diff options
| author | Simon Pilgrim <llvm-dev@redking.me.uk> | 2026-02-03 17:50:59 +0000 |
|---|---|---|
| committer | GitHub <noreply@github.com> | 2026-02-03 17:50:59 +0000 |
| commit | 5586d4ace9a9d12b58f5fbae736d1164c9ed44a4 (patch) | |
| tree | 7967ca22fb90810d9324a459444b9727f02c7fbe /llvm/lib/CodeGen/InsertCodePrefetch.cpp | |
| parent | abdf66d60bab8ad8eeed6e91f63c90616e6ae232 (diff) | |
| download | llvm-5586d4ace9a9d12b58f5fbae736d1164c9ed44a4.zip llvm-5586d4ace9a9d12b58f5fbae736d1164c9ed44a4.tar.gz llvm-5586d4ace9a9d12b58f5fbae736d1164c9ed44a4.tar.bz2 | |
[X86] mayFoldIntoVector - recognise larger than legal logic ops may fold to vectors (#179503)
Inspired by the hack to #174761 - move the custom operation handling
inside mayFoldIntoVector where we can more accurately predict ops that
can be moved to the vector unit
Diffstat (limited to 'llvm/lib/CodeGen/InsertCodePrefetch.cpp')
0 files changed, 0 insertions, 0 deletions