diff options
| author | Guray Ozen <guray.ozen@gmail.com> | 2023-12-04 16:48:39 +0100 |
|---|---|---|
| committer | GitHub <noreply@github.com> | 2023-12-04 16:48:39 +0100 |
| commit | 3a03da37a3c1e146ce7af1a1bbf8a2d3a0bf53df (patch) | |
| tree | 6c9701fcfaa20cf4db009ab2b03d84ced24e4d07 /clang/lib/Frontend/CompilerInvocation.cpp | |
| parent | 0c568c2535848d1596a612c15248f299ec8c42be (diff) | |
| download | llvm-3a03da37a3c1e146ce7af1a1bbf8a2d3a0bf53df.zip llvm-3a03da37a3c1e146ce7af1a1bbf8a2d3a0bf53df.tar.gz llvm-3a03da37a3c1e146ce7af1a1bbf8a2d3a0bf53df.tar.bz2 | |
[mlir][nvgpu] Add address space attribute converter in nvgpu-to-nvvm pass (#74075)
GPU dialect has `#gpu.address_space<workgroup>` for shared memory of
NVGPU (address space =3). Howeverm when IR combine NVGPU and GPU
dialect, `nvgpu-to-nvvm` pass fails due to missing attribute conversion.
This PR adds `populateGpuMemorySpaceAttributeConversions` to
nvgou-to-nvvm lowering, so we can use `#gpu.address_space<workgroup>`
`nvgpu-to-nvvm` pass
Diffstat (limited to 'clang/lib/Frontend/CompilerInvocation.cpp')
0 files changed, 0 insertions, 0 deletions