Skip to content

Add IntelliKit (Metrix) profiling results for Example 7: GEMM all-scatter - #90

Draft
JoseSantosAMD with Copilot wants to merge 2 commits into
JoseSantosAMD/intellikitfrom
copilot/sub-pr-24-4498d0b4-9746-40ff-bcdd-b589bd702c82
Draft

Add IntelliKit (Metrix) profiling results for Example 7: GEMM all-scatter#90
JoseSantosAMD with Copilot wants to merge 2 commits into
JoseSantosAMD/intellikitfrom
copilot/sub-pr-24-4498d0b4-9746-40ff-bcdd-b589bd702c82

Conversation

Copilot AI commented Mar 12, 2026

Copy link
Copy Markdown

Runs Example 7 (persistent_gemm_all_scatter) on 2× AMD MI300X (gfx942) and profiles it with Metrix from IntelliKit.

Changes

  • examples/07_gemm_all_scatter/README.md — new file documenting:
    • Benchmark: ~611 TFLOPS, ~4.55 ms (M=8192, N=4608, K=36864, FP16, 2 GPUs)
    • Memory profile: 3.2% HBM utilization, 81.3% L2 hit rate, 52.6% L1 hit rate, zero LDS bank conflicts and atomic latency
    • Compute profile: ~315K GFLOPS, 267.7 FLOP/Byte HBM arithmetic intensity → compute-bound
    • Analysis section interpreting the key metrics
  • .gitignore — added .rocprofv3/ (Metrix profiling artifact directory)

📱 Kick off Copilot coding agent tasks wherever you are with GitHub Mobile, available on iOS and Android.

…catter)

Co-authored-by: JoseSantosAMD <87447437+JoseSantosAMD@users.noreply.github.com>
Copilot AI changed the title [WIP] Add IntelliKit support for profiling Add IntelliKit (Metrix) profiling results for Example 7: GEMM all-scatter Mar 12, 2026
Copilot AI requested a review from JoseSantosAMD March 12, 2026 19:36
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants