Repository navigation
[CI] Exclude nvidia-cuda-cccl from the ROCm test lockfile - #1412
Merged
Merged
Conversation
Signed-off-by: Zohaib Moti <zohaib.moti@amd.com>
eble-amd
approved these changes
Oct 28, 2026
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Overview
Port of vllm-project#64302 to
gfx11. Thepip-compile-rocmpre-commit hook currently fails on every PR againstgfx11(for example #1411), because CI's regeneration ofrequirements/test/rocm.txtno longer matches the committed file.Root cause: torch 2.15.0 reached PyPI on 2026-10-28 (15:16 UTC). The ROCm test lockfile resolves torch and its CUDA dependencies from PyPI only to exclude them via
--no-emit-package. The new torch moves those dependencies to CUDA 13.2, wherenvidia-nvshmem-cu133.7.2 depends onnvidia-cuda-cccl. That package was not on the exclude list, so it leaked into the ROCm lockfile, andnvidia-curanddropped out of the resolution.Changes
.pre-commit-config.yaml: add--no-emit-package nvidia-cuda-cccltopip-compile-rocm. This is the same line as upstream; that part cherry-picks cleanly.requirements/test/rocm.txt: regenerated with the hook using its pinned uv 0.11.1, rather than taking upstream's lockfile hunk, which does not apply because thegfx11lockfile has diverged. The only changes are the embedded command line and the excluded-package comment list (+nvidia-cuda-cccl,-nvidia-curand). No pinned versions change.Validation
pip-compile-rocmpasses locally on this commit. The lockfile tracks what PyPI serves, so a later upstream release could require regenerating it again.