https://gcc.gnu.org/bugzilla/show_bug.cgi?id=91573
avieira at gcc dot gnu.org changed:
What|Removed |Added
Status|NEW |RESOLVED
CC|
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=91573
Tamar Christina changed:
What|Removed |Added
CC||tnfchris at gcc dot gnu.org
--- Commen
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=91573
--- Comment #5 from Hao Liu ---
Great. It seems really a SLP issue.
I've learnt a lot about vectorization, dump info and -march. Thanks for your
help.
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=91573
Richard Biener changed:
What|Removed |Added
Keywords||missed-optimization
Status|
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=91573
--- Comment #3 from Andrew Pinski ---
t5.c:10:5: missed: cost model: the vector iteration cost = 12372 divided by
the scalar iteration cost = 236 is greater or equal to the vectorization factor
= 16.
t5.c:10:5: missed: not vectorized: vectoriza
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=91573
Hongtao.liu changed:
What|Removed |Added
CC||crazylht at gmail dot com
--- Comment #2 f
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=91573
--- Comment #1 from Andrew Pinski ---
>2) If WIDTH is larger. e.g.
#define WIDTH 16
Either we are unrolling the loop and not doing SLP on the following loop or it
is a cost issue saying it is not profitable for some reason. Use
-fno-vect-c
