[Clang] Do not attempt to directly link arch tools in 32-bit mode
authorJoseph Huber <jhuber6@vols.utk.edu>
Tue, 31 Jan 2023 17:55:05 +0000 (11:55 -0600)
committerJoseph Huber <jhuber6@vols.utk.edu>
Tue, 31 Jan 2023 17:56:24 +0000 (11:56 -0600)
commit9f64fbb882dc3c76712ec70316253ca396baa824
tree5d5462c51d58a46a2ce5432f2d2f66be085d6ccb
parent67ee18cc7a308ae8418689c4f03a1be73b8b874a
[Clang] Do not attempt to directly link arch tools in 32-bit mode

Summary:
We offer almost no support for offloading on 32-bit systems. This causes
some problems when cross-compiling for 32-bit machines as it will find
the CUDA from the host that is incompatible. Instead we force these to
always use the dynamically loaded version, which should always compile.
clang/tools/amdgpu-arch/CMakeLists.txt
clang/tools/nvptx-arch/CMakeLists.txt