[PATCH] D141051: [CUDA][HIP] Add support for `--offload-arch=native` to CUDA and refactor

[Joseph Huber via Phabricator via cfe-commits]

jhuber6 added a comment.

In D141051#4048456 <https://reviews.llvm.org/D141051#4048456>, @srj wrote:

> For reasons that aren't yet clear to me, this change is failing to compile when using gcc-7 and targeting 32-bit targets; the error is of the form
>
>   AMDGPU.cpp:773:10: error: could not convert ‘GPUArchs’ from ‘llvm::SmallVector<std::__cxx11::basic_string<char>, 1>’ to ‘llvm::Expected<llvm::SmallVector<std::__cxx11::basic_string<char> > >’
>      return GPUArchs;
>
> it's not (yet) clear to me whether this is specific to gcc-7 (which I realize is fairly old -- is it still supported?) or what -- investigating further.

Let me know if rG26d62674cf50 <https://reviews.llvm.org/rG26d62674cf5013249f99323d15b6044586e6b3a8> solved it. I can't reproduce it on my system.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D141051/new/

https://reviews.llvm.org/D141051