This is the mail archive of the gcc-patches@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [PATCH] Fix Fortran DO loop fallback


On Mon, Jul 11, 2016 at 4:44 PM, Jeff Law <law@redhat.com> wrote:
> On 07/08/2016 08:26 AM, Martin Liška wrote:
>>
>> Hello
>>
>> Following patch fixes fallout caused by the patch set:
>> https://gcc.gnu.org/ml/gcc-regression/2016-07/msg00097.html
>>
>> Ready after it finished regression tests?
>> Thanks,
>> Martin
>>
>>
>> 0001-Fix-Fortran-DO-loop-fallback.patch
>>
>>
>> From c5dd7ad62f795cce560c7f1bb8767b7ed9298d8a Mon Sep 17 00:00:00 2001
>> From: marxin <mliska@suse.cz>
>> Date: Fri, 8 Jul 2016 15:51:54 +0200
>> Subject: [PATCH] Fix Fortran DO loop fallback
>>
>> gcc/testsuite/ChangeLog:
>>
>> 2016-07-08  Martin Liska  <mliska@suse.cz>
>>
>>         * gfortran.dg/ldist-1.f90: Update expected dump scan.
>>         * gfortran.dg/pr42108.f90: Likewise.
>>         * gfortran.dg/vect/pr62283.f: Likewise.
>
> Shouldn't ldist-1.f90 be scan-tree-dump-not?  It seems like you change it
> from that just last week, but it wasn't mentioned in the ChangeLog.

gfortran.dg/pr42108.f90 also looks pointless now?  I suppose there is nothing
to hoist after the change?  Do we now have an option to revert back to old
behavior?  If so it would be better to use that flag here.

diff --git a/gcc/testsuite/gfortran.dg/vect/pr62283.f
b/gcc/testsuite/gfortran.dg/vect/pr62283.f
index 7df3d99..2933f51 100644
--- a/gcc/testsuite/gfortran.dg/vect/pr62283.f
+++ b/gcc/testsuite/gfortran.dg/vect/pr62283.f
@@ -13,4 +13,4 @@ C { dg-additional-options "-fvect-cost-model=dynamic" }
       beta=3.141593
       y=y+beta*x
       end
-C { dg-final { scan-tree-dump-times "vectorized 1 loops" 2 "vect" {
target { vect_hw_misalign } } } }
+C { dg-final { scan-tree-dump "vectorized 1 loops" "vect" { target {
vect_hw_misalign } } } }

so why do we no longer vectorize 1 loops in two functions?  The
testcase works for me
unchanged.

Richard.

> OK with that change.
>
> jeff
>
>


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]