Why is CMAKE_CUDA_ARCHITECTURES not enough
Nobody has claimed this yet.
- Dominant language
- C++
- Stars
- 296
- Forks
- 66
- Avg merge
- 1h 3m
- Merged PRs (30d)
- 1
Description
In all host-configs we set both `CMAKE_CUDA_ARCHITECTURES` and add the `-arch` flag manually to the `CMAKE_CUDA_FLAGS`. Why is does this seem to be necessary? Is it possibly our reliance on the old FindCUDA, which is soon to be removed?
Does this flag need to be set if CMAKE_CUDA_ARCHITECTURES is set?
_Originally posted by @adayton1 in https://github.com/LLNL/blt/pull/570#r830399325_
Contributor guide
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Research direction
Start by reviewing the repository's host-configs and how they set CMAKE_CUDA_ARCHITECTURES and CMAKE_CUDA_FLAGS, then inspect the old FindCUDA usage mentioned in the issue. Reproduce the configuration with and without the manual -arch flag across the relevant CMake setup; document whether it is redundant and identify the host-config changes needed.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- cmake, cpp
- Domain
- build-system, hpc
- Issue type
- Refactor
- Difficulty
- 4/5
- Estimated time
- 3-5 days
- Activity status
- Stale
- Clarity
- Needs clarification
- Newbie friendliness
- 25/100