public inbox for gcc-bugs@sourceware.org
help / color / mirror / Atom feed
* [Bug tree-optimization/112892] New: MAX<VEC_PERM<a,a,c>, VEC_PERM<b,b,c>> could be just VEC_PERM<MAX<a,b>,...,c>
@ 2023-12-07 2:02 pinskia at gcc dot gnu.org
2023-12-07 7:29 ` [Bug tree-optimization/112892] " pinskia at gcc dot gnu.org
0 siblings, 1 reply; 2+ messages in thread
From: pinskia at gcc dot gnu.org @ 2023-12-07 2:02 UTC (permalink / raw)
To: gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=112892
Bug ID: 112892
Summary: MAX<VEC_PERM<a,a,c>, VEC_PERM<b,b,c>> could be just
VEC_PERM<MAX<a,b>,...,c>
Product: gcc
Version: 14.0
Status: UNCONFIRMED
Keywords: missed-optimization
Severity: enhancement
Priority: P3
Component: tree-optimization
Assignee: unassigned at gcc dot gnu.org
Reporter: pinskia at gcc dot gnu.org
Target Milestone: ---
Take:
```
static inline int MAX(int a, int b)
{
if (a < b) return b;
return a;
}
void f(int *a, int n)
{
n = 128;
int i;
for(i = n-1; i >= 1; i--)
{
a[i] = MAX(a[i], a[i-1]);
}
}
```
Currently we vectorize to:
```
vect__4.9_43 = MEM <vector(4) int> [(int *)vectp_a.7_41];
vect__4.10_44 = VEC_PERM_EXPR <vect__4.9_43, vect__4.9_43, { 3, 2, 1, 0 }>;
vect__7.13_48 = MEM <vector(4) int> [(int *)vectp_a.7_41 + -4B];
vect__7.14_49 = VEC_PERM_EXPR <vect__7.13_48, vect__7.13_48, { 3, 2, 1, 0 }>;
vect__12.15_50 = MAX_EXPR <vect__4.10_44, vect__7.14_49>;
vect__12.18_54 = VEC_PERM_EXPR <vect__12.15_50, vect__12.15_50, { 3, 2, 1, 0
}>;
MEM <vector(4) int> [(int *)vectp_a.7_41] = vect__12.18_54;
```
But the VEC_PERMs here can be removed ...
as we are generating the opposite vector and such.
^ permalink raw reply [flat|nested] 2+ messages in thread
* [Bug tree-optimization/112892] MAX<VEC_PERM<a,a,c>, VEC_PERM<b,b,c>> could be just VEC_PERM<MAX<a,b>,...,c>
2023-12-07 2:02 [Bug tree-optimization/112892] New: MAX<VEC_PERM<a,a,c>, VEC_PERM<b,b,c>> could be just VEC_PERM<MAX<a,b>,...,c> pinskia at gcc dot gnu.org
@ 2023-12-07 7:29 ` pinskia at gcc dot gnu.org
0 siblings, 0 replies; 2+ messages in thread
From: pinskia at gcc dot gnu.org @ 2023-12-07 7:29 UTC (permalink / raw)
To: gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=112892
Andrew Pinski <pinskia at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|UNCONFIRMED |RESOLVED
Resolution|--- |DUPLICATE
--- Comment #1 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
Turns out this is a dup of bug 61338 .
*** This bug has been marked as a duplicate of bug 61338 ***
^ permalink raw reply [flat|nested] 2+ messages in thread
end of thread, other threads:[~2023-12-07 7:29 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2023-12-07 2:02 [Bug tree-optimization/112892] New: MAX<VEC_PERM<a,a,c>, VEC_PERM<b,b,c>> could be just VEC_PERM<MAX<a,b>,...,c> pinskia at gcc dot gnu.org
2023-12-07 7:29 ` [Bug tree-optimization/112892] " pinskia 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).