Reset dynamo state in test_memory_compile_regions#3443
Open
amd-maradosa wants to merge 1 commit into
Open
Conversation
|
Jenkins build for ad54c5992eda08520997db8ec488eb78d7542809 commit finished as FAILURE |
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
ROCm/TheRock#6173
Motivation
test_memory_compile_regionsintest/test_cuda.pyasserts on Dynamo compile-region counts. When the test runs in the same process as other compile tests (e.g. a fulltest_cuda.pyrun), compiled artifacts cached by earlier tests leak into this test, making the asserted counts non-deterministic and the test flaky. This PR isolates the test from prior Dynamo state.CI fail
Technical Details
Added
torch._dynamo.reset()at the start oftest_memory_compile_regions, beforetorch.cuda.memory.empty_cache(), so the test starts from a clean Dynamo cache regardless of what ran before it.Test Plan
Reproduced locally with a small prelude test that runs
torch.compile(...)on CUDA before the affected test, in the same process:python -m pytest .py test_cuda.py -k "<prelude_test> or test_memory_compile_regions" -v
Test Result
Before fix:
- standalone: PASSED
- with a compiling test first: FAILED
After fix:
- standalone: PASSED
- with a compiling test first: PASSED
Submission Checklist