[PATCH] D99949: [AMDGPU][OpenMP] Add amdgpu-arch tool to list AMD GPUs installed

Jon Chesterfield via Phabricator via cfe-commits cfe-commits at lists.llvm.org
Wed May 5 10:52:42 PDT 2021


JonChesterfield added a comment.

We may need more robust error handling around this. I'm seeing test failures when building clang,

`clang-13: error: Cannot determine AMDGPU architecture: $HOME/llvm-build/llvm/./bin/amdgpu-arch: Execute failed: No such file or directory. Consider passing it via --march.`

The executable exists, but when run directly it fails with `error while loading shared libraries: libhsa-runtime64.so.1: cannot open shared object file: No such file or directory`
It's built with RUNPATH = [$ORIGIN/../lib], which doesn't necessarily work from the build directory.

Two somewhat orthogonal things here I think.

- skip running the corresponding clang tests when ./amdgpu-arch fails, whatever reason it fails for (which I thought was already the case, but can't find that in D99656 <https://reviews.llvm.org/D99656>)
- be able to run from the build directory, which seems to be the default location for the tests

Changing clang_target_link_libraries to target_link_libraries made no difference. There is some cmake handling available around rpath but I haven't been able to determine what it should be set to.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D99949/new/

https://reviews.llvm.org/D99949



More information about the cfe-commits mailing list