[DEV] fix(megatron-fsdp): compute SWiGLU/GDN split in item coordinates for non-DTensor optimizer states#4424
Closed
xuwchen wants to merge 2 commits into
Closed
[DEV] fix(megatron-fsdp): compute SWiGLU/GDN split in item coordinates for non-DTensor optimizer states#4424xuwchen wants to merge 2 commits into
xuwchen wants to merge 2 commits into
Conversation
This was referenced May 13, 2026
Contributor
Author
|
/ok to test 7ec906c |
shjwudp
approved these changes
May 15, 2026
Autumn1998
approved these changes
May 19, 2026
Contributor
Author
|
The changes here have been incorporated into #4748, which is now merged. Closing this PR. |
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
main PR: #4423
What does this PR do ?
split_swiglu_linear_fc1andsplit_gdn_fusedoriginally useddata.numel() // tp_mesh.mesh.numel()to compute component slices (w_slice/v_slicefor SWiGLU,comp_slicefor GDN), then intersected them withfsdp_sliceto determine each rank's portion. Whendatais a DTensor,data.numel()returns the global element count (=param.numel()), so component slices andfsdp_sliceare both in item coordinates and the intersection is valid. However, whendatais a plain Tensor (optimizer states created by FusedAdam),data.numel()returns the local shard size on this rank, making the component slices's range far smaller than the item coordinate space and resulting in an intersection across incompatible coordinate systems.The fix replaces
data.numel()withdist_param.numel()(dist_paramis always a DTensor and therefore reflects the global element count), ensuringfsdp_sliceand all component slices are in item coordinates.Contribution process
Pre-checks
Code review
Feel free to message or comment the @mcore-oncall to help accelerate your merge into main. The less complex your PR is, the faster it will be approved and merged!
All PRs start as draft. If you open a non-draft PR, it will be automatically converted to draft.
Step 1: Mark PR as "Ready for Review"
.github/CODEOWNERS.Final Review might get declined if these requirements are not fulfilled.
Step 2: Final Review
For PRs that change
megatron/core, once all expert reviewers have approved, theFinal Reviewlabel is applied automatically and final reviewers are assigned.For PRs outside
megatron/core, this step is skipped.Step 3: Approved
Once all required reviewers have approved, the
Approvedlabel is applied automatically.Merge
Any member of mcore-engineers will be able to merge your PR.
For MRs into `dev` branch
The proposed review process for `dev` branch is under active discussion.MRs are mergable after one approval by either
eharper@nvidia.comorzijiey@nvidia.com.