Re: PING^2 [PATCH v2] rs6000: Use direct move for char/short vector CTOR [PR96933]

[David Edelsohn <dje.gcc@gmail.com>]

On Mon, Nov 2, 2020 at 4:11 AM Kewen.Lin <linkw@linux.ibm.com> wrote:

> >> Hi,
> >>
> >> As Segher's suggestion in the PR, for 128bit_direct_move, this new
> >> version leverages vector pack insns instead of vector perms with
> >> one control vector.  The performance evaluation shows that it's on
> >> par with the previous version for char, while it's better than the
> >> previous for short.
> >>
> >> Bootstrapped/regtested again on powerpc64{,le}-linux-gnu P8 and
> >> powerpc64le-linux-gnu P9.
> >>
> >> Is it ok for trunk?

The patch looks fine to me, but I'll let Segher decide if it addresses
his requested changes.

I'm trying to be stricter about the test cases.

+++ b/gcc/testsuite/gcc.target/powerpc/pr96933-1.c
@@ -0,0 +1,14 @@
+/* { dg-do compile { target { lp64 && has_arch_pwr9 } } } */
+/* { dg-require-effective-target powerpc_p9vector_ok } */
+/* { dg-options "-O2" } */

Why does this test has_arch_pwr9 instead of adding -mdejagnu-cpu=power9?

+++ b/gcc/testsuite/gcc.target/powerpc/pr96933-2.c
@@ -0,0 +1,14 @@
+/* { dg-do compile { target powerpc_p8vector_ok } } */
+/* { dg-options "-O2 -mdejagnu-cpu=power8" } */

Please place powerpc_p8vector_ok on a separate dg-require-effective-target line.

+++ b/gcc/testsuite/gcc.target/powerpc/pr96933-3.c
@@ -0,0 +1,63 @@
+/* { dg-do run } */
+/* { dg-require-effective-target p8vector_hw } */
+/* { dg-options "-O2" } */

Doesn't this need -mdejagnu-cpu=power8?

Thanks, David