https://gcc.gnu.org/bugzilla/show_bug.cgi?id=120823
Richard Biener <rguenth at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Blocks| |53947
Keywords| |missed-optimization
See Also| |https://gcc.gnu.org/bugzill
| |a/show_bug.cgi?id=109892
--- Comment #1 from Richard Biener <rguenth at gcc dot gnu.org> ---
It seems to me that for .FMA we'd want to generate a vector multiplication
and then a chain of scalar adds - but that's incorrect in case the .FMA
wasn't introduced by GCC (which it might be with -ffp-contract=on).
So possibly we cannot handle .FMA as open-coded fold-left reduction after all.
Referenced Bugs:
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=53947
[Bug 53947] [meta-bug] vectorizer missed-optimizations
