This repository has been archived by the owner on Sep 18, 2024. It is now read-only.
Disable GPU in unit test #4025
Merged
Disable GPU in unit test #4025
Conversation
This file contains 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
QuanluZhang
approved these changes
Aug 5, 2021
|
I triggered an IT to test this PR. Please ensure it passes before merge. https://msrasrg.visualstudio.com/NNIOpenSource/_build/results?buildId=25918 |
ultmaster
approved these changes
Aug 5, 2021
Sign up for free
to subscribe to this conversation on GitHub.
Already have an account?
Sign in.
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Outdated suggestions cannot be applied.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.
Currently the UT will not exit if the machine has GPU. This is because local training service test case will open GPU metric file and will never close it. We could add closing logic to fix it, but I think it's not the right way.
Our default UT environment is CPU-only, and by design an experiment with GPU number > 1 should fail on this type of machine. The test case was passing by accident because the error of GPU scheduler is silently ignored. And since it failed to generate GPU metric file, there is no closing issue.
This does not harm real experiments because they are stopped by
process.exit(), which will not be blocked by file handlers.