Harden pinned NUMA mempool tests against constructor OOM flakes#2096
Merged
rwgk merged 1 commit intoMay 16, 2026
Merged
Conversation
Route pinned NUMA-ID constructor tests through the existing Windows MCDM mempool OOM helper so they stay focused on NUMA semantics instead of failing on the known constructor flake.
rwgk
added
P0
Contributor
|
Auto-sync is disabled for draft pull requests in this repository. Workflows must be run manually. Contributors can view more details about this message here. |
Contributor
Author
|
/ok to test |
This comment has been minimized.
This comment has been minimized.
Member
|
btw nightly also failed for the OOM issue (which I don't quite understand why would happen): https://github.com/NVIDIA/cuda-python/actions/runs/25947528269/job/76279804206#step:26:17561 Merging not because I agree we need to address cuda-core test issues during CTK bring-up, but because this actually blocks our CI and thus development, as seen in #2087. |
leofang
approved these changes
May 16, 2026
|
rwgk
deleted the
nvbugs5815123_addl_create_pinned_memory_resource_or_xfail
branch
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
2 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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
This is a small follow-up to PR #2084.
The
cuda_core_5.15.logattached tonvbug 5815123shows two remaining intermittentcuda_corefailures after the PR #2084 patch:tests/test_memory.py::test_pinned_mr_numa_id_default_no_ipctests/test_memory.py::test_pinned_mr_numa_id_explicitBoth tests are meant to validate
PinnedMemoryResource.numa_idbehavior, but they instantiatePinnedMemoryResource(PinnedMemoryResourceOptions(...))directly. Those constructor paths create a real pinned memory pool immediately, so they can still hit the same constructor-timeCUDA_ERROR_OUT_OF_MEMORYfailure mode that is already handled elsewhere in the Windows mempool workaround coverage.This change routes those constructor sites through the existing
create_pinned_memory_resource_or_xfail(...)helper. It also updates the nearbytest_pinned_mr_numa_id_default_with_ipccase for consistency, since it exercises the same pool-creation path.The goal is to keep these tests focused on NUMA-ID semantics instead of failing on the known Windows MCDM mempool-constructor flake.
xref:
nvbug 5815123