Skip to content

Fix CUDA graph correctness issues due to memory bugs - #5975

Merged
jiemingz merged 7 commits into
NVIDIA:mainfrom
jiemingz:jiemingz/cudagraph_fix
Jul 28, 2026
Merged

Fix CUDA graph correctness issues due to memory bugs#5975
jiemingz merged 7 commits into
NVIDIA:mainfrom
jiemingz:jiemingz/cudagraph_fix

Conversation

@jiemingz

@jiemingz jiemingz commented Jul 22, 2026 •
edited
Loading

Copy link
Copy Markdown
Contributor
  • I, the PR author, have personally reviewed every line of this PR.

What does this PR do?

This adds explicit control for activations saved for the backward pass via an observer. Tensors that need to be maintained until after the backward are explicitly marked. This also cleans up the cudagraph IO reuse code.

⚠️ For major changes (either in lines of code or in its impact), please make sure to first share a design doc with the team. If you're unsure what's the best way to do so, contact @NVIDIA/mcore-oncall.

Issue tracking

For PRs from open-source community contributors:

  • New features: a linked issue is required. Please open a feature request and reference it here before submitting the PR.
  • Small updates (bug fixes, minor improvements): a linked issue is recommended and will accelerate the PR review process.

Linked issue:

Contribution process

Pre-checks

  • I have added relevant unit tests
  • I have added relevant functional tests
  • I have added proper typing to my code Typing guidelines
  • I have added relevant documentation
  • I have run the autoformatter.sh on my PR

Code review

Feel free to message or comment @NVIDIA/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"

  1. When your PR is ready, click Ready for Review.
  2. An oncall reviewer is auto-assigned and expert reviewers are notified based on your changes.
    • Some PRs may jump straight to step 2. This is determined by .github/CODEOWNERS.

⚠️ Only mark as ready once merge-conflicts are resolved and the CI is passing.
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, the Final Review label 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 Approved label is applied automatically.

Merge

Any member of mcore-engineers will be able to merge your PR.

@jiemingz
jiemingz requested review from a team as code owners July 22, 2026 19:41
@copy-pr-bot

copy-pr-bot Bot commented Jul 22, 2026

Copy link
Copy Markdown

This pull request requires additional validation before any workflows can run on NVIDIA's runners.

Pull request vetters can view their responsibilities here.

Contributors can view more details about this message here.

@svcnvidia-nemo-ci
svcnvidia-nemo-ci marked this pull request as draft July 22, 2026 19:41
@github-actions

github-actions Bot commented Jul 22, 2026

Copy link
Copy Markdown
Contributor

This PR has been automatically converted to draft because all PRs must start as drafts.

When you are ready for review, click Ready for Review to begin the review process. This will:

  1. Add the oncall reviewer (optional reviewer)
  2. Add required review teams based on your changes

See the contribution guide for more details.

@jiemingz
jiemingz marked this pull request as ready for review July 23, 2026 14:25
@jiemingz

jiemingz commented Jul 23, 2026

Copy link
Copy Markdown
Contributor Author

/claude review

claude[bot]
claude Bot approved these changes Jul 23, 2026
View reviewed changes

@claude claude Bot left a comment

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

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

LGTM

mathemakitten
mathemakitten reviewed Jul 23, 2026
View reviewed changes
Comment thread tests/unit_tests/transformer/test_cuda_graphs.py Outdated
Comment thread megatron/core/transformer/cuda_graphs.py Outdated
Comment thread megatron/core/transformer/cuda_graphs.py
Comment thread megatron/core/transformer/cuda_graphs.py Outdated
@jiemingz
jiemingz requested a review from a team as a code owner July 24, 2026 16:11
@jiemingz
jiemingz force-pushed the jiemingz/cudagraph_fix branch 4 times, most recently from 772beb3 to 14dc52b Compare July 24, 2026 16:36
@jiemingz
fix cudagraph correctness introduced by previous fwd weakrefing
e43d623 
Signed-off-by: Jimmy Zhang <jiemingz@nvidia.com>
@jiemingz
jiemingz force-pushed the jiemingz/cudagraph_fix branch from 14dc52b to e43d623 Compare July 24, 2026 16:37
jiemingz added 2 commits July 24, 2026 09:57
@jiemingz
fix typo
189596e 
Signed-off-by: Jimmy Zhang <jiemingz@nvidia.com>
@jiemingz
loosen can_skip_reply_copy assertion
84b393d 
Signed-off-by: Jimmy Zhang <jiemingz@nvidia.com>
@svcnvidia-nemo-ci svcnvidia-nemo-ci added the Final Review PR is in the "final review" stage label Jul 24, 2026
@jiemingz
fix mtp convergnece
1acbca2 
Signed-off-by: Jimmy Zhang <jiemingz@nvidia.com>
deepakn94
deepakn94 reviewed Jul 27, 2026
View reviewed changes
Comment thread megatron/core/transformer/cuda_graphs.py Outdated
deepakn94
deepakn94 reviewed Jul 27, 2026
View reviewed changes
Comment thread megatron/core/transformer/cuda_graphs.py Outdated
@deepakn94 deepakn94 changed the title Fix cudagraph correctness due to memory bugs Fix CUDA graph correctness issues due to memory bugs Jul 27, 2026
@svcnvidia-nemo-ci svcnvidia-nemo-ci removed the Final Review PR is in the "final review" stage label Jul 27, 2026
@svcnvidia-nemo-ci svcnvidia-nemo-ci added the Approved All necessary approvals have been made label Jul 27, 2026
@jiemingz
cleanup docstrings
28f65b7 
Signed-off-by: Jimmy Zhang <jiemingz@nvidia.com>
@jiemingz

jiemingz commented Jul 27, 2026

Copy link
Copy Markdown
Contributor Author

/ok to test 28f65b7

@jiemingz
fix lint
d1265fd 
Signed-off-by: Jimmy Zhang <jiemingz@nvidia.com>
@jiemingz

jiemingz commented Jul 27, 2026

Copy link
Copy Markdown
Contributor Author

/ok to test d1265fd

@jiemingz
jiemingz enabled auto-merge July 27, 2026 22:14
@jiemingz
jiemingz disabled auto-merge July 27, 2026 22:24
@jiemingz
jiemingz added this pull request to the merge queue Jul 27, 2026
@svcnvidia-nemo-ci

svcnvidia-nemo-ci commented Jul 27, 2026

Copy link
Copy Markdown
Contributor

🔄 Merge queue validation started!

You can track the progress here: https://github.com/NVIDIA/Megatron-LM/actions/runs/30315203782

@svcnvidia-nemo-ci

svcnvidia-nemo-ci commented Jul 28, 2026

Copy link
Copy Markdown
Contributor

🔄 Merge queue validation started!

You can track the progress here: https://github.com/NVIDIA/Megatron-LM/actions/runs/30317256451

@svcnvidia-nemo-ci

svcnvidia-nemo-ci commented Jul 28, 2026

Copy link
Copy Markdown
Contributor

🔄 Merge queue validation started!

You can track the progress here: https://github.com/NVIDIA/Megatron-LM/actions/runs/30326597767

Merged via the queue into NVIDIA:main with commit 4b18b26 Jul 28, 2026
316 of 328 checks passed
@jiemingz
jiemingz deleted the jiemingz/cudagraph_fix branch July 28, 2026 06:34
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

@deepakn94 deepakn94 deepakn94 approved these changes
@mathemakitten mathemakitten mathemakitten approved these changes
@yaoyu-33 yaoyu-33 yaoyu-33 approved these changes
@claude claude[bot] claude[bot] approved these changes

Assignees

No one assigned

Labels

Approved All necessary approvals have been made complexity: medium

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

5 participants

@jiemingz @svcnvidia-nemo-ci @deepakn94 @mathemakitten @yaoyu-33