Skip to content

implement collective all_to_all op #9442

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
Open
wants to merge 3 commits into
base: master
Choose a base branch
from
Open

implement collective all_to_all op #9442

wants to merge 3 commits into from
+48 −4

Conversation

bfolie
Copy link
Collaborator

@bfolie bfolie commented Jul 2, 2025

@bfolie bfolie mentioned this pull request Jul 1, 2025
Open
@bfolie bfolie requested review from pgmoka and benawilson July 2, 2025 19:29
pgmoka
pgmoka approved these changes Jul 8, 2025
View reviewed changes
Copy link
Collaborator

@pgmoka pgmoka left a comment

Choose a reason for hiding this comment

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

LGTM

yaoshiang
yaoshiang reviewed Jul 8, 2025
View reviewed changes
Copy link
Collaborator

@yaoshiang yaoshiang left a comment •
edited
Loading

Choose a reason for hiding this comment

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

it seems like the implementation has always been there but just not exposed? Any thoughts on why that might be?

yaoshiang
yaoshiang reviewed Jul 8, 2025
View reviewed changes
yaoshiang
yaoshiang reviewed Jul 8, 2025
View reviewed changes
@bfolie
Copy link
Collaborator Author

bfolie commented Jul 8, 2025

it seems like the implementation has always been there but just not exposed? Any thoughts on why that might be?

How do you mean? There are two torch.distributed functions: all_to_all_single and all_to_all. The former is already implemented and exposed. This PR implements the latter. Both use the same underlying xla function, xla_model.all_to_all.

@bfolie bfolie requested a review from pgmoka July 11, 2025 18:13
@bfolie bfolie requested a review from yaoshiang July 14, 2025 16:56
@bfolie bfolie requested a review from pgmoka July 15, 2025 14:40
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers

@qihqi qihqi qihqi approved these changes

@pgmoka pgmoka pgmoka approved these changes

@benawilson benawilson Awaiting requested review from benawilson

@yaoshiang yaoshiang Awaiting requested review from yaoshiang

Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.
4 participants
@bfolie @qihqi @yaoshiang @pgmoka