https://gcc.gnu.org/bugzilla/show_bug.cgi?id=122587
Richard Biener <rguenth at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Ever confirmed|0 |1
Last reconfirmed| |2025-11-06
Status|UNCONFIRMED |NEW
--- Comment #1 from Richard Biener <rguenth at gcc dot gnu.org> ---
Without masked epilogs we also create a VF == 8 epilog (plus VF == 4 if
cascaded), not realizing the VF == 8 epilog is never needed. In the end
jump threading will resolve things optimally when not masking and when
using cascaed epilogs.
Sofar a target cost model issue, but some generic vectorizer work might
be there as well.
