update workloads]

Signed-off-by: Jose Santos <josantos@amd.com>
This commit is contained in:
Jose Santos
2024-03-07 15:19:48 -06:00
committed by Karl W. Schulz
parent afea510551
commit 35ef8c0707
1472 changed files with 5798 additions and 630 deletions
@@ -0,0 +1,5 @@
pmc: SQ_INSTS_VALU_ADD_F16 SQ_INSTS_VALU_MUL_F16 SQ_INSTS_VALU_FMA_F16 SQ_INSTS_VALU_TRANS_F16 SQ_INSTS_VALU_ADD_F32 SQ_INSTS_VALU_MUL_F32 SQ_INSTS_VALU_FMA_F32 SQ_INSTS_VALU_TRANS_F32 TCP_TCC_READ_REQ_sum TCP_TOTAL_CACHE_ACCESSES_sum TCP_TCC_WRITE_REQ_sum TCP_TCC_ATOMIC_WITH_RET_REQ_sum TCC_EA_RDREQ_32B_sum TCC_EA_RDREQ_sum TCC_EA_WRREQ_64B_sum TCC_EA_WRREQ_sum
gpu:
range:
kernel:
@@ -0,0 +1,5 @@
pmc: SQ_INSTS_VALU_ADD_F64 SQ_INSTS_VALU_MUL_F64 SQ_INSTS_VALU_FMA_F64 SQ_INSTS_VALU_TRANS_F64 SQ_INSTS_VALU_MFMA_MOPS_F16 SQ_INSTS_VALU_MFMA_MOPS_BF16 SQ_INSTS_VALU_MFMA_MOPS_F32 SQ_INSTS_VALU_MFMA_MOPS_F64 TCP_TCC_ATOMIC_WITHOUT_RET_REQ_sum
gpu:
range:
kernel:
@@ -0,0 +1,5 @@
pmc: SQ_LDS_IDX_ACTIVE SQ_LDS_BANK_CONFLICT
gpu:
range:
kernel:
@@ -0,0 +1,5 @@
pmc:
gpu:
range:
kernel:
@@ -0,0 +1,4 @@
Dispatch_ID,Kernel_Name,GPU_ID,Grid_Size,Workgroup_Size,LDS_Per_Workgroup,Scratch_Per_Workitem,Arch_VGPR,Accum_VGPR,SGPR,wave_size,obj,SQ_INSTS_VALU_ADD_F64,SQ_INSTS_VALU_MUL_F64,SQ_INSTS_VALU_FMA_F64,SQ_INSTS_VALU_TRANS_F64,SQ_INSTS_VALU_MFMA_MOPS_F16,SQ_INSTS_VALU_MFMA_MOPS_BF16,SQ_INSTS_VALU_MFMA_MOPS_F32,SQ_INSTS_VALU_MFMA_MOPS_F64,TCP_TCC_ATOMIC_WITHOUT_RET_REQ_sum,wave_size_1,obj_1,SQ_LDS_IDX_ACTIVE,SQ_LDS_BANK_CONFLICT,wave_size_2,obj_2,SQ_INSTS_VALU_ADD_F16,SQ_INSTS_VALU_MUL_F16,SQ_INSTS_VALU_FMA_F16,SQ_INSTS_VALU_TRANS_F16,SQ_INSTS_VALU_ADD_F32,SQ_INSTS_VALU_MUL_F32,SQ_INSTS_VALU_FMA_F32,SQ_INSTS_VALU_TRANS_F32,TCP_TCC_READ_REQ_sum,TCP_TOTAL_CACHE_ACCESSES_sum,TCP_TCC_WRITE_REQ_sum,TCP_TCC_ATOMIC_WITH_RET_REQ_sum,TCC_EA_RDREQ_32B_sum,TCC_EA_RDREQ_sum,TCC_EA_WRREQ_64B_sum,TCC_EA_WRREQ_sum,Start_Timestamp,End_Timestamp
0,"vecCopy(double*, double*, double*, int, int) ",2,1048576,256,0,0,8,0,16,64,0x7fc43245cec0,0,0,0,0,0,0,0,0,0.0,64,0x7f8e30790ec0,0,0,64,0x7f3100404ec0,0,0,0,0,0,0,0,0,131072.0,524288.0,131072.0,0.0,0.0,131083.0,131072.0,131072.0,199148231834526,199148231855006
1,"vecCopy(double*, double*, double*, int, int) ",2,1048576,256,0,0,8,0,16,64,0x7fc43245cec0,0,0,0,0,0,0,0,0,0.0,64,0x7f8e30790ec0,0,0,64,0x7f3100404ec0,0,0,0,0,0,0,0,0,131072.0,524288.0,131072.0,0.0,0.0,131085.0,91176.0,91176.0,199148231874206,199148231889886
2,"vecCopy(double*, double*, double*, int, int) ",2,1048576,256,0,0,8,0,16,64,0x7fc43245cec0,0,0,0,0,0,0,0,0,0.0,64,0x7f8e30790ec0,0,0,64,0x7f3100404ec0,0,0,0,0,0,0,0,0,131072.0,524288.0,131072.0,0.0,0.0,131085.0,90976.0,90976.0,199148231956286,199148231972286
1 Dispatch_ID Kernel_Name GPU_ID Grid_Size Workgroup_Size LDS_Per_Workgroup Scratch_Per_Workitem Arch_VGPR Accum_VGPR SGPR wave_size obj SQ_INSTS_VALU_ADD_F64 SQ_INSTS_VALU_MUL_F64 SQ_INSTS_VALU_FMA_F64 SQ_INSTS_VALU_TRANS_F64 SQ_INSTS_VALU_MFMA_MOPS_F16 SQ_INSTS_VALU_MFMA_MOPS_BF16 SQ_INSTS_VALU_MFMA_MOPS_F32 SQ_INSTS_VALU_MFMA_MOPS_F64 TCP_TCC_ATOMIC_WITHOUT_RET_REQ_sum wave_size_1 obj_1 SQ_LDS_IDX_ACTIVE SQ_LDS_BANK_CONFLICT wave_size_2 obj_2 SQ_INSTS_VALU_ADD_F16 SQ_INSTS_VALU_MUL_F16 SQ_INSTS_VALU_FMA_F16 SQ_INSTS_VALU_TRANS_F16 SQ_INSTS_VALU_ADD_F32 SQ_INSTS_VALU_MUL_F32 SQ_INSTS_VALU_FMA_F32 SQ_INSTS_VALU_TRANS_F32 TCP_TCC_READ_REQ_sum TCP_TOTAL_CACHE_ACCESSES_sum TCP_TCC_WRITE_REQ_sum TCP_TCC_ATOMIC_WITH_RET_REQ_sum TCC_EA_RDREQ_32B_sum TCC_EA_RDREQ_sum TCC_EA_WRREQ_64B_sum TCC_EA_WRREQ_sum Start_Timestamp End_Timestamp
2 0 vecCopy(double*, double*, double*, int, int) 2 1048576 256 0 0 8 0 16 64 0x7fc43245cec0 0 0 0 0 0 0 0 0 0.0 64 0x7f8e30790ec0 0 0 64 0x7f3100404ec0 0 0 0 0 0 0 0 0 131072.0 524288.0 131072.0 0.0 0.0 131083.0 131072.0 131072.0 199148231834526 199148231855006
3 1 vecCopy(double*, double*, double*, int, int) 2 1048576 256 0 0 8 0 16 64 0x7fc43245cec0 0 0 0 0 0 0 0 0 0.0 64 0x7f8e30790ec0 0 0 64 0x7f3100404ec0 0 0 0 0 0 0 0 0 131072.0 524288.0 131072.0 0.0 0.0 131085.0 91176.0 91176.0 199148231874206 199148231889886
4 2 vecCopy(double*, double*, double*, int, int) 2 1048576 256 0 0 8 0 16 64 0x7fc43245cec0 0 0 0 0 0 0 0 0 0.0 64 0x7f8e30790ec0 0 0 64 0x7f3100404ec0 0 0 0 0 0 0 0 0 131072.0 524288.0 131072.0 0.0 0.0 131085.0 90976.0 90976.0 199148231956286 199148231972286
@@ -0,0 +1,5 @@
device,HBMBw,HBMBwLow,hbmBwHigh,L2Bw,L2BwLow,L2BwHigh,L1Bw,L1BwLow,L1BwHigh,LDSBw,LDSBwLow,LDSBwHigh,FP32Flops,FP32FlopsLow,FP32FlopsHigh,FP64Flops,FP64FlopsLow,FP64FlopsHigh,MFMABF16Flops,MFMABF16FlopsLow,MFMABF16FlopsHigh,MFMAF16Flops,MFMAF16FlopsLow,MFMAF16FlopsHigh,MFMAF32Flops,MFMAF32FlopsLow,MFMAF32FlopsHigh,MFMAF64Flops,MFMAF64FlopsLow,MFMAF64FlopsHigh,MFMAI8Ops,MFMAFI8OpsLow,MFMAI8OpsHigh
0,1388.817,1388.1855,1389.4485,5012.6865,5009.5811,5015.792,9220.6367,9220.0068,9221.2666,18045.574,18044.607,18046.541,20922.65,20862.299,20983.002,20255.824,20255.203,20256.445,170366.09,170360.36,170371.83,164790.75,164786.69,164794.81,41401.891,41400.898,41402.883,41459.949,41459.016,41460.883,166269.62,165773.78,166765.47
1,1388.777,1388.2977,1389.2562,5028.2456,5026.2231,5030.2681,9234.8936,9234.1396,9235.6475,18095.113,18090.842,18099.385,20973.869,20906.225,21041.514,20293.096,20292.455,20293.736,170831.62,170827.34,170835.91,165106.66,165103.06,165110.25,41486.219,41485.355,41487.082,41542.781,41542.02,41543.543,166885.89,166882.86,166888.92
2,1388.8604,1388.2374,1389.4833,5037.2046,5034.9478,5039.4614,9261.4053,9260.6631,9262.1475,18107.312,18106.568,18108.057,21049.461,21049.072,21049.85,20353.656,20353.066,20354.246,171148.52,171144.45,171152.58,165531.92,165527.5,165536.34,41590.082,41588.914,41591.25,41652.598,41651.301,41653.895,166988.36,166396.97,167579.75
3,1389.4712,1388.8257,1390.1167,5033.3506,5032.0864,5034.6147,9241.7871,9241.1338,9242.4404,18025.9,18022.322,18029.479,21038.676,21038.127,21039.225,20316.174,20315.494,20316.854,171045.52,171041.98,171049.05,165247.22,165242.25,165252.19,41445.641,41290.402,41600.879,41588.078,41587.602,41588.555,167041.83,167039.48,167044.17
1 device HBMBw HBMBwLow hbmBwHigh L2Bw L2BwLow L2BwHigh L1Bw L1BwLow L1BwHigh LDSBw LDSBwLow LDSBwHigh FP32Flops FP32FlopsLow FP32FlopsHigh FP64Flops FP64FlopsLow FP64FlopsHigh MFMABF16Flops MFMABF16FlopsLow MFMABF16FlopsHigh MFMAF16Flops MFMAF16FlopsLow MFMAF16FlopsHigh MFMAF32Flops MFMAF32FlopsLow MFMAF32FlopsHigh MFMAF64Flops MFMAF64FlopsLow MFMAF64FlopsHigh MFMAI8Ops MFMAFI8OpsLow MFMAI8OpsHigh
2 0 1388.817 1388.1855 1389.4485 5012.6865 5009.5811 5015.792 9220.6367 9220.0068 9221.2666 18045.574 18044.607 18046.541 20922.65 20862.299 20983.002 20255.824 20255.203 20256.445 170366.09 170360.36 170371.83 164790.75 164786.69 164794.81 41401.891 41400.898 41402.883 41459.949 41459.016 41460.883 166269.62 165773.78 166765.47
3 1 1388.777 1388.2977 1389.2562 5028.2456 5026.2231 5030.2681 9234.8936 9234.1396 9235.6475 18095.113 18090.842 18099.385 20973.869 20906.225 21041.514 20293.096 20292.455 20293.736 170831.62 170827.34 170835.91 165106.66 165103.06 165110.25 41486.219 41485.355 41487.082 41542.781 41542.02 41543.543 166885.89 166882.86 166888.92
4 2 1388.8604 1388.2374 1389.4833 5037.2046 5034.9478 5039.4614 9261.4053 9260.6631 9262.1475 18107.312 18106.568 18108.057 21049.461 21049.072 21049.85 20353.656 20353.066 20354.246 171148.52 171144.45 171152.58 165531.92 165527.5 165536.34 41590.082 41588.914 41591.25 41652.598 41651.301 41653.895 166988.36 166396.97 167579.75
5 3 1389.4712 1388.8257 1390.1167 5033.3506 5032.0864 5034.6147 9241.7871 9241.1338 9242.4404 18025.9 18022.322 18029.479 21038.676 21038.127 21039.225 20316.174 20315.494 20316.854 171045.52 171041.98 171049.05 165247.22 165242.25 165252.19 41445.641 41290.402 41600.879 41588.078 41587.602 41588.555 167041.83 167039.48 167044.17
@@ -0,0 +1,2 @@
workload_name,command,ip_blocks,timestamp,version,hostname,cpu_model,sbios,linux_distro,linux_kernel_version,amd_gpu_kernel_version,cpu_memory,gpu_memory,rocm_version,vbios,compute_partition,memory_partition,gpu_model,gpu_arch,gpu_l1,gpu_l2,cu_per_gpu,simd_per_cu,se_per_gpu,wave_size,workgroup_max_size,max_waves_per_cu,max_sclk,max_mclk,cur_sclk,cur_mclk,total_l2_chan,lds_banks_per_cu,sqc_per_gpu,pipes_per_gpu,hbm_bw,num_xcd
sort_kernels,./sample/vcopy -n 1048576 -b 256 -i 3,SQ|LDS|SQC|TA|TD|TCP|TCC|SPI|CPC|CPF|roofline,Thu 07 Mar 2024 02:30:13 PM (CST),2,t007-002.hpcfund,AMD EPYC 7V13 64-Core Processor,American Megatrends Inc.0602,Rocky Linux 9.1 (Blue Onyx),5.14.0-162.18.1.el9_1.x86_64,,527650760,,6.0.2-115,113-D67301-059,NA,NA,MI200,gfx90a,16,8192,104,4,8,64,1024,32,1700,1600,1700,1600,32,32,56,4,1638.4,1
1 workload_name command ip_blocks timestamp version hostname cpu_model sbios linux_distro linux_kernel_version amd_gpu_kernel_version cpu_memory gpu_memory rocm_version vbios compute_partition memory_partition gpu_model gpu_arch gpu_l1 gpu_l2 cu_per_gpu simd_per_cu se_per_gpu wave_size workgroup_max_size max_waves_per_cu max_sclk max_mclk cur_sclk cur_mclk total_l2_chan lds_banks_per_cu sqc_per_gpu pipes_per_gpu hbm_bw num_xcd
2 sort_kernels ./sample/vcopy -n 1048576 -b 256 -i 3 SQ|LDS|SQC|TA|TD|TCP|TCC|SPI|CPC|CPF|roofline Thu 07 Mar 2024 02:30:13 PM (CST) 2 t007-002.hpcfund AMD EPYC 7V13 64-Core Processor American Megatrends Inc.0602 Rocky Linux 9.1 (Blue Onyx) 5.14.0-162.18.1.el9_1.x86_64 527650760 6.0.2-115 113-D67301-059 NA NA MI200 gfx90a 16 8192 104 4 8 64 1024 32 1700 1600 1700 1600 32 32 56 4 1638.4 1
@@ -0,0 +1,4 @@
Dispatch_ID,Kernel_Name,GPU_ID,queue-id,queue-index,pid,tid,Grid_Size,Workgroup_Size,LDS_Per_Workgroup,Scratch_Per_Workitem,Arch_VGPR,Accum_VGPR,SGPR,wave_size,sig,obj,DispatchNs,Start_Timestamp,End_Timestamp,CompleteNs
0,"vecCopy(double*, double*, double*, int, int) [clone .kd]",2,0,0,335825,335825,1048576,256,0,0,8,0,16,64,0x0,0x7f0145064ec0,199148231809224,199148231834526,199148231855006,199148231871421
1,"vecCopy(double*, double*, double*, int, int) [clone .kd]",2,0,2,335825,335825,1048576,256,0,0,8,0,16,64,0x0,0x7f0145064ec0,199148231865981,199148231874206,199148231889886,199148231953617
2,"vecCopy(double*, double*, double*, int, int) [clone .kd]",2,0,4,335825,335825,1048576,256,0,0,8,0,16,64,0x0,0x7f0145064ec0,199148231900226,199148231956286,199148231972286,199148231973294
1 Dispatch_ID Kernel_Name GPU_ID queue-id queue-index pid tid Grid_Size Workgroup_Size LDS_Per_Workgroup Scratch_Per_Workitem Arch_VGPR Accum_VGPR SGPR wave_size sig obj DispatchNs Start_Timestamp End_Timestamp CompleteNs
2 0 vecCopy(double*, double*, double*, int, int) [clone .kd] 2 0 0 335825 335825 1048576 256 0 0 8 0 16 64 0x0 0x7f0145064ec0 199148231809224 199148231834526 199148231855006 199148231871421
3 1 vecCopy(double*, double*, double*, int, int) [clone .kd] 2 0 2 335825 335825 1048576 256 0 0 8 0 16 64 0x0 0x7f0145064ec0 199148231865981 199148231874206 199148231889886 199148231953617
4 2 vecCopy(double*, double*, double*, int, int) [clone .kd] 2 0 4 335825 335825 1048576 256 0 0 8 0 16 64 0x0 0x7f0145064ec0 199148231900226 199148231956286 199148231972286 199148231973294