https://gcc.gnu.org/bugzilla/show_bug.cgi?id=49442
Richard Biener <rguenth at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Resolution|--- |FIXED
Status|NEW |RESOLVED
--- Comment #27 from Richard Biener <rguenth at gcc dot gnu.org> ---
I'd say we should close this now, at least -mtune=generic has long stopped
avoiding movupd and splitting up moves instead. -mtune=athlon64 still does.
In the end it would be a cost model thing, but we don't really care for such
old uarchs anymore.
On Zen2 vectorizing gives the intended > 2x speedup for me with SSE2/generic.
"Fixed" as by adjusting the generic cost model.
