This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: RFA: Fix some vector testsuite failures
- From: Janis Johnson <janis187 at us dot ibm dot com>
- To: Richard Sandiford <rsandifo at nildram dot co dot uk>, dorit at il dot ibm dot com
- Cc: gcc-patches at gcc dot gnu dot org
- Date: Fri, 21 Sep 2007 13:41:41 -0700
- Subject: Re: RFA: Fix some vector testsuite failures
- References: <87ir659j9g.fsf@firetop.home>
- Reply-to: janis187 at us dot ibm dot com
Dorit, will you please review this one?
On Thu, 2007-09-20 at 21:25 +0100, Richard Sandiford wrote:
> These tests are failing on MIPS64 targets:
>
> gcc.dg/vect/pr21591.c
> gcc.dg/vect/vect-outer-2.c
> gcc.dg/vect/vect-outer-2a.c
> gcc.dg/vect/vect-outer-2c.c
>
> The first is testing the vectorisation of"int" operations, so I think
> it needs to depend on vect_int. The others are testing for cases
> where an integer is assigned to a float array, so I think it needs
> to depend on vect_intfloat_cvt. (I notice that the outer tests
> are also failing in HJ's ia64 results, which is another target
> with float vectors but no vectorised int->float conversions.)
>
> Tested on x86_64-linux-gnu and mipsisa64-elfoabi. OK to apply?
>
> (Maybe obvious, but I'm not too familiar with the vectorisation stuff.)
>
> Richard
>
>
> gcc/testsuite/
> * gcc.dg/vect/pr21591.c: Require vect_int.
> * gcc.dg/vect/vect-outer-2.c: Require vect_intfloat_cvt.
> * gcc.dg/vect/vect-outer-2a.c: Likewise.
> * gcc.dg/vect/vect-outer-2c.c: Likewise.
>
> Index: gcc/testsuite/gcc.dg/vect/pr21591.c
> ===================================================================
> --- gcc/testsuite/gcc.dg/vect/pr21591.c 2007-09-20 20:54:38.000000000 +0100
> +++ gcc/testsuite/gcc.dg/vect/pr21591.c 2007-09-20 21:18:15.000000000 +0100
> @@ -1,4 +1,5 @@
> /* { dg-do compile } */
> +/* { dg-require-effective-target vect_int } */
>
> struct a
> {
> Index: gcc/testsuite/gcc.dg/vect/vect-outer-2.c
> ===================================================================
> --- gcc/testsuite/gcc.dg/vect/vect-outer-2.c 2007-09-20 20:54:38.000000000 +0100
> +++ gcc/testsuite/gcc.dg/vect/vect-outer-2.c 2007-09-20 21:18:15.000000000 +0100
> @@ -1,4 +1,5 @@
> /* { dg-require-effective-target vect_float } */
> +/* { dg-require-effective-target vect_intfloat_cvt } */
> #include <stdarg.h>
> #include "tree-vect.h"
>
> Index: gcc/testsuite/gcc.dg/vect/vect-outer-2a.c
> ===================================================================
> --- gcc/testsuite/gcc.dg/vect/vect-outer-2a.c 2007-09-20 20:54:38.000000000 +0100
> +++ gcc/testsuite/gcc.dg/vect/vect-outer-2a.c 2007-09-20 21:18:15.000000000 +0100
> @@ -1,4 +1,5 @@
> /* { dg-require-effective-target vect_float } */
> +/* { dg-require-effective-target vect_intfloat_cvt } */
> #include <stdarg.h>
> #include "tree-vect.h"
>
> Index: gcc/testsuite/gcc.dg/vect/vect-outer-2c.c
> ===================================================================
> --- gcc/testsuite/gcc.dg/vect/vect-outer-2c.c 2007-09-20 20:54:38.000000000 +0100
> +++ gcc/testsuite/gcc.dg/vect/vect-outer-2c.c 2007-09-20 21:18:15.000000000 +0100
> @@ -1,4 +1,5 @@
> /* { dg-require-effective-target vect_float } */
> +/* { dg-require-effective-target vect_intfloat_cvt } */
> #include <stdarg.h>
> #include "tree-vect.h"
>