/* { dg-do run } */
|
/* { dg-do run } */
|
/* { dg-require-effective-target fma4 } */
|
/* { dg-require-effective-target fma4 } */
|
/* { dg-options "-O0 -mfma4" } */
|
/* { dg-options "-O0 -mfma4" } */
|
|
|
#include "fma4-check.h"
|
#include "fma4-check.h"
|
|
|
#include <x86intrin.h>
|
#include <x86intrin.h>
|
#include <string.h>
|
#include <string.h>
|
|
|
#define NUM 20
|
#define NUM 20
|
|
|
union
|
union
|
{
|
{
|
__m128 x[NUM];
|
__m128 x[NUM];
|
float f[NUM * 4];
|
float f[NUM * 4];
|
__m128d y[NUM];
|
__m128d y[NUM];
|
double d[NUM * 2];
|
double d[NUM * 2];
|
} dst, res, src1, src2, src3;
|
} dst, res, src1, src2, src3;
|
|
|
/* Note that in macc*,msub*,mnmacc* and mnsub* instructions, the intermdediate
|
/* Note that in macc*,msub*,mnmacc* and mnsub* instructions, the intermdediate
|
product is not rounded, only the addition is rounded. */
|
product is not rounded, only the addition is rounded. */
|
|
|
static void
|
static void
|
init_nmaccps ()
|
init_nmaccps ()
|
{
|
{
|
int i;
|
int i;
|
for (i = 0; i < NUM * 4; i++)
|
for (i = 0; i < NUM * 4; i++)
|
{
|
{
|
src1.f[i] = i;
|
src1.f[i] = i;
|
src2.f[i] = i + 10;
|
src2.f[i] = i + 10;
|
src3.f[i] = i + 20;
|
src3.f[i] = i + 20;
|
}
|
}
|
}
|
}
|
|
|
static void
|
static void
|
init_nmaccpd ()
|
init_nmaccpd ()
|
{
|
{
|
int i;
|
int i;
|
for (i = 0; i < NUM * 4; i++)
|
for (i = 0; i < NUM * 4; i++)
|
{
|
{
|
src1.d[i] = i;
|
src1.d[i] = i;
|
src2.d[i] = i + 10;
|
src2.d[i] = i + 10;
|
src3.d[i] = i + 20;
|
src3.d[i] = i + 20;
|
}
|
}
|
}
|
}
|
|
|
static int
|
static int
|
check_nmaccps ()
|
check_nmaccps ()
|
{
|
{
|
int i, j, check_fails = 0;
|
int i, j, check_fails = 0;
|
for (i = 0; i < NUM * 4; i = i + 4)
|
for (i = 0; i < NUM * 4; i = i + 4)
|
for (j = 0; j < 4; j++)
|
for (j = 0; j < 4; j++)
|
{
|
{
|
res.f[i + j] = - (src1.f[i + j] * src2.f[i + j]) + src3.f[i + j];
|
res.f[i + j] = - (src1.f[i + j] * src2.f[i + j]) + src3.f[i + j];
|
if (dst.f[i + j] != res.f[i + j])
|
if (dst.f[i + j] != res.f[i + j])
|
check_fails++;
|
check_fails++;
|
}
|
}
|
return check_fails++;
|
return check_fails++;
|
}
|
}
|
|
|
static int
|
static int
|
check_nmaccpd ()
|
check_nmaccpd ()
|
{
|
{
|
int i, j, check_fails = 0;
|
int i, j, check_fails = 0;
|
for (i = 0; i < NUM * 2; i = i + 2)
|
for (i = 0; i < NUM * 2; i = i + 2)
|
for (j = 0; j < 2; j++)
|
for (j = 0; j < 2; j++)
|
{
|
{
|
res.d[i + j] = - (src1.d[i + j] * src2.d[i + j]) + src3.d[i + j];
|
res.d[i + j] = - (src1.d[i + j] * src2.d[i + j]) + src3.d[i + j];
|
if (dst.d[i + j] != res.d[i + j])
|
if (dst.d[i + j] != res.d[i + j])
|
check_fails++;
|
check_fails++;
|
}
|
}
|
return check_fails++;
|
return check_fails++;
|
}
|
}
|
|
|
|
|
static int
|
static int
|
check_nmaccss ()
|
check_nmaccss ()
|
{
|
{
|
int i, j, check_fails = 0;
|
int i, j, check_fails = 0;
|
for (i = 0; i < NUM * 4; i = i + 4)
|
for (i = 0; i < NUM * 4; i = i + 4)
|
{
|
{
|
res.f[i] = - (src1.f[i] * src2.f[i]) + src3.f[i];
|
res.f[i] = - (src1.f[i] * src2.f[i]) + src3.f[i];
|
if (dst.f[i] != res.f[i])
|
if (dst.f[i] != res.f[i])
|
check_fails++;
|
check_fails++;
|
}
|
}
|
return check_fails++;
|
return check_fails++;
|
}
|
}
|
|
|
static int
|
static int
|
check_nmaccsd ()
|
check_nmaccsd ()
|
{
|
{
|
int i, j, check_fails = 0;
|
int i, j, check_fails = 0;
|
for (i = 0; i < NUM * 2; i = i + 2)
|
for (i = 0; i < NUM * 2; i = i + 2)
|
{
|
{
|
res.d[i] = - (src1.d[i] * src2.d[i]) + src3.d[i];
|
res.d[i] = - (src1.d[i] * src2.d[i]) + src3.d[i];
|
if (dst.d[i] != res.d[i])
|
if (dst.d[i] != res.d[i])
|
check_fails++;
|
check_fails++;
|
}
|
}
|
return check_fails++;
|
return check_fails++;
|
}
|
}
|
|
|
static void
|
static void
|
fma4_test (void)
|
fma4_test (void)
|
{
|
{
|
int i;
|
int i;
|
|
|
init_nmaccps ();
|
init_nmaccps ();
|
|
|
for (i = 0; i < NUM; i++)
|
for (i = 0; i < NUM; i++)
|
dst.x[i] = _mm_nmacc_ps (src1.x[i], src2.x[i], src3.x[i]);
|
dst.x[i] = _mm_nmacc_ps (src1.x[i], src2.x[i], src3.x[i]);
|
|
|
if (check_nmaccps ())
|
if (check_nmaccps ())
|
abort ();
|
abort ();
|
|
|
|
|
for (i = 0; i < NUM; i++)
|
for (i = 0; i < NUM; i++)
|
dst.x[i] = _mm_nmacc_ss (src1.x[i], src2.x[i], src3.x[i]);
|
dst.x[i] = _mm_nmacc_ss (src1.x[i], src2.x[i], src3.x[i]);
|
|
|
if (check_nmaccss ())
|
if (check_nmaccss ())
|
abort ();
|
abort ();
|
|
|
init_nmaccpd ();
|
init_nmaccpd ();
|
|
|
for (i = 0; i < NUM; i++)
|
for (i = 0; i < NUM; i++)
|
dst.y[i] = _mm_nmacc_pd (src1.y[i], src2.y[i], src3.y[i]);
|
dst.y[i] = _mm_nmacc_pd (src1.y[i], src2.y[i], src3.y[i]);
|
|
|
if (check_nmaccpd ())
|
if (check_nmaccpd ())
|
abort ();
|
abort ();
|
|
|
|
|
for (i = 0; i < NUM; i++)
|
for (i = 0; i < NUM; i++)
|
dst.y[i] = _mm_nmacc_sd (src1.y[i], src2.y[i], src3.y[i]);
|
dst.y[i] = _mm_nmacc_sd (src1.y[i], src2.y[i], src3.y[i]);
|
|
|
if (check_nmaccsd ())
|
if (check_nmaccsd ())
|
abort ();
|
abort ();
|
|
|
}
|
}
|
|
|
