public inbox for gcc-bugs@sourceware.org
help / color / mirror / Atom feed
* [Bug tree-optimization/92834] misssed SLP vectorization in LightPixel
[not found] <bug-92834-4@http.gcc.gnu.org/bugzilla/>
@ 2021-08-25 8:00 ` pinskia at gcc dot gnu.org
2021-08-25 9:30 ` rguenth at gcc dot gnu.org
1 sibling, 0 replies; 2+ messages in thread
From: pinskia at gcc dot gnu.org @ 2021-08-25 8:00 UTC (permalink / raw)
To: gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=92834
--- Comment #5 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
I think this is now the standard GCC SLP does not vectorize starting at loads
but rather starts at stores issue.
^ permalink raw reply [flat|nested] 2+ messages in thread
* [Bug tree-optimization/92834] misssed SLP vectorization in LightPixel
[not found] <bug-92834-4@http.gcc.gnu.org/bugzilla/>
2021-08-25 8:00 ` [Bug tree-optimization/92834] misssed SLP vectorization in LightPixel pinskia at gcc dot gnu.org
@ 2021-08-25 9:30 ` rguenth at gcc dot gnu.org
1 sibling, 0 replies; 2+ messages in thread
From: rguenth at gcc dot gnu.org @ 2021-08-25 9:30 UTC (permalink / raw)
To: gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=92834
--- Comment #6 from Richard Biener <rguenth at gcc dot gnu.org> ---
So there is a store and we vectorize it:
_188 = {_10, _14, _18, _22};
MEM[(union ._anon_7 *)&color] = _188;
_43 = color.bgra;
color ={v} {CLOBBER};
return _43;
but we fail to vectorize more because there's a mix of min/max operations
here. In fact we even recognize the reduction on the 'A' component but
again we fail to be clever enough to exploit it (it's also 3 lanes and
thus difficult to handle).
^ permalink raw reply [flat|nested] 2+ messages in thread
end of thread, other threads:[~2021-08-25 9:30 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
[not found] <bug-92834-4@http.gcc.gnu.org/bugzilla/>
2021-08-25 8:00 ` [Bug tree-optimization/92834] misssed SLP vectorization in LightPixel pinskia at gcc dot gnu.org
2021-08-25 9:30 ` rguenth at gcc dot gnu.org
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).