feat: add NCCL flight recorder configuration support#3806
Open
sbak5 wants to merge 3 commits intoNVIDIA:mainfrom
Open
feat: add NCCL flight recorder configuration support#3806sbak5 wants to merge 3 commits intoNVIDIA:mainfrom
sbak5 wants to merge 3 commits intoNVIDIA:mainfrom
Conversation
Add flight recorder config fields to DistributedInitConfig and set the corresponding env vars (TORCH_FR_DUMP_TEMP_FILE, TORCH_NCCL_TRACE_BUFFER_SIZE, etc.) before distributed init when flight_recorder_dump_path is specified. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
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:
See the contribution guide for more details. |
Contributor
Author
|
/claude review |
![claude[bot]](https://avatars.githubusercontent.com/in/1236702?s=60&v=4){kind=small}
…log effective values Use os.environ.setdefault semantics for optional flight recorder env vars (TORCH_NCCL_TRACE_BUFFER_SIZE, TORCH_NCCL_DUMP_ON_TIMEOUT, etc.) so that values already set in the environment are respected. Emit a warning via warn_rank_0 for each var that is skipped, and print all effective flight recorder env var values via print_rank_0 once at startup. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Contributor
Author
|
/claude review |
Contributor
There was a problem hiding this comment.
Looks good overall — clean implementation with nice setdefault semantics for the optional env vars.
One issue: TORCH_FR_DUMP_TEMP_FILE and TORCH_NCCL_DEBUG_INFO_TEMP_FILE are unconditionally overwritten (lines 342-343), unlike the other vars which warn when a pre-existing value is found. Left an inline comment with a suggestion to at least warn before overriding.
… environment Apply setdefault semantics to TORCH_FR_DUMP_TEMP_FILE and TORCH_NCCL_DEBUG_INFO_TEMP_FILE consistently with all other flight recorder vars. The block now also fires when either path var is already present in the environment (without the MLM arg), so the remaining defaults (buffer size, dump-on-timeout, etc.) are applied in all cases. Priority order: pre-existing environment variable > MLM argument. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Contributor
Author
|
/claude review |
deepakn94
approved these changes
Mar 11, 2026
ericharper
approved these changes
Mar 12, 2026
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
4 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.
Add flight recorder config fields to DistributedInitConfig and set the corresponding env vars (TORCH_FR_DUMP_TEMP_FILE, TORCH_NCCL_TRACE_BUFFER_SIZE, etc.) before distributed init when flight_recorder_dump_path is specified.
What does this PR do ?
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.