diff options
| author | David Sherwood <david.sherwood@arm.com> | 2023-04-25 08:46:41 +0000 |
|---|---|---|
| committer | David Sherwood <david.sherwood@arm.com> | 2023-05-18 10:35:57 +0000 |
| commit | c7dbe326dff81273eabe339fe69cd7bef947619c (patch) | |
| tree | 600d5c957ee03fb2dd93f83e8c69a1ae4252aed9 /lldb/source/Commands/CommandObjectExpression.cpp | |
| parent | 01efcec6dbd1431d2ac112f537d5639a9eab18b2 (diff) | |
| download | llvm-c7dbe326dff81273eabe339fe69cd7bef947619c.zip llvm-c7dbe326dff81273eabe339fe69cd7bef947619c.tar.gz llvm-c7dbe326dff81273eabe339fe69cd7bef947619c.tar.bz2 | |
[AArch64][LoopVectorize] Enable tail-folding of simple loops on neoverse-v1
This patch enables the tail-folding of simple loops by default
when targeting the neoverse-v1 CPU. Simple loops exclude those
with recurrences or reductions or loops that are reversed.
New tests have been added here:
Transforms/LoopVectorize/AArch64/sve-tail-folding-option.ll
In terms of SPEC2017 only one benchmark is really affected when
building with "-Ofast -mcpu=neoverse-v1 -flto", which is
(+ faster, - slower):
525.x264: +7.0%
Differential Revision: https://reviews.llvm.org/D130618
Diffstat (limited to 'lldb/source/Commands/CommandObjectExpression.cpp')
0 files changed, 0 insertions, 0 deletions