diff options
| author | Krzysztof Drewniak <Krzysztof.Drewniak@amd.com> | 2025-10-20 10:42:22 -0700 |
|---|---|---|
| committer | GitHub <noreply@github.com> | 2025-10-20 12:42:22 -0500 |
| commit | d37141776f6b6d0116f4566d4ee56b29ecdc7d0e (patch) | |
| tree | d877a05603864e6302d4e5d7bc6cd34671f1aeba /clang/lib/Frontend/CompilerInvocation.cpp | |
| parent | e10afe01ec7c79f45f62d9a42b07a7d984bdc345 (diff) | |
| download | llvm-d37141776f6b6d0116f4566d4ee56b29ecdc7d0e.zip llvm-d37141776f6b6d0116f4566d4ee56b29ecdc7d0e.tar.gz llvm-d37141776f6b6d0116f4566d4ee56b29ecdc7d0e.tar.bz2 | |
[AMDGPU] Enable volatile and non-temporal for loads to LDS (#153244)
The primary purpose of this commit is to enable marking loads to LDS
(global.load.lds, buffer.*.load.lds) volatile (using bit 31 of the aux
as with normal buffer loads) and to ensure that their !nontemporal
annotations translate to appropriate settings of te cache control bits.
However, in the process of implementing this feature, we also fixed
- Incorrect handling of buffer loads to LDS in GlobalISel
- Updating the handling of volatile on buffers in SIMemoryLegalizer:
previously, the mapping of address spaces would cause volatile on buffer
loads to be silently dropped on at least gfx10.
---------
Co-authored-by: Matt Arsenault <arsenm2@gmail.com>
Diffstat (limited to 'clang/lib/Frontend/CompilerInvocation.cpp')
0 files changed, 0 insertions, 0 deletions