--- Comment #7 from pinskia at gcc dot gnu dot org 2010-04-08 18:33 ---
Fixed in 4.5.0 which should be released soon (in the next couple of days not
months :) ).
--
pinskia at gcc dot gnu dot org changed:
What|Removed |Added
---
--- Comment #6 from wbrana at gmail dot com 2010-04-08 18:31 ---
Loop is vectorized with GCC 4.4 if these flags are used:
-O1 -ftree-vectorize -march=core2
.LFB0:
.cfi_startproc
movaps b(%rip), %xmm0
addps c(%rip), %xmm0
movaps %xmm0, a(%rip)
--- Comment #5 from irar at il dot ibm dot com 2010-04-08 17:59 ---
In GCC 4.4 the smaller loop gets completely unrolled before the vectorizer.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43692
--- Comment #4 from wbrana at gmail dot com 2010-04-08 17:50 ---
I have got same code with -O2 and O3 on Gentoo
Using built-in specs.
Target: x86_64-pc-linux-gnu
Configured with:
/mnt/md3/cache/portage/sys-devel/gcc-4.4.3/work/gcc-4.4.3/configure
--prefix=/usr --bindir=/usr/x86_64-pc-li
--- Comment #3 from irar at il dot ibm dot com 2010-04-08 17:33 ---
Both loops get vectorized for me with -O3 on x86_64-suse-linux.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43692
--- Comment #2 from wbrana at gmail dot com 2010-04-08 17:22 ---
-fno-vect-cost-model has no effect
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43692
--- Comment #1 from irar at il dot ibm dot com 2010-04-08 17:14 ---
It probably happens because the vectorization is not profitable. Try
-fno-vect-cost-model flag.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43692
