diff options
| author | Charlie Turner <charlie.turner@arm.com> | 2015-10-27 17:54:16 +0000 |
|---|---|---|
| committer | Charlie Turner <charlie.turner@arm.com> | 2015-10-27 17:54:16 +0000 |
| commit | cd6e8cf8c250609c7339ef83b3e8844cfbf00966 (patch) | |
| tree | c34a20c94ef24f45d1e668a78a97e7a3e9c49547 /clang/lib/CodeGen/CodeGenAction.cpp | |
| parent | 74c387feb7023dae60c760b5e98dcf617466a434 (diff) | |
| download | bcm5719-llvm-cd6e8cf8c250609c7339ef83b3e8844cfbf00966.tar.gz bcm5719-llvm-cd6e8cf8c250609c7339ef83b3e8844cfbf00966.zip | |
[SLP] Try a bit harder to find reduction PHIs
Summary:
Currently, when the SLP vectorizer considers whether a phi is part of a reduction, it dismisses phi's whose incoming blocks are not the same as the block containing the phi. For the patterns I'm looking at, extending this rule to allow phis whose incoming block is a containing loop latch allows me to vectorize certain workloads.
There is no significant compile-time impact, and combined with D13949, no performance improvement measured in ARM/AArch64 in any of SPEC2000, SPEC2006 or LNT.
Reviewers: jmolloy, mcrosier, nadav
Subscribers: mssimpso, nadav, aemerson, llvm-commits
Differential Revision: http://reviews.llvm.org/D14063
llvm-svn: 251425
Diffstat (limited to 'clang/lib/CodeGen/CodeGenAction.cpp')
0 files changed, 0 insertions, 0 deletions
