projects / platform / upstream / llvm.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: 3560ed0) | patch
[ARM] Permit auto-vectorization using MVE
authorDavid Green <david.green@arm.com>
Sun, 11 Aug 2019 08:42:57 +0000 (08:42 +0000)
committerDavid Green <david.green@arm.com>
Sun, 11 Aug 2019 08:42:57 +0000 (08:42 +0000)
commit44f8d635e20429365782b831ef59ee0903982d02
tree97702b6b050efa90f8b973201fad7b28bc6710dftree | snapshot
parent3560ed052354940756e87a3372be819b765b3c45commit | diff
[ARM] Permit auto-vectorization using MVE

With enough codegen complete, we can now correctly report the number and size
of vector registers for MVE, allowing auto vectorisation. This also allows FP
auto-vectorization for MVE without -Ofast/-ffast-math, due to support for IEEE
FP arithmetic and parity between scalar and vector FP behaviour.

Patch by David Sherwood.

Differential Revision: https://reviews.llvm.org/D63728

llvm-svn: 368529
llvm/lib/Target/ARM/ARMTargetTransformInfo.h diff | blob | history
llvm/test/Transforms/LoopVectorize/ARM/arm-ieee-vectorize.ll diff | blob | history
Domain: System / Toolchain;