[NCCL] create NCCL communicator for send/recv on demand #44922
Closed
Conversation
This file contains 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
For NCCL send/recv operations, we will create NCCL communicator on demand following the same design as how it's currently done for collective operations. Differential Revision: [D23773726](https://our.internmc.facebook.com/intern/diff/D23773726/) [ghstack-poisoned]
mingzhe09088
requested review from
mrshenli,
pietern,
pritamdamania87,
rohan-varma and
zhaojuanmao
as code owners
mingzhe09088
pushed a commit
that referenced
this pull request
Sep 18, 2020
For NCCL send/recv operations, we will create NCCL communicator on demand following the same design as how it's currently done for collective operations. Differential Revision: [D23773726](https://our.internmc.facebook.com/intern/diff/D23773726/) ghstack-source-id: 112338493 Pull Request resolved: #44922
💊 CI failures summary and remediationsAs of commit 50d40ea (more details on the Dr. CI page): Commit 50d40ea was recently pushed. Waiting for builds... This comment was automatically generated by Dr. CI (expand for details).Follow this link to opt-out of these comments for your Pull Requests.Please report bugs/suggestions on the GitHub issue tracker or post in the (internal) Dr. CI Users group. This comment has been revised 18 times. |
pritamdamania87
suggested changes
Sep 18, 2020
For NCCL send/recv operations, we will create NCCL communicator on demand following the same design as how it's currently done for collective operations. Differential Revision: [D23773726](https://our.internmc.facebook.com/intern/diff/D23773726/) [ghstack-poisoned]
mingzhe09088
pushed a commit
that referenced
this pull request
Sep 21, 2020
Pull Request resolved: #44922 For NCCL send/recv operations, we will create NCCL communicator on demand following the same design as how it's currently done for collective operations. ghstack-source-id: 112511958 Differential Revision: [D23773726](https://our.internmc.facebook.com/intern/diff/D23773726/)
pritamdamania87
approved these changes
Sep 23, 2020
There was a problem hiding this comment.
LGTM! Although, I just realized that we should add a test to validate what happens when we send to self? For example, does NCCL support sending to rank 0 from rank 0? If so, we should add a test to validate our APIs support this correctly. If NCCL doesn't support this, we should add tests validating that our APIs raise an appropriate error.
|
NCCL supports sending to self. Let me add a test for that. |
For NCCL send/recv operations, we will create NCCL communicator on demand following the same design as how it's currently done for collective operations. Differential Revision: [D23773726](https://our.internmc.facebook.com/intern/diff/D23773726/) [ghstack-poisoned]
mingzhe09088
pushed a commit
that referenced
this pull request
Sep 24, 2020
Pull Request resolved: #44922 For NCCL send/recv operations, we will create NCCL communicator on demand following the same design as how it's currently done for collective operations. ghstack-source-id: 112836264 Differential Revision: [D23773726](https://our.internmc.facebook.com/intern/diff/D23773726/)
For NCCL send/recv operations, we will create NCCL communicator on demand following the same design as how it's currently done for collective operations. Differential Revision: [D23773726](https://our.internmc.facebook.com/intern/diff/D23773726/) [ghstack-poisoned]
This was referenced Sep 25, 2020
For NCCL send/recv operations, we will create NCCL communicator on demand following the same design as how it's currently done for collective operations. Differential Revision: [D23773726](https://our.internmc.facebook.com/intern/diff/D23773726/) [ghstack-poisoned]
mingzhe09088
pushed a commit
that referenced
this pull request
Oct 5, 2020
Pull Request resolved: #44922 For NCCL send/recv operations, we will create NCCL communicator on demand following the same design as how it's currently done for collective operations. ghstack-source-id: 113592757 Differential Revision: [D23773726](https://our.internmc.facebook.com/intern/diff/D23773726/)
|
This pull request has been merged in 10d86d1. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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.
Stack from ghstack:
For NCCL send/recv operations, we will create NCCL communicator on demand following the same design as how it's currently done for collective operations.
Differential Revision: D23773726