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

Switch to cuda native #74

Merged
merged 2 commits into from
Feb 12, 2025
Merged

Switch to cuda native #74

merged 2 commits into from
Feb 12, 2025

Conversation

mkstoyanov
Copy link
Collaborator

@mkstoyanov mkstoyanov commented Feb 11, 2025 •
edited
Loading

  • in some new versions of CUDA the JIT seems to be misbehaving causing crashes
    • change the default CUDA architecture to native fixes the issue
    • applies when using CMake 3.24 or newer (released in 2022)
    • CMAKE_CUDA_ARCHITECTURES must be set manually when doing a cross-compile
  • found a small bug where CUDA error was generated on nodes using empty input or output boxes
    • affects only the subcomm use case or very irregular user input
    • does not alter the correctness of the result but yields a CUDA error that can have side-effects

Should fix: #73 (comment)

@mkstoyanov mkstoyanov force-pushed the switch_to_cuda_native branch from 18b545b to bcd131e Compare February 11, 2025 20:09
@mkstoyanov mkstoyanov mentioned this pull request Feb 11, 2025
Closed
@mkstoyanov mkstoyanov merged commit b7afa48 into icl-utk-edu:master Feb 12, 2025
10 checks passed
@mkstoyanov mkstoyanov deleted the switch_to_cuda_native branch February 12, 2025 15:09
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers
No reviews
Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.

Failing tests on AMPERE80 with gcc-13 and cuda-12.6
1 participant
@mkstoyanov