From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <gcc-patches-return-292813-listarch-gcc-patches=gcc.gnu.org@gcc.gnu.org>
Received: (qmail 3631 invoked by alias); 26 May 2011 09:27:15 -0000
Received: (qmail 3620 invoked by uid 22791); 26 May 2011 09:27:13 -0000
X-SWARE-Spam-Status: No, hits=-1.0 required=5.0	tests=AWL,BAYES_50,DKIM_SIGNED,DKIM_VALID,DKIM_VALID_AU,FREEMAIL_FROM,RCVD_IN_DNSWL_LOW,RFC_ABUSE_POST
X-Spam-Check-By: sourceware.org
Received: from mail-ww0-f51.google.com (HELO mail-ww0-f51.google.com) (74.125.82.51)    by sourceware.org (qpsmtpd/0.43rc1) with ESMTP; Thu, 26 May 2011 09:26:48 +0000
Received: by wwf26 with SMTP id 26so463800wwf.8        for <gcc-patches@gcc.gnu.org>; Thu, 26 May 2011 02:26:46 -0700 (PDT)
MIME-Version: 1.0
Received: by 10.227.62.207 with SMTP id y15mr572274wbh.86.1306402005240; Thu, 26 May 2011 02:26:45 -0700 (PDT)
Received: by 10.227.37.152 with HTTP; Thu, 26 May 2011 02:26:45 -0700 (PDT)
In-Reply-To: <1306352744.4821.65.camel@L3G5336.ibm.com>
References: <1306352744.4821.65.camel@L3G5336.ibm.com>
Date: Thu, 26 May 2011 10:45:00 -0000
Message-ID: <BANLkTinAWRjHF7_X3sO3SKYEJ62fQqv9iQ@mail.gmail.com>
Subject: Re: [PATCH, testsuite] Additional tests for PR46728 (PR46728 patch 4)
From: Richard Guenther <richard.guenther@gmail.com>
To: "William J. Schmidt" <wschmidt@linux.vnet.ibm.com>
Cc: gcc-patches@gcc.gnu.org
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable
X-IsSubscribed: yes
Mailing-List: contact gcc-patches-help@gcc.gnu.org; run by ezmlm
Precedence: bulk
List-Id: <gcc-patches.gcc.gnu.org>
List-Archive: <http://gcc.gnu.org/ml/gcc-patches/>
List-Post: <mailto:gcc-patches@gcc.gnu.org>
List-Help: <mailto:gcc-patches-help@gcc.gnu.org>
Sender: gcc-patches-owner@gcc.gnu.org
X-SW-Source: 2011-05/txt/msg01984.txt.bz2

On Wed, May 25, 2011 at 9:45 PM, William J. Schmidt
<wschmidt@linux.vnet.ibm.com> wrote:
> Since I'm in process of moving the lowering of pow and powi calls from
> expand into gimple, I wrote some tests to improve coverage in this area.
> Most of these look for specific code generation patterns in PowerPC
> assembly where the existence of a hardware floating square root can be
> guaranteed.
>
> This patch is conditional on patch 3 of the PR46728 series; without it,
> test pr46728-16.c will fail, since the FMA will not be generated. =A0All
> other tests currently pass.
>
> OK to add to test suite on trunk?

Ok.

Richard.

> Thanks,
> Bill
>
>
> 2011-05-25 =A0Bill Schmidt =A0<wschmidt@linux.vnet.ibm.com>
>
> =A0 =A0 =A0 =A0* gcc.target/powerpc/pr46728-1.c: New.
> =A0 =A0 =A0 =A0* gcc.target/powerpc/pr46728-2.c: New.
> =A0 =A0 =A0 =A0* gcc.target/powerpc/pr46728-3.c: New.
> =A0 =A0 =A0 =A0* gcc.target/powerpc/pr46728-4.c: New.
> =A0 =A0 =A0 =A0* gcc.target/powerpc/pr46728-5.c: New.
> =A0 =A0 =A0 =A0* gcc.dg/pr46728-6.c: New.
> =A0 =A0 =A0 =A0* gcc.target/powerpc/pr46728-7.c: New.
> =A0 =A0 =A0 =A0* gcc.target/powerpc/pr46728-8.c: New.
> =A0 =A0 =A0 =A0* gcc.dg/pr46728-9.c: New.
> =A0 =A0 =A0 =A0* gcc.target/powerpc/pr46728-10.c: New.
> =A0 =A0 =A0 =A0* gcc.target/powerpc/pr46728-11.c: New.
> =A0 =A0 =A0 =A0* gcc.dg/pr46728-12.c: New.
> =A0 =A0 =A0 =A0* gcc.target/powerpc/pr46728-13.c: New.
> =A0 =A0 =A0 =A0* gcc.target/powerpc/pr46728-14.c: New.
> =A0 =A0 =A0 =A0* gcc.target/powerpc/pr46728-15.c: New.
> =A0 =A0 =A0 =A0* gcc.target/powerpc/pr46728-16.c: New.
>
> Index: gcc/testsuite/gcc.target/powerpc/pr46728-13.c
> =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
> --- gcc/testsuite/gcc.target/powerpc/pr46728-13.c =A0 =A0 =A0 (revision 0)
> +++ gcc/testsuite/gcc.target/powerpc/pr46728-13.c =A0 =A0 =A0 (revision 0)
> @@ -0,0 +1,27 @@
> +/* { dg-do run } */
> +/* { dg-options "-O2 -ffast-math -fno-inline -fno-unroll-loops -lm -mpow=
erpc-gpopt" } */
> +
> +#include <math.h>
> +
> +extern void abort (void);
> +
> +#define NVALS 6
> +
> +static double
> +convert_it (double x)
> +{
> + =A0return pow (x, 1.0 / 6.0);
> +}
> +
> +int
> +main (int argc, char *argv[])
> +{
> + =A0double values[NVALS] =3D { 3.0, 1.95, 2.227, 729.0, 64.0, .0008797 };
> + =A0unsigned i;
> +
> + =A0for (i =3D 0; i < NVALS; i++)
> + =A0 =A0if (convert_it (values[i]) !=3D cbrt (sqrt (values[i])))
> + =A0 =A0 =A0abort ();
> +
> + =A0return 0;
> +}
> Index: gcc/testsuite/gcc.target/powerpc/pr46728-3.c
> =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
> --- gcc/testsuite/gcc.target/powerpc/pr46728-3.c =A0 =A0 =A0 =A0(revision=
 0)
> +++ gcc/testsuite/gcc.target/powerpc/pr46728-3.c =A0 =A0 =A0 =A0(revision=
 0)
> @@ -0,0 +1,31 @@
> +/* { dg-do compile } */
> +/* { dg-options "-O2 -ffast-math -fno-inline -fno-unroll-loops -lm -mpow=
erpc-gpopt" } */
> +
> +#include <math.h>
> +
> +extern void abort (void);
> +
> +#define NVALS 6
> +
> +static double
> +convert_it (double x)
> +{
> + =A0return pow (x, 0.75);
> +}
> +
> +int
> +main (int argc, char *argv[])
> +{
> + =A0double values[NVALS] =3D { 3.0, 1.95, 2.227, 4.0, 256.0, .0008797 };
> + =A0unsigned i;
> +
> + =A0for (i =3D 0; i < NVALS; i++)
> + =A0 =A0if (convert_it (values[i]) !=3D sqrt(values[i]) * sqrt (sqrt (va=
lues[i])))
> + =A0 =A0 =A0abort ();
> +
> + =A0return 0;
> +}
> +
> +
> +/* { dg-final { scan-assembler-times "sqrt" 4 { target powerpc*-*-* } } =
} */
> +/* { dg-final { scan-assembler-not "pow" { target powerpc*-*-* } } } */
> Index: gcc/testsuite/gcc.target/powerpc/pr46728-14.c
> =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
> --- gcc/testsuite/gcc.target/powerpc/pr46728-14.c =A0 =A0 =A0 (revision 0)
> +++ gcc/testsuite/gcc.target/powerpc/pr46728-14.c =A0 =A0 =A0 (revision 0)
> @@ -0,0 +1,78 @@
> +/* { dg-do run } */
> +/* { dg-options "-O2 -ffast-math -fno-inline -fno-unroll-loops -lm -mpow=
erpc-gpopt" } */
> +
> +#include <math.h>
> +
> +extern void abort (void);
> +
> +#define NVALS 6
> +
> +static double
> +convert_it_1 (double x)
> +{
> + =A0return pow (x, 1.5);
> +}
> +
> +static double
> +convert_it_2 (double x)
> +{
> + =A0return pow (x, 2.5);
> +}
> +
> +static double
> +convert_it_3 (double x)
> +{
> + =A0return pow (x, -0.5);
> +}
> +
> +static double
> +convert_it_4 (double x)
> +{
> + =A0return pow (x, 10.5);
> +}
> +
> +static double
> +convert_it_5 (double x)
> +{
> + =A0return pow (x, -3.5);
> +}
> +
> +int
> +main (int argc, char *argv[])
> +{
> + =A0double values[NVALS] =3D { 3.0, 1.95, 2.227, 4.0, 256.0, .0008797 };
> + =A0double PREC =3D .999999;
> + =A0unsigned i;
> +
> + =A0for (i =3D 0; i < NVALS; i++)
> + =A0 =A0{
> + =A0 =A0 =A0volatile double x, y;
> +
> + =A0 =A0 =A0x =3D sqrt (values[i]);
> + =A0 =A0 =A0y =3D __builtin_powi (values[i], 1);
> + =A0 =A0 =A0if (fabs (convert_it_1 (values[i]) / (x * y)) < PREC)
> + =A0 =A0 =A0 abort ();
> +
> + =A0 =A0 =A0x =3D sqrt (values[i]);
> + =A0 =A0 =A0y =3D __builtin_powi (values[i], 2);
> + =A0 =A0 =A0if (fabs (convert_it_2 (values[i]) / (x * y)) < PREC)
> + =A0 =A0 =A0 abort ();
> +
> + =A0 =A0 =A0x =3D sqrt (values[i]);
> + =A0 =A0 =A0y =3D __builtin_powi (values[i], -1);
> + =A0 =A0 =A0if (fabs (convert_it_3 (values[i]) / (x * y)) < PREC)
> + =A0 =A0 =A0 abort ();
> +
> + =A0 =A0 =A0x =3D sqrt (values[i]);
> + =A0 =A0 =A0y =3D __builtin_powi (values[i], 10);
> + =A0 =A0 =A0if (fabs (convert_it_4 (values[i]) / (x * y)) < PREC)
> + =A0 =A0 =A0 abort ();
> +
> + =A0 =A0 =A0x =3D sqrt (values[i]);
> + =A0 =A0 =A0y =3D __builtin_powi (values[i], -4);
> + =A0 =A0 =A0if (fabs (convert_it_5 (values[i]) / (x * y)) < PREC)
> + =A0 =A0 =A0 abort ();
> + =A0 =A0}
> +
> + =A0return 0;
> +}
> Index: gcc/testsuite/gcc.target/powerpc/pr46728-4.c
> =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
> --- gcc/testsuite/gcc.target/powerpc/pr46728-4.c =A0 =A0 =A0 =A0(revision=
 0)
> +++ gcc/testsuite/gcc.target/powerpc/pr46728-4.c =A0 =A0 =A0 =A0(revision=
 0)
> @@ -0,0 +1,31 @@
> +/* { dg-do compile } */
> +/* { dg-options "-O2 -ffast-math -fno-inline -fno-unroll-loops -lm -mpow=
erpc-gpopt" } */
> +
> +#include <math.h>
> +
> +extern void abort (void);
> +
> +#define NVALS 6
> +
> +static double
> +convert_it (double x)
> +{
> + =A0return pow (x, 1.0 / 3.0);
> +}
> +
> +int
> +main (int argc, char *argv[])
> +{
> + =A0double values[NVALS] =3D { 3.0, 1.95, 2.227, 729.0, 64.0, .0008797 };
> + =A0unsigned i;
> +
> + =A0for (i =3D 0; i < NVALS; i++)
> + =A0 =A0if (convert_it (values[i]) !=3D cbrt (values[i]))
> + =A0 =A0 =A0abort ();
> +
> + =A0return 0;
> +}
> +
> +
> +/* { dg-final { scan-assembler-times "cbrt" 2 { target powerpc*-*-* } } =
} */
> +/* { dg-final { scan-assembler-not "pow" { target powerpc*-*-* } } } */
> Index: gcc/testsuite/gcc.target/powerpc/pr46728-15.c
> =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
> --- gcc/testsuite/gcc.target/powerpc/pr46728-15.c =A0 =A0 =A0 (revision 0)
> +++ gcc/testsuite/gcc.target/powerpc/pr46728-15.c =A0 =A0 =A0 (revision 0)
> @@ -0,0 +1,67 @@
> +/* { dg-do run } */
> +/* { dg-options "-O2 -ffast-math -fno-inline -fno-unroll-loops -lm -mpow=
erpc-gpopt" } */
> +
> +#include <math.h>
> +
> +extern void abort (void);
> +
> +#define NVALS 6
> +
> +static double
> +convert_it_1 (double x)
> +{
> + =A0return pow (x, 10.0 / 3.0);
> +}
> +
> +static double
> +convert_it_2 (double x)
> +{
> + =A0return pow (x, 11.0 / 3.0);
> +}
> +
> +static double
> +convert_it_3 (double x)
> +{
> + =A0return pow (x, -7.0 / 3.0);
> +}
> +
> +static double
> +convert_it_4 (double x)
> +{
> + =A0return pow (x, -8.0 / 3.0);
> +}
> +
> +int
> +main (int argc, char *argv[])
> +{
> + =A0double values[NVALS] =3D { 3.0, 1.95, 2.227, 4.0, 256.0, .0008797 };
> + =A0double PREC =3D .999999;
> + =A0unsigned i;
> +
> + =A0for (i =3D 0; i < NVALS; i++)
> + =A0 =A0{
> + =A0 =A0 =A0volatile double x, y;
> +
> + =A0 =A0 =A0x =3D __builtin_powi (values[i], 3);
> + =A0 =A0 =A0y =3D __builtin_powi (cbrt (values[i]), 1);
> + =A0 =A0 =A0if (fabs (convert_it_1 (values[i]) / (x * y)) < PREC)
> + =A0 =A0 =A0 abort ();
> +
> + =A0 =A0 =A0x =3D __builtin_powi (values[i], 3);
> + =A0 =A0 =A0y =3D __builtin_powi (cbrt (values[i]), 2);
> + =A0 =A0 =A0if (fabs (convert_it_2 (values[i]) / (x * y)) < PREC)
> + =A0 =A0 =A0 abort ();
> +
> + =A0 =A0 =A0x =3D __builtin_powi (values[i], -3);
> + =A0 =A0 =A0y =3D __builtin_powi (cbrt (values[i]), 2);
> + =A0 =A0 =A0if (fabs (convert_it_3 (values[i]) / (x * y)) < PREC)
> + =A0 =A0 =A0 abort ();
> +
> + =A0 =A0 =A0x =3D __builtin_powi (values[i], -3);
> + =A0 =A0 =A0y =3D __builtin_powi (cbrt (values[i]), 1);
> + =A0 =A0 =A0if (fabs (convert_it_4 (values[i]) / (x * y)) < PREC)
> + =A0 =A0 =A0 abort ();
> + =A0 =A0}
> +
> + =A0return 0;
> +}
> Index: gcc/testsuite/gcc.target/powerpc/pr46728-5.c
> =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
> --- gcc/testsuite/gcc.target/powerpc/pr46728-5.c =A0 =A0 =A0 =A0(revision=
 0)
> +++ gcc/testsuite/gcc.target/powerpc/pr46728-5.c =A0 =A0 =A0 =A0(revision=
 0)
> @@ -0,0 +1,31 @@
> +/* { dg-do compile } */
> +/* { dg-options "-O2 -ffast-math -fno-inline -fno-unroll-loops -lm -mpow=
erpc-gpopt" } */
> +
> +#include <math.h>
> +
> +extern void abort (void);
> +
> +#define NVALS 6
> +
> +static double
> +convert_it (double x)
> +{
> + =A0return pow (x, 1.0 / 6.0);
> +}
> +
> +int
> +main (int argc, char *argv[])
> +{
> + =A0double values[NVALS] =3D { 3.0, 1.95, 2.227, 729.0, 64.0, .0008797 };
> + =A0unsigned i;
> +
> + =A0for (i =3D 0; i < NVALS; i++)
> + =A0 =A0if (convert_it (values[i]) !=3D cbrt (sqrt (values[i])))
> + =A0 =A0 =A0abort ();
> +
> + =A0return 0;
> +}
> +
> +
> +/* { dg-final { scan-assembler-times "cbrt" 2 { target powerpc*-*-* } } =
} */
> +/* { dg-final { scan-assembler-not " pow " { target powerpc*-*-* } } } */
> Index: gcc/testsuite/gcc.target/powerpc/pr46728-16.c
> =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
> --- gcc/testsuite/gcc.target/powerpc/pr46728-16.c =A0 =A0 =A0 (revision 0)
> +++ gcc/testsuite/gcc.target/powerpc/pr46728-16.c =A0 =A0 =A0 (revision 0)
> @@ -0,0 +1,10 @@
> +/* { dg-do compile } */
> +/* { dg-options "-O2 -ffast-math -mcpu=3Dpower6" } */
> +
> +double foo (double x, double y)
> +{
> + =A0return __builtin_pow (x, 0.75) + y;
> +}
> +
> +
> +/* { dg-final { scan-assembler "fmadd" { target powerpc*-*-* } } } */
> Index: gcc/testsuite/gcc.target/powerpc/pr46728-7.c
> =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
> --- gcc/testsuite/gcc.target/powerpc/pr46728-7.c =A0 =A0 =A0 =A0(revision=
 0)
> +++ gcc/testsuite/gcc.target/powerpc/pr46728-7.c =A0 =A0 =A0 =A0(revision=
 0)
> @@ -0,0 +1,57 @@
> +/* { dg-do compile } */
> +/* { dg-options "-O2 -ffast-math -fno-inline -fno-unroll-loops -lm -mpow=
erpc-gpopt" } */
> +
> +#include <math.h>
> +
> +extern void abort (void);
> +
> +#define NVALS 6
> +
> +static double
> +convert_it_1 (double x)
> +{
> + =A0return pow (x, 1.5);
> +}
> +
> +static double
> +convert_it_2 (double x)
> +{
> + =A0return pow (x, 2.5);
> +}
> +
> +static double
> +convert_it_3 (double x)
> +{
> + =A0return pow (x, -0.5);
> +}
> +
> +static double
> +convert_it_4 (double x)
> +{
> + =A0return pow (x, 10.5);
> +}
> +
> +int
> +main (int argc, char *argv[])
> +{
> + =A0double values[NVALS] =3D { 3.0, 1.95, 2.227, 4.0, 256.0, .0008797 };
> + =A0unsigned i;
> +
> + =A0for (i =3D 0; i < NVALS; i++)
> + =A0 =A0{
> + =A0 =A0 =A0if (convert_it_1 (values[i]) !=3D sqrt (values[i]) * __built=
in_powi (values[i], 1))
> + =A0 =A0 =A0 abort ();
> + =A0 =A0 =A0if (convert_it_2 (values[i]) !=3D sqrt (values[i]) * __built=
in_powi (values[i], 2))
> + =A0 =A0 =A0 abort ();
> + =A0 =A0 =A0if (convert_it_3 (values[i]) !=3D sqrt (values[i]) * __built=
in_powi (values[i], -1))
> + =A0 =A0 =A0 abort ();
> + =A0 =A0 =A0if (convert_it_4 (values[i]) !=3D sqrt (values[i]) * __built=
in_powi (values[i], 10))
> + =A0 =A0 =A0 abort ();
> + =A0 =A0}
> +
> + =A0return 0;
> +}
> +
> +
> +/* { dg-final { scan-assembler-times "sqrt" 5 { target powerpc*-*-* } } =
} */
> +/* { dg-final { scan-assembler-not "pow" { target powerpc*-*-* } } } */
> Index: gcc/testsuite/gcc.target/powerpc/pr46728-10.c
> =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
> --- gcc/testsuite/gcc.target/powerpc/pr46728-10.c =A0 =A0 =A0 (revision 0)
> +++ gcc/testsuite/gcc.target/powerpc/pr46728-10.c =A0 =A0 =A0 (revision 0)
> @@ -0,0 +1,28 @@
> +/* { dg-do run } */
> +/* { dg-options "-O2 -ffast-math -fno-inline -fno-unroll-loops -lm -mpow=
erpc-gpopt" } */
> +
> +#include <math.h>
> +
> +extern void abort (void);
> +
> +#define NVALS 6
> +
> +static double
> +convert_it (double x)
> +{
> + =A0return pow (x, 0.25);
> +}
> +
> +int
> +main (int argc, char *argv[])
> +{
> + =A0double values[NVALS] =3D { 3.0, 1.95, 2.227, 4.0, 256.0, .0008797 };
> + =A0unsigned i;
> +
> + =A0for (i =3D 0; i < NVALS; i++)
> + =A0 =A0if (convert_it (values[i]) !=3D sqrt (sqrt (values[i])))
> + =A0 =A0 =A0abort ();
> +
> + =A0return 0;
> +}
> +
> Index: gcc/testsuite/gcc.target/powerpc/pr46728-8.c
> =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
> --- gcc/testsuite/gcc.target/powerpc/pr46728-8.c =A0 =A0 =A0 =A0(revision=
 0)
> +++ gcc/testsuite/gcc.target/powerpc/pr46728-8.c =A0 =A0 =A0 =A0(revision=
 0)
> @@ -0,0 +1,61 @@
> +/* { dg-do compile } */
> +/* { dg-options "-O2 -ffast-math -fno-inline -fno-unroll-loops -lm -mpow=
erpc-gpopt" } */
> +
> +#include <math.h>
> +
> +extern void abort (void);
> +
> +#define NVALS 6
> +
> +static double
> +convert_it_1 (double x)
> +{
> + =A0return pow (x, 10.0 / 3.0);
> +}
> +
> +static double
> +convert_it_2 (double x)
> +{
> + =A0return pow (x, 11.0 / 3.0);
> +}
> +
> +static double
> +convert_it_3 (double x)
> +{
> + =A0return pow (x, -7.0 / 3.0);
> +}
> +
> +static double
> +convert_it_4 (double x)
> +{
> + =A0return pow (x, -8.0 / 3.0);
> +}
> +
> +int
> +main (int argc, char *argv[])
> +{
> + =A0double values[NVALS] =3D { 3.0, 1.95, 2.227, 4.0, 256.0, .0008797 };
> + =A0unsigned i;
> +
> + =A0for (i =3D 0; i < NVALS; i++)
> + =A0 =A0{
> + =A0 =A0 =A0if (convert_it_1 (values[i]) !=3D
> + =A0 =A0 =A0 =A0 __builtin_powi (values[i], 3) * __builtin_powi (cbrt (v=
alues[i]), 1))
> + =A0 =A0 =A0 abort ();
> + =A0 =A0 =A0if (convert_it_2 (values[i]) !=3D
> + =A0 =A0 =A0 =A0 __builtin_powi (values[i], 3) * __builtin_powi (cbrt (v=
alues[i]), 2))
> + =A0 =A0 =A0 abort ();
> + =A0 =A0 =A0if (convert_it_3 (values[i]) !=3D
> + =A0 =A0 =A0 =A0 __builtin_powi (values[i], -3) * __builtin_powi (cbrt (=
values[i]), 2))
> + =A0 =A0 =A0 abort ();
> + =A0 =A0 =A0if (convert_it_4 (values[i]) !=3D
> + =A0 =A0 =A0 =A0 __builtin_powi (values[i], -3) * __builtin_powi (cbrt (=
values[i]), 1))
> + =A0 =A0 =A0 abort ();
> + =A0 =A0}
> +
> + =A0return 0;
> +}
> +
> +
> +/* { dg-final { scan-assembler-times "cbrt" 5 { target powerpc*-*-* } } =
} */
> +/* { dg-final { scan-assembler-not "pow" { target powerpc*-*-* } } } */
> Index: gcc/testsuite/gcc.target/powerpc/pr46728-11.c
> =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
> --- gcc/testsuite/gcc.target/powerpc/pr46728-11.c =A0 =A0 =A0 (revision 0)
> +++ gcc/testsuite/gcc.target/powerpc/pr46728-11.c =A0 =A0 =A0 (revision 0)
> @@ -0,0 +1,34 @@
> +/* { dg-do run } */
> +/* { dg-options "-O2 -ffast-math -fno-inline -fno-unroll-loops -lm -mpow=
erpc-gpopt" } */
> +
> +#include <math.h>
> +
> +extern void abort (void);
> +
> +#define NVALS 6
> +
> +static double
> +convert_it (double x)
> +{
> + =A0return pow (x, 0.75);
> +}
> +
> +int
> +main (int argc, char *argv[])
> +{
> + =A0double values[NVALS] =3D { 3.0, 1.95, 2.227, 4.0, 256.0, .0008797 };
> + =A0double PREC =3D 0.999999;
> + =A0unsigned i;
> +
> + =A0for (i =3D 0; i < NVALS; i++)
> + =A0 =A0{
> + =A0 =A0 =A0volatile double x, y;
> + =A0 =A0 =A0x =3D sqrt (values[i]);
> + =A0 =A0 =A0y =3D sqrt (sqrt (values[i]));
> +
> + =A0 =A0 =A0if (fabs (convert_it (values[i]) / (x * y)) < PREC)
> + =A0 =A0 =A0 abort ();
> + =A0 =A0}
> +
> + =A0return 0;
> +}
> Index: gcc/testsuite/gcc.target/powerpc/pr46728-1.c
> =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
> --- gcc/testsuite/gcc.target/powerpc/pr46728-1.c =A0 =A0 =A0 =A0(revision=
 0)
> +++ gcc/testsuite/gcc.target/powerpc/pr46728-1.c =A0 =A0 =A0 =A0(revision=
 0)
> @@ -0,0 +1,31 @@
> +/* { dg-do compile } */
> +/* { dg-options "-O2 -ffast-math -fno-inline -fno-unroll-loops -lm -mpow=
erpc-gpopt" } */
> +
> +#include <math.h>
> +
> +extern void abort (void);
> +
> +#define NVALS 6
> +
> +static double
> +convert_it (double x)
> +{
> + =A0return pow (x, 0.5);
> +}
> +
> +int
> +main (int argc, char *argv[])
> +{
> + =A0double values[NVALS] =3D { 3.0, 1.95, 2.227, 4.0, 256.0, .0008797 };
> + =A0unsigned i;
> +
> + =A0for (i =3D 0; i < NVALS; i++)
> + =A0 =A0if (convert_it (values[i]) !=3D sqrt (values[i]))
> + =A0 =A0 =A0abort ();
> +
> + =A0return 0;
> +}
> +
> +
> +/* { dg-final { scan-assembler-times "fsqrt" 2 { target powerpc*-*-* } }=
 } */
> +/* { dg-final { scan-assembler-not "pow" { target powerpc*-*-* } } } */
> Index: gcc/testsuite/gcc.target/powerpc/pr46728-2.c
> =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
> --- gcc/testsuite/gcc.target/powerpc/pr46728-2.c =A0 =A0 =A0 =A0(revision=
 0)
> +++ gcc/testsuite/gcc.target/powerpc/pr46728-2.c =A0 =A0 =A0 =A0(revision=
 0)
> @@ -0,0 +1,31 @@
> +/* { dg-do compile } */
> +/* { dg-options "-O2 -ffast-math -fno-inline -fno-unroll-loops -lm -mpow=
erpc-gpopt" } */
> +
> +#include <math.h>
> +
> +extern void abort (void);
> +
> +#define NVALS 6
> +
> +static double
> +convert_it (double x)
> +{
> + =A0return pow (x, 0.25);
> +}
> +
> +int
> +main (int argc, char *argv[])
> +{
> + =A0double values[NVALS] =3D { 3.0, 1.95, 2.227, 4.0, 256.0, .0008797 };
> + =A0unsigned i;
> +
> + =A0for (i =3D 0; i < NVALS; i++)
> + =A0 =A0if (convert_it (values[i]) !=3D sqrt (sqrt (values[i])))
> + =A0 =A0 =A0abort ();
> +
> + =A0return 0;
> +}
> +
> +
> +/* { dg-final { scan-assembler-times "fsqrt" 4 { target powerpc*-*-* } }=
 } */
> +/* { dg-final { scan-assembler-not "pow" { target powerpc*-*-* } } } */
> Index: gcc/testsuite/gcc.dg/pr46728-9.c
> =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
> --- gcc/testsuite/gcc.dg/pr46728-9.c =A0 =A0(revision 0)
> +++ gcc/testsuite/gcc.dg/pr46728-9.c =A0 =A0(revision 0)
> @@ -0,0 +1,29 @@
> +/* { dg-do run } */
> +/* { dg-options "-O2 -ffast-math -fno-inline -fno-unroll-loops -lm" } */
> +
> +#include <math.h>
> +
> +extern void abort (void);
> +
> +#define NVALS 6
> +
> +static double
> +convert_it (double x)
> +{
> + =A0return pow (x, 0.5);
> +}
> +
> +int
> +main (int argc, char *argv[])
> +{
> + =A0double values[NVALS] =3D { 3.0, 1.95, 2.227, 4.0, 256.0, .0008797 };
> + =A0double PREC =3D 0.999999;
> + =A0unsigned i;
> +
> + =A0for (i =3D 0; i < NVALS; i++)
> + =A0 =A0if (fabs (convert_it (values[i]) / sqrt (values[i])) < PREC)
> + =A0 =A0 =A0abort ();
> +
> + =A0return 0;
> +}
> +
> Index: gcc/testsuite/gcc.dg/pr46728-12.c
> =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
> --- gcc/testsuite/gcc.dg/pr46728-12.c =A0 (revision 0)
> +++ gcc/testsuite/gcc.dg/pr46728-12.c =A0 (revision 0)
> @@ -0,0 +1,28 @@
> +/* { dg-do run } */
> +/* { dg-options "-O2 -ffast-math -fno-inline -fno-unroll-loops -lm" } */
> +
> +#include <math.h>
> +
> +extern void abort (void);
> +
> +#define NVALS 6
> +
> +static double
> +convert_it (double x)
> +{
> + =A0return pow (x, 1.0 / 3.0);
> +}
> +
> +int
> +main (int argc, char *argv[])
> +{
> + =A0double values[NVALS] =3D { 3.0, 1.95, 2.227, 729.0, 64.0, .0008797 };
> + =A0double PREC =3D 0.999999;
> + =A0unsigned i;
> +
> + =A0for (i =3D 0; i < NVALS; i++)
> + =A0 =A0if (fabs (convert_it (values[i]) / cbrt (values[i])) < PREC)
> + =A0 =A0 =A0abort ();
> +
> + =A0return 0;
> +}
> Index: gcc/testsuite/gcc.dg/pr46728-6.c
> =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
> --- gcc/testsuite/gcc.dg/pr46728-6.c =A0 =A0(revision 0)
> +++ gcc/testsuite/gcc.dg/pr46728-6.c =A0 =A0(revision 0)
> @@ -0,0 +1,21 @@
> +/* { dg-do compile } */
> +/* { dg-options "-O2 -ffast-math -lm" } */
> +
> +#include <math.h>
> +
> +int
> +main (int argc, char *argv[])
> +{
> + =A0volatile double result;
> +
> + =A0result =3D pow (-0.0, 3.0);
> + =A0result =3D pow (26.47, -2.0);
> + =A0result =3D pow (0.0, 0.0);
> + =A0result =3D pow (22.3, 1.0);
> + =A0result =3D pow (33.2, -1.0);
> +
> + =A0return 0;
> +}
> +
> +
> +/* { dg-final { scan-assembler-not "pow" } } */
>
>
>