llvm / llvm/offload-test-suite

[Metal] WaveActiveMax/Min fp16 tests crash the metal driver

Open
#1,200 0 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

Dominant language
C++
Stars
18
Forks
39
Avg merge
2d 18h
Merged PRs (30d)
40

Description

Since we enabled metal testing with explicit root signatures, the WaveActiveMax.fp16 and WaveActiveMin.fp16 tests are consistently failing with a driver crash on the m2-mbp machine in the lab. Report here: https://github.com/llvm/offload-test-suite/pull/1061#issuecomment-4455026120

This doesn't seem to be an issue on M3 and later, but looks like it is likely a driver bug.

Contributor guide

No contributing guide indexed for this repository

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

Start by reproducing the WaveActiveMax.fp16 and WaveActiveMin.fp16 failures on the M2-MBP lab machine, then compare their behavior with M3 and later systems. Read the linked report and inspect the explicit root-signature testing change. Done means the driver crash is confirmed and an agreed test handling or fix is identified.

Written by the indexing model from the issue text.

Assessment

Tech stack
cpp
Domain
testing-qa
Issue type
Bug
Difficulty
4/5
Estimated time
3-5 days
Activity status
Quiet
Clarity
Mostly clear
Newbie friendliness
35/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.