Support multiple cuDecomp handles.#139
Merged
Merged
Conversation
Signed-off-by: romerojosh <joshr@nvidia.com>
Collaborator
Author
|
/build |
|
🚀 Build workflow triggered! View run |
|
✅ Build workflow passed! View run |
Signed-off-by: Josh Romero <joshr@nvidia.com>
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Currently, cuDecomp prohibits users from having more than one active library handle. Because of this, users have always been restricted to running cuDecomp operations on the single set of processes associated with the MPI communicator passed to
cudecompInitfor then entire job duration.In order in increase the flexibility of the library, this PR relaxes this restriction and enables users to create multiple cuDecomp library handles, each potentially associated with a different MPI subcommunicator.
One restriction to this feature is that handles using NVSHMEM backends must be initialized with congruent sets of ranks. This is to work with existing restrictions on NVSHMEM finalization/reinitialization and a lack of teams-based
nvshmem_malloc.