diff options
| author | Florian Hahn <flo@fhahn.com> | 2025-07-31 19:20:05 +0100 |
|---|---|---|
| committer | GitHub <noreply@github.com> | 2025-07-31 19:20:05 +0100 |
| commit | 078d214672e23691566137fa88b851c7022666b7 (patch) | |
| tree | c886f264c930def397460d7154f4319d873b2b6e /llvm/lib/Transforms/Utils/LoopUtils.cpp | |
| parent | 5482ef76f5b3d5ffcaded397fa924569e83f0b2d (diff) | |
| download | llvm-078d214672e23691566137fa88b851c7022666b7.zip llvm-078d214672e23691566137fa88b851c7022666b7.tar.gz llvm-078d214672e23691566137fa88b851c7022666b7.tar.bz2 | |
[TailDup] Delay aggressive computed-goto taildup to after RegAlloc. (#150911)
https://github.com/llvm/llvm-project/pull/114990 allowed more aggressive
tail duplication for computed-gotos in both pre- and post-regalloc tail
duplication.
In some cases, performing tail-duplication too early can lead to worse
results, especially if we duplicate blocks with a number of phi nodes.
This is causing a ~3% performance regression in some workloads using
Python 3.12.
This patch updates TailDup to delay aggressive tail-duplication for
computed gotos to after register allocation.
This means we can keep the non-duplicated version for a bit longer
throughout the backend, which should reduce compile-time as well as
allowing a number of optimizations and simplifications to trigger before
drastically expanding the CFG.
For the case in https://github.com/llvm/llvm-project/issues/106846, I
get the same performance with and without this patch on Skylake.
PR: https://github.com/llvm/llvm-project/pull/150911
Diffstat (limited to 'llvm/lib/Transforms/Utils/LoopUtils.cpp')
0 files changed, 0 insertions, 0 deletions