On Wed, Apr 8, 2020 at 10:17 AM Kewen.Lin via Gcc-patches
<gcc-patches@gcc.gnu.org> wrote:
>
> Hi,
>
> This is another vect case which requires special handling with
> vect_hw_misalign. The alignment of the second part requires
> misaligned vector access supports. This patch is to adjust
> the related guard condition and comments.
>
> Verified it on ppc64-redhat-linux (Power7 BE).
>
> Is it ok for trunk?
OK.
> BR,
> Kewen
> --------
>
> gcc/testsuite/ChangeLog
>
> 2020-04-08 Kewen Lin <li...@gcc.gnu.org>
>
> PR testsuite/94023
> * gfortran.dg/vect/pr83232.f90: Expect 2 rather than 3 times SLP on
> non-vect_hw_misalign targets.
>
> diff --git a/gcc/testsuite/gfortran.dg/vect/pr83232.f90
> b/gcc/testsuite/gfortran.dg/vect/pr83232.f90
> index a353578..af88342 100644
> --- a/gcc/testsuite/gfortran.dg/vect/pr83232.f90
> +++ b/gcc/testsuite/gfortran.dg/vect/pr83232.f90
> @@ -27,5 +27,7 @@
> call foo (Einc)
> END SUBROUTINE
>
> -! We should vectorize (1), (2) and (3)
> -! { dg-final { scan-tree-dump-times "vectorizing stmts using SLP" 3 "slp1" }
> }
> +! We should vectorize (1), (2) and (3) under vect_hw_misalign.
> +! { dg-final { scan-tree-dump-times "vectorizing stmts using SLP" 3 "slp1" {
> target vect_hw_misalign } } }
> +! But only (1) and (3) under !vect_hw_misalign due to the alignment of (2).
> +! { dg-final { scan-tree-dump-times "vectorizing stmts using SLP" 2 "slp1" {
> target { ! vect_hw_misalign } } } }
>
