Fix bucket sizes for AutoParallel 1D #1545
Merged
Conversation
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
meta-cla
bot
added
the
CLA Signed
fmassa
commented
Aug 8, 2025
| (global_batch_size, job_config.training.seq_len), | ||
| device=torch.device("cuda"), | ||
| ), | ||
| return ( |
Member
Author
There was a problem hiding this comment.
Sorry for the unrelated lint changes, my editor decided to annoy me here
wconstab
approved these changes
Aug 8, 2025
| assert parallel_dims.pp_enabled is False, "PP not supported yet" | ||
|
|
||
| torch._inductor.config.bucket_all_gathers_fx_bucket_size_determinator = ( | ||
| lambda bucket_idx: 500 / parallel_dims.tp |
Contributor
There was a problem hiding this comment.
We may want to consolidate at some point, the other inductor configs that control which passes and pass modes live in titan today, are cli driven, and could change even which bucketing pass is used.
Contributor
There was a problem hiding this comment.
Oops just realized this is totchtitan..
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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Outdated suggestions cannot be applied.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.
This PR makes bucket sizes for all-gather and reduce-scatter to be of the same size for 1d FSDP.