diff options
| author | pvanhout <pierre.vanhoutryve@amd.com> | 2023-06-27 16:50:01 +0200 |
|---|---|---|
| committer | pvanhout <pierre.vanhoutryve@amd.com> | 2023-06-29 14:49:18 +0200 |
| commit | 026fc9e9c41db6f20df275e7e405cadfecb1b6e6 (patch) | |
| tree | 8ace13b4727007ca7e078827d7a109c763b2e446 /llvm/lib/CodeGen/MachineCopyPropagation.cpp | |
| parent | a68dcd09e8084dc179fb3218e8953417c4b7474a (diff) | |
| download | llvm-026fc9e9c41db6f20df275e7e405cadfecb1b6e6.zip llvm-026fc9e9c41db6f20df275e7e405cadfecb1b6e6.tar.gz llvm-026fc9e9c41db6f20df275e7e405cadfecb1b6e6.tar.bz2 | |
[AMDGPU] Handle Additional Cases in tryFoldPhiAGPR
Sometimes PHI have different incoming values, such as:
```
%1:vgpr_256 = COPY %0:agpr_256
%2:vgpr_32 = COPY %1:vgpr_256.sub0
```
Those weren't handled, which could lead to massive performance issues if break-large-PHIs kicked in + AGPRs were used (MFMA)
Fixes SWDEV-407986
Reviewed By: #amdgpu, arsenm
Differential Revision: https://reviews.llvm.org/D153879
Diffstat (limited to 'llvm/lib/CodeGen/MachineCopyPropagation.cpp')
0 files changed, 0 insertions, 0 deletions