https://gcc.gnu.org/bugzilla/show_bug.cgi?id=108953
Richard Biener <rguenth at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Last reconfirmed| |2023-02-28
Status|UNCONFIRMED |NEW
Ever confirmed|0 |1
--- Comment #2 from Richard Biener <rguenth at gcc dot gnu.org> ---
I think it only makes sense to combine the compares when the operand(s) are
from memory which means the machinery we have in load/store merging becomes
useful
which means we eventually want to do this in this very pass (or in the bswap
pass).
The alternative would be ifcombine but then the order of the compares in the
source could be arbitrary, so it doesn't really fit that one for the most
optimal result.
There's also a related vectorization bug in case the SLP vectorizer is the
target to look at.
