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.

Sign up for GitHub

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

Jump to bottom

legate runtime maps distributed data to single partition #69

Open
mfoerste4 opened this issue Oct 19, 2023 · 1 comment
Open

legate runtime maps distributed data to single partition #69

mfoerste4 opened this issue Oct 19, 2023 · 1 comment

Comments

@mfoerste4
Copy link
Contributor

When running legateboost on multiple GPUs, the legate runtime maps followup cunumeric tasks to individual instances despite the data being distributed. The issue is being described and discussed in more detail here.

This causes performance to degrade as some code portions are being executed sequentially.

Current workaround is to run distributed code with the environment variable LEGATE_MIN_GPU_CHUNK=1 or LEGATE_TEST=1.
@RAMitchell
Copy link
Contributor

Is this potentially the same issue as the cunumeric repeat issue we just found?

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
2 participants
@RAMitchell @mfoerste4