Skip to content

SparseArray in-place send/recv #624

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.

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

Open
wants to merge 2 commits into
base: yg/faster-mpi
Choose a base branch
from

Conversation

yanzin00
Copy link

This PR improves in-place send/receive capabilities for SparseMatrixCSC types within our MPI communication utilities. It allows data to be directly written into pre-allocated buffers on the receiving end.

Key Changes:

  1. InplaceSparseInfo Struct: A new struct, InplaceSparseInfo, has been added to encapsulate the metadata required for the in-place transfer of SparseMatrixCSC objects.

  2. Performance and Memory Optimization: Enabling in-place transfers for SparseMatrixCSC significantly reduces overhead from serializing and deserializing sparse matrix data. This is particularly beneficial for large sparse matrix operations in distributed memory environments.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

1 participant