Skip to content

Fix hip memadvise events regression #12333

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Closed

Conversation

GeorgeWeb
Copy link
Contributor

Updated the event discarding lit tests that check the piextUSMEnqueueMemAdvise entry point. The return code can also be PI_ERROR_PLUGIN_SPECIFIC_ERROR (-996) if the driver returns that managed memory is not supported for the device.

@sergey-semenov
Copy link
Contributor

Could you please add a comment to the test explaining why -996 is an acceptable result?

@GeorgeWeb
Copy link
Contributor Author

Could you please add a comment to the test explaining why -996 is an acceptable result?

Yeah I've just added it. It's used as a warning in mem advise similarly to how it was previously used in the prefetch implementation for CUDA.

@GeorgeWeb
Copy link
Contributor Author

Closing as I will be moving those test changes to the relevant PR #10697. I confused myself thinking it was already merged (what was actually merged was its unified-runtime dependency), hence I created this PR to fix the regression. First day after holidays haha. Thanks for the review though @sergey-semenov!

@GeorgeWeb GeorgeWeb closed this Jan 9, 2024
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

2 participants