Race condition fix in checkpoint loading util #3001
Merged
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
mvpatel2000
previously approved these changes
Feb 13, 2024
mvpatel2000
reviewed
Feb 13, 2024
eracah
reviewed
Feb 13, 2024
mvpatel2000
reviewed
Feb 13, 2024
Co-authored-by: Mihir Patel <mihir.v.patel7@gmail.com>
Co-authored-by: Mihir Patel <mihir.v.patel7@gmail.com>
Co-authored-by: Mihir Patel <mihir.v.patel7@gmail.com>
mvpatel2000
approved these changes
Feb 13, 2024
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Outdated suggestions cannot be applied.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.
What does this PR do?
With HSDP, there are multiple replicas of a model, and multiple shards per replica. There is a race condition when multiple shards try to download the same file, which causes higher network latency, resulting in a longer time needed for checkpoint resumption.
We fix the issue by deduplicating across the shards in the first replica (only the first replica downloads the files) by first gathering all the required files, and then checking if a lesser ranked shard is already downloading the file.