https://gcc.gnu.org/bugzilla/show_bug.cgi?id=124339
Andrew Pinski <pinskia at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Resolution|--- |INVALID
Keywords| |missed-optimization
Status|UNCONFIRMED |RESOLVED
--- Comment #1 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
I don't see anything wrong with this and there is a check first to see if we
can load 16 bytes at a time first.
There is an early exit vectorization happening (happens on x86_64 with -mavx
too).
Moving s++; into the loop changes the loop to be uncountable but that is
vectorized on the trunk.