diff options
| author | Jeffrey Byrnes <jeffrey.byrnes@amd.com> | 2025-01-23 17:08:23 -0800 |
|---|---|---|
| committer | GitHub <noreply@github.com> | 2025-01-23 17:08:23 -0800 |
| commit | acb7859f075f91b1105c04c37c6aa85db27a898a (patch) | |
| tree | 3649d5c5dee305add91d3727d6d9337319f32cb8 /llvm/lib/CodeGen/MachineFunctionPass.cpp | |
| parent | 24f177df61f673804a612dc48279c517bdecd696 (diff) | |
| download | llvm-acb7859f075f91b1105c04c37c6aa85db27a898a.zip llvm-acb7859f075f91b1105c04c37c6aa85db27a898a.tar.gz llvm-acb7859f075f91b1105c04c37c6aa85db27a898a.tar.bz2 | |
[MachineSink] Extend loop sinking capability (#117247)
The current MIR cycle sinking capabilities are rather limited. It only
support sinking copies into a single successor block while obeying
limits.
This opt-in feature adds a more aggressive option, that is not limited
to the above concerns. The feature will try to "sink" by duplicating any
top-level preheader instruction (that we are sure is safe to sink) into
any user block, then does some dead code cleanup. In particular, this is
useful for high RP situations when loop bodies have control flow.
Diffstat (limited to 'llvm/lib/CodeGen/MachineFunctionPass.cpp')
0 files changed, 0 insertions, 0 deletions