diff options
author | Andrea Di Biagio <Andrea_DiBiagio@sn.scee.net> | 2014-06-25 17:41:58 +0000 |
---|---|---|
committer | Andrea Di Biagio <Andrea_DiBiagio@sn.scee.net> | 2014-06-25 17:41:58 +0000 |
commit | cae1ea691d79ce786e9209cb5917036f42f13d7c (patch) | |
tree | 746c84fc614d04044dc59ea1e309b8a9adc8f2c8 /lib/Transforms/Vectorize | |
parent | 75f1d664b426f8d4892609f4c5f2e1308f6492d8 (diff) | |
download | llvm-cae1ea691d79ce786e9209cb5917036f42f13d7c.tar.gz llvm-cae1ea691d79ce786e9209cb5917036f42f13d7c.tar.bz2 llvm-cae1ea691d79ce786e9209cb5917036f42f13d7c.tar.xz |
[X86] Always prefer to lower a VECTOR_SHUFFLE into a BLENDI instead of SHUFP (or VPERM2X128).
This patch teaches method 'LowerVECTOR_SHUFFLE' to give higher precedence to
the check for 'isBlendMask'; the idea is that, when possible, we should firstly
check if a shuffle performs a blend, and in case, try to lower it into a BLENDI
instead of selecting a SHUFP or (worse) a VPERM2X128.
In general:
- AVX VBLENDPS/D always have better latency and throughput than VPERM2F128;
- BLENDPS/D instructions tend to always have better 'reciprocal throughput'
than the equivalent SHUFPS/D;
- Both BLENDPS/D and SHUFPS/D are often decoded into the same number of
m-ops; however, a m-op obtained from a BLENDPS/D can be scheduled to more
than one execution port.
This patch:
- Moves the check for 'isBlendMask' immediately before the check for
'isSHUFPMask' within method 'LowerVECTOR_SHUFFLE';
- Updates existing tests for sse/avx shuffle/blend instructions to verify
that we select (v)blendps/d when possible (instead of (v)shufps/d or
vperm2f128).
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@211720 91177308-0d34-0410-b5e6-96231b3b80d8
Diffstat (limited to 'lib/Transforms/Vectorize')
0 files changed, 0 insertions, 0 deletions