Open
Conversation
Add `mx.distributed.all_to_all(x)` — splits input along axis 0, sends chunk i to rank i, and concatenates received chunks. - CPU eval via backend-specific GroupImpl::all_to_all - MPI backend (MPI_Alltoall) and JACCL backend (RDMA pipelined) - VJP support (all_to_all is its own transpose) - GPU/CUDA stubs (not-implemented exceptions) Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
7 tasks
Member
|
Great I 'll take a look soon |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Add
all_to_allcollective primitive for exchanging equal-sized chunks between all ranks. First in a series of PRs adding Expert Parallelism support.mx.distributed.all_to_all(x)— splits input along axis 0, sends chunk i to rank i, concatenates received chunksGroupImpl::all_to_allall_to_allis its own transpose)Test plan
pytest python/tests/mlx_distributed_tests.py -k all_to_all— 4 passed, 1 skipped (shape validation requires ws>1)🤖 Generated with Claude Code