Skip to content

Removed unnecessary branching in dot products - #469

Closed
shakedregev wants to merge 2 commits into
developfrom
shaked/cuda_gmres_fix
Closed

Removed unnecessary branching in dot products#469
shakedregev wants to merge 2 commits into
developfrom
shaked/cuda_gmres_fix

Conversation

@shakedregev

@shakedregev shakedregev commented Aug 11, 2026

Copy link
Copy Markdown
Collaborator

Description

The current branching statement introduces errors on some machine.
Partially addresses: #468

Proposed changes

Removed the branching.

Checklist

  • All tests pass (make test and make test_install per testing instructions). Code tested on
    • CPU backend
    • CUDA backend
    • HIP backend
  • I have manually run the non-experimental examples and verified that residuals are close to machine precision. (In your build directory run: ./examples/<your_example>.exe -h to get instructions how to run examples). Code tested on:
    • CPU backend
    • CUDA backend
    • HIP backend
  • Code compiles cleanly with flags -Wall -Wpedantic -Wconversion -Wextra.
  • The new code follows Re::Solve style guidelines.
  • There are unit tests for the new code.
  • The new code is documented.
  • The feature branch is rebased with respect to the target branch.
  • Minor PR, no changelog

@shakedregev
shakedregev requested a review from pelesh August 11, 2026 14:56
@pelesh
pelesh requested a review from kswirydo August 11, 2026 17:55
@pelesh

pelesh commented Aug 11, 2026

Copy link
Copy Markdown
Collaborator

There was a performance reason why @kswirydo has implemented a separate kernel for certain sizes of multiple dot products. I would rather check what errors are showing up and on what machines, first. It may be a fixable bug.

@pelesh pelesh closed this Aug 11, 2026
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.

2 participants