diff options
| author | Jay Foad <jay.foad@amd.com> | 2024-08-22 17:14:53 +0100 |
|---|---|---|
| committer | GitHub <noreply@github.com> | 2024-08-22 17:14:53 +0100 |
| commit | 2012b25420160c3d4e595b29910afffa6c5f3fc2 (patch) | |
| tree | 4c81916c13c50a4123ecb791c569055f72adbfe4 /llvm/lib/CodeGen/MachineSink.cpp | |
| parent | 46707b0a83b7769965f9b1b3d08b2cc6bd26c469 (diff) | |
| download | llvm-2012b25420160c3d4e595b29910afffa6c5f3fc2.zip llvm-2012b25420160c3d4e595b29910afffa6c5f3fc2.tar.gz llvm-2012b25420160c3d4e595b29910afffa6c5f3fc2.tar.bz2 | |
[AMDGPU][GlobalISel] Disable fixed-point iteration in all Combiners (#105517)
Disable fixed-point iteration in all AMDGPU Combiners after #102163.
This saves around 2% compile time in ad hoc testing on some large
graphics shaders. I did not notice any regressions in the generated
code, just a bunch of harmless differences in instruction selection and
register allocation.
Diffstat (limited to 'llvm/lib/CodeGen/MachineSink.cpp')
0 files changed, 0 insertions, 0 deletions