Skip to content

Conversation

@crcrpar
Copy link
Collaborator

@crcrpar crcrpar commented Jun 25, 2025

apex.parallel was removed in #1896

Signed-off-by: Masaki Kozuki <[email protected]>
@crcrpar crcrpar requested a review from Copilot June 25, 2025 23:35
Copy link
Contributor

Copilot AI left a comment

Choose a reason for hiding this comment

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

Pull Request Overview

This PR updates the implementation of flat_dist_call within the test suite to replace the old distributed flat call mechanism with one that leverages _coalescing_manager for managing asynchronous operations.

  • Removed the old dependency on apex.parallel.distributed.flat_dist_call
  • Introduced a new flat_dist_call function using _coalescing_manager and a subsequent cm.wait() call to ensure completion of operations
Comments suppressed due to low confidence (1)

apex/contrib/test/optimizers/test_distributed_fused_lamb.py:12

  • [nitpick] Consider adding a brief docstring to flat_dist_call that explains its role and the asynchronous behavior managed by _coalescing_manager for future maintainability.
def flat_dist_call(param_list: list[torch.Tensor], op, args):

@crcrpar crcrpar changed the title replace with coalescing manager [contrib][DistributedFusedLamb] replace apex.parallel.flat_dist_call with coalescing manager in test Jun 25, 2025
@crcrpar crcrpar merged commit d8200a1 into master Jun 26, 2025
@crcrpar crcrpar deleted the replace-removed-flat_dist_call branch June 26, 2025 03:52
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.

3 participants