Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Use nccl group calls to prevent from dead lock. #4113

Merged
merged 1 commit into from
Feb 7, 2019
Merged

Use nccl group calls to prevent from dead lock. #4113

merged 1 commit into from
Feb 7, 2019

Conversation

trivialfis
Copy link
Member

  • launch all reduce sequentially.
  • Fix gpu_exact test memory leak.

close #4107

@trivialfis
Use nccl group calls to prevent from dead lock.
52afe3f 
* launch all reduce sequentially.
* Fix gpu_exact test memory leak.
@trivialfis trivialfis requested a review from RAMitchell February 7, 2019 13:25
@trivialfis
Copy link
Member Author

#4095 should fix this in a better way. But before that PR is fully baked, here is a quick fix.

@trivialfis trivialfis merged commit f8ca296 into dmlc:master Feb 7, 2019
@trivialfis trivialfis deleted the fix/nccl-hangs branch February 9, 2019 21:45
@hcho3 hcho3 mentioned this pull request Mar 4, 2019
Merged
@lock lock bot locked as resolved and limited conversation to collaborators May 10, 2019
Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.
Reviewers

@RAMitchell RAMitchell RAMitchell approved these changes

Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.

xgboost for GPU hangs on AllReduce and CPU Barrier
2 participants
@trivialfis @RAMitchell