[https://nvbugs/6437342][fix] Add --dist=loadfile to the child pytest cmd list so all tests in one file…#16329
Open
trtllm-agent wants to merge 1 commit into
Open
Conversation
The outer test_run_unit_tests spawns a pytest -n 4 subprocess over 1800+ child tests. Under xdist's default --dist=load, tests are dispatched dynamically to workers, so module-level torch.manual_seed(0) no longer produces a stable RNG state at each test entry. This intermittently trips tight numerical tolerances in flashinfer_mla_op, gemm_fusion, and piecewise_memory child tests, and the outer 'assert num_failed == 0' at test_standalone_package.py:313 fires with a different failing test on each run. Add --dist=loadfile so all tests in one file execute on the same worker in file-declaration order. Module-level RNG seeding then reliably precedes the tests it is meant to seed, and per-test input state becomes reproducible across runs. Parallelism is preserved (still 4 workers, each running a distinct file at a time), so wall-clock stays close to the current ~260s. The same class of flakiness is documented in tests/unittest/auto_deploy/singlegpu/custom_ops/rope/test_rope_op_variants.py:28, which notes that pytest-xdist dynamic distribution 'used order-dependent inputs and failed intermittently'. Signed-off-by: trtllm-agent <296075020+trtllm-agent@users.noreply.github.com>
Contributor
|
No actionable comments were generated in the recent review. 🎉 ℹ️ Recent review info⚙️ Run configurationConfiguration used: Path: .coderabbit.yaml Review profile: CHILL Plan: Enterprise Run ID: 📒 Files selected for processing (1)
📝 WalkthroughWalkthroughThe standalone package test runner now passes ChangesStandalone test execution
Estimated code review effort: 2 (Simple) | ~5 minutes Suggested reviewers: 🚥 Pre-merge checks | ✅ 5✅ Passed checks (5 passed)
✨ Finishing Touches🧪 Generate unit tests (beta)
Comment |
Collaborator
|
/bot run |
Collaborator
|
PR_Github #59107 [ run ] triggered by Bot. Commit: |
Collaborator
|
PR_Github #59107 [ run ] completed with state
|
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
3 participants
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.
Summary
--dist=loadwhich dispatches tests dynamically across workers, breaking module-level torch.manual_seed(0) determinism and intermittently tripping tight tolerances in numerical child tests (flashinfer_mla, gemm_fusion) and mem_get_info-based assertions (piecewise_memory).--dist=loadfileto the child pytest cmd list so all tests in one file run on the same worker in file-declaration order; keep-n 4for wall-clock, but restore deterministic in-file execution so per-test RNG state is reproducible.Test plan
Links
Summary by CodeRabbit