diff options
| author | Simon Pilgrim <llvm-dev@redking.me.uk> | 2025-10-28 13:05:52 +0000 |
|---|---|---|
| committer | GitHub <noreply@github.com> | 2025-10-28 13:05:52 +0000 |
| commit | e588c7fa713d8bdd5c424831ca42136b560ff66b (patch) | |
| tree | 4288ef4bb3d58a0d47acafddaf6bfac30c904ca8 /llvm/lib/Bitcode/Writer/BitcodeWriter.cpp | |
| parent | 4678f16f6d9089ce7d8de8a7fa31b70190208ab2 (diff) | |
| download | llvm-e588c7fa713d8bdd5c424831ca42136b560ff66b.zip llvm-e588c7fa713d8bdd5c424831ca42136b560ff66b.tar.gz llvm-e588c7fa713d8bdd5c424831ca42136b560ff66b.tar.bz2 | |
[X86] Attempt to fold trunc(srl(load(p),amt) -> load(p+amt/8) (#165266)
As reported on #164853 - we only attempt to reduce shifted loads for constant shift amounts, but we could do more with non-constant values if value tracking can confirm basic alignments.
This patch determines if a truncated shifted load of scalar integer shifts by a byte aligned amount and replaces the non-constant shift amount with a pointer offset instead.
I had hoped to make this a generic DAG fold, but reduceLoadWidth isn't ready to be converted to a KnownBits value tracking mechanism, and other targets don't have complex address math like X86.
Fixes #164853
Diffstat (limited to 'llvm/lib/Bitcode/Writer/BitcodeWriter.cpp')
0 files changed, 0 insertions, 0 deletions