diff options
| author | Matt Arsenault <Matthew.Arsenault@amd.com> | 2024-08-01 09:42:04 +0400 |
|---|---|---|
| committer | GitHub <noreply@github.com> | 2024-08-01 09:42:04 +0400 |
| commit | 86815a1842d308521f46048bb9ed08e47c0d8357 (patch) | |
| tree | 40426d28d6cb9cdfbe0781b8cb5862f606a80900 /clang/lib/CodeGen/CodeGenModule.cpp | |
| parent | 9d068f7137a2ad732d008df46eb71aadb0cd8a8e (diff) | |
| download | llvm-86815a1842d308521f46048bb9ed08e47c0d8357.zip llvm-86815a1842d308521f46048bb9ed08e47c0d8357.tar.gz llvm-86815a1842d308521f46048bb9ed08e47c0d8357.tar.bz2 | |
AMDGPU/GlobalISel: Permit mapping G_FRAME_INDEX to sgprs (#101325)
eliminateFrameIndex should now properly handle materializing
frame indices in SGPRs, so treat this like the other constant
operand types.
On average this will produce worse code; we need to detect
VGPR uses, and improve SGPR->VGPR frame index folds.
Diffstat (limited to 'clang/lib/CodeGen/CodeGenModule.cpp')
0 files changed, 0 insertions, 0 deletions