https://gcc.gnu.org/bugzilla/show_bug.cgi?id=70321
--- Comment #11 from H.J. Lu ---
(In reply to Jakub Jelinek from comment #10)
> Are the changes I've made sufficient enough that we can postpone the rest of
> this for stage1?
Now we got
[hjl@gnu-6 pr70321]$ cat z.i
void
foo (long long ixi)
{
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=70321
--- Comment #10 from Jakub Jelinek ---
Are the changes I've made sufficient enough that we can postpone the rest of
this for stage1?
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=70321
--- Comment #9 from Jakub Jelinek ---
Author: jakub
Date: Wed Mar 23 09:49:12 2016
New Revision: 234416
URL: https://gcc.gnu.org/viewcvs?rev=234416&root=gcc&view=rev
Log:
PR target/70321
* config/i386/i386.md (*anddi3_doubleword,
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=70321
--- Comment #8 from H.J. Lu ---
(In reply to Ilya Enkovich from comment #7)
> (In reply to Jakub Jelinek from comment #6)
> > Why couldn't STV just "vectorize" AND and NOT patterns and let the combiner
> > combine that in the vectorized code?
>
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=70321
--- Comment #7 from Ilya Enkovich ---
(In reply to Jakub Jelinek from comment #6)
> Why couldn't STV just "vectorize" AND and NOT patterns and let the combiner
> combine that in the vectorized code?
I think the only thing we miss for that is cor
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=70321
--- Comment #6 from Jakub Jelinek ---
(In reply to Ilya Enkovich from comment #5)
> (In reply to Jakub Jelinek from comment #4)
> > For stage1, I wonder if it can't move earlier, say before the combiner. If
> > it could, then we could split the
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=70321
Ilya Enkovich changed:
What|Removed |Added
CC||ienkovich at gcc dot gnu.org
--- Comment
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=70321
Jakub Jelinek changed:
What|Removed |Added
CC||jakub at gcc dot gnu.org
--- Comment #4
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=70321
--- Comment #3 from H.J. Lu ---
Should STV split DImode early if STV is known to not profitable?
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=70321
--- Comment #2 from H.J. Lu ---
or and xor have the same issue:
[hjl@gnu-tools-1 bitwise-1]$ cat or.i
extern long long x;
void
foo (long long ixi)
{
x = ixi | 14348907;
}
[hjl@gnu-tools-1 bitwise-1]$ make or.s
/export/build/gnu/gcc/build-x86
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=70321
--- Comment #1 from H.J. Lu ---
"and" is also less optimized:
[hjl@gnu-tools-1 bitwise-1]$ cat and.i
extern long long x;
void
foo (long long ixi)
{
x = ixi & 14348907;
}
[hjl@gnu-tools-1 bitwise-1]$ make and.s
/export/build/gnu/gcc/build-x86
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=70321
H.J. Lu changed:
What|Removed |Added
Status|UNCONFIRMED |NEW
Last reconfirmed|
12 matches
Mail list logo
