Skip to content

[DCP] Add DefaultStager example to distributed async checkpoint recipe #3711

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.

Sign up for GitHub

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

Jump to bottom
Open
niyunsheng wants to merge 6 commits into
base: main
Choose a base branch
from
Open

[DCP] Add DefaultStager example to distributed async checkpoint recipe #3711

niyunsheng wants to merge 6 commits into from
+155 −7

Conversation

@niyunsheng
Copy link

@niyunsheng niyunsheng commented Dec 29, 2025 •
edited by pytorch-bot bot
Loading

Fixes #3710

Description

This PR updates the distributed_async_checkpoint_recipe to include the DefaultStager functionality introduced in PyTorch 2.9.

Motivation:
In large-scale training, even with standard async_save, the initial memory copy (Staging phase, GPU -> CPU) occurs on the main thread. This blocks the training loop. This PR introduces DefaultStager, which offloads this copy to a background thread, enabling full computation-communication overlap.

Key Changes:

  1. New Section: Added "Fully Asynchronous Staging with DefaultStager" to the recipe.
  2. Version Note: Added .. versionadded:: 2.9 to indicate version requirements.
  3. Advanced Example: Provided a code example demonstrating how to overlap the D2H copy with the entire Forward and Backward pass:
    • Check staging_completion after backward but before optimizer.step() to ensure data consistency while maximizing parallel execution.
    • Check upload_completion before the next save to manage memory backpressure.
  4. Authors: Added myself to the author list.

Checklist

  • The issue that is being fixed is referred in the description (see above "Fixes #ISSUE_NUMBER")
  • Only one issue is addressed in this pull request
  • Labels from the issue that this PR is fixing are added to this pull request
  • No unnecessary issues are included into this pull request.

cc @LucasLLC @MeetVadakkanchery @mhorowitz @pradeepfn @ekr0 @haochengsong @Saiteja64

@pytorch-bot
Copy link

pytorch-bot bot commented Dec 29, 2025 •
edited
Loading

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/tutorials/3711

Note: Links to docs will display an error until the docs builds have been completed.

❗ 1 Active SEVs

There are 1 currently active SEVs. If your PR is affected, please view them below:

This comment was automatically generated by Dr. CI and updates every 15 minutes.

@meta-cla
Copy link

meta-cla bot commented Dec 29, 2025

Hi @niyunsheng!

Thank you for your pull request and welcome to our community.

Action Required

In order to merge any pull request (code, docs, etc.), we require contributors to sign our Contributor License Agreement, and we don't seem to have one on file for you.

Process

In order for us to review and merge your suggested changes, please sign at https://code.facebook.com/cla. If you are contributing on behalf of someone else (eg your employer), the individual CLA may not be sufficient and your employer may need to sign the corporate CLA.

Once the CLA is signed, our tooling will perform checks and validations. Afterwards, the pull request will be tagged with CLA signed. The tagging process may take up to 1 hour after signing. Please give it that time before contacting us about it.

If you have received this in error or have any questions, please contact us at [email protected]. Thanks!

@meta-cla meta-cla bot added the cla signed label Dec 29, 2025
@meta-cla
Copy link

meta-cla bot commented Dec 29, 2025

Thank you for signing our Contributor License Agreement. We can now accept your code for this (and any) Meta Open Source project. Thanks!

@sekyondaMeta sekyondaMeta requested a review from LucasLLC January 5, 2026 15:30
@niyunsheng
Copy link
Author

niyunsheng commented Jan 13, 2026

Hi @LucasLLC, gentle ping on this.

Just to clarify context: This tutorial covers an existing feature (DefaultStager) and is independent of my other open PRs. It does not require any upstream changes.

It has been passing CI for two weeks. Could you please take a look when you have a moment? Thanks!

@AlannaBurke
Copy link
Contributor

AlannaBurke commented Jan 14, 2026

Preview link: https://docs-preview.pytorch.org/pytorch/tutorials/3711/recipes/distributed_async_checkpoint_recipe.html

AlannaBurke
AlannaBurke requested changes Jan 14, 2026
View reviewed changes
Copy link
Contributor

@AlannaBurke AlannaBurke left a comment

Choose a reason for hiding this comment

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

Just a few grammar and minor changes. Otherwise this looks good to me.

@niyunsheng
Copy link
Author

niyunsheng commented Jan 14, 2026

Hi @AlannaBurke , thanks for the review!

I've applied your grammar and other suggestions.

Extra Fix: While previewing the file locally (make html-noplot), I noticed another small display issue in this file that was missing from the review, so I went ahead and fixed that as well.

The file should be perfect now. Ready for another look!

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

@LucasLLC LucasLLC Awaiting requested review from LucasLLC

@AlannaBurke AlannaBurke Awaiting requested review from AlannaBurke

Assignees

No one assigned

Labels

cla signed module: distributed_checkpoint

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

[DCP] Add DefaultStager example to distributed async checkpoint recipe

3 participants

@niyunsheng @AlannaBurke @svekars