Skip to content

Populate dp process group in auto-built ProcessGroupCollection in pipeline schedules - #5901

Merged
ilml merged 2 commits into
NVIDIA:mainfrom
ilml:split/3386-01-schedules-pg-dp
Jul 28, 2026
Merged

Populate dp process group in auto-built ProcessGroupCollection in pipeline schedules#5901
ilml merged 2 commits into
NVIDIA:mainfrom
ilml:split/3386-01-schedules-pg-dp

Conversation

@ilml

@ilml ilml commented Jul 20, 2026

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

What does this PR do?

Populates pg_collection.dp in the pg_collection is None fallback of the three forward-backward schedules (no-pipelining, interleaved, non-interleaved), matching the MPU default groups. Groundwork for sequence-packing stats collectives; no behavior change for existing callers.

Part 01/10 of splitting #3386 (Add E2E support for THD format; dev-branch PR #2924). Original changes by @xiaoyao0115 in #3386 — split into functionally self-contained PRs to ease review. Hard dependencies (must merge first): none — independently mergeable.

Split series (#3386)

Part PR Hard deps (merge first)
01 #5901
02 #5902
03 #5903
04 #5904 #5903
05 #5905 #5904
06 #5906 #5902, #5903
07 #5907 #5902, #5904, #5905
08 #5908
09 #5909 #5906, #5908
10 #5910 #5907, #5909

Branches are stacked linearly (each on the previous) so every PR shows a clean own-diff once its base is retargeted to the copy-pr-bot pull-request/<parent> ref; until then the Files-changed view of a stacked PR includes its ancestors — its own change is the last commit.

Issue tracking

Linked issue: Related to #3386

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

🤖 Generated with Claude Code

…eline schedules

Split from NVIDIA#3386 (sequence packing / THD E2E support).
Original changes by @xiaoyao0115 in NVIDIA#3386.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Signed-off-by: ilml <tolong@nvidia.com>
@copy-pr-bot

copy-pr-bot Bot commented Jul 20, 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.

@ilml
ilml marked this pull request as ready for review July 22, 2026 01:02
@ilml
ilml requested review from a team as code owners July 22, 2026 01:02
@svcnvidia-nemo-ci svcnvidia-nemo-ci added the Final Review PR is in the "final review" stage label Jul 22, 2026
Comment thread megatron/core/pipeline_parallel/schedules.py Outdated
…nto a helper

The fallback pg_collection construction was repeated in all three
schedule entry points (no-pipelining, interleaved, non-interleaved).
Consolidate it into _build_default_pg_collection(); the interleaved and
non-interleaved sites keep their local tp_group/cp_group/cp_size
bindings by reading them off the returned collection.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Signed-off-by: Tom Long <tolong@nvidia.com>
@svcnvidia-nemo-ci svcnvidia-nemo-ci added Approved All necessary approvals have been made and removed Final Review PR is in the "final review" stage labels Jul 27, 2026
@ilml

ilml commented Jul 28, 2026

Copy link
Copy Markdown
Contributor Author

/ok to test be08e66

@svcnvidia-nemo-ci

Copy link
Copy Markdown
Contributor

🔄 Merge queue validation started!

You can track the progress here: https://github.com/NVIDIA/Megatron-LM/actions/runs/30339663477

Merged via the queue into NVIDIA:main with commit 6f7bcd4 Jul 28, 2026
85 checks passed
@ilml
ilml deleted the split/3386-01-schedules-pg-dp branch July 28, 2026 10:12
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

Approved All necessary approvals have been made complexity: low

Projects

None yet

Development

Successfully merging this pull request may close these issues.

5 participants