https://gcc.gnu.org/bugzilla/show_bug.cgi?id=70251
--- Comment #14 from Richard Biener ---
Author: rguenth
Date: Wed Mar 23 13:40:50 2016
New Revision: 234427
URL: https://gcc.gnu.org/viewcvs?rev=234427&root=gcc&view=rev
Log:
2016-03-23 Richard Biener
PR middle-end/70251
* ma
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=70251
--- Comment #13 from Richard Biener ---
Author: rguenth
Date: Tue Mar 22 14:38:42 2016
New Revision: 234405
URL: https://gcc.gnu.org/viewcvs?rev=234405&root=gcc&view=rev
Log:
2016-03-22 Richard Biener
PR middle-end/70251
* ge
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=70251
--- Comment #12 from Richard Biener ---
Thanks - I'm testing a fix for that and will adjust the match.pd patterns
accordingly (without adding a cmp ? -1 : 0 -> VIEW_CONVERT_EXPR pattern).
I verified the original aarch64 testcase still works after
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=70251
--- Comment #11 from Marc Glisse ---
(In reply to rguent...@suse.de from comment #10)
> > Unless we start transforming x?-1:0 into
> > view_convert_expr(x) under the same condition (VECTOR_MODE_P(...)), it
> > doesn't
> > make sense to me.
>
>
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=70251
--- Comment #10 from rguenther at suse dot de ---
On Sat, 19 Mar 2016, glisse at gcc dot gnu.org wrote:
> https://gcc.gnu.org/bugzilla/show_bug.cgi?id=70251
>
> --- Comment #9 from Marc Glisse ---
> I don't find the current situation very cons
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=70251
--- Comment #9 from Marc Glisse ---
I don't find the current situation very consistent. We seem to consider that in
gimple, the output of a vector comparison can only be used directly in a
vec_cond_expr (possibly also {and,ior,xor,bit_not}_expr?)
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=70251
--- Comment #2 from Ilya Enkovich ---
Here is a responsible match.pd pattern:
/* A + (B vcmp C ? 1 : 0) -> A - (B vcmp C), since vector comparisons
return all-1 or all-0 results. */
/* ??? We could instead convert all instances of the vec_co
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=70251
--- Comment #5 from rguenther at suse dot de ---
On Wed, 16 Mar 2016, glisse at gcc dot gnu.org wrote:
> https://gcc.gnu.org/bugzilla/show_bug.cgi?id=70251
>
> --- Comment #4 from Marc Glisse ---
> (In reply to Ilya Enkovich from comment #2)
>
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=70251
--- Comment #6 from Ilya Enkovich ---
(In reply to rguent...@suse.de from comment #5)
> I think that would be an odd transform. But yes, The pattern is probably
> dead now and can as well be removed...
This test proves pattern is not dead and s
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=70251
Ilya Enkovich changed:
What|Removed |Added
Status|NEW |RESOLVED
Resolution|---
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=70251
--- Comment #3 from rguenther at suse dot de ---
On Wed, 16 Mar 2016, ienkovich at gcc dot gnu.org wrote:
> https://gcc.gnu.org/bugzilla/show_bug.cgi?id=70251
>
> --- Comment #2 from Ilya Enkovich ---
> Here is a responsible match.pd pattern:
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=70251
Ilya Enkovich changed:
What|Removed |Added
Status|UNCONFIRMED |NEW
Last reconfirmed|
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=70251
--- Comment #7 from rguenther at suse dot de ---
On Thu, 17 Mar 2016, ienkovich at gcc dot gnu.org wrote:
> https://gcc.gnu.org/bugzilla/show_bug.cgi?id=70251
>
> --- Comment #6 from Ilya Enkovich ---
> (In reply to rguent...@suse.de from comm
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=70251
--- Comment #4 from Marc Glisse ---
(In reply to Ilya Enkovich from comment #2)
> Here is a responsible match.pd pattern:
>
> /* A + (B vcmp C ? 1 : 0) -> A - (B vcmp C), since vector comparisons
>return all-1 or all-0 results. */
Since we
14 matches
Mail list logo
