Skip to content
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

Scrub use of explicit PJRT_DEVICE=CUDA for Python CI/CD and rely on autodetection #6623

Closed
wants to merge 3 commits into from

Conversation

changm
Copy link
Collaborator

@changm changm commented Feb 27, 2024

Can't do it for CPP tests since CPP uses a different PJRT C++ logic to directly read the PJRT_DEVICE to initialize.

@changm changm self-assigned this Feb 27, 2024
@changm changm changed the title Scrub use of explicit PJRT_DEVICE=CUDA everywhere and rely on autodetection Scrub use of explicit PJRT_DEVICE=CUDA for Python CI/CD and rely on autodetection Feb 27, 2024
@changm changm requested a review from will-cromar February 29, 2024 14:26
Copy link
Collaborator

@will-cromar will-cromar left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

We should leave PJRT_DEVICE=CUDA set for almost all of our tests. We'd rather our CI crash if there was an issue with the GPU client vs silently falling back to CPU. See my comment here: #6605 (comment)

@changm
Copy link
Collaborator Author

changm commented Feb 29, 2024

Ahh ack ok sorry I misread that versus a different comment. I'll close this then. Thanks!

@changm changm closed this Feb 29, 2024
@changm changm deleted the changm/gpu branch February 29, 2024 17:46
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