I recently updated my local workstation, and found that CUDA compilation wasn't working properly due to CUDA_ARCH not being detected properly. It turns out the deviceQuery utility is no longer distributed with the toolkit since CUDA 12 or so. nvidia-smi can get that info, so we can fall back to that if deviceQuery isn't present.
Additional background
Checklist
The proposed changes:
[ ] fix a bug or incorrect behavior in AMReX
[ ] add new capabilities to AMReX
[ ] changes answers in the test suite to more than roundoff level
[ ] are likely to significantly affect the results of downstream AMReX users
[ ] include documentation in the code and/or rst files, if appropriate
Summary
I recently updated my local workstation, and found that CUDA compilation wasn't working properly due to
CUDA_ARCH
not being detected properly. It turns out thedeviceQuery
utility is no longer distributed with the toolkit since CUDA 12 or so.nvidia-smi
can get that info, so we can fall back to that if deviceQuery isn't present.Additional background
Checklist
The proposed changes: