AI/rocm-systems
2
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Merge pull request #155 from arghdos/bf16_flops

Browse Source 
fix max BF16 flop rate on CDNA2

[ROCm/rocprofiler-compute commit: c809346b1b]
This commit is contained in:
Cole Ramos
2023-08-03 15:36:45 -07:00
committed by GitHub
parent 60c0dbd7f4 49f5647608
commit 00e4558fca
2 changed files with 3 additions and 3 deletions
Download Patch File Download Diff File Expand all files Collapse all files
projects/rocprofiler-compute/dashboards/Omniperf_v1.0.8_pub.json
+1 -1
View File
File diff suppressed because one or more lines are too long
projects/rocprofiler-compute/src/omniperf_analyze/configs/gfx90a/0200_system-speed-of-light.yaml
+2 -2
View File
@@ -45,9 +45,9 @@ Panel Config:
MFMA FLOPs (BF16):
value: AVG(((SQ_INSTS_VALU_MFMA_MOPS_BF16 * 512) / (EndNs - BeginNs)))
unit: GFLOP
peak: ((($sclk * $numCU) * 512) / 1000)
peak: ((($sclk * $numCU) * 1024) / 1000)
pop: ((100 * AVG(((SQ_INSTS_VALU_MFMA_MOPS_BF16 * 512) / (EndNs - BeginNs))))
/ ((($sclk * $numCU) * 512) / 1000))
/ ((($sclk * $numCU) * 1024) / 1000))
tips:
MFMA FLOPs (F16):
value: AVG(((SQ_INSTS_VALU_MFMA_MOPS_F16 * 512) / (EndNs - BeginNs)))