This is the mail archive of the
gcc-bugs@gcc.gnu.org
mailing list for the GCC project.
[Bug target/71722] incorrect code for test pr64252.c for -mcpu=power9 -mpower9-vector -ftree-vectorize -O3
- From: "wschmidt at gcc dot gnu.org" <gcc-bugzilla at gcc dot gnu dot org>
- To: gcc-bugs at gcc dot gnu dot org
- Date: Fri, 01 Jul 2016 22:28:59 +0000
- Subject: [Bug target/71722] incorrect code for test pr64252.c for -mcpu=power9 -mpower9-vector -ftree-vectorize -O3
- Auto-submitted: auto-generated
- References: <bug-71722-4 at http dot gcc dot gnu dot org/bugzilla/>
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=71722
Bill Schmidt <wschmidt at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
CC| |meissner at gcc dot gnu.org
--- Comment #3 from Bill Schmidt <wschmidt at gcc dot gnu.org> ---
For some reason I can't yet reproduce this from the original test case. I'm
getting an unvectorized version of the code instead. Also, I'm having trouble
compiling the preprocessed test case. Perhaps you could point me to a copy
offline.
That said, the only suspicious P9 instruction I see in here is an mtvsrws.
There are a couple of xxspltib instructions, but they are obviously just
splatting 0 and -1 and it doesn't seem like much can go wrong there. I'd guess
the direct move to be worth looking at.