Add reconfigurable support to match MCCL behavior (#2077)#2077
Open
d4l3k wants to merge 1 commit intometa-pytorch:mainfrom
Open
Add reconfigurable support to match MCCL behavior (#2077)#2077d4l3k wants to merge 1 commit intometa-pytorch:mainfrom
d4l3k wants to merge 1 commit intometa-pytorch:mainfrom
Conversation
meta-cla
bot
added
the
CLA Signed
Contributor
|
@d4l3k has exported this pull request. If you are a Meta employee, you can view the originating Diff in D100904495. |
meta-codesync
bot
changed the title
d4l3k
added a commit
to d4l3k/torchcomms-1
that referenced
this pull request
Apr 15, 2026
Summary:
Implements fault-tolerant reconfiguration for the Gloo backend by adding store-based atomic rank assignment during `reconfigure()`. Each process now obtains a unique rank via `rankStore->add("rank_counter", 1) - 1`, scoped to a per-reconfigure UUID prefix, matching MCCL's dynamic regime behavior. The implementation defers context initialization when `enable_reconfigure` is set and resets state during reconfiguration.
Adds comprehensive test coverage including basic reconfigure, unordered handle sets, send/recv after reconfigure, and allreduce verification.
---
AI generated Summary & Test Plan from DEV175861935
Reviewed By: dolpm
Differential Revision: D100904495
d4l3k
force-pushed
the
export-D100904495
branch
from
ec099d9 to
039db9b
Compare
Summary:
Implements fault-tolerant reconfiguration for the Gloo backend by adding store-based atomic rank assignment during `reconfigure()`. Each process now obtains a unique rank via `rankStore->add("rank_counter", 1) - 1`, scoped to a per-reconfigure UUID prefix, matching MCCL's dynamic regime behavior. The implementation defers context initialization when `enable_reconfigure` is set and resets state during reconfiguration.
Adds comprehensive test coverage including basic reconfigure, unordered handle sets, send/recv after reconfigure, and allreduce verification.
---
AI generated Summary & Test Plan from DEV175861935
Reviewed By: dolpm
Differential Revision: D100904495
d4l3k
force-pushed
the
export-D100904495
branch
from
039db9b to
bfcabfe
Compare
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
1 participant
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.
Summary:
Implements fault-tolerant reconfiguration for the Gloo backend by adding store-based atomic rank assignment during
reconfigure(). Each process now obtains a unique rank viarankStore->add("rank_counter", 1) - 1, scoped to a per-reconfigure UUID prefix, matching MCCL's dynamic regime behavior. The implementation defers context initialization whenenable_reconfigureis set and resets state during reconfiguration.Adds comprehensive test coverage including basic reconfigure, unordered handle sets, send/recv after reconfigure, and allreduce verification.
AI generated Summary & Test Plan from DEV175861935
Reviewed By: dolpm
Differential Revision: D100904495