文件
rocm-systems/tests/workloads/dev1/mi200/prev_analysis/201.csv
T
colramos-amd 62d130b458 Initial commit
2022-11-04 14:49:36 -05:00

1.5 KiB

1MetricValueUnitPeakPoP
2VALU FLOPs7440.165497034273Gflop22630.432.87686252578069
3VALU IOPs2298.891682714651Giop22630.410.158422664710525
4MFMA FLOPs (BF16)0.0Gflop90521.60.0
5MFMA FLOPs (F16)0.0Gflop181043.20.0
6MFMA FLOPs (F32)0.0Gflop45260.80.0
7MFMA FLOPs (F64)0.0Gflop45260.80.0
8MFMA IOPs (Int8)0.0Giop181043.20.0
9Active CUs103Cus10499.03846153846153
10SALU Util2.5928524008424265Pct1002.5928524008424265
11VALU Util57.676542261213726Pct10057.676542261213726
12MFMA Util0.0Pct1000.0
13VALU Active Threads/Wave64.0Threads64100.0
14IPC - Issue1.0Instr/cycle520.0
15LDS BW0.0Gb/sec22630.40.0
16LDS Bank ConflictConflicts/access32
17Instr Cache Hit Rate99.99260892771422Pct10099.99260892771422
18Instr Cache BW1672.988845263023Gb/s6092.827.45845662524657
19Scalar L1D Cache Hit Rate99.34855886103351Pct10099.34855886103351
20Scalar L1D Cache BW57.053677451411716Gb/s6092.80.9364114602713319
21Vector L1D Cache Hit Rate20.35928143712575Pct10020.35928143712575
22Vector L1D Cache BW1691.605226500164Gb/s11315.19999999999914.949848226281146
23L2 Cache Hit Rate3.8459589299426997Pct1003.8459589299426997
24L2-Fabric Read BW1158.3220712925001Gb/s1638.470.69836860916138
25L2-Fabric Write BW6.646011203399653Gb/s1638.40.40564033223874835
26L2-Fabric Read Latency531.9029600851349Cycles
27L2-Fabric Write Latency412.8785129123264Cycles
28Wave Occupancy2590.8817173172824Wavefronts332877.85101314054333
29Instr Fetch BW413.31420363396796Gb/s3046.413.567299226430146
30Instr Fetch Latency17.23791657227002Cycles