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

Bml whypre #752

Merged
merged 1 commit into from
Mar 20, 2024
Merged

Bml whypre #752

merged 1 commit into from
Mar 20, 2024

Conversation

oseikuffuor1
Copy link
Collaborator

This PR enables integrating hypre into BML to allow BML to access hypre's GPU capable sparse linear algebra routines, particularly, matrix multiplication and matrix addition.

  • The approach should allow BML to use hypre for sparse linear algebra operations on NVIDIA, AMD and Intel GPUs. However, the current implementation targets NVIDIA GPUs. Minor edits/ additions are necessary to support AMD.
  • Currently, hypre can be built in one precision at a time, so bml tests for multiple precisions will only pass for the precision for which hypre was built (if BML is built with hypre support).

nicolasbock
nicolasbock requested changes Mar 20, 2024
View reviewed changes
Copy link
Collaborator

@nicolasbock nicolasbock left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Could you

  1. squash the commits
  2. Clean up the commit message such that [1]
    a. The first line is at most 72 characters long
    b. There is an empty line
    c. Wrap the "body" of the message to 72 characters per line

[1] https://gist.github.com/robertpainsi/b632364184e70900af4ab688decf6f53

@oseikuffuor1
Copy link
Collaborator Author

@nicolasbock you want just a single commit for this PR?

@oseikuffuor1
Enable BML to use hypre's GPU-capable sparse matrix routines
36b6522 
* Enables sparse matrix multiplication and addition on GPUs
* Uses openMP offload to transfer BML data to hypre
* Current implementation supports NVIDIA GPUs only
* Supports only one precision build of hypre (single  of double)
@nicolasbock
Copy link
Collaborator

Yes, ideally 😄

@oseikuffuor1
Copy link
Collaborator Author

@nicolasbock Ok done in the latest commit. Thanks

@nicolasbock
Copy link
Collaborator

PPAs are currently offline for maintenance [1]. I will wait for them to be back online and re-run CI before merging.

[1] https://ubuntu.social/@launchpadstatus

nicolasbock
nicolasbock approved these changes Mar 20, 2024
View reviewed changes
Copy link
Collaborator

@nicolasbock nicolasbock left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thanks!

@nicolasbock nicolasbock merged commit 393f5ad into lanl:master Mar 20, 2024
31 checks passed
@oseikuffuor1
Copy link
Collaborator Author

Thanks @nicolasbock!

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers

@nicolasbock nicolasbock nicolasbock approved these changes

@jeanlucf22 jeanlucf22 Awaiting requested review from jeanlucf22 jeanlucf22 is a code owner

@cnegre cnegre Awaiting requested review from cnegre cnegre is a code owner

@suemni suemni Awaiting requested review from suemni suemni is a code owner

@tokshgithub tokshgithub Awaiting requested review from tokshgithub tokshgithub is a code owner

@jmohdyusof jmohdyusof Awaiting requested review from jmohdyusof jmohdyusof is a code owner

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

Successfully merging this pull request may close these issues.
2 participants
@oseikuffuor1 @nicolasbock