f9aa7be97c
* Add MI 350 hardware information
* Refactor MI GPU YAML file and corresponding interface
* Add SoC file for gfx950 architecture
* Add analysis report configs for MI 350 containing existing metrics
* Add placeholder None valued metrics for previous architectures to make
baseline comparison work
* Enable testing on MI 350
* Analysis config metric changes
- SPI changes
- Update metric formula for default SPI pipe counter
- Use efficiently collected pipe wise SPI counters
- Add SPI Wave Occupancy
- Add Scheduler-Pipe Wave Utilization
- Update formula for VGPR Writes
- Add Scheduler-Pipe FIFO Full Rate
- CPC changes
- Add CPC SYNC FIFO Full Rate
- Add CPC CANE Stall Rate
- Add CPC ADC Utilization
- SQ changes
- Add VALU co-issue efficiency
- Add F6F4 datatype metrics
- Update formula for total FLOPs by adding F6F4 counters
- Add LDS STORE / LOAD / ATOMIC metrics
- Add LDS STORE / LOAD / ATOMIC bandwidth
- Add LDS FIFO and TA ADDR / CMD / DATA FIFO full rates
* Collect TCP_TCP_LATENCY_sum only for gfx950 (MI 350)
* Do not inject SQ_ACCUM_PREV_HIRES unnecesarily
* Do not hardcode memory and shader clock speeds
* Write num_hbm_channels to sysinfo.csv instead of hbm_bw while profiling
* Move generate sysinfo.csv to pre processing step of profiling
* Add warnings to use --specs-correction for missing sysinfo.csv values during analysis phase
* Update CHANGELOG
* Analysis phase warning to use --specs-correction when needed
7.7 KiB
7.7 KiB
| 1 | Dispatch_ID | GPU_ID | Queue_ID | PID | TID | Grid_Size | Workgroup_Size | LDS_Per_Workgroup | Scratch_Per_Workitem | Arch_VGPR | Accum_VGPR | SGPR | Wave_Size | Kernel_Name | Start_Timestamp | End_Timestamp | Correlation_ID | Kernel_ID | CPC_CPC_TCIU_IDLE | CPC_TG_SEND | CPF_CPF_TCIU_IDLE | SPI_CS0_BUSY | SPI_CS1_BUSY | SPI_CS2_BUSY | SPI_CS3_BUSY | SQ_INSTS_LDS | SQ_INSTS_LDS_LOAD_BANDWIDTH | SQ_INSTS_VALU_ADD_F16 | SQ_INSTS_VALU_FMA_F16 | SQ_LDS_IDX_ACTIVE | SQ_LEVEL_WAVES | SQ_WAVE_CYCLES | TA_BUFFER_ATOMIC_WAVEFRONTS_sum | TA_FLAT_ATOMIC_WAVEFRONTS_sum | TCC_BUBBLE_sum | TCC_EA0_ATOMIC_LEVEL[0] | TCC_EA0_ATOMIC_LEVEL[11] | TCC_EA0_ATOMIC_LEVEL[12] | TCC_EA0_ATOMIC_LEVEL[13] | TCC_EA0_ATOMIC_LEVEL[14] | TCC_EA0_ATOMIC_LEVEL[16] | TCC_EA0_ATOMIC_LEVEL[17] | TCC_EA0_ATOMIC_LEVEL[18] | TCC_EA0_ATOMIC_LEVEL[19] | TCC_EA0_ATOMIC_LEVEL[1] | TCC_EA0_ATOMIC_LEVEL[22] | TCC_EA0_ATOMIC_LEVEL[23] | TCC_EA0_ATOMIC_LEVEL[24] | TCC_EA0_ATOMIC_LEVEL[25] | TCC_EA0_ATOMIC_LEVEL[27] | TCC_EA0_ATOMIC_LEVEL[28] | TCC_EA0_ATOMIC_LEVEL[29] | TCC_EA0_ATOMIC_LEVEL[2] | TCC_EA0_ATOMIC_LEVEL[30] | TCC_EA0_ATOMIC_LEVEL[32] | TCC_EA0_ATOMIC_LEVEL[33] | TCC_EA0_ATOMIC_LEVEL[34] | TCC_EA0_ATOMIC_LEVEL[35] | TCC_EA0_ATOMIC_LEVEL[37] | TCC_EA0_ATOMIC_LEVEL[38] | TCC_EA0_ATOMIC_LEVEL[39] | TCC_EA0_ATOMIC_LEVEL[3] | TCC_EA0_ATOMIC_LEVEL[40] | TCC_EA0_ATOMIC_LEVEL[43] | TCC_EA0_ATOMIC_LEVEL[44] | TCC_EA0_ATOMIC_LEVEL[45] | TCC_EA0_ATOMIC_LEVEL[46] | TCC_EA0_ATOMIC_LEVEL[48] | TCC_EA0_ATOMIC_LEVEL[49] | TCC_EA0_ATOMIC_LEVEL[4] | TCC_EA0_ATOMIC_LEVEL[50] | TCC_EA0_ATOMIC_LEVEL[51] | TCC_EA0_ATOMIC_LEVEL[53] | TCC_EA0_ATOMIC_LEVEL[54] | TCC_EA0_ATOMIC_LEVEL[55] | TCC_EA0_ATOMIC_LEVEL[56] | TCC_EA0_ATOMIC_LEVEL[58] | TCC_EA0_ATOMIC_LEVEL[59] | TCC_EA0_ATOMIC_LEVEL[60] | TCC_EA0_ATOMIC_LEVEL[61] | TCC_EA0_ATOMIC_LEVEL[6] | TCC_EA0_ATOMIC_LEVEL[7] | TCC_EA0_ATOMIC_LEVEL[8] | TCC_EA0_ATOMIC_LEVEL[9] | TCC_EA0_RDREQ_LEVEL[0] | TCC_EA0_RDREQ_LEVEL[11] | TCC_EA0_RDREQ_LEVEL[12] | TCC_EA0_RDREQ_LEVEL[13] | TCC_EA0_RDREQ_LEVEL[14] | TCC_EA0_RDREQ_LEVEL[16] | TCC_EA0_RDREQ_LEVEL[17] | TCC_EA0_RDREQ_LEVEL[18] | TCC_EA0_RDREQ_LEVEL[19] | TCC_EA0_RDREQ_LEVEL[1] | TCC_EA0_RDREQ_LEVEL[22] | TCC_EA0_RDREQ_LEVEL[23] | TCC_EA0_RDREQ_LEVEL[24] | TCC_EA0_RDREQ_LEVEL[25] | TCC_EA0_RDREQ_LEVEL[27] | TCC_EA0_RDREQ_LEVEL[28] | TCC_EA0_RDREQ_LEVEL[29] | TCC_EA0_RDREQ_LEVEL[2] | TCC_EA0_RDREQ_LEVEL[30] | TCC_EA0_RDREQ_LEVEL[32] | TCC_EA0_RDREQ_LEVEL[33] | TCC_EA0_RDREQ_LEVEL[34] | TCC_EA0_RDREQ_LEVEL[35] | TCC_EA0_RDREQ_LEVEL[37] | TCC_EA0_RDREQ_LEVEL[38] | TCC_EA0_RDREQ_LEVEL[39] | TCC_EA0_RDREQ_LEVEL[3] | TCC_EA0_RDREQ_LEVEL[40] | TCC_EA0_RDREQ_LEVEL[43] | TCC_EA0_RDREQ_LEVEL[44] | TCC_EA0_RDREQ_LEVEL[45] | TCC_EA0_RDREQ_LEVEL[46] | TCC_EA0_RDREQ_LEVEL[48] | TCC_EA0_RDREQ_LEVEL[49] | TCC_EA0_RDREQ_LEVEL[4] | TCC_EA0_RDREQ_LEVEL[50] | TCC_EA0_RDREQ_LEVEL[51] | TCC_EA0_RDREQ_LEVEL[53] | TCC_EA0_RDREQ_LEVEL[54] | TCC_EA0_RDREQ_LEVEL[55] | TCC_EA0_RDREQ_LEVEL[56] | TCC_EA0_RDREQ_LEVEL[58] | TCC_EA0_RDREQ_LEVEL[59] | TCC_EA0_RDREQ_LEVEL[60] | TCC_EA0_RDREQ_LEVEL[61] | TCC_EA0_RDREQ_LEVEL[6] | TCC_EA0_RDREQ_LEVEL[7] | TCC_EA0_RDREQ_LEVEL[8] | TCC_EA0_RDREQ_LEVEL[9] | TCC_TOO_MANY_EA_WRREQS_STALL[0] | TCC_TOO_MANY_EA_WRREQS_STALL[11] | TCC_TOO_MANY_EA_WRREQS_STALL[12] | TCC_TOO_MANY_EA_WRREQS_STALL[13] | TCC_TOO_MANY_EA_WRREQS_STALL[14] | TCC_TOO_MANY_EA_WRREQS_STALL[16] | TCC_TOO_MANY_EA_WRREQS_STALL[17] | TCC_TOO_MANY_EA_WRREQS_STALL[18] | TCC_TOO_MANY_EA_WRREQS_STALL[19] | TCC_TOO_MANY_EA_WRREQS_STALL[1] | TCC_TOO_MANY_EA_WRREQS_STALL[22] | TCC_TOO_MANY_EA_WRREQS_STALL[23] | TCC_TOO_MANY_EA_WRREQS_STALL[24] | TCC_TOO_MANY_EA_WRREQS_STALL[25] | TCC_TOO_MANY_EA_WRREQS_STALL[27] | TCC_TOO_MANY_EA_WRREQS_STALL[28] | TCC_TOO_MANY_EA_WRREQS_STALL[29] | TCC_TOO_MANY_EA_WRREQS_STALL[2] | TCC_TOO_MANY_EA_WRREQS_STALL[30] | TCC_TOO_MANY_EA_WRREQS_STALL[32] | TCC_TOO_MANY_EA_WRREQS_STALL[33] | TCC_TOO_MANY_EA_WRREQS_STALL[34] | TCC_TOO_MANY_EA_WRREQS_STALL[35] | TCC_TOO_MANY_EA_WRREQS_STALL[37] | TCC_TOO_MANY_EA_WRREQS_STALL[38] | TCC_TOO_MANY_EA_WRREQS_STALL[39] | TCC_TOO_MANY_EA_WRREQS_STALL[3] | TCC_TOO_MANY_EA_WRREQS_STALL[40] | TCC_TOO_MANY_EA_WRREQS_STALL[43] | TCC_TOO_MANY_EA_WRREQS_STALL[44] | TCC_TOO_MANY_EA_WRREQS_STALL[45] | TCC_TOO_MANY_EA_WRREQS_STALL[46] | TCC_TOO_MANY_EA_WRREQS_STALL[48] | TCC_TOO_MANY_EA_WRREQS_STALL[49] | TCC_TOO_MANY_EA_WRREQS_STALL[4] | TCC_TOO_MANY_EA_WRREQS_STALL[50] | TCC_TOO_MANY_EA_WRREQS_STALL[51] | TCC_TOO_MANY_EA_WRREQS_STALL[53] | TCC_TOO_MANY_EA_WRREQS_STALL[54] | TCC_TOO_MANY_EA_WRREQS_STALL[55] | TCC_TOO_MANY_EA_WRREQS_STALL[56] | TCC_TOO_MANY_EA_WRREQS_STALL[58] | TCC_TOO_MANY_EA_WRREQS_STALL[59] | TCC_TOO_MANY_EA_WRREQS_STALL[60] | TCC_TOO_MANY_EA_WRREQS_STALL[61] | TCC_TOO_MANY_EA_WRREQS_STALL[6] | TCC_TOO_MANY_EA_WRREQS_STALL[7] | TCC_TOO_MANY_EA_WRREQS_STALL[8] | TCC_TOO_MANY_EA_WRREQS_STALL[9] | TCP_GATE_EN2_sum | TCP_TCC_RW_WRITE_REQ_sum | TCP_TCC_UC_ATOMIC_REQ_sum | TCP_TOTAL_CACHE_ACCESSES_sum | TD_TD_BUSY_sum |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 2 | 0 | 0 | 1 | 15765 | 15765 | 1048576 | 256 | 0 | 0 | 4 | 0 | 16 | 64 | vecCopy(double*, double*, double*, int, int) | 1248226695430472 | 1248226695452112 | 1 | 13 | 76230.0 | 4096.0 | 74700.0 | 146248.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 6406528.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 366347.0 | 357772.0 | 356963.0 | 349090.0 | 376766.0 | 737844.0 | 674665.0 | 709219.0 | 678639.0 | 690427.0 | 343990.0 | 324671.0 | 343293.0 | 346854.0 | 381616.0 | 358883.0 | 363344.0 | 703259.0 | 361957.0 | 743500.0 | 699149.0 | 680115.0 | 684746.0 | 338927.0 | 323941.0 | 341659.0 | 662566.0 | 341692.0 | 342216.0 | 362540.0 | 345698.0 | 325196.0 | 670504.0 | 706680.0 | 336552.0 | 689433.0 | 668474.0 | 338628.0 | 351132.0 | 343139.0 | 331632.0 | 367394.0 | 363667.0 | 344667.0 | 324349.0 | 343830.0 | 339430.0 | 345257.0 | 339020.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 823402.0 | 131072.0 | 0.0 | 524288.0 | 816209.0 |
| 3 | 1 | 0 | 1 | 15765 | 15765 | 1048576 | 256 | 0 | 0 | 4 | 0 | 16 | 64 | vecCopy(double*, double*, double*, int, int) | 1248226696169603 | 1248226696186603 | 2 | 13 | 66030.0 | 4096.0 | 64509.0 | 108920.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 4413902.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 168628.0 | 188771.0 | 180374.0 | 166979.0 | 187051.0 | 341252.0 | 347378.0 | 366453.0 | 358634.0 | 349942.0 | 166901.0 | 185540.0 | 167789.0 | 186890.0 | 195821.0 | 191987.0 | 161167.0 | 349386.0 | 173327.0 | 351353.0 | 362889.0 | 349396.0 | 340194.0 | 166112.0 | 162174.0 | 179731.0 | 326950.0 | 183560.0 | 167731.0 | 191528.0 | 159788.0 | 174031.0 | 333041.0 | 347181.0 | 191522.0 | 361391.0 | 371374.0 | 166146.0 | 185780.0 | 171148.0 | 183365.0 | 196992.0 | 178416.0 | 168290.0 | 169604.0 | 172663.0 | 165574.0 | 182558.0 | 177592.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 647210.0 | 131072.0 | 0.0 | 524288.0 | 639099.0 |
| 4 | 2 | 0 | 1 | 15765 | 15765 | 1048576 | 256 | 0 | 0 | 4 | 0 | 16 | 64 | vecCopy(double*, double*, double*, int, int) | 1248226696801619 | 1248226696818339 | 3 | 13 | 61716.0 | 4096.0 | 60350.0 | 107896.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 4337140.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 192457.0 | 179283.0 | 183607.0 | 168303.0 | 188445.0 | 363907.0 | 353513.0 | 336506.0 | 358454.0 | 357389.0 | 166113.0 | 175264.0 | 165655.0 | 198953.0 | 179626.0 | 182508.0 | 166395.0 | 349820.0 | 187371.0 | 369566.0 | 387313.0 | 340202.0 | 353747.0 | 167292.0 | 163935.0 | 187961.0 | 331568.0 | 189609.0 | 166929.0 | 186667.0 | 164998.0 | 190238.0 | 336870.0 | 380054.0 | 177919.0 | 357408.0 | 351545.0 | 166698.0 | 185523.0 | 186549.0 | 188624.0 | 178890.0 | 183734.0 | 165658.0 | 163188.0 | 167422.0 | 177959.0 | 191058.0 | 197686.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 648321.0 | 131072.0 | 0.0 | 524288.0 | 640253.0 |