,2560,2560,0.032204800844192506
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,2560,2048,0.014984533190727234
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,2560,2048,0.0290175994237264
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,2560,51200,0.21262720425923667
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,2560,1536,0.011923199892044068
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,2560,1536,0.02686079939206441
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,2560,2048,0.020787199338277183
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,2560,1024,0.009181867043177288
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,2560,2560,0.0235807995001475
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,2560,1024,0.024596265951792397
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,2560,768,0.007833600044250488
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,2560,1536,0.018921599785486857
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,2560,768,0.02174826661745707
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,2560,512,0.0063509335120519
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,2560,512,0.02099840044975281
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,2560,768,0.016578132907549538
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,2560,256,0.005062399804592133
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,2560,512,0.016717867056528727
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,2560,256,0.01806400020917257
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,2560,1024,0.01814613342285156
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,2560,128,0.004241066674391428
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,2560,128,0.01789439916610718
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,2560,64,0.003905066599448522
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,2560,64,0.01655359963575999
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,2560,128,0.016551466782887776
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,2560,256,0.01632533371448517
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,2560,32,0.016823466618855795
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,2048,65536,0.32229013442993165
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,2048,65536,0.5183808008829753
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,2048,51200,0.2543615976969401
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,2560,32,0.004212266703446706
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,2048,51200,0.408356253306071
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,2048,16384,0.08399893442789713
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,2048,16384,0.12638933658599855
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,2048,16384,0.05396906534830729
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,2560,65536,0.26617600123087565
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,2048,12288,0.06387946605682374
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,2048,12288,0.09225920041402182
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,2048,10240,0.05434240102767944
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,2048,10240,0.07718506654103598
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,2048,10240,0.0382805347442627
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,2048,8192,0.043304534753163655
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,2048,8192,0.06048426628112793
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,2048,12288,0.04244053363800049
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,2048,7168,0.03705493211746216
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,2048,7168,0.05425920089085897
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,2048,7168,0.029361067215601604
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,2048,6144,0.03131733338038127
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,2048,6144,0.04725013176600139
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,2048,8192,0.031378134091695147
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,2048,5120,0.026237867275873822
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,2048,5120,0.043297068277994795
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,2048,5120,0.025164800882339477
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,2048,4096,0.0213536004225413
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,2048,4096,0.037089065710703535
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,2048,6144,0.026816000541051228
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,2048,3584,0.019051732619603474
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,2048,3584,0.03615146478017171
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,2048,3584,0.020294400056203206
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,2048,3072,0.01695786714553833
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,2048,3072,0.03266026576360066
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,2048,4096,0.02190933426221212
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,2048,2560,0.014971733093261719
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,2048,2560,0.029127466678619384
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,2048,2560,0.01776960094769796
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,2048,2048,0.01248426636060079
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,2048,3072,0.01853333314259847
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,2048,2048,0.026705066363016765
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,2048,1536,0.010227200388908387
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,2048,1536,0.025036799907684325
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,2048,1536,0.014149333039919535
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,2048,1024,0.007650133470694225
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,2048,1024,0.022990934054056802
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,2048,2048,0.01585599978764852
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,2048,768,0.006693333387374878
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,2048,768,0.020949333906173706
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,2048,1024,0.013656533757845559
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,2048,512,0.0055754666527112326
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,2048,512,0.01936639944712321
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,2048,768,0.012430933117866517
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,2048,256,0.004681600133577982
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,2048,256,0.01679253379503886
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,2048,256,0.012072533369064331
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,2048,128,0.003788800040880839
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,2048,128,0.016846932967503867
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,2048,128,0.01244586706161499
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,2048,64,0.003432533393303553
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,2048,64,0.015247999628384908
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,2048,32,0.003828266759713491
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,2048,32,0.015396266182263692
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,1536,65536,0.2335103988647461
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,2048,512,0.01251520017782847
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,1536,65536,0.48268267313639324
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,1536,51200,0.18319679896036783
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,2048,51200,0.1378719965616862
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,1536,51200,0.3821717262268066
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,1536,16384,0.06018773317337036
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,1536,16384,0.11285119851430256
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,1536,16384,0.05467306772867838
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,1536,12288,0.04582826693852742
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,1536,12288,0.08419840335845948
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,2048,65536,0.17079572677612304
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,1536,10240,0.03864533503850301
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,1536,10240,0.06954027016957601
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,1536,12288,0.04312533140182495
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,1536,8192,0.030958932638168336
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,1536,8192,0.054731734593709315
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,1536,10240,0.03895359834035238
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,1536,7168,0.027530666192372637
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,1536,7168,0.05045333305994669
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,1536,8192,0.032568534215291337
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,1536,6144,0.023959465821584067
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,1536,6144,0.04385066827138265
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,1536,6144,0.027508266766866046
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,1536,5120,0.020537600914637247
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,1536,5120,0.04079893430074056
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,1536,7168,0.03006933331489563
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,1536,4096,0.01702079971631368
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,1536,4096,0.035554134845733644
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,1536,5120,0.02600746750831604
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,1536,3584,0.015522133310635886
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,1536,3584,0.03364799817403157
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,1536,4096,0.023165865739186605
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,1536,3072,0.013748266299565635
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,1536,3072,0.03107306758562724
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,1536,3584,0.021514666080474854
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,1536,2560,0.012948266665140786
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,1536,2560,0.028614399830500285
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,1536,3072,0.0202293336391449
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,1536,2048,0.010043733318646749
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,1536,2048,0.025700267155965167
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,1536,2560,0.019423999389012656
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,1536,1536,0.008875733613967896
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,1536,1536,0.024715733528137208
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,1536,51200,0.1382143974304199
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,1536,1536,0.015687466661135355
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,1536,1024,0.006687999765078227
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,1536,1024,0.020850133895874024
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,1536,2048,0.017679999272028603
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,1536,768,0.0064298664530118305
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,1536,768,0.02090346614519755
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,1536,1024,0.0150218665599823
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,1536,512,0.005072000126043955
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,1536,512,0.018782933553059898
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,1536,768,0.01409173309803009
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,1536,256,0.00415040006240209
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,1536,256,0.017663999398549398
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,1536,512,0.013633066415786743
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,1536,128,0.004273066421349844
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,1536,256,0.014141866564750671
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,1536,128,0.016993065675099693
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,1536,64,0.00382080003619194
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,1536,64,0.016239999731381734
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,1536,32,0.004233600199222564
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,1536,128,0.01448853313922882
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,1536,32,0.01641386648019155
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,1024,65536,0.18063999811808268
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,1024,65536,0.4413440068562825
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,1024,51200,0.13314560254414876
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,1024,51200,0.3498218536376953
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,1024,16384,0.04628159999847412
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,1024,16384,0.1055402676264445
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,1024,16384,0.05215040047963461
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,1024,12288,0.03490453163782756
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,1024,12288,0.07556053002675375
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,1536,65536,0.17088534037272135
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,1024,10240,0.029275733232498168
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,1024,10240,0.060022401809692386
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,1024,12288,0.03956693410873413
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,1024,8192,0.023324799537658692
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,1024,8192,0.04927359819412232
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,1024,10240,0.03547733227411906
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,1024,7168,0.021401600042978922
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,1024,7168,0.045740799109141035
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,1024,8192,0.02945706645647685
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,1024,6144,0.01773973306020101
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,1024,6144,0.040591998895009355
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,1024,7168,0.028117332855860395
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,1024,5120,0.015336533387502035
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,1024,5120,0.037003731727600096
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,1024,6144,0.025242666403452556
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,1024,5120,0.023933867613474526
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,1024,4096,0.014365866780281067
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,1024,4096,0.03174613316853841
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,1024,3584,0.01169706682364146
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,1024,3584,0.031085866689682006
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,1024,3584,0.01904426614443461
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,1024,3072,0.010439466436703999
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,1024,3072,0.0295306662718455
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,1024,4096,0.021053866545359293
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,1024,2560,0.009720533092816671
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,1024,2560,0.02686506708463033
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,1024,2560,0.016587733229001363
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,1024,2048,0.008397866288820903
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,1024,3072,0.017762132485707603
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,1024,2048,0.0243776003519694
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,1024,1536,0.007136000196139018
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,1024,1536,0.023491199811299643
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,1024,1536,0.013418666521708169
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,1024,1024,0.005464533468087515
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,1024,2048,0.015155200163523355
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,1024,1024,0.02062293291091919
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,1024,768,0.004987733562787374
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,1024,768,0.01946559945742289
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,1024,1024,0.012847999731699625
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,1024,512,0.004267733295758565
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,1024,768,0.01162453293800354
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,1024,512,0.018576000134150186
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,1024,256,0.003979733337958654
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,1024,256,0.016495999693870545
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,1024,512,0.011662933230400085
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,1024,128,0.003432533393303553
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,1024,51200,0.13514026006062824
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,1024,128,0.014590932925542196
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,1024,64,0.002996266633272171
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,1024,64,0.01463573376337687
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,1024,128,0.012116266290346782
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,1024,32,0.00342399999499321
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,1024,32,0.01572480003039042
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,768,65536,0.13709012667338055
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,768,65536,0.4252757390340169
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,768,51200,0.10712959766387939
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,1024,256,0.011611732840538024
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,768,51200,0.3382730801900228
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,768,16384,0.03682560125986735
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,768,16384,0.10484693050384522
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,768,16384,0.053656534353892005
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,768,12288,0.026676267385482788
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,768,12288,0.07306986649831136
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,1024,65536,0.16833492914835613
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,768,10240,0.021669334173202513
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,768,10240,0.06167893409729004
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,768,12288,0.0410314679145813
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,768,8192,0.01816640098889669
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,768,8192,0.04744746685028076
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,768,10240,0.03622506856918335
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,768,7168,0.016080000003178916
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,768,7168,0.04338560104370117
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,768,7168,0.02895786762237549
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,768,6144,0.014145066340764364
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,768,6144,0.038152531782786055
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,768,8192,0.030905600388844805
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,768,5120,0.012078932921091716
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,768,5120,0.03516586621602376
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,768,5120,0.02480640014012655
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,768,6144,0.026401066780090333
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,768,4096,0.010005333026250203
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,768,4096,0.03105173309644063
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,768,3584,0.009307733178138733
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,768,3584,0.030561065673828124
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,768,4096,0.022338134050369263
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,768,3072,0.008364799618721008
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,768,3072,0.02874133388201396
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,768,51200,0.13578559557596842
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,768,2560,0.007503999769687653
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,768,2560,0.026761599381764728
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,768,3584,0.020707199970881142
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,768,2048,0.006506666541099548
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,768,2048,0.024936532974243163
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,768,3072,0.01898026665051778
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,768,2560,0.0181386669476827
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,768,1536,0.0059008002281188965
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,768,1536,0.02273599902788798
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,768,1024,0.005883733431498209
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,768,1024,0.020639999707539877
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,768,1536,0.01483519971370697
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,768,768,0.005064533154169718
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,768,1024,0.014108799894650779
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,768,768,0.020127999782562255
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,768,512,0.004264533519744873
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,768,512,0.018947199980417887
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,768,2048,0.01654293338457743
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,768,768,0.012931199868520102
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,768,256,0.003806933263937632
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,768,256,0.016515200336774193
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,768,512,0.012968533237775168
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,768,128,0.003433600068092346
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,768,128,0.01713599960009257
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,768,64,0.0030741333961486817
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,768,128,0.013249066472053529
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,768,256,0.012839466333389282
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,768,32,0.003067733347415924
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,768,64,0.014683733383814493
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,512,65536,0.09467306931813559
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,768,32,0.015646933515866598
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,512,51200,0.07465706666310629
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,512,65536,0.40518614451090496
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,512,51200,0.32260799407958984
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,512,16384,0.03068160017331441
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,512,16384,0.09706453482309976
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,768,65536,0.16884907086690265
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,512,12288,0.02243946592013041
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,512,12288,0.06567039887110392
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,512,16384,0.051209600766499844
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,512,10240,0.019092265764872232
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,512,10240,0.053540265560150145
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,512,12288,0.03752426703770955
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,512,8192,0.01663040022055308
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,512,10240,0.03420480092366536
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,512,8192,0.044123733043670656
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,512,7168,0.01394773324330648
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,512,7168,0.04318293333053589
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,512,7168,0.027221333980560303
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,512,6144,0.012061867117881774
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,512,6144,0.037299199899037676
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,512,8192,0.02891626755396525
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,512,5120,0.010904533664385478
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,512,5120,0.03518720070521037
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,512,6144,0.025062400102615356
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,512,4096,0.010956799983978272
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,512,4096,0.030554666121800737
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,512,4096,0.020682666699091592
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,512,5120,0.023181867599487305
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,512,3584,0.01035093367099762
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,512,3584,0.03095039923985799
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,512,3072,0.008881066242853801
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,512,3072,0.02777600089708964
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,512,3584,0.01880533297856649
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,512,2560,0.007754666606585185
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,512,2560,0.026423466205596925
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,512,51200,0.13469972610473632
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,512,2048,0.0062496001521746315
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,512,2048,0.023050665855407715
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,512,3072,0.01738986571629842
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,512,2560,0.01654293338457743
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,512,1536,0.005548800031344095
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,512,1536,0.023401600122451783
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,512,1024,0.004640000065167745
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,512,1024,0.02032426595687866
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,512,2048,0.014934399724006652
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,512,1536,0.013246933619181315
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,512,768,0.004287999868392944
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,512,1024,0.0125109334786733
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,512,512,0.0038005332152048744
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,512,768,0.020107734203338622
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,512,512,0.017961599429448447
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,512,768,0.011614933609962463
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,512,256,0.0030239999294281008
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,512,512,0.011185066898663838
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,512,256,0.016845866044362386
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,512,128,0.0026602665583292644
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,512,128,0.016819200913111367
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,512,256,0.01127893328666687
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,512,64,0.002624000112215678
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,512,128,0.011645866433779399
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,512,64,0.014986667037010192
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,512,32,0.002621866762638092
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,512,32,0.015641599893569946
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,256,65536,0.0599616010983785
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,256,65536,0.3867232004801432
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,256,51200,0.047884798049926756
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,256,51200,0.30441385904947915
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,256,16384,0.01769066651662191
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,256,16384,0.09435306390126547
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,512,65536,0.1671829382578532
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,256,12288,0.01660480002562205
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,256,12288,0.061205331484476724
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,256,16384,0.049831465880076094
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,256,10240,0.02036906679471334
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,256,10240,0.049890132745107015
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,256,12288,0.0371018648147583
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,256,8192,0.011961600184440613
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,256,8192,0.04183040062586467
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,256,10240,0.03374933401743571
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,256,7168,0.01046399970849355
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,256,7168,0.03912746508916219
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,256,8192,0.028455466032028198
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,256,6144,0.009634133179982502
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,256,6144,0.03495039939880371
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,256,7168,0.027293866872787474
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,256,5120,0.008343467116355896
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,256,5120,0.03309333324432373
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,256,6144,0.024758400519688924
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,256,4096,0.0071050668756167095
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,256,4096,0.028756266832351683
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,256,5120,0.023110399643580117
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,256,3584,0.006818133095900218
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,256,65536,0.16728639602661133
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,256,3584,0.02872106631596883
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,256,3072,0.006414933502674103
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,256,3072,0.026895999908447266
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,256,4096,0.020406399170557657
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,256,2560,0.0054730668663978575
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,256,2560,0.02462293306986491
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,256,51200,0.13330559730529784
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,256,2048,0.004702933132648468
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,256,3584,0.018769067525863648
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,256,2048,0.02218986749649048
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,256,1536,0.0042357335488001505
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,256,1536,0.022567466894785563
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,256,2560,0.01611199975013733
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,256,1024,0.003442133218050003
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,256,2048,0.014923733472824097
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,256,1024,0.019222400585810342
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,256,768,0.0034005333979924522
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,256,3072,0.017352533340454102
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,256,1536,0.012875733772913614
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,256,768,0.018669867515563966
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,256,512,0.0030154667794704436
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,256,1024,0.012457600235939026
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,256,512,0.016591999928156534
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,256,256,0.002664533257484436
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,256,256,0.01670080025990804
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,256,512,0.011612799763679505
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,256,128,0.0026186667382717133
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,256,256,0.011205333471298217
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,256,768,0.0116565336783727
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,256,128,0.01447466711203257
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,256,64,0.002363733450571696
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,256,32,0.0025973332424958544
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,256,128,0.01112320025761922
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,256,64,0.014800000190734863
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,128,65536,0.050170667966206874
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,128,51200,0.03993279933929443
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,128,65536,0.3802741368611654
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,128,51200,0.3002410570780436
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,128,16384,0.0130431999762853
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,128,16384,0.08828799724578858
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,256,32,0.015610667069753012
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,128,12288,0.012054399649302164
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,128,12288,0.05890453259150187
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,128,16384,0.04601706663767497
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,128,10240,0.010823466380437215
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,128,10240,0.04732053279876709
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,128,12288,0.03629546562830607
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,128,8192,0.00959999958674113
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,128,8192,0.041317331790924075
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,128,10240,0.03331413269042969
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,128,7168,0.01002239982287089
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,128,7168,0.038711468378702804
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,128,8192,0.028014934062957762
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,128,6144,0.009153067072232565
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,128,6144,0.035079467296600345
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,128,7168,0.026345600684483845
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,128,5120,0.008019199967384339
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,128,6144,0.02393280069033305
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,128,51200,0.13069547017415364
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,128,5120,0.032814933856328325
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,128,4096,0.006729599833488464
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,128,4096,0.02908160090446472
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,128,5120,0.02214720050493876
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,128,3584,0.006371200084686279
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,128,4096,0.019516799847284952
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,128,3584,0.018222934007644652
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,128,3072,0.005886933207511902
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,128,3072,0.02667199969291687
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,128,3072,0.016569599509239197
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,128,2560,0.005161599814891815
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,128,2560,0.025144533316294355
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,128,3584,0.02768426736195882
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,128,2048,0.004683733483155568
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,128,65536,0.1628565311431885
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,128,2048,0.02109439969062805
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,128,1536,0.003910399973392487
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,128,2560,0.015850667158762613
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,128,1536,0.021893332401911415
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,128,1024,0.003437866767247518
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,128,1536,0.012059733271598816
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,128,1024,0.018707199891408285
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,128,768,0.003005866706371307
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,128,2048,0.013860266407330832
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,128,768,0.018999467293421425
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,128,1024,0.011618133385976155
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,128,768,0.010435199737548828
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,128,512,0.018565332889556883
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,128,256,0.00257493332028389
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,128,512,0.010423466563224792
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,128,256,0.01576746702194214
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,128,128,0.0023775999744733173
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,128,256,0.010362666845321656
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,128,128,0.014546133081118264
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,128,64,0.002164266755183538
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2049,128,128,0.010692266623179118
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,128,32,0.0021845333278179167
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,128,64,0.014459733168284097
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,64,65536,0.04729919830958049
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,128,32,0.014158933361371358
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,64,51200,0.03869119882583618
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,64,16384,0.011128532886505126
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,64,16384,0.08742506504058838
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,64,51200,0.2986581484476725
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,64,12288,0.01160533328851064
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,64,10240,0.010327466328938802
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,64,12288,0.05731840133666992
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,64,8192,0.008753066261609394
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,64,10240,0.04732799927393595
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,64,7168,0.008313600222269695
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,64,8192,0.041006934642791745
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,128,512,0.003092266619205475
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,64,7168,0.03783040046691895
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,64,5120,0.007916800181070964
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,64,5120,0.0312394658724467
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,64,6144,0.035019731521606444
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,64,4096,0.006697600086530049
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,64,3584,0.006247466802597046
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,64,4096,0.028994133075078327
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,64,3072,0.005568000177542368
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,64,3072,0.026866134007771807
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,64,2560,0.0051594664653142296
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,64,65536,0.37955945332845054
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,64,2048,0.004583466549714406
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,64,2560,0.024753065903981526
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,64,1536,0.003962666789690653
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,64,1536,0.02148800094922384
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,64,6144,0.007909333209196727
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,64,1024,0.0034154665966828666
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,64,768,0.003018666555484136
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,64,1024,0.020523732900619505
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,64,512,0.0029781334102153777
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,64,768,0.019223467508951823
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,64,512,0.018040533860524496
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,64,256,0.0025600001215934755
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,64,128,0.0022367998957633973
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,64,256,0.015999999642372132
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,64,64,0.002259200066328049
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,64,128,0.015546666582425437
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,64,32,0.002328533430894216
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,64,64,0.014636799693107605
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,32,65536,0.04444266557693481
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,64,32,0.015579733252525329
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,32,51200,0.03588159879048665
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,64,2048,0.02262506683667501
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,32,65536,0.3769866625467936
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,32,16384,0.009630933403968811
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,32,51200,0.29800427754720055
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,32,12288,0.008328533172607422
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,32,10240,0.007541333138942718
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,32,16384,0.0897877295811971
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,32,8192,0.011638399958610535
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,32,10240,0.04934399922688802
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,32,12288,0.05903893311818441
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,32,7168,0.010839466253916423
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,32,8192,0.042029865582784015
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,32,6144,0.009830400347709656
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,32,5120,0.009960533181826273
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,32,7168,0.03950506846110026
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,32,6144,0.03561280171076457
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,32,4096,0.008642133076985676
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,64,3584,0.027237333854039508
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,32,5120,0.03402560154596965
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,32,4096,0.02879040042559306
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,32,3072,0.006930133203665416
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,32,3584,0.028241066137949626
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,32,3072,0.026395734151204425
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,32,2560,0.00625600020090739
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,32,2048,0.004376533130804697
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,32,2048,0.02244266668955485
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,32,1536,0.003916800022125244
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,32,1536,0.020997333526611327
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,32,1024,0.0032778667906920114
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,32,1024,0.019820799430211387
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,32,768,0.0029877332349618276
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,32,768,0.018761599063873292
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,32,512,0.0026880001028378804
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,32,512,0.01800000071525574
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,32,256,0.0026261332134405774
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,32,2560,0.024035199483235677
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,32,256,0.01546986699104309
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,32,128,0.0022528000175952912
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,32,64,0.002136533210674922
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,32,128,0.01602026621500651
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,32,64,0.014630400141080222
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,32,32,0.0022815999885400133
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2049,32,32,0.014641066392262777
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2049,32,3584,0.007737599809964498
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,65536,16384,2.503062438964844
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,65536,51200,4.292271931966146
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,65536,16384,1.2843648274739583
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,65536,12288,1.8560682932535808
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,65536,51200,7.879148864746094
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,65536,12288,0.9634239832560221
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,65536,10240,1.4492596944173177
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,65536,10240,0.8142101287841796
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,65536,10240,0.47661759058634445
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,65536,8192,1.1015093485514322
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,65536,8192,0.6366207758585612
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,65536,12288,0.5356821060180664
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,65536,7168,0.9524181365966797
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,65536,7168,0.5521194458007812
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,65536,16384,0.6961589177449544
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,65536,6144,0.7892981211344401
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,65536,6144,0.4610176086425781
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,65536,8192,0.39383573532104493
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,65536,5120,0.6810368220011394
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,65536,5120,0.42875626881917317
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,65536,7168,0.3634112040201823
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,65536,4096,0.5355146408081055
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,65536,4096,0.3422762552897135
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,65536,6144,0.30896854400634766
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,65536,3584,0.4654975891113281
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,65536,3584,0.27693332036336266
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,65536,5120,0.2873525301615397
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,65536,3072,0.4124874750773112
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,65536,3072,0.2549802621205648
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,65536,4096,0.2578975995381673
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,65536,2560,0.3438709259033203
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,65536,2560,0.20485332806905113
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,65536,3584,0.22710186640421548
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,65536,2048,0.2739797274271647
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,65536,2048,0.1672202746073405
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,65536,2560,0.19518292744954427
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,65536,1536,0.21280213991800942
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,65536,1536,0.1357653299967448
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,65536,3072,0.2069983959197998
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,65536,1024,0.14810345967610677
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,65536,1024,0.1018229325612386
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,65536,1024,0.13876266479492189
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,65536,768,0.11609280109405518
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,65536,768,0.08710506757100424
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,65536,1536,0.14554559389750163
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,65536,512,0.08432426452636718
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,65536,512,0.07517973581949869
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,65536,512,0.1312490701675415
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,65536,256,0.05760426521301269
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,65536,256,0.06259626547495524
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,65536,768,0.13114773432413737
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,65536,128,0.04708906809488932
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,65536,128,0.05788266658782959
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,65536,2048,0.17195520401000977
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,65536,64,0.04383786519368489
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,65536,64,0.056858666737874354
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,65536,32,0.04355839888254802
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,65536,32,0.05769493182500204
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,65536,128,0.13162346680959064
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,65536,256,0.1309290647506714
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,51200,65536,4.219557444254557
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,51200,65536,7.576384989420573
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,51200,51200,3.283280944824219
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,51200,51200,5.97748057047526
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,51200,16384,1.7991007486979167
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,51200,16384,0.9738933563232421
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,51200,16384,0.5717311859130859
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,51200,12288,1.3073407491048177
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,51200,12288,0.744712511698405
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,51200,12288,0.4350005467732747
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,51200,10240,1.1423936208089194
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,51200,10240,0.6232970555623372
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,51200,10240,0.3867402712504069
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,51200,8192,0.9223946889241537
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,51200,8192,0.503110408782959
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,65536,51200,2.1589088439941406
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,51200,7168,0.7654698689778645
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,51200,7168,0.4349098523457845
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,51200,8192,0.3199850718180338
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,51200,6144,0.659332275390625
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,51200,6144,0.38565759658813475
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,51200,7168,0.29316479365030923
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,51200,5120,0.5520256042480469
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,51200,5120,0.32357333501180013
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,51200,5120,0.23066239356994628
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,51200,4096,0.438266658782959
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,51200,4096,0.25887786547342934
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,51200,6144,0.25360746383666993
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,51200,3584,0.38755734761555993
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,51200,3584,0.23296106656392418
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,51200,3584,0.18603946367899576
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,51200,3072,0.3336703936258952
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,51200,3072,0.20850772857666017
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,51200,4096,0.20749546686808268
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,51200,2560,0.2797130584716797
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,51200,2560,0.17211732864379883
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,51200,51200,1.748430887858073
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,51200,2048,0.22046507199605309
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,51200,2048,0.14639147122701007
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,51200,2560,0.15559040705362956
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,51200,1536,0.1731701374053955
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,51200,1536,0.11856106917063396
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,51200,3072,0.16688639322916668
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,51200,1024,0.12858880360921224
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,51200,1024,0.08508266607920328
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,51200,1536,0.11681173642476399
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,51200,768,0.09898773034413656
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,51200,768,0.08596693674723307
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,51200,1024,0.11403520107269287
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,51200,512,0.06756800015767415
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,51200,512,0.07806399663289389
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,51200,2048,0.13909866015116373
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,51200,256,0.05955520073572794
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,51200,256,0.06662613153457642
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,51200,256,0.10525226593017578
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,51200,128,0.04119893312454224
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,51200,128,0.050629333655039466
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,51200,128,0.10653973420461019
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,51200,64,0.03592960039774577
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,51200,64,0.051475199063618984
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,51200,32,0.03814719915390015
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,51200,32,0.05150186618169149
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,16384,65536,2.5642166137695312
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,51200,512,0.10635733604431152
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,16384,65536,1.5779871622721353
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,16384,51200,1.8972501118977863
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,16384,51200,1.2298314412434896
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,51200,768,0.10555093288421631
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,16384,16384,0.607267189025879
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,16384,16384,0.3964490572611491
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,16384,16384,0.20253440539042153
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,16384,12288,0.4588768005371094
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,16384,12288,0.29635305404663087
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,16384,12288,0.1554207960764567
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,16384,10240,0.3559274673461914
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,16384,10240,0.23074560165405272
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,16384,10240,0.1392800013224284
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,16384,8192,0.284282652537028
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,16384,8192,0.1940597375233968
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,16384,8192,0.11407039960225422
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,16384,7168,0.24824959437052407
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,16384,7168,0.16933973630269367
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,16384,51200,0.5819658915201823
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,16384,6144,0.20520853996276855
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,16384,6144,0.14409173329671224
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,16384,7168,0.10572266578674316
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,16384,5120,0.17553280194600424
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,16384,5120,0.1216490666071574
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,16384,5120,0.0840714693069458
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,16384,4096,0.14056639671325682
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,16384,4096,0.09811093012491862
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,16384,6144,0.09245759646097819
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,16384,3584,0.12101973692576092
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,16384,3584,0.09191892941792806
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,16384,3584,0.06630400021870932
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,16384,3072,0.10432000160217285
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,16384,3072,0.08061013221740723
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,16384,4096,0.07402773698170981
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,16384,2560,0.08944746653238932
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,16384,2560,0.06910293102264405
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,16384,2560,0.053998935222625735
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,16384,2048,0.0725045363108317
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,16384,2048,0.058039466540018715
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,16384,3072,0.0577951987584432
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,16384,1536,0.05500160058339437
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,16384,1536,0.052324267228444424
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,16384,1536,0.04075733423233032
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,16384,1024,0.03868800004323324
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,16384,1024,0.04130133390426636
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,16384,2048,0.046521600087483725
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,16384,768,0.03057493368784587
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,16384,768,0.036524800459543864
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,16384,1024,0.039961600303649904
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,16384,512,0.02218559980392456
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,16384,512,0.034653866291046144
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,16384,768,0.03438933293024699
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,16384,256,0.016473600268363954
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,16384,256,0.02791573405265808
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,16384,256,0.03334506750106812
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,16384,128,0.015269333124160766
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,16384,512,0.033905065059661864
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,16384,128,0.029090134302775066
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,16384,64,0.016940800348917644
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,16384,64,0.029726932446161907
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,16384,32,0.017940266927083334
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,16384,32,0.03006400068600972
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,16384,128,0.034264532725016277
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,12288,65536,1.8559605916341144
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,12288,65536,1.24923947652181
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,12288,51200,1.4081429799397787
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,12288,51200,0.9953919728597006
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,16384,65536,0.7423210779825846
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,12288,16384,0.45253868103027345
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,12288,16384,0.32262614568074544
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,51200,65536,2.2068639119466145
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,12288,12288,0.3398538589477539
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,12288,12288,0.23505172729492188
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,12288,12288,0.13401172955830892
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,12288,10240,0.2899519920349121
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,12288,10240,0.19282986323038737
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,12288,16384,0.1750698725382487
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,12288,8192,0.23344213167826333
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,12288,8192,0.16575039227803548
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,12288,10240,0.11978240013122558
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,12288,7168,0.20072959264119467
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,12288,7168,0.13859626452128093
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,12288,8192,0.09940799872080484
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,12288,6144,0.1721717357635498
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,12288,6144,0.1209877332051595
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,12288,7168,0.09235093593597413
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,12288,5120,0.1391541322072347
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,12288,5120,0.10129280090332031
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,12288,6144,0.08082773685455322
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,12288,4096,0.1127669334411621
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,12288,4096,0.08305706977844238
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,12288,5120,0.07291946411132813
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,12288,3584,0.09643733501434326
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,12288,3584,0.07839146455128988
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,12288,4096,0.06347946723302206
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,12288,3072,0.08526079654693604
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,12288,3072,0.07019946575164795
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,12288,3072,0.05074666738510132
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,12288,2560,0.07059733072916666
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,12288,2560,0.06019733349482218
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,12288,3584,0.056637867291768396
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,12288,2048,0.07141013145446777
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,12288,2048,0.051089068253835045
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,12288,2048,0.04448000192642212
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,12288,1536,0.0435477336247762
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,12288,1536,0.04487359921137492
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,12288,2560,0.049319465955098465
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,12288,1024,0.03070080081621806
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,12288,1024,0.03619199991226196
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,12288,1024,0.040202665328979495
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,12288,768,0.02395520011583964
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,12288,768,0.03298133412996928
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,12288,1536,0.0406826655069987
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,12288,512,0.017800533771514894
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,12288,512,0.02991360028584798
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,12288,768,0.0362335999806722
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,12288,256,0.01341546674569448
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,12288,256,0.024893865982691447
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,12288,512,0.036183468500773114
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,12288,128,0.011963733037312825
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,12288,128,0.02302186687787374
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,12288,256,0.03601173162460327
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,12288,64,0.009960533181826273
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,12288,64,0.02318506638209025
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,12288,32,0.010401067137718201
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,12288,32,0.024513065814971924
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,12288,128,0.03597013155619304
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,10240,65536,1.0666154861450194
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,10240,65536,1.5194410959879556
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,10240,51200,1.1858570098876953
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,10240,51200,0.8514901479085287
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,12288,51200,0.5044416109720866
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,10240,16384,0.3478432019551595
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,10240,16384,0.2604128042856852
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,10240,16384,0.15131947199503581
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,10240,12288,0.2627072016398112
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,10240,12288,0.20021546681722008
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,12288,65536,0.6397162755330403
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,10240,10240,0.24554880460103354
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,10240,10240,0.1722080071767171
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,10240,12288,0.1174997329711914
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,10240,8192,0.19531413714090984
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,10240,8192,0.13757440249125164
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,10240,10240,0.10538986523946126
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,10240,7168,0.17262187004089355
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,10240,7168,0.125055996576945
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,10240,8192,0.08736639817555746
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,10240,6144,0.1429909388224284
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,10240,6144,0.1069386641184489
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,10240,6144,0.07024319966634116
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,10240,7168,0.08172159989674886
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,10240,5120,0.1178986628850301
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,10240,5120,0.09495573043823242
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,10240,4096,0.09581973552703857
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,10240,4096,0.07756906350453695
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,10240,4096,0.055166931947072353
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,10240,3584,0.08423146406809488
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,10240,3584,0.07377280394236246
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,10240,5120,0.0640992005666097
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,10240,3072,0.07175467014312745
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,10240,3072,0.06552746693293253
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,10240,3584,0.050025598208109534
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,10240,2560,0.06149013439814249
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,10240,2560,0.05659093459447225
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,10240,3072,0.045915734767913816
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,10240,2048,0.05010773340861002
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,10240,51200,0.43520320256551104
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,10240,2048,0.04750293493270874
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,10240,1536,0.03863253196080525
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,10240,1536,0.04267093340555827
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,10240,1536,0.03712426821390788
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,10240,1024,0.02703253428141276
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,10240,1024,0.03511999845504761
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,10240,2048,0.04133439858754476
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,10240,768,0.022026666005452476
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,10240,768,0.031112533807754517
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,10240,2560,0.044411734739939375
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,10240,512,0.016516266266504924
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,10240,512,0.027083732684453327
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,10240,1024,0.03627093235651652
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,10240,256,0.012025599678357441
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,10240,256,0.022906666994094847
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,10240,256,0.03373119831085205
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,10240,128,0.00983893374602
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,10240,128,0.021574399868647256
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,10240,768,0.033402665456136064
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,10240,512,0.03333333333333333
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,10240,64,0.009226666887601216
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,10240,32,0.009538132945696514
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,10240,64,0.0228383998076121
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,10240,32,0.021588265895843506
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,10240,128,0.033690667152404784
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,8192,65536,0.9570773442586263
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,8192,65536,1.23668581644694
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,8192,51200,0.7661216100056966
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,8192,51200,0.9474005381266275
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,8192,16384,0.32208852767944335
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,8192,16384,0.24752853711446127
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,8192,16384,0.12695893446604412
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,8192,12288,0.22424747149149576
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,8192,12288,0.17898027102152508
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,10240,65536,0.547381337483724
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,8192,10240,0.19286293983459474
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,8192,10240,0.14772799809773762
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,8192,12288,0.09848960240681967
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,8192,8192,0.1583072026570638
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,8192,8192,0.12171413103739422
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,8192,10240,0.0879701296488444
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,8192,7168,0.13276373545328776
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,8192,7168,0.10992106596628826
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,8192,8192,0.07348053455352783
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,8192,6144,0.1161461353302002
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,8192,6144,0.09518400033315023
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,8192,7168,0.06802240212758383
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,8192,5120,0.09377493063608805
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,8192,5120,0.08469226360321044
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,8192,5120,0.05264533360799154
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,8192,4096,0.0757205327351888
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,8192,4096,0.06495466629664103
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,8192,6144,0.05922133525212606
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,8192,3584,0.0676522652308146
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,8192,3584,0.060268799463907875
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,8192,3584,0.04124159812927246
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,8192,3072,0.05873493353525797
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,8192,3072,0.0531765341758728
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,8192,4096,0.04639039834340413
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,8192,2560,0.048877867062886556
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,8192,2560,0.046325333913167316
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,8192,3072,0.03674986759821574
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,8192,2048,0.038809601465861
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,8192,2048,0.04073066711425781
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,8192,2560,0.03557759920756022
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,8192,1536,0.029229867458343505
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,8192,1536,0.03751039902369181
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,8192,1536,0.026792534192403156
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,8192,1024,0.021203200022379555
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,8192,1024,0.03269439935684204
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,8192,2048,0.031065599123636885
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,8192,768,0.017876267433166504
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,8192,768,0.02922559976577759
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,8192,1024,0.026101332902908326
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,8192,512,0.01446399986743927
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,8192,512,0.026850134134292603
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,8192,768,0.0233024001121521
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,8192,256,0.010955733060836793
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,8192,256,0.02392639915148417
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,8192,512,0.023084799448649086
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,8192,128,0.008950400352478027
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,8192,128,0.023115734259287514
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,8192,128,0.02307413419087728
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,8192,64,0.00951573352018992
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,8192,64,0.020793600877126058
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,8192,32,0.009594666957855224
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,8192,256,0.02291733423868815
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,8192,32,0.022665599981943764
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,7168,65536,0.9503456115722656
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,7168,65536,0.8550432205200196
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,7168,51200,0.7543071746826172
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,7168,51200,0.7137280146280924
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,8192,51200,0.3608448028564453
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,7168,16384,0.2428426742553711
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,7168,16384,0.20549227396647135
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,7168,16384,0.10436480045318604
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,7168,12288,0.18337066968282062
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,7168,12288,0.15145920117696127
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,8192,65536,0.45340693791707354
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,7168,10240,0.15075519879659016
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,7168,10240,0.12604906558990478
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,7168,12288,0.08173973560333252
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,7168,8192,0.12489173412322999
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,7168,8192,0.10278933048248291
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,7168,10240,0.07351360321044922
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,7168,7168,0.10629013379414876
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,7168,7168,0.09515946706136068
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,7168,7168,0.056935465335845946
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,7168,6144,0.09022933642069499
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,7168,6144,0.08175786336263022
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,7168,8192,0.06179733276367187
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,7168,5120,0.07667946815490723
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,7168,5120,0.07250560124715169
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,7168,5120,0.044557865460713705
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,7168,4096,0.06179626782735189
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,7168,4096,0.05876906712849935
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,7168,6144,0.049882666269938154
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,7168,3584,0.05364586512247721
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,7168,3584,0.055225598812103274
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,7168,3584,0.035545599460601804
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,7168,3072,0.046136534214019774
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,7168,3072,0.04974826574325562
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,7168,4096,0.03918933471043905
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,7168,2560,0.039545599619547525
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,7168,2560,0.04274666706720988
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,7168,2560,0.03078719973564148
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,7168,2048,0.032177066802978514
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,7168,2048,0.03709333340326945
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,7168,3072,0.032636799414952594
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,7168,1536,0.024846933285395303
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,7168,1536,0.03410346508026123
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,7168,1536,0.025175466140111284
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,7168,1024,0.01794346570968628
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,7168,1024,0.02866133252779643
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,7168,2048,0.02805333336194356
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,7168,768,0.014233600099881491
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,7168,768,0.026681600014368693
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,7168,768,0.026281599203745527
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,7168,512,0.010662399729092916
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,7168,512,0.023337600628534953
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,7168,1024,0.024753065903981526
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,7168,256,0.008001066744327545
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,7168,256,0.021283199389775596
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,7168,256,0.02616853316624959
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,7168,128,0.0071733335653940845
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,7168,128,0.01904746691385905
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,7168,512,0.02632746696472168
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,7168,64,0.006379733482996623
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,7168,64,0.018639999628067016
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,7168,32,0.006613333523273468
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,7168,32,0.019768534104029338
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,7168,128,0.025948800643285114
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,6144,65536,0.7988597234090169
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,6144,65536,0.9093290964762369
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,6144,51200,0.7006687800089518
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,6144,51200,0.6294911702473958
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,7168,51200,0.2897738774617513
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,6144,16384,0.2245024045308431
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,6144,16384,0.19810454050699872
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,6144,16384,0.10366079807281495
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,6144,12288,0.17102400461832684
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,6144,12288,0.14728213946024576
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,7168,65536,0.36221866607666015
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,6144,10240,0.14241706530253093
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,6144,10240,0.12372586727142335
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,6144,12288,0.08105279604593912
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,6144,8192,0.11645973523457845
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,6144,8192,0.09860479831695557
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,6144,10240,0.07354346911112467
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,6144,7168,0.0996789296468099
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,6144,7168,0.09069440364837647
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,6144,8192,0.06052693525950113
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,6144,6144,0.0908245325088501
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,6144,6144,0.07884053389231363
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,6144,7168,0.05588693221410116
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,6144,5120,0.07124693393707275
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,6144,5120,0.0701311985651652
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,6144,6144,0.04981546799341838
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,6144,4096,0.057684266567230226
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,6144,4096,0.05696426630020142
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,6144,5120,0.04459093411763509
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,6144,3584,0.04975786606470744
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,6144,3584,0.05242986679077148
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,6144,3584,0.03548479874928792
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,6144,3072,0.04291093349456787
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,6144,3072,0.04719039996465047
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,6144,4096,0.039018666744232176
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,6144,2560,0.03742186625798543
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,6144,2560,0.042173866430918375
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,6144,2560,0.030374399820963544
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,6144,2048,0.03031253417332967
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,6144,2048,0.0356661319732666
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,6144,3072,0.031921066840489704
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,6144,1536,0.023602133989334105
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,6144,1536,0.03462186654408773
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,6144,1536,0.024737066030502318
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,6144,1024,0.016810667514801026
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,6144,1024,0.029317333300908403
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,6144,2048,0.02725013295809428
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,6144,768,0.01404906709988912
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,6144,768,0.02686079939206441
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,6144,768,0.025967999299367266
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,6144,512,0.011422933141390482
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,6144,512,0.02486506700515747
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,6144,1024,0.02437439958254496
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,6144,256,0.009211732943852743
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,6144,256,0.022782933712005616
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,6144,51200,0.28764588038126626
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,6144,128,0.009139200051625569
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,6144,128,0.02108586629231771
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,6144,512,0.025998934110005693
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,6144,64,0.008456533153851826
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,6144,64,0.02062826752662659
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,6144,32,0.00874133308728536
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,6144,128,0.024925865729649863
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,6144,32,0.021542400121688843
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,6144,256,0.025675733884175617
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,5120,65536,0.7809066772460938
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,5120,65536,0.7556757609049479
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,5120,51200,0.6465973536173503
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,5120,51200,0.5925973256429036
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,5120,16384,0.1987306594848633
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,5120,16384,0.19032106399536133
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,5120,16384,0.10292692979176839
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,5120,12288,0.1511029402414958
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,5120,12288,0.13717439969380696
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,6144,65536,0.3582207997639974
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,5120,10240,0.11520533561706543
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,5120,10240,0.10708800156911213
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,5120,12288,0.08123626708984374
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,5120,8192,0.09437759717305502
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,5120,8192,0.09005333582560221
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,5120,10240,0.07310933272043864
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,5120,7168,0.07941653728485107
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,5120,7168,0.0823466698328654
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,5120,8192,0.05999786853790283
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,5120,6144,0.06920959949493408
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,5120,6144,0.07056746482849122
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,5120,7168,0.05594346523284912
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,5120,5120,0.0565717339515686
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,5120,5120,0.06124693155288696
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,5120,6144,0.048920532067616776
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,5120,4096,0.046945067246754964
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,5120,4096,0.04995306730270386
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,5120,5120,0.04442133506139119
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,5120,3584,0.040357331434885665
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,5120,3584,0.04753066698710124
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,5120,3584,0.03583466609319051
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,5120,3072,0.03502293427785237
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,5120,3072,0.04307946761449178
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,5120,4096,0.03929813305536906
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,5120,2560,0.030478932460149127
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,5120,2560,0.038629333178202316
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,5120,2560,0.031098665793736775
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,5120,2048,0.02505279978116353
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,5120,2048,0.03347520033518474
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,5120,3072,0.03282240033149719
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,5120,1536,0.019399466117223103
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,5120,1536,0.031038933992385866
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,5120,1536,0.02617173393567403
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,5120,1024,0.013892267147699991
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,5120,1024,0.02685760060946147
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,5120,2048,0.02850773334503174
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,5120,768,0.011406933267911274
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,5120,768,0.02422399918238322
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,5120,1024,0.025282132625579833
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,5120,768,0.02309760053952535
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,5120,512,0.021237333615620933
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,5120,512,0.008374399940172831
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,5120,256,0.006670933465162914
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,5120,256,0.019809067249298096
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,5120,512,0.023151999711990355
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,5120,128,0.005926399926344554
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,5120,128,0.01817173361778259
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,5120,256,0.02315946618715922
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,5120,64,0.005479466418425242
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,5120,128,0.023062400023142495
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,5120,32,0.005481599768002828
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,5120,64,0.017340799172719322
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,5120,32,0.01806079943974813
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,4096,65536,0.5701653162638347
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,4096,65536,0.6635989507039388
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,4096,51200,0.4435146649678548
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,4096,51200,0.5058538754781087
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,5120,51200,0.28457492192586265
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,4096,16384,0.14549652735392254
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,4096,16384,0.15623572667439778
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,4096,16384,0.07654613653818766
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,4096,12288,0.11006826559702557
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,4096,12288,0.11771413485209148
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,5120,65536,0.3589440027872721
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,4096,10240,0.09517013231913249
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,4096,10240,0.1013482650121053
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,4096,12288,0.06052159865697225
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,4096,8192,0.07720639705657958
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,4096,8192,0.08338987032572429
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,4096,10240,0.0550005316734314
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,4096,7168,0.06758080323537191
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,4096,7168,0.07663573424021403
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,4096,8192,0.04528319835662842
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,4096,6144,0.05780693292617798
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,4096,6144,0.06432533264160156
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,4096,7168,0.042878933747609455
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,4096,5120,0.04891626834869385
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,4096,5120,0.05715946753819784
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,4096,6144,0.03726293245951335
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,4096,4096,0.03952320019404094
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,4096,4096,0.04670933485031128
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,4096,5120,0.03433813254038493
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,4096,3584,0.0352127989133199
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,4096,3584,0.04565973281860351
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,4096,4096,0.030076799790064494
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,4096,3072,0.03066026568412781
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,4096,3072,0.04061546723047892
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,4096,3584,0.026893866062164307
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,4096,2560,0.02637760043144226
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,4096,2560,0.0380949338277181
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,4096,3072,0.024091732501983643
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,4096,2048,0.021989333629608154
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,4096,2048,0.031658667325973514
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,4096,2560,0.022797866662343343
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,4096,1536,0.01787733236948649
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,4096,1536,0.029025065898895263
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,4096,51200,0.2131242593129476
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,4096,1024,0.012929067015647888
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,4096,1024,0.025014400482177734
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,4096,2048,0.020259199539820354
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,4096,768,0.009961600104967754
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,4096,768,0.02333973248799642
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,4096,768,0.01530026694138845
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,4096,512,0.007941333452860515
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,4096,1024,0.017280000448226928
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,4096,512,0.020760534207026164
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,4096,256,0.006307200094064077
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,4096,256,0.01946880022684733
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,4096,256,0.015051733454068503
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,4096,128,0.005942399799823761
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,4096,128,0.018706132968266807
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,4096,512,0.01493333379427592
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,4096,64,0.0060149331887563075
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,4096,64,0.019934932390848793
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,4096,32,0.006265600025653839
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,4096,128,0.015340800086657206
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,4096,1536,0.01795733372370402
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,3584,65536,0.5431765238444011
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,3584,65536,0.63012269337972
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,3584,51200,0.42182292938232424
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,4096,32,0.018051199118296304
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,3584,51200,0.4935488065083821
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,3584,16384,0.13817493120829266
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,3584,16384,0.1520799954732259
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,3584,16384,0.07804053624471029
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,3584,12288,0.10461333592732747
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,3584,12288,0.11532586415608723
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,4096,65536,0.2669610659281413
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,3584,10240,0.09695359865824381
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,3584,10240,0.10108373165130616
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,3584,10240,0.056012801329294835
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,3584,8192,0.0779263973236084
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,3584,8192,0.08356160322825114
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,3584,12288,0.06257386604944865
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,3584,7168,0.06696000099182128
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,3584,7168,0.07581226825714112
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,3584,7168,0.04415253400802612
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,3584,6144,0.05762773354848226
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,3584,6144,0.0638869325319926
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,3584,8192,0.04713279803593953
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,3584,5120,0.04905920028686524
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,3584,5120,0.05559999942779541
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,3584,5120,0.035537068049112955
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,3584,4096,0.03980799913406372
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,3584,6144,0.03864320119222005
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,3584,4096,0.0469322681427002
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,3584,3584,0.03562133312225342
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,3584,3584,0.0446997324625651
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,3584,3584,0.028536534309387206
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,3584,3072,0.030931200583775836
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,3584,3072,0.03996479908625285
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,3584,4096,0.031830400228500366
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,3584,2560,0.026391466458638508
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,3584,2560,0.0371018648147583
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,3584,3072,0.02654186685880025
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,3584,2048,0.021516799926757812
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,3584,2048,0.032656000057856245
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,3584,2560,0.025035732984542848
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,3584,1536,0.01726079980532328
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,3584,1536,0.029090134302775066
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,3584,2048,0.023042132457097374
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,3584,1024,0.012959999839464822
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,3584,1024,0.02490239938100179
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,3584,1536,0.02111146648724874
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,3584,768,0.010055466492970785
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,3584,768,0.023423999547958374
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,3584,1024,0.020012799898783365
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,3584,512,0.007961600025494894
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,3584,512,0.020875734090805054
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,3584,512,0.018389334281285606
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,3584,256,0.005924266576766968
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,3584,256,0.018626133600870766
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,3584,768,0.018633600076039633
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,3584,128,0.005524266759554545
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,3584,128,0.018554667631785073
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,3584,256,0.018361600240071614
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,3584,64,0.004679466784000397
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,3584,64,0.01722666621208191
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,3584,32,0.005026133358478546
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,3584,32,0.01811093290646871
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,3584,128,0.018554667631785073
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,3072,65536,0.43476158777872725
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,3072,65536,0.5723125457763671
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,3072,51200,0.33658132553100584
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,3072,51200,0.4559157371520996
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,3584,51200,0.21368533770243325
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,3072,16384,0.10997227032979329
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,3072,16384,0.13951999346415203
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,3072,16384,0.07649813493092855
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,3072,12288,0.0835914691289266
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,3072,12288,0.10514346758524577
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,3584,65536,0.26972160339355467
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,3072,10240,0.07343146800994874
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,3072,10240,0.08708053429921468
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,3072,12288,0.061256531874338785
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,3072,8192,0.05726186831792196
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,3072,8192,0.07205013434092203
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,3072,10240,0.054885331789652506
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,3072,7168,0.05053226550420126
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,3072,7168,0.06408106486002604
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,3072,8192,0.045716265837351486
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,3072,6144,0.042344534397125246
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,3072,6144,0.05499840180079142
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,3072,7168,0.0425162672996521
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,3072,5120,0.03596373399098714
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,3072,5120,0.048342398802439374
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,3072,5120,0.03431253433227539
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,3072,4096,0.029373866319656373
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,3072,4096,0.04174613157908122
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,3072,6144,0.03760533332824707
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,3072,3584,0.025707733631134034
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,3072,3584,0.0402890682220459
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,3072,3584,0.02799466649691264
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,3072,3072,0.022784000635147093
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,3072,3072,0.03693973223368327
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,3072,4096,0.03025493423144023
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,3072,2560,0.019782400131225585
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,3072,2560,0.03238079945246379
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,3072,2560,0.024392533302307128
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,3072,2048,0.016424533724784852
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,3072,2048,0.02887679934501648
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,3072,3072,0.025608533620834352
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,3072,1536,0.013187199831008911
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,3072,1536,0.02751680016517639
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,3072,1536,0.020198400815327963
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,3072,1024,0.009610666831334432
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,3072,1024,0.024700800577799477
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,3072,2048,0.02257386644681295
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,3072,768,0.007998933394749958
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,3072,768,0.022580265998840332
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,3072,1024,0.01986453334490458
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,3072,512,0.0067104001839955645
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,3072,768,0.01816426714261373
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,3072,512,0.01816533406575521
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,3072,256,0.006252799928188324
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,3072,256,0.01853653391202291
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,3072,512,0.021171200275421142
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,3072,128,0.005610666672388713
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,3072,128,0.01950613260269165
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,3072,256,0.017762132485707603
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,3072,64,0.005446400245030721
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,3072,128,0.018106667200724284
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,3072,32,0.0055178667108217875
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,3072,64,0.017513600985209148
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,3072,32,0.018147200345993042
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,2560,65536,0.37827841440836585
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,2560,65536,0.5441525141398112
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,2560,51200,0.29191786448160806
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,2560,51200,0.4344650586446126
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,3072,51200,0.2127936045328776
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,2560,16384,0.09765120347340903
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,2560,16384,0.13421546618143718
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,2560,16384,0.07628373305002847
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,2560,12288,0.07469653288523356
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,2560,12288,0.09947946866353354
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,3072,65536,0.26891520818074544
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,2560,10240,0.06310720046361287
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,2560,10240,0.0817909320195516
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,2560,10240,0.05347733497619629
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,2560,8192,0.04982613325119019
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,2560,8192,0.06590826511383056
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,2560,12288,0.05994666814804077
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,2560,7168,0.04191466569900513
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,2560,7168,0.059139200051625575
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,2560,7168,0.04150079886118571
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,2560,6144,0.03619733254114787
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,2560,6144,0.05074666738510132
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,2560,8192,0.04450773398081462
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,2560,5120,0.030744532744089764
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,2560,5120,0.04541226625442505
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,2560,5120,0.03376426696777344
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,2560,4096,0.025229867299397784
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,2560,4096,0.03954133192698161
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,2560,6144,0.036866132418314615
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,2560,3584,0.022129066785176597
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,2560,3584,0.03824746608734131
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,2560,3584,0.02700693408648173
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,2560,3072,0.019307732582092285
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,2560,3072,0.034279465675354004
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,2560,4096,0.02985493342081706
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,2560,2560,0.016834133863449098
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,2560,2560,0.03086400032043457
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,2560,3072,0.02445759971936544
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,2560,2048,0.014152533809343972
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,2560,2560,0.023125332593917847
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,2560,2048,0.027717334032058717
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,2560,1536,0.011195733149846395
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,2560,1536,0.027038933833440144
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,2560,2048,0.020848000049591066
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,2560,1024,0.008372267087300617
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,2560,1024,0.023219199975331624
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,2560,1536,0.01864853302637736
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,2560,768,0.00730560024579366
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,2560,768,0.022420267264048256
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,2560,1024,0.01817173361778259
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,2560,512,0.006296533346176148
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,2560,512,0.02061226765314738
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,2560,51200,0.2116447925567627
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,2560,768,0.016752000649770102
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,2560,256,0.005374933282534281
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,2560,256,0.01867626706759135
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,2560,128,0.00469760000705719
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,2560,128,0.017131733894348144
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,2560,256,0.016526933511098227
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,2560,64,0.004491733511288961
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,2560,512,0.01655359963575999
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,2560,64,0.016667733589808144
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,2560,32,0.0046293333172798155
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,2560,128,0.016973867019017538
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,2048,65536,0.29820054372151694
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,2048,65536,0.5017290751139323
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,2048,51200,0.2288053353627523
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,2048,51200,0.3936629295349121
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,2560,32,0.016291200121243795
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,2048,16384,0.0750986655553182
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,2048,16384,0.12288959821065266
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,2048,16384,0.05364053249359131
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,2048,12288,0.05766079823176066
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,2048,12288,0.0913045326868693
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,2560,65536,0.2655445257822672
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,2048,10240,0.048699732621510824
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,2048,10240,0.07630720138549804
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,2048,12288,0.04251199960708618
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,2048,8192,0.03936320145924886
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,2048,8192,0.06108266512552897
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,2048,10240,0.03755946556727092
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,2048,7168,0.03481813271840413
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,2048,7168,0.05438719987869263
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,2048,8192,0.0312991996606191
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,2048,6144,0.03020159999529521
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,2048,6144,0.048342398802439374
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,2048,7168,0.029283199707667035
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,2048,5120,0.02581973274548848
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,2048,5120,0.04327359994252523
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,2048,6144,0.026408533255259197
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,2048,4096,0.021602133909861244
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,2048,4096,0.03789120117823283
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,2048,5120,0.023995733261108397
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,2048,3584,0.01953386664390564
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,2048,3584,0.03501226504643758
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,2048,3584,0.01988160014152527
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,2048,3072,0.017100799083709716
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,2048,3072,0.032789333661397295
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,2048,4096,0.021891200542449953
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,2048,2560,0.015187199910481772
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,2048,2560,0.030315732955932616
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,2048,2560,0.017410133282343546
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,2048,2048,0.012865066528320312
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,2048,2048,0.026732800404230754
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,2048,3072,0.018220800161361694
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,2048,1536,0.010067199667294819
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,2048,1536,0.0248799999554952
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,2048,2048,0.016148266196250916
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,2048,1024,0.008025600016117096
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,2048,1024,0.02259733279546102
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,2048,1536,0.013742933670679728
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,2048,768,0.006742399930953979
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,2048,1024,0.013264000415802002
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,2048,768,0.021115734179814657
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,2048,512,0.0055167997876803074
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,2048,512,0.020078933238983153
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,2048,512,0.012414933244387309
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,2048,768,0.012030933300654094
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,2048,256,0.004699733356634776
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,2048,256,0.017372800906499227
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,2048,128,0.004750933249791463
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,2048,128,0.016572800278663636
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,2048,256,0.011707733074824016
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,2048,64,0.00384853333234787
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,2048,128,0.012023466825485229
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,2048,64,0.01649493376413981
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,2048,32,0.004194133480389913
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,2048,32,0.01602026621500651
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,1536,65536,0.24060479799906412
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,1536,65536,0.4793279965718587
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,1536,51200,0.18584319750467937
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,1536,51200,0.3734410603841146
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,2048,51200,0.13704106012980144
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,1536,16384,0.06309760014216105
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,1536,16384,0.11637012958526612
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,1536,16384,0.054598399003346766
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,1536,12288,0.04814720153808594
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,1536,12288,0.08557120164235434
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,2048,65536,0.16946239471435548
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,1536,10240,0.040888532002766924
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,1536,10240,0.07004160086313883
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,1536,12288,0.04301333427429199
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,1536,8192,0.03237760066986084
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,1536,8192,0.05540800094604492
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,1536,10240,0.0380181352297465
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,1536,7168,0.028675200541814168
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,1536,7168,0.05020266771316528
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,1536,8192,0.03177493413289388
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,1536,6144,0.02486506700515747
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,1536,6144,0.044392534097035724
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,1536,7168,0.030220800638198854
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,1536,5120,0.021600000063578286
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,1536,5120,0.04007360140482585
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,1536,6144,0.027457066377003986
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,1536,4096,0.01783999999364217
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,1536,4096,0.0349727988243103
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,1536,5120,0.025917865832646686
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,1536,3584,0.015986133615175882
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,1536,3584,0.033641600608825685
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,1536,4096,0.023130667209625245
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,1536,3072,0.013773866494496665
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,1536,3072,0.03105386694272359
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,1536,3584,0.0214901328086853
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,1536,2560,0.011628799637158712
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,1536,2560,0.027771733204523724
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,1536,2560,0.019003732999165853
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,1536,2048,0.010182399551073711
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,1536,2048,0.024938666820526124
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,1536,3072,0.02027626633644104
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,1536,1536,0.008416000008583068
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,1536,1536,0.025677865743637084
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,1536,2048,0.01734293301900228
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,1536,1024,0.006740266581376393
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,1536,1024,0.02132693330446879
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,1536,1536,0.015667200088500977
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,1536,768,0.005967999994754791
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,1536,768,0.020977065960566203
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,1536,1024,0.014919466773668923
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,1536,512,0.0050911997755368555
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,1536,512,0.01872319976488749
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,1536,768,0.013682132959365845
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,1536,256,0.004263466596603394
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,1536,256,0.016964266697565712
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,1536,512,0.01390506625175476
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,1536,128,0.0039061332742373147
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,1536,128,0.01588159998257955
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,1536,128,0.014521599809328715
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,1536,64,0.0034122665723164878
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,1536,256,0.013799466689427695
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,1536,64,0.014612266421318054
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,1536,32,0.003487999985615412
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,1536,32,0.015595733126004537
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,1024,65536,0.16207359631856283
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,1536,51200,0.13771093686421712
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,1024,51200,0.12748586336771647
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,1024,65536,0.4447914759318034
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,1024,51200,0.3500394821166992
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,1024,16384,0.04289386669794719
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,1024,16384,0.10470399856567383
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,1024,16384,0.052222931385040285
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,1024,12288,0.032177066802978514
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,1024,12288,0.07474773724873861
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,1536,65536,0.1698410669962565
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,1024,10240,0.026199465990066527
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,1024,10240,0.05955520073572794
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,1024,12288,0.04001493453979492
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,1024,8192,0.021504000822703043
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,1024,8192,0.04938773314158122
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,1024,10240,0.03526826699574788
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,1024,7168,0.019462400674819948
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,1024,7168,0.045160531997680664
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,1024,7168,0.02770666678746541
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,1024,6144,0.017017600933710735
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,1024,8192,0.028973867495854694
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,1024,6144,0.03938239812850952
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,1024,5120,0.014670933286348978
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,1024,5120,0.036397866408030194
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,1024,5120,0.02313813368479411
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,1024,4096,0.012124799688657125
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,1024,4096,0.03121386567751567
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,1024,6144,0.025220266977945965
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,1024,3584,0.011368532975514729
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,1024,3584,0.030930133660634358
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,1024,3584,0.019040000438690186
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,1024,3072,0.01002346674601237
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,1024,4096,0.0206112007300059
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,1024,3072,0.028996266921361286
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,1024,2560,0.008775466680526733
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,1024,2560,0.027011199792226152
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,1024,2560,0.016618667046229045
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,1024,2048,0.007570133109887441
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,1024,3072,0.017794134219487508
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,1024,2048,0.024628265698750814
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,1024,1536,0.006363733112812043
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,1024,1536,0.022832000255584718
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,1024,51200,0.13466240564982096
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,1024,1024,0.005177600185076395
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,1024,1024,0.02069013317426046
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,1024,1536,0.013403733571370443
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,1024,768,0.004661333560943603
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,1024,2048,0.014913066228230795
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,1024,768,0.0202890674273173
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,1024,512,0.004082133372624716
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,1024,512,0.018718934059143065
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,1024,1024,0.012444800138473511
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,1024,256,0.003421866645415624
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,1024,256,0.016364799936612447
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,1024,768,0.011587199568748475
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,1024,128,0.0030720000465710956
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,1024,512,0.011609599987665812
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,1024,256,0.011296000083287556
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,1024,128,0.015219199657440185
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,1024,64,0.0029834667841593427
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,1024,32,0.002994133283694585
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,1024,64,0.014613333344459533
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,1024,128,0.012035199999809265
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,1024,32,0.015275733669598899
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,768,65536,0.14637333552042645
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,768,51200,0.1151039997736613
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,768,65536,0.41500053405761717
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,768,51200,0.327567990620931
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,768,16384,0.03938666582107544
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,768,16384,0.10330560207366943
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,768,16384,0.05289813280105591
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,768,12288,0.029653332630793255
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,768,12288,0.07312746842702231
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,1024,65536,0.1677888075510661
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,768,10240,0.02540160020192464
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,768,10240,0.05849599838256836
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,768,12288,0.039868799845377605
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,768,8192,0.020951465765635172
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,768,8192,0.04730133215586345
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,768,10240,0.036508798599243164
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,768,7168,0.01885226567586263
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,768,7168,0.04434773524602254
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,768,8192,0.030425600210825604
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,768,6144,0.01660266617933909
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,768,7168,0.02887360056241353
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,768,6144,0.039082666238149
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,768,5120,0.01395626664161682
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,768,5120,0.03542186816533406
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,768,5120,0.02433919906616211
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,768,4096,0.011453866958618164
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,768,4096,0.032789333661397295
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,768,6144,0.02640639940897624
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,768,3584,0.010441600282986959
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,768,3584,0.03105386694272359
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,768,4096,0.021860265731811525
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,768,3072,0.009342933694521587
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,768,3072,0.028778666257858278
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,768,3072,0.0190175990263621
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,768,2560,0.008355200290679932
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,768,2560,0.026841600735982258
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,768,3584,0.020351999998092653
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,768,2048,0.007536000013351441
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,768,2048,0.02387946645418803
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,768,2560,0.017771732807159425
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,768,1536,0.006289066871007283
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,768,1536,0.022835199038187662
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,768,51200,0.13550400733947754
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,768,1024,0.005474133292833964
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,768,1024,0.021112533410390218
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,768,1536,0.014465066790580749
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,768,768,0.00468800018231074
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,768,768,0.018835200866063436
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,768,2048,0.016539733608563742
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,768,512,0.004247466723124186
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,768,1024,0.013771733641624451
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,768,768,0.01286186675230662
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,768,512,0.01861226757367452
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,768,256,0.003769599894682566
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,768,256,0.016785067319869996
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,768,512,0.012850133577982583
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,768,128,0.0033962666988372804
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,768,256,0.013091199596722922
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,768,128,0.015617066621780395
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,768,64,0.0030421334008375804
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,768,64,0.014619732896486918
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,768,128,0.013314132889111837
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,768,32,0.0032768001159032187
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,768,32,0.014648532867431641
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,512,65536,0.09113279978434244
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,512,65536,0.40619198481241864
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,512,51200,0.07212159633636475
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,512,51200,0.3232991854349772
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,512,16384,0.030754133065541582
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,512,16384,0.0982421318689982
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,768,65536,0.16853866577148438
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,512,16384,0.05123733282089234
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,512,12288,0.022264534235000612
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,512,12288,0.0661631981531779
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,512,10240,0.018889600038528444
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,512,10240,0.053107198079427084
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,512,12288,0.03790613412857056
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,512,8192,0.016657066345214844
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,512,8192,0.04437013467152913
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,512,10240,0.03407786687215169
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,512,7168,0.013899733622868856
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,512,7168,0.04121066729227702
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,512,8192,0.029066665967305498
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,512,6144,0.012294399738311767
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,512,6144,0.037061333656311035
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,512,7168,0.027245867252349853
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,512,5120,0.010472533106803895
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,512,5120,0.03426666657129924
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,512,6144,0.024371200799942018
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,512,4096,0.008941866954167684
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,512,4096,0.028845866521199543
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,512,4096,0.0206389327843984
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,512,3584,0.008354133367538452
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,512,5120,0.02312320073445638
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,512,3584,0.02883946696917216
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,512,3072,0.007523199915885926
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,512,3072,0.026779733101526898
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,512,51200,0.13368639945983887
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,512,2560,0.006797866523265838
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,512,2560,0.024959999322891235
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,512,3584,0.019030400117238364
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,512,2048,0.005890133480230967
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,512,3072,0.017762132485707603
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,512,2048,0.022829866409301756
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,512,1536,0.005494399865468343
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,512,1536,0.02272426684697469
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,512,2560,0.016557866334915163
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,512,1024,0.004279466470082601
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,512,2048,0.014860799908638
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,512,1024,0.0200053334236145
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,512,768,0.0038719999293486277
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,512,768,0.019232000907262167
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,512,1024,0.012466133634249369
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,512,512,0.0033952000240484873
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,512,768,0.011596799890200297
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,512,1536,0.013108266393343606
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,512,256,0.003047466774781545
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,512,256,0.017349332571029663
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,512,512,0.011326932907104492
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,512,128,0.002616533388694127
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,512,128,0.01490239997704824
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,512,256,0.011220266421635944
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,512,64,0.0025898667673269907
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,512,128,0.011669333775838215
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,512,32,0.0026496000587940215
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,512,64,0.015095466375350952
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,512,32,0.014536533753077188
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,256,65536,0.062232534090677895
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,256,65536,0.38517227172851565
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,256,51200,0.047713065147399904
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,256,51200,0.3047573407491048
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,512,512,0.016664533813794454
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,256,16384,0.01715839902559916
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,256,16384,0.09402453104654948
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,512,65536,0.1664576053619385
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,256,12288,0.01605226695537567
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,256,12288,0.061080535252889
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,256,16384,0.048785066604614256
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,256,10240,0.014475733041763306
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,256,10240,0.04931306838989258
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,256,12288,0.03709866603215535
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,256,8192,0.0116266667842865
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,256,8192,0.04133760134379069
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,256,10240,0.03332053422927857
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,256,7168,0.01042133371035258
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,256,7168,0.039074134826660153
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,256,8192,0.028460800647735596
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,256,6144,0.00958079993724823
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,256,6144,0.03494506676991781
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,256,7168,0.027191466093063353
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,256,6144,0.02427840034166972
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,256,5120,0.008449066678682964
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,256,5120,0.03280213276545207
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,256,51200,0.13308693567911783
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,256,4096,0.007166933516661327
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,256,4096,0.029002666473388672
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,256,5120,0.022706133127212525
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,256,3584,0.006733866532643636
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,256,3584,0.027460267146428425
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,256,4096,0.020232532421747842
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,256,3072,0.006006399790445963
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,256,3584,0.018594133853912353
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,256,3072,0.027080533901850383
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,256,2560,0.005520000060399374
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,256,2560,0.02483946681022644
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,256,2048,0.004859733581542969
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,256,3072,0.01694933374722799
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,256,2048,0.022975999116897582
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,256,1536,0.004208000004291534
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,256,2560,0.016515200336774193
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,256,1536,0.02235520084698995
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,256,1024,0.0034527999659379324
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,256,2048,0.014522666732470194
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,256,1024,0.019924267133076986
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,256,768,0.0034111998975276947
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,256,1536,0.012854400277137756
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,256,768,0.018888533115386963
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,256,512,0.0029418667157491045
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,256,1024,0.01209386686484019
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,256,512,0.016643200318018594
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,256,768,0.011252267162005107
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,256,256,0.002584533393383026
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,256,256,0.016979199647903443
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,256,512,0.011196800072987874
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,256,128,0.0025770666698614756
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,256,256,0.011191466450691223
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,256,128,0.01455466647942861
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,256,64,0.0022261333962281544
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,256,128,0.011239467064539592
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,256,64,0.014455466469128927
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,128,65536,0.05258026520411173
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,128,65536,0.38054078420003257
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,256,65536,0.16643840471903484
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,128,51200,0.04038933515548706
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,256,32,0.0021749332547187803
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,256,32,0.014432000120480857
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,128,16384,0.011358933647473653
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,128,51200,0.30141334533691405
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,128,16384,0.08829120000203451
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,128,12288,0.01167039970556895
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,128,12288,0.05799573262532552
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,128,12288,0.036212265491485596
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,128,10240,0.010500267148017883
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,128,10240,0.048520533243815105
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,128,16384,0.047446401913960774
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,128,8192,0.01167466640472412
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,128,8192,0.04130239884058635
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,128,10240,0.03291306694348653
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,128,7168,0.010443733135859171
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,128,7168,0.03912426630655925
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,128,8192,0.02765226761500041
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,128,6144,0.00916373332341512
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,128,6144,0.03574506839116414
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,128,51200,0.12995093663533527
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,128,5120,0.007993599772453308
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,128,5120,0.032407466570536295
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,128,6144,0.023576533794403075
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,128,4096,0.0067104001839955645
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,128,4096,0.029177600145339967
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,128,5120,0.021895466248194377
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,128,3584,0.0063274666666984555
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,128,3584,0.027698133389155073
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,128,7168,0.026371200879414875
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,128,3072,0.005862399935722351
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,128,3072,0.026868265867233277
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,128,3584,0.017760000626246133
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,128,2560,0.0050687998533248905
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,128,4096,0.01953279972076416
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,128,3072,0.016590933005015053
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,128,2560,0.024170666933059692
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,128,2048,0.004610133171081543
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,128,2048,0.021682133277257286
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,128,1536,0.0038677332301934562
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,128,65536,0.16231999397277833
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,128,1536,0.020947200059890748
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,128,1024,0.0034282666941483817
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,128,1536,0.01206826666990916
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,128,2560,0.015398400028546652
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,128,768,0.003193599979082743
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,128,1024,0.011560533444086711
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,128,768,0.0187882661819458
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,128,512,0.0028522667785485585
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,128,768,0.010377599795659383
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,128,512,0.018016000588734947
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,128,256,0.0026549334327379864
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,128,512,0.010002133250236512
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,128,256,0.01572053333123525
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,128,256,0.01033066709836324
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,128,128,0.0023936000963052113
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,128,128,0.01456000010172526
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,128,2048,0.013697066903114319
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,2048,128,128,0.010436266660690308
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,128,64,0.002148266633351644
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,128,32,0.002194133400917053
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,128,1024,0.019040000438690186
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,128,64,0.014541866381963095
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,128,32,0.014428800344467163
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,64,65536,0.046453332901000975
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,64,51200,0.03885866800944011
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,64,16384,0.01393066644668579
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,64,16384,0.08712106545766195
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,64,51200,0.29913174311319984
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,64,65536,0.3799541473388672
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,64,10240,0.010314666231473287
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,64,12288,0.05584319829940796
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,64,8192,0.008806399504343669
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,64,10240,0.04739306767781575
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,64,8192,0.040889600912729904
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,64,7168,0.008328533172607422
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,64,6144,0.007773866752783458
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,64,7168,0.03794879913330078
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,64,5120,0.007117866476376851
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,64,6144,0.03502613306045532
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,64,5120,0.03178986708323161
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,64,4096,0.006850133339564006
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,64,3584,0.006302933394908905
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,64,4096,0.028233599662780762
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,64,3072,0.0054613331953684485
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,64,3584,0.028663466374079388
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,64,3072,0.02499306599299113
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,64,2560,0.0050794666012128195
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,64,2048,0.004622933268547058
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,64,2560,0.023819732666015624
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,64,1536,0.0038133333126703895
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,64,12288,0.01157866617043813
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,64,1024,0.0034175999462604523
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,64,2048,0.02202026645342509
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,64,1536,0.0208512008190155
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,64,768,0.0030559999247392017
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,64,1024,0.01999893387158712
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,64,512,0.0026410666604836782
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,64,768,0.018580265839894614
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,64,256,0.0025653332471847535
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,64,512,0.018054399887720743
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,64,128,0.002234666546185811
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,64,256,0.014934399724006652
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,64,64,0.0022026665508747103
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,64,128,0.015464533368746439
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,64,32,0.002197333425283432
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,64,64,0.015098667144775391
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,32,65536,0.04616959889729817
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,64,32,0.014435199896494546
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,32,51200,0.03746453523635864
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,32,16384,0.009573333462079366
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,32,12288,0.007820799946784973
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,32,65536,0.37707945505777996
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,32,51200,0.29845333099365234
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,32,10240,0.007187200089295705
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,32,16384,0.08850239912668864
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,32,8192,0.010578133662541707
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,32,12288,0.05859520037968954
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,32,10240,0.04771626790364583
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,32,7168,0.009699199597040813
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,32,6144,0.009142399827639262
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,32,8192,0.04173440138498942
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,32,5120,0.00831573357184728
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,32,6144,0.034251733620961504
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,32,4096,0.007541333138942718
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,32,7168,0.03901760180791219
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,32,3584,0.0062730665008227035
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,32,4096,0.02794666687647502
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,32,5120,0.03174613316853841
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,32,3072,0.005474133292833964
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,32,2560,0.005006933212280273
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,32,3072,0.02537920077641805
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,32,2048,0.0042250668009122215
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,32,2560,0.024683733781178795
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,32,3584,0.02860693335533142
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,32,1536,0.003804799914360046
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,32,1024,0.0030805334448814393
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,32,1536,0.021630932887395225
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,32,1024,0.019102933009465535
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,32,768,0.003134933362404505
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,32,512,0.0025792000194390613
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,32,768,0.018627200524012247
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,32,256,0.0022304000953833262
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,32,128,0.0021546666820844015
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,32,256,0.01619733373324076
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,32,512,0.018345600366592406
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,32,128,0.01453013320763906
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,32,32,0.002178133279085159
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,32,64,0.014868266383806863
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,32,2048,0.020869332551956176
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,2048,32,64,0.0021930667261282604
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,2048,32,32,0.014379733800888061
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,65536,16384,2.301890055338542
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,65536,16384,1.2030517578125
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,65536,51200,3.950443776448568
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,65536,12288,1.6960437774658204
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,65536,51200,7.286409505208333
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,65536,12288,0.8938079833984375
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,65536,10240,1.38974609375
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,65536,10240,0.7148149490356446
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,65536,10240,0.4358528137207031
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,65536,8192,1.083899688720703
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,65536,8192,0.6166912078857422
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,65536,12288,0.4956021308898926
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,65536,7168,0.9452426910400391
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,65536,7168,0.520796807607015
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,65536,16384,0.6567914962768555
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,65536,6144,0.8508725484212241
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,65536,6144,0.4878335952758789
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,65536,8192,0.3570805231730143
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,65536,5120,0.6364842732747396
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,65536,5120,0.3708287874857584
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,65536,7168,0.32902294794718423
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,65536,4096,0.5715967814127605
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,65536,4096,0.3213909467061361
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,65536,6144,0.27966400782267253
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,65536,3584,0.4679999987284343
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,65536,3584,0.28358081181844075
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,65536,4096,0.22630507151285806
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,65536,3072,0.3946197191874186
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,65536,3072,0.24917227427164712
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,65536,5120,0.2550869305928548
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,65536,2560,0.32685972849527994
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,65536,2560,0.2036896069844564
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,65536,2560,0.17042346000671388
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,65536,2048,0.26429972648620603
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,65536,2048,0.16271573702494305
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,65536,3584,0.20360107421875
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,65536,1536,0.2019733270009359
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,65536,1536,0.12949013710021973
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,65536,3072,0.1785813331604004
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,65536,1024,0.14186879793802898
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,65536,1024,0.09807999928792319
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,65536,1024,0.12383786837259929
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,65536,768,0.11121386686960857
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,65536,768,0.0833845297495524
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,65536,1536,0.12776213486989338
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,65536,512,0.0806506633758545
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,65536,512,0.07849067052205404
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,65536,768,0.11540266672770183
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,65536,256,0.06375146706899007
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,65536,256,0.0637183984120687
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,65536,512,0.11508266925811768
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,65536,128,0.04564053217569987
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,65536,128,0.06269439856211344
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,65536,256,0.11399679978688557
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,65536,64,0.04762453238169352
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,65536,64,0.06425600051879883
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,65536,32,0.05050453344980875
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,65536,128,0.1167413314183553
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,65536,32,0.06372053225835164
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,51200,65536,4.022597249348959
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,65536,2048,0.1519488016764323
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,51200,65536,7.10641581217448
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,51200,51200,3.169513702392578
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,51200,51200,5.532762654622396
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,51200,16384,1.7428267161051434
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,51200,16384,0.9259455998738607
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,51200,16384,0.5154176076253255
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,51200,12288,1.2914581298828125
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,51200,12288,0.6918826421101888
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,51200,12288,0.3901269276936849
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,51200,10240,1.1655349731445312
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,51200,10240,0.6330730438232421
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,51200,10240,0.3434687932332357
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,51200,8192,0.9111263910929361
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,51200,8192,0.5525322596232096
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,65536,51200,2.0553653717041014
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,51200,7168,0.7700735727945964
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,51200,7168,0.41646401087443036
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,51200,7168,0.2590112050374349
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,51200,6144,0.6736597061157227
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,51200,6144,0.4214549382527669
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,51200,8192,0.28060159683227537
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,51200,5120,0.5623722712198893
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,51200,5120,0.3278623898824056
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,51200,5120,0.20128000577290855
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,51200,4096,0.40581439336140945
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,51200,4096,0.24257599512736
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,51200,6144,0.22121173540751138
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,51200,3584,0.3517855962117513
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,51200,3584,0.21252479553222656
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,51200,3584,0.1595882733662923
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,51200,3072,0.30391146341959635
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,51200,3072,0.18548693656921386
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,51200,51200,1.5749312082926432
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,51200,4096,0.17931200663248698
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,51200,2560,0.26766719818115237
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,51200,2560,0.1736021359761556
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,51200,2048,0.2076192061106364
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,51200,2048,0.14388160705566405
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,51200,2048,0.12181546688079833
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,51200,1536,0.15940799713134765
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,51200,1536,0.11728213628133137
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,51200,2560,0.13533439636230468
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,51200,1024,0.12309013207753498
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,51200,1024,0.08862720330556234
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,51200,3072,0.1419274648030599
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,51200,768,0.09593173662821451
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,51200,768,0.07690773010253907
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,51200,768,0.09437440236409506
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,51200,512,0.07077866395314535
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,51200,512,0.06588160196940104
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,51200,1024,0.10276052951812745
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,51200,256,0.04943893353144328
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,51200,1536,0.10563093026479084
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,51200,256,0.05362773338953654
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,51200,128,0.03584213256835937
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,51200,128,0.045613865057627365
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,51200,128,0.0956384023030599
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,51200,64,0.03125653266906738
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,51200,64,0.04594773451487223
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,51200,32,0.03300053278605143
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,51200,32,0.04639039834340413
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,16384,65536,2.267516835530599
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,16384,65536,1.4043572743733725
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,51200,256,0.09423786799112956
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,16384,51200,1.698635737101237
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,16384,51200,1.0920832316080729
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,51200,512,0.09401173591613769
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,16384,16384,0.5324255943298339
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,16384,16384,0.3553226788838705
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,16384,16384,0.19679147402445477
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,16384,12288,0.40176000595092776
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,16384,12288,0.26287360191345216
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,16384,12288,0.1503071943918864
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,16384,10240,0.33351573944091795
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,16384,10240,0.23180799484252929
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,16384,10240,0.13391253153483074
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,16384,8192,0.265116802851359
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,16384,8192,0.1762986660003662
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,16384,8192,0.11000426610310872
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,16384,7168,0.23021653493245445
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,16384,7168,0.15304746627807617
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,16384,7168,0.10201919873555501
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,16384,6144,0.19573973019917806
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,16384,6144,0.13493013381958008
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,16384,51200,0.5805119832356771
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,16384,5120,0.16434879302978517
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,16384,5120,0.11322666803995769
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,16384,6144,0.08817706902821859
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,16384,4096,0.13076480229695636
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,16384,4096,0.09722987016042074
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,16384,5120,0.08179732958475748
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,16384,3584,0.11572799682617188
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,16384,3584,0.08728106816609701
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,16384,3584,0.06204906702041626
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,16384,3072,0.10008959770202637
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,16384,3072,0.07657173474629721
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,16384,4096,0.06959893703460693
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,16384,2560,0.08343146642049154
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,16384,2560,0.06635839939117431
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,16384,2560,0.051541332403818765
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,16384,2048,0.06774933338165283
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,16384,2048,0.058805334568023684
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,16384,3072,0.05351253350575765
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,16384,1536,0.05303786595662435
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,16384,1536,0.05095146497090658
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,16384,1536,0.03993173440297444
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,16384,1024,0.037536001205444335
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,16384,1024,0.03933759927749634
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,16384,2048,0.045201067129770914
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,16384,768,0.02916693290074666
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,16384,768,0.03913813432057699
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,16384,1024,0.038737066586812335
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,16384,512,0.02097599903742472
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,16384,512,0.033086933692296344
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,16384,768,0.03336533308029175
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,16384,256,0.017032533884048462
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,16384,256,0.027115732431411743
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,51200,65536,2.0590848286946613
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,16384,256,0.03276159962018331
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,16384,128,0.02646399935086568
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,16384,128,0.016131200393040977
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,16384,64,0.01330880026022593
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,16384,64,0.027784534295399982
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,16384,32,0.011657599608103435
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,16384,32,0.022717867294947305
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,16384,128,0.03356159925460815
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,12288,65536,1.6504725138346354
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,16384,512,0.032671999931335446
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,12288,65536,1.1069727579752604
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,12288,51200,0.8692714691162109
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,12288,51200,1.2678496042887368
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,12288,16384,0.4416629473368327
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,12288,16384,0.28161598841349283
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,16384,65536,0.7344117482503255
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,12288,12288,0.3054666519165039
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,12288,12288,0.21226773262023926
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,12288,16384,0.1507413387298584
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,12288,10240,0.2563103993733724
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,12288,10240,0.18103893597920734
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,12288,12288,0.11626666386922199
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,12288,8192,0.2106112003326416
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,12288,8192,0.16453439394632977
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,12288,8192,0.08587626616160074
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,12288,7168,0.18408746719360353
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,12288,7168,0.1430741310119629
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,12288,10240,0.10413440068562825
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,12288,6144,0.15950934092203778
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,12288,6144,0.12700586318969725
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,12288,7168,0.08086720307668051
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,12288,5120,0.13496534029642743
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,12288,5120,0.1050645351409912
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,12288,6144,0.07000959714253743
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,12288,4096,0.12788159847259523
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,12288,4096,0.08993173440297444
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,12288,4096,0.05392640034357706
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,12288,3584,0.11202027002970379
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,12288,5120,0.06313493251800537
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,12288,3584,0.08068053722381592
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,12288,3072,0.09305173556009928
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,12288,3072,0.07170133590698242
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,12288,3072,0.04410133361816406
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,12288,2560,0.07531306743621827
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,12288,2560,0.06212906837463379
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,12288,3584,0.04899626572926839
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,12288,2048,0.06047466595967611
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,12288,2048,0.0553706685702006
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,12288,2048,0.039401598771413165
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,12288,1536,0.04759680032730103
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,12288,1536,0.04612693389256795
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,12288,2560,0.04324053525924683
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,12288,1024,0.03457493384679158
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,12288,1024,0.036635732650756835
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,12288,1024,0.03535679976145427
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,12288,768,0.026855466763178508
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,12288,768,0.03286293347676595
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,12288,1536,0.03592960039774577
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,12288,512,0.01862293283144633
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,12288,512,0.029420799016952513
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,12288,768,0.032762666543324784
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,12288,256,0.013301333785057068
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,12288,256,0.024228266874949136
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,12288,512,0.03218773404757182
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,12288,128,0.011587199568748475
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,12288,128,0.020696532726287842
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,12288,128,0.03257813254992167
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,12288,64,0.00995199978351593
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,12288,64,0.02065599958101908
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,12288,32,0.010692266623179118
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,12288,256,0.03207893371582031
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,12288,32,0.022683733701705934
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,10240,65536,0.990236790974935
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,10240,65536,1.420855458577474
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,10240,51200,1.092153549194336
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,10240,51200,0.7756234486897786
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,12288,51200,0.4310858726501465
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,10240,16384,0.33448851903279625
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,10240,16384,0.25326293309529624
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,10240,16384,0.12784639994303387
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,10240,12288,0.2532778739929199
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,10240,12288,0.18547199567159017
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,12288,65536,0.5471925099690755
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,10240,10240,0.229368527730306
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,10240,10240,0.16324267387390137
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,10240,12288,0.09970666567484537
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,10240,8192,0.17925225893656413
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,10240,8192,0.1283018668492635
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,10240,10240,0.09027199745178223
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,10240,7168,0.16171733538309735
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,10240,7168,0.11504213015238444
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,10240,8192,0.07400853633880615
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,10240,6144,0.1381450653076172
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,10240,6144,0.10074666341145833
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,10240,7168,0.06957440376281739
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,10240,5120,0.11767786343892414
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,10240,5120,0.08560960292816162
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,10240,6144,0.060329600175221765
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,10240,4096,0.09211093584696452
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,10240,4096,0.07405973275502523
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,10240,51200,0.3625472068786621
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,10240,3584,0.08130026658376058
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,10240,5120,0.05458986759185791
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,10240,3584,0.06834133466084799
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,10240,3072,0.06857386430104574
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,10240,3072,0.059742931524912515
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,10240,4096,0.04692266782124837
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,10240,2560,0.0557151993115743
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,10240,2560,0.05186560153961182
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,10240,3072,0.03959999879201253
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,10240,2048,0.046196266015370684
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,10240,2048,0.04559573332468669
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,10240,3584,0.043380268414815265
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,10240,1536,0.03634666601816813
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,10240,1536,0.039435732364654544
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,10240,1536,0.032758400837580366
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,10240,1024,0.025414399305979413
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,10240,1024,0.03284693360328674
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,10240,2560,0.03846826553344727
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,10240,2048,0.03563626607259114
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,10240,768,0.019916800657908122
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,10240,768,0.031489066282908124
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,10240,512,0.015154133240381876
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,10240,512,0.02687573234240214
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,10240,512,0.02919573386510213
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,10240,256,0.011273599664370219
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,10240,1024,0.03205546736717224
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,10240,768,0.02959146698315938
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,10240,256,0.023090134064356484
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,10240,128,0.010435199737548828
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,10240,128,0.022730666399002075
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,10240,128,0.030139732360839843
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,10240,64,0.009145599603652955
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,10240,64,0.022230400641759237
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,10240,32,0.009497599800427754
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,10240,32,0.021811199188232423
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,10240,256,0.02882453401883443
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,8192,65536,0.8433749516805014
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,8192,65536,1.0874463399251302
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,8192,51200,0.8163679758707681
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,8192,51200,0.7060341517130534
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,8192,16384,0.27223787307739256
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,8192,16384,0.21233065923055014
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,10240,65536,0.4555786768595378
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,8192,12288,0.20273280143737793
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,8192,12288,0.1605077266693115
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,8192,16384,0.12411839962005615
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,8192,10240,0.17215040524800618
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,8192,10240,0.13880426088968914
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,8192,10240,0.08634880383809408
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,8192,8192,0.13994773228963214
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,8192,8192,0.11529066562652587
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,8192,12288,0.09513599872589111
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,8192,7168,0.12175253232320149
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,8192,7168,0.09992640018463135
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,8192,8192,0.07103040218353271
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,8192,6144,0.10650986830393475
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,8192,6144,0.09074772993723551
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,nvfp4,1793,8192,7168,0.06509653329849244
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,float16,1793,8192,5120,0.08920640150705973
TRTLLM,1.2.0rc6.post3,NVIDIA GB300,gemm,torch_flow,fp8,1793,8192,5120,0.07660799821217855
