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

[RFC] Standardize Collective (AllGatherOp, AllReduceOp, AllToAllOp) ops to support variadic operand/result #2099

Merged
merged 12 commits into from
Jun 26, 2024

Conversation

abhigunj
Copy link
Member

@abhigunj abhigunj commented Mar 15, 2024

This RFC proposes standardizing collective (AllGatherOp, AllReduceOp, AllToAllOp) ops to enable support for multi-operand and multi-result

Please review and provide you feedback.

@abhigunj abhigunj changed the title [RFC] Standardize Tuple-collective (AllGatherOp, AllReduceOp, AllToAllOp) ops [RFC] Standardize Collective (AllGatherOp, AllReduceOp, AllToAllOp) ops to support variadic operand/result Mar 15, 2024
@abhigunj abhigunj added the RFC label Mar 15, 2024
docs/spec.md Outdated Show resolved Hide resolved
Copy link
Member

@ghpvnist ghpvnist left a comment

Choose a reason for hiding this comment

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

Could you update the examples to showcase how this works?

Copy link
Member

@sdasgup3 sdasgup3 left a comment

Choose a reason for hiding this comment

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

LGTM once the remaining request related to providing example is fixed.

rfcs/20240312-standardize-collective-ops.md Outdated Show resolved Hide resolved
rfcs/20240312-standardize-collective-ops.md Outdated Show resolved Hide resolved
rfcs/20240312-standardize-collective-ops.md Outdated Show resolved Hide resolved
@abhigunj abhigunj requested a review from ghpvnist June 26, 2024 00:57
@GleasonK GleasonK merged commit 6b69e21 into openxla:main Jun 26, 2024
10 checks passed
abhigunj added a commit that referenced this pull request Jul 22, 2024
The PR allows each Process from the ProcessGrid to contribute more than
one tensor at `ProcessGrid::rendezvous`.
This is the first set of change, will be followed by interpreter updates
to collectives.

Note: The PR does **not** make collectives interpreter variadic. 

Tested: 
1. No existing test failures indicate no change in behavior for ops
using the `rendezvous`
2. The diff is tested with new variadic interpreter for `all_reduce` op.
Will upload the PR soon.


ref: #2099
abhigunj added a commit that referenced this pull request Jul 29, 2024
…2450)

RFC : #2099

* updated the op spec and relevant tests
* updated interpreter
* required VHLO changes
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
Projects
None yet
Development

Successfully merging this pull request may close these issues.

6 participants