🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/175710

• 📄 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.

❌ 1 New Failure, 3 Unrelated Failures

As of commit d1a08d4 with merge base c15f33d ():

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

@pytorchbot merge

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

Merge failed

Reason: 1 mandatory check(s) failed. The first few are:

• Lint / lintrunner-noclang-all / linux-job

Dig deeper by viewing the failures on hud

@claude fix lint and trigger merge

Claude finished @ezyang's task in 2m 44s —— View job

Fixing lint and triggering merge

• Investigate lint failure — F841: unused variable fake_pg on line 610 of test/distributed/test_local_tensor.py
• Fix lint issue — removed the unused fake_pg variable (it was assigned but never passed to dist.reduce_scatter)
• Push fix — committed and pushed to gh/ezyang/3274/head
• Trigger merge — I'm unable to post comments on the PR to invoke @pytorchbot merge due to permissions. Please comment @pytorchbot merge to trigger the merge once CI passes.

The committers listed above are authorized under a signed CLA.

• ✅ login: ezyang / name: Edward Z. Yang (02c3e0b, 8ed38de, c6435b8, d1a08d4)

@pytorchbot merge

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

@pytorchbot revert -m "Reverted automatically by pytorch's autorevert, to avoid this behaviour add the tag autorevert: disable" -c autorevert

This PR is attributed to have caused regression in:

• Lint: lintrunner-noclang-all / linux-job (hud)

Please investigate and fix the issues.
@claude Can you please read this revert comment, follow the links and read the errors, to then give a brief diagnostics on the cause of the error? If you judge the error to be legitimate reason for a revert, please provide brief guidance on how the author could fix it.

Claude finished @pytorch-auto-revert[bot]'s task in 1m 1s —— View job

Revert Diagnosis

• Read CI failure logs from the lint job
• Diagnose root cause
• Assess if revert is legitimate and provide fix guidance

Root Cause

The lint failure is RUFF S101 ("Use of assert detected") in torch/distributed/_local_tensor/_c10d.py. The new reduce_scatter_ function uses bare assert statements on lines 619, 620, 628, 629, and 641, which violates the project's lint rules.

Is the revert legitimate?

Yes, this is a legitimate revert. The lint failure is a real issue — the rest of the file consistently uses raise AssertionError(...) instead of bare assert statements. This is the established pattern throughout _c10d.py (there are 40+ instances of raise AssertionError in the file, and zero assert statements prior to this PR).

How to fix

Replace the bare assert statements with explicit raise AssertionError(...), matching the existing code style. Specifically, change:

# Lines 619-620
assert len(output_tensors) == 1
assert len(input_tensors) == 1

to:

if len(output_tensors) != 1:
    raise AssertionError
if len(input_tensors) != 1:
    raise AssertionError

# Line 628
assert isinstance(output_tensor, LocalTensor), "Output tensor must be a LocalTensor"

to:

if not isinstance(output_tensor, LocalTensor):
    raise AssertionError("Output tensor must be a LocalTensor")

# Lines 629-631
assert len(input_list) == len(ranks), (
    f"Number of input chunks ({len(input_list)}) must match number of ranks ({len(ranks)})"
)

to:

if len(input_list) != len(ranks):
    raise AssertionError(
        f"Number of input chunks ({len(input_list)}) must match number of ranks ({len(ranks)})"
    )

# Line 641
assert isinstance(chunk, LocalTensor), "Input chunk must be a LocalTensor"

to:

if not isinstance(chunk, LocalTensor):
    raise AssertionError("Input chunk must be a LocalTensor")

This matches the established pattern used by every other function in the file (e.g., _local_all_reduce_, _local_reduce_scatter_, _local_all_gather_, etc.).

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

@ezyang your PR has been successfully reverted.

@pytorchbot merge

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

Merge failed

Reason: 1 jobs have failed, first few of them are: trunk / linux-jammy-rocm-py3.10 / test (default, 5, 6, linux.rocm.gpu.gfx950.1)