On 23 June 2016 at 21:56, Uros Bizjak <ubiz...@gmail.com> wrote:
> Hello!
>
> 2016-06-23 Uros Bizjak <ubiz...@gmail.com>
>
> * g++.dg/vect/pr33834_2.cc: Use dg-additional-options instead of
> dg-options and remove default vector testsuite compile flags.
> * g++.dg/vect/pr33860a.cc: Ditto.
> * g++.dg/vect/pr45470-a.cc: Ditto.
> * g++.dg/vect/pr45470-b.cc: Ditto.
> * g++.dg/vect/pr60896.cc: Ditto.
> * gcc.dg/vect/no-tree-pre-pr45241.c: Ditto.
> * gcc.dg/vect/pr18308.c: Ditto.
> * gcc.dg/vect/pr24049.c: Ditto.
> * gcc.dg/vect/pr33373.c: Ditto.
> * gcc.dg/vect/pr36228.c: Ditto.
> * gcc.dg/vect/pr42395.c: Ditto.
> * gcc.dg/vect/pr42604.c: Ditto.
> * gcc.dg/vect/pr46663.c: Ditto.
> * gcc.dg/vect/pr48765.c: Ditto.
> * gcc.dg/vect/pr49093.c: Ditto.
> * gcc.dg/vect/pr49352.c: Ditto.
> * gcc.dg/vect/pr52298.c: Ditto.
> * gcc.dg/vect/pr52870.c: Ditto.
> * gcc.dg/vect/pr53185.c: Ditto.
> * gcc.dg/vect/pr53773.c: Ditto.
> * gcc.dg/vect/pr56695.c: Ditto.
> * gcc.dg/vect/pr62171.c: Ditto.
> * gcc.dg/vect/pr63530.c: Ditto.
> * gcc.dg/vect/pr68339.c: Ditto.
> * gcc.dg/vect/pr71259.c: Ditto.
> * gcc.dg/vect/vect-82_64.c: Ditto.
> * gcc.dg/vect/vect-83_64.c: Ditto.
> * gcc.dg/vect/vect-debug-pr41926.c: Ditto.
> * gcc.dg/vect/vect-shift-2-big-array.c: Ditto.
> * gcc.dg/vect/vect-shift-2.c: Ditto.
> * gfortran.dg/vect/fast-math-mgrid-resid.f: Ditto.
> * gfortran.dg/vect/pr39318.f90: Ditto.
> * gfortran.dg/vect/pr45714-a.f: Ditto.
> * gfortran.dg/vect/pr45714-b.f: Ditto.
> * gfortran.dg/vect/pr46213.f90: Ditto.
>
> Tested on x86_64-linux-gnu {,-m32} and committed to mainline SVN.
>
Hi,
I can see this is a cleaner version of the patch I proposed, since you
also remove
the default vector testsuite compilation flags.
(https://gcc.gnu.org/ml/gcc-patches/2016-06/msg01104.html)
I've also noticed that you updated pr63530.c to call check_vect().
Regarding vect-debug-pr41926.c, it seems to me that your patch removes
-msse2 on x86 targets: is that expected?
My patch also modified vect-fold-1.c which uses dg-options instead of
dg-additional-options: did you forget it or did you
skip it on purpose?
I'm still looking at improving check_vect for arm cores without Neon.
Christophe
