address DDE in matmul decomp

[laithsakka]

Stack from ghstack (oldest at bottom):

• -> address DDE in matmul decomp #166541

Address #165081

[pytorch-bot]

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/166541

• 📄 Preview Python docs built from this PR
• 📄 Preview C++ docs built from this PR
• ❓ Need help or want to give feedback on the CI? Visit the bot commands wiki

Note: Links to docs will display an error until the docs builds have been completed.

✅ You can merge normally! (2 Unrelated Failures)

As of commit 896fa02 with merge base 32fe4f6 ():

FLAKY - The following job failed but was likely due to flakiness present on trunk:

• inductor / unit-test / inductor-test / test (inductor_distributed, 1, 1, linux.g5.12xlarge.nvidia.gpu) (gh) (matched linux rule in flaky-rules.json)
  The runner has received a shutdown signal. This can happen when the runner service is stopped, or a manually started runner is canceled.

BROKEN TRUNK - The following job failed but was present on the merge base:

👉 Rebase onto the `viable/strict` branch to avoid these failures

• trunk / linux-jammy-cuda12.8-py3.10-gcc11 / test (default, 2, 5, lf.linux.g6.4xlarge.experimental.nvidia.gpu) (gh) (trunk failure)
  ul

This comment was automatically generated by Dr. CI and updates every 15 minutes.

[mlazos]

lgtm, one question when you guard_or_false are we just optimizing the case when it equals 1? I'm guessing we can do something when the value is 1, and recompile if it's not 1 later. But in the other direction we'll just use the unpsecialized code for all cases.

[laithsakka]

lgtm, one question when you guard_or_false are we just optimizing the case when it equals 1? I'm guessing we can do something when the value is 1, and recompile if it's not 1 later. But in the other direction we'll just use the unpsecialized code for all cases.

yeh when we know one of them is for sure 1 we take the opt path , not that we already know they are not both 1s.
now in the case that we do not know we take the general path and that would be our unbacked semantics for this.

[laithsakka]

@pytorchbot merge

[pytorchmergebot]

Merge started

Your change will be merged once all checks pass (ETA 0-4 Hours).

Learn more about merging in the wiki.

Questions? Feedback? Please reach out to the PyTorch DevX Team

Advanced Debugging

Check the merge workflow status
here

[atalman]

@pytorchmergebot revert -c ghfirst -m "breaks internal test"

[pytorchmergebot]

@pytorchbot successfully started a revert job. Check the current status here.
Questions? Feedback? Please reach out to the PyTorch DevX Team

[pytorchmergebot]

@laithsakka your PR has been successfully reverted.