Skip to content

Shaked/random error fix#424

Merged
shakedregev merged 4 commits into
tamar/sccg-gpu-cpufrom
shaked/random_error_fix
May 28, 2026
Merged

Shaked/random error fix#424
shakedregev merged 4 commits into
tamar/sccg-gpu-cpufrom
shaked/random_error_fix

Conversation

@shakedregev
Copy link
Copy Markdown
Collaborator

@shakedregev shakedregev commented May 28, 2026

Description

The Cholesky factorization test stochastically fails due to an outdated random number generator, a bad seed, and roundoff.

Closes #423.

Proposed changes

I fixed the random seed and started using C++ style random numbers.

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.
  • I have updated CHANGELOG.md to reflect the changes in this PR. If this is a minor PR that is part of a larger fix already included in the file, state so.

@shakedregev shakedregev requested a review from pelesh May 28, 2026 13:11
@shakedregev shakedregev self-assigned this May 28, 2026
@shakedregev shakedregev merged commit e76a845 into tamar/sccg-gpu-cpu May 28, 2026
6 checks passed
@shakedregev shakedregev deleted the shaked/random_error_fix branch May 28, 2026 13:32
shakedregev added a commit that referenced this pull request May 29, 2026
* use caller-provided backend-specific handlers in SCCG

* load SCCG data on host before device sync

* refresh CUDA SpMV cache when matrix changes

* fix HyKKT SCCG HIP target setup

* document SCCG backend setup and CUDA SpMV cache handling

* mirror SpMV cache reset for HIP

* Shaked/random error fix (#424)

Fixed stochastic failures due to a faulty random number generator and asynchronous operations.

Co-authored-by: shakedregev <shakedregev@users.noreply.github.com>

---------

Co-authored-by: tamar-dewilde <tamar-dewilde@users.noreply.github.com>
Co-authored-by: Shaked Regev <35384901+shakedregev@users.noreply.github.com>
Co-authored-by: shakedregev <shakedregev@users.noreply.github.com>
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