+2015-07-13 Maxim Blumenthal <maxim.blumenthal@intel.com>
+
+ * testsuite/libgomp.c++/examples-4/e.53.2.C: Renamed to...
+ * testsuite/libgomp.c++/examples-4/declare_target-2.C: ...this.
+ * testsuite/libgomp.c++/examples-4/e.51.5.C: Renamed to...
+ * testsuite/libgomp.c++/examples-4/target_data-5.C: ...this.
+ * testsuite/libgomp.c/examples-4/e.56.3.c: Renamed to...
+ * testsuite/libgomp.c/examples-4/array_sections-3.c: ...this.
+ * testsuite/libgomp.c/examples-4/e.56.4.c: Renamed to...
+ * testsuite/libgomp.c/examples-4/array_sections-4.c: ...this.
+ * testsuite/libgomp.c/examples-4/e.55.1.c: Renamed to...
+ * testsuite/libgomp.c/examples-4/async_target-1.c: ...this.
+ * testsuite/libgomp.c/examples-4/e.55.2.c: Renamed to...
+ * testsuite/libgomp.c/examples-4/async_target-2.c: ...this.
+ (vec_mult_ref): Remove v1 and v2 arguments, turn them into local
+ variables.
+ (vec_mult): Likewise. Add #pragma omp taskwait.
+ (main): Adjust caller.
+ * testsuite/libgomp.c/examples-4/e.53.1.c: Renamed to...
+ * testsuite/libgomp.c/examples-4/declare_target-1.c: ...this.
+ * testsuite/libgomp.c/examples-4/e.53.3.c: Renamed to...
+ * testsuite/libgomp.c/examples-4/declare_target-3.c: ...this.
+ * testsuite/libgomp.c/examples-4/e.53.4.c: Renamed to...
+ * testsuite/libgomp.c/examples-4/declare_target-4.c: ...this.
+ * testsuite/libgomp.c/examples-4/e.53.5.c: Renamed to...
+ * testsuite/libgomp.c/examples-4/declare_target-5.c: ...this.
+ * testsuite/libgomp.c/examples-4/e.57.1.c: Renamed to...
+ * testsuite/libgomp.c/examples-4/device-1.c: ...this.
+ * testsuite/libgomp.c/examples-4/e.57.2.c: Renamed to...
+ * testsuite/libgomp.c/examples-4/device-2.c: ...this.
+ * testsuite/libgomp.c/examples-4/e.57.3.c: Renamed to...
+ * testsuite/libgomp.c/examples-4/device-3.c: ...this.
+ * testsuite/libgomp.c/examples-4/simd-1.c: New file.
+ * testsuite/libgomp.c/examples-4/simd-2.c: New file.
+ * testsuite/libgomp.c/examples-4/simd-3.c: New file.
+ * testsuite/libgomp.c/examples-4/simd-4.c: New file.
+ * testsuite/libgomp.c/examples-4/simd-5.c: New file.
+ * testsuite/libgomp.c/examples-4/simd-6.c: New file.
+ * testsuite/libgomp.c/examples-4/simd-7.c: New file.
+ * testsuite/libgomp.c/examples-4/simd-8.c: New file.
+ * testsuite/libgomp.c/examples-4/e.50.1.c: Renamed to...
+ * testsuite/libgomp.c/examples-4/target-1.c: ...this.
+ * testsuite/libgomp.c/examples-4/e.50.2.c: Renamed to...
+ * testsuite/libgomp.c/examples-4/target-2.c: ...this.
+ * testsuite/libgomp.c/examples-4/e.50.3.c: Renamed to...
+ * testsuite/libgomp.c/examples-4/target-3.c: ...this.
+ * testsuite/libgomp.c/examples-4/e.50.4.c: Renamed to...
+ * testsuite/libgomp.c/examples-4/target-4.c: ...this.
+ * testsuite/libgomp.c/examples-4/e.50.5.c: Renamed to...
+ * testsuite/libgomp.c/examples-4/target-5.c: ...this.
+ * testsuite/libgomp.c/examples-4/e.51.1.c: Renamed to...
+ * testsuite/libgomp.c/examples-4/target_data-1.c: ...this.
+ * testsuite/libgomp.c/examples-4/e.51.2.c: Renamed to...
+ * testsuite/libgomp.c/examples-4/target_data-2.c: ...this.
+ * testsuite/libgomp.c/examples-4/e.51.3.c: Renamed to...
+ * testsuite/libgomp.c/examples-4/target_data-3.c: ...this.
+ * testsuite/libgomp.c/examples-4/e.51.4.c: Renamed to...
+ * testsuite/libgomp.c/examples-4/target_data-4.c: ...this.
+ * testsuite/libgomp.c/examples-4/e.51.6.c: Renamed to...
+ * testsuite/libgomp.c/examples-4/target_data-6.c: ...this.
+ * testsuite/libgomp.c/examples-4/e.51.7.c: Renamed to...
+ * testsuite/libgomp.c/examples-4/target_data-7.c: ...this.
+ * testsuite/libgomp.c/examples-4/e.52.1.c: Renamed to...
+ * testsuite/libgomp.c/examples-4/target_update-1.c: ...this.
+ * testsuite/libgomp.c/examples-4/e.52.2.c: Renamed to...
+ * testsuite/libgomp.c/examples-4/target_update-2.c: ...this.
+ * testsuite/libgomp.c/examples-4/task_dep-1.c: New file.
+ * testsuite/libgomp.c/examples-4/task_dep-2.c: New file.
+ * testsuite/libgomp.c/examples-4/task_dep-3.c: New file.
+ * testsuite/libgomp.c/examples-4/task_dep-4.c: New file.
+ * testsuite/libgomp.c/examples-4/task_dep-5.c: New file.
+ * testsuite/libgomp.c/examples-4/e.54.2.c: Renamed to...
+ * testsuite/libgomp.c/examples-4/teams-2.c: ...this.
+ * testsuite/libgomp.c/examples-4/e.54.3.c: Renamed to...
+ * testsuite/libgomp.c/examples-4/teams-3.c: ...this.
+ * testsuite/libgomp.c/examples-4/e.54.4.c: Renamed to...
+ * testsuite/libgomp.c/examples-4/teams-4.c: ...this.
+ * testsuite/libgomp.c/examples-4/e.54.5.c: Renamed to...
+ * testsuite/libgomp.c/examples-4/teams-5.c: ...this.
+ * testsuite/libgomp.c/examples-4/e.54.6.c: Renamed to...
+ * testsuite/libgomp.c/examples-4/teams-6.c: ...this.
+ * testsuite/libgomp.fortran/examples-4/e.56.3.f90: Renamed to...
+ * testsuite/libgomp.fortran/examples-4/array_sections-3.f90: ...this.
+ * testsuite/libgomp.fortran/examples-4/e.56.4.f90: Renamed to...
+ * testsuite/libgomp.fortran/examples-4/array_sections-4.f90: ...this.
+ * testsuite/libgomp.fortran/examples-4/e.55.1.f90: Renamed to...
+ * testsuite/libgomp.fortran/examples-4/async_target-1.f90: ...this.
+ * testsuite/libgomp.fortran/examples-4/e.55.2.f90: Renamed to...
+ * testsuite/libgomp.fortran/examples-4/async_target-2.f90: ...this.
+ (vec_mult): Add !$omp taskwait.
+ * testsuite/libgomp.fortran/examples-4/e.53.1.f90: Renamed to...
+ * testsuite/libgomp.fortran/examples-4/declare_target-1.f90: ...this.
+ * testsuite/libgomp.fortran/examples-4/e.53.2.f90: Renamed to...
+ * testsuite/libgomp.fortran/examples-4/declare_target-2.f90: ...this.
+ * testsuite/libgomp.fortran/examples-4/e.53.3.f90: Renamed to...
+ * testsuite/libgomp.fortran/examples-4/declare_target-3.f90: ...this.
+ * testsuite/libgomp.fortran/examples-4/e.53.4.f90: Renamed to...
+ * testsuite/libgomp.fortran/examples-4/declare_target-4.f90: ...this.
+ * testsuite/libgomp.fortran/examples-4/e.53.5.f90: Renamed to...
+ * testsuite/libgomp.fortran/examples-4/declare_target-5.f90: ...this.
+ * testsuite/libgomp.fortran/examples-4/e.57.1.f90: Renamed to...
+ * testsuite/libgomp.fortran/examples-4/device-1.f90: ...this.
+ * testsuite/libgomp.fortran/examples-4/e.57.2.f90: Renamed to...
+ * testsuite/libgomp.fortran/examples-4/device-2.f90: ...this.
+ * testsuite/libgomp.fortran/examples-4/e.57.3.f90: Renamed to...
+ * testsuite/libgomp.fortran/examples-4/device-3.f90: ...this.
+ * testsuite/libgomp.fortran/examples-4/simd-1.f90: New file.
+ * testsuite/libgomp.fortran/examples-4/simd-2.f90: New file.
+ * testsuite/libgomp.fortran/examples-4/simd-3.f90: New file.
+ * testsuite/libgomp.fortran/examples-4/simd-4.f90: New file.
+ * testsuite/libgomp.fortran/examples-4/simd-5.f90: New file.
+ * testsuite/libgomp.fortran/examples-4/simd-6.f90: New file.
+ * testsuite/libgomp.fortran/examples-4/simd-7.f90: New file.
+ * testsuite/libgomp.fortran/examples-4/simd-8.f90: New file.
+ * testsuite/libgomp.fortran/examples-4/e.50.1.f90: Renamed to...
+ * testsuite/libgomp.fortran/examples-4/target-1.f90: ...this.
+ * testsuite/libgomp.fortran/examples-4/e.50.2.f90: Renamed to...
+ * testsuite/libgomp.fortran/examples-4/target-2.f90: ...this.
+ * testsuite/libgomp.fortran/examples-4/e.50.3.f90: Renamed to...
+ * testsuite/libgomp.fortran/examples-4/target-3.f90: ...this.
+ * testsuite/libgomp.fortran/examples-4/e.50.4.f90: Renamed to...
+ * testsuite/libgomp.fortran/examples-4/target-4.f90: ...this.
+ * testsuite/libgomp.fortran/examples-4/e.50.5.f90: Renamed to...
+ * testsuite/libgomp.fortran/examples-4/target-5.f90: ...this.
+ * testsuite/libgomp.fortran/examples-4/e.51.1.f90: Renamed to...
+ * testsuite/libgomp.fortran/examples-4/target_data-1.f90: ...this.
+ * testsuite/libgomp.fortran/examples-4/e.51.2.f90: Renamed to...
+ * testsuite/libgomp.fortran/examples-4/target_data-2.f90: ...this.
+ * testsuite/libgomp.fortran/examples-4/e.51.3.f90: Renamed to...
+ * testsuite/libgomp.fortran/examples-4/target_data-3.f90: ...this.
+ * testsuite/libgomp.fortran/examples-4/e.51.4.f90: Renamed to...
+ * testsuite/libgomp.fortran/examples-4/target_data-4.f90: ...this.
+ * testsuite/libgomp.fortran/examples-4/e.51.5.f90: Renamed to...
+ * testsuite/libgomp.fortran/examples-4/target_data-5.f90: ...this.
+ * testsuite/libgomp.fortran/examples-4/e.51.6.f90: Renamed to...
+ * testsuite/libgomp.fortran/examples-4/target_data-6.f90: ...this.
+ * testsuite/libgomp.fortran/examples-4/e.51.7.f90: Renamed to...
+ * testsuite/libgomp.fortran/examples-4/target_data-7.f90: ...this.
+ * testsuite/libgomp.fortran/examples-4/e.52.1.f90: Renamed to...
+ * testsuite/libgomp.fortran/examples-4/target_update-1.f90: ...this.
+ * testsuite/libgomp.fortran/examples-4/e.52.2.f90: Renamed to...
+ * testsuite/libgomp.fortran/examples-4/target_update-2.f90: ...this.
+ * testsuite/libgomp.fortran/examples-4/task_dep-1.f90: New file.
+ * testsuite/libgomp.fortran/examples-4/task_dep-2.f90: New file.
+ * testsuite/libgomp.fortran/examples-4/task_dep-3.f90: New file.
+ * testsuite/libgomp.fortran/examples-4/task_dep-4.f90: New file.
+ * testsuite/libgomp.fortran/examples-4/task_dep-5.f90: New file.
+ * testsuite/libgomp.fortran/examples-4/e.54.2.f90: Renamed to...
+ * testsuite/libgomp.fortran/examples-4/teams-2.f90: ...this.
+ * testsuite/libgomp.fortran/examples-4/e.54.3.f90: Renamed to...
+ * testsuite/libgomp.fortran/examples-4/teams-3.f90: ...this.
+ * testsuite/libgomp.fortran/examples-4/e.54.4.f90: Renamed to...
+ * testsuite/libgomp.fortran/examples-4/teams-4.f90: ...this.
+ * testsuite/libgomp.fortran/examples-4/e.54.5.f90: Renamed to...
+ * testsuite/libgomp.fortran/examples-4/teams-5.f90: ...this.
+ * testsuite/libgomp.fortran/examples-4/e.54.6.f90: Renamed to...
+ * testsuite/libgomp.fortran/examples-4/teams-6.f90: ...this.
+
2015-07-10 Tom de Vries <tom@codesourcery.com>
* testsuite/libgomp.fortran/parloops-exit-first-loop-alt-2.f95: New test.
}
#pragma omp end declare target
-void vec_mult_ref (float *p, float *v1, float *v2, int n)
+void vec_mult_ref(float *p, int n)
{
+ float *v1, *v2;
int i;
v1 = (float *) malloc (n * sizeof (float));
free (v2);
}
-void vec_mult (float *p, float *v1, float *v2, int n)
+void vec_mult(float *p, int n)
{
+ float *v1, *v2;
int i;
#pragma omp task shared(v1, v2) depend(out: v1, v2)
free (v1);
free (v2);
}
+
+ #pragma omp taskwait
}
void check (float *a, float *b, int n)
{
float *p1 = (float *) malloc (N * sizeof (float));
float *p2 = (float *) malloc (N * sizeof (float));
- float *v1, *v2;
- vec_mult_ref (p1, v1, v2, N);
- vec_mult (p2, v1, v2, N);
+ vec_mult_ref (p1, N);
+ vec_mult (p2, N);
check (p1, p2, N);
--- /dev/null
+/* { dg-do run } */
+/* { dg-additional-options "-msse2" { target sse2_runtime } } */
+/* { dg-additional-options "-mavx" { target avx_runtime } } */
+
+#define N 100
+#define OFF 32
+#define EPS 0.0000000000000001
+
+#include <stdlib.h>
+
+void init(double *a, double *a_ref, double *b, double *c, int n, int ioff)
+{
+ int i;
+ for ( i = 0; i < n; i++ )
+ {
+ a[i] = i*i;
+ a_ref[i] = a[i];
+ b[i] = i+i;
+ }
+
+ int s = -1;
+ for ( i = 0; i < n+ioff; i++ )
+ {
+ c[i] = s*3;
+ s = -s;
+ }
+}
+
+void star( double *a, double *b, double *c, int n, int *ioff )
+{
+ int i;
+ #pragma omp simd
+ for ( i = 0; i < n; i++ )
+ a[i] *= b[i] * c[i+ *ioff];
+}
+
+void star_ref( double *a, double *b, double *c, int n, int *ioff )
+{
+ int i;
+ for ( i = 0; i < n; i++ )
+ a[i] *= b[i] * c[i+ *ioff];
+}
+
+void check (double *a, double *b)
+{
+ int i;
+ for (i = 0; i < N; i++)
+ if (a[i] - b[i] > EPS || b[i] - a[i] > EPS)
+ abort ();
+}
+
+int main ()
+{
+ double a[N], a_ref[N], b[N], c[N+OFF];
+ int ioff = OFF;
+
+ init(a, a_ref, b, c, N, ioff);
+
+ star(a, b, c, N, &ioff);
+ star_ref(a_ref, b, c, N, &ioff);
+
+ check(a, a_ref);
+
+ return 0;
+}
--- /dev/null
+/* { dg-do run { target vect_simd_clones } } */
+/* { dg-additional-options "-msse2" { target sse2_runtime } } */
+/* { dg-additional-options "-mavx" { target avx_runtime } } */
+
+#define N 100
+#define EPS 0.0000000000000001
+
+#include <stdlib.h>
+
+void init(double *a, double *a_ref, double *b, int n)
+{
+ int i;
+ for ( i=0; i<N; i++ )
+ {
+ a[i] = i;
+ a_ref[i] = i;
+ b[i] = N-i;
+ }
+}
+
+#pragma omp declare simd uniform(fact)
+double add1(double a, double b, double fact)
+{
+ double c;
+ c = a + b + fact;
+ return c;
+}
+
+#pragma omp declare simd uniform(a,b,fact) linear(i:1)
+double add2(double *a, double *b, int i, double fact)
+{
+ double c;
+ c = a[i] + b[i] + fact;
+ return c;
+}
+
+#pragma omp declare simd uniform(fact) linear(a,b:1)
+double add3(double *a, double *b, double fact)
+{
+ double c;
+ c = *a + *b + fact;
+ return c;
+}
+
+void work( double *a, double *b, int n )
+{
+ int i;
+ double tmp;
+ #pragma omp simd private(tmp)
+ for ( i = 0; i < n; i++ ) {
+ tmp = add1( a[i], b[i], 1.0);
+ a[i] = add2( a, b, i, 1.0) + tmp;
+ a[i] = add3(&a[i], &b[i], 1.0);
+ }
+}
+
+void work_ref( double *a, double *b, int n )
+{
+ int i;
+ double tmp;
+ for ( i = 0; i < n; i++ ) {
+ tmp = add1( a[i], b[i], 1.0);
+ a[i] = add2( a, b, i, 1.0) + tmp;
+ a[i] = add3(&a[i], &b[i], 1.0);
+ }
+}
+
+void check (double *a, double *b)
+{
+ int i;
+ for (i = 0; i < N; i++)
+ if (a[i] - b[i] > EPS || b[i] - a[i] > EPS)
+ abort ();
+}
+
+
+int main ()
+{
+ int i;
+ double a[N], a_ref[N], b[N];
+
+ init(a, a_ref, b, N);
+
+ work(a, b, N );
+ work_ref(a_ref, b, N );
+
+ check(a, a_ref);
+
+ return 0;
+}
--- /dev/null
+/* { dg-do run } */
+/* { dg-additional-options "-msse2" { target sse2_runtime } } */
+/* { dg-additional-options "-mavx" { target avx_runtime } } */
+
+#define N 100
+#define EPS 0.0000000000000001
+
+#include <stdlib.h>
+
+void init(double *a, double *a_ref, double *b, int n)
+{
+ int i, s = -1;
+ for ( i = 0; i < n; i++ )
+ {
+ a[i] = i*i*s;
+ a_ref[i] = a[i];
+ b[i] = i+i;
+ s = -s;
+ }
+}
+
+double work( double *a, double *b, int n )
+{
+ int i;
+ double tmp, sum;
+ sum = 0.0;
+ #pragma omp simd private(tmp) reduction(+:sum)
+ for (i = 0; i < n; i++) {
+ tmp = a[i] + b[i];
+ sum += tmp;
+ }
+ return sum;
+}
+
+double work_ref( double *a, double *b, int n )
+{
+ int i;
+ double tmp, sum;
+ sum = 0.0;
+ for (i = 0; i < n; i++) {
+ tmp = a[i] + b[i];
+ sum += tmp;
+ }
+ return sum;
+}
+
+int main ()
+{
+ double a[N], a_ref[N], b[N];
+ int res, ref;
+
+ init(a, a_ref, b, N);
+
+ res = work(a, b, N);
+ ref = work_ref(a_ref, b, N);
+
+ if (res != ref)
+ abort ();
+
+ return 0;
+}
--- /dev/null
+/* { dg-do run } */
+/* { dg-additional-options "-msse2" { target sse2_runtime } } */
+/* { dg-additional-options "-mavx" { target avx_runtime } } */
+
+#define N 128
+#define M 16
+#define EPS 0.0000000000000001
+#define SAFELEN 16
+
+#include <stdlib.h>
+
+void init(double *a, double *b, int n)
+{
+ int i, s = -1;
+ for ( i = 0; i < n; i++ )
+ {
+ a[i] = i*i*s;
+ b[i] = a[i];
+ s = -s;
+ }
+}
+
+void work( double *b, int n, int m )
+{
+ int i;
+ #pragma omp simd safelen(SAFELEN)
+ for (i = m; i < n; i++)
+ b[i] = b[i-m] - 1.0f;
+}
+
+void work_ref( double *b, int n, int m )
+{
+ int i;
+ for (i = m; i < n; i++)
+ b[i] = b[i-m] - 1.0f;
+}
+
+void check (double *a, double *b)
+{
+ int i;
+ for (i = 0; i < N; i++)
+ if (a[i] - b[i] > EPS || b[i] - a[i] > EPS)
+ abort ();
+}
+
+int main ()
+{
+ double b[N], b_ref[N];
+
+ init(b, b_ref, N);
+
+ work(b, N, M);
+ work(b_ref, N, M);
+
+ check(b, b_ref);
+
+ return 0;
+}
--- /dev/null
+/* { dg-do run } */
+/* { dg-additional-options "-msse2" { target sse2_runtime } } */
+/* { dg-additional-options "-mavx" { target avx_runtime } } */
+
+#define N 128
+#define M 16
+#define EPS 0.0000000000000001
+#define SAFELEN 16
+
+#include <stdlib.h>
+
+void init(double a[N][N], double b[N][N], int n)
+{
+ int i, j, s = -1;
+ for (i = 0; i < n; i++)
+ {
+ for (j = 0; j < n; j++)
+ {
+ a[i][j] = i * j * s;
+ b[i][j] = i + j + s;
+ s = -s;
+ }
+ }
+}
+
+void work( double a[N][N], double b[N][N], double c[N][N], int n )
+{
+ int i, j;
+ double tmp;
+ #pragma omp for simd collapse(2) private(tmp)
+ for (i = 0; i < n; i++)
+ {
+ for (j = 0; j < n; j++)
+ {
+ tmp = a[i][j] + b[i][j];
+ c[i][j] = tmp;
+ }
+ }
+}
+
+void work_ref( double a[N][N], double b[N][N], double c[N][N], int n )
+{
+ int i, j;
+ double tmp;
+ for (i = 0; i < n; i++)
+ {
+ for (j = 0; j < n; j++)
+ {
+ tmp = a[i][j] + b[i][j];
+ c[i][j] = tmp;
+ }
+ }
+}
+
+void check (double a[N][N], double b[N][N])
+{
+ int i, j;
+ for (i = 0; i < N; i++)
+ for (j = 0; j < N; j++)
+ if (a[i][j] - b[i][j] > EPS || b[i][j] - a[i][j] > EPS)
+ abort ();
+}
+
+int main ()
+{
+ double a[N][N], b[N][N], c[N][N], c_ref[N][N];
+
+ init(a, b, N);
+
+ work(a, b, c, N);
+ work_ref(a, b, c_ref, N);
+
+ check(c, c_ref);
+
+ return 0;
+}
--- /dev/null
+/* { dg-do run { target vect_simd_clones } } */
+/* { dg-additional-options "-msse2" { target sse2_runtime } } */
+/* { dg-additional-options "-mavx" { target avx_runtime } } */
+
+#define N 100
+#define EPS 0.000001
+
+#include <stdlib.h>
+#include <stdio.h>
+
+void init(int *b, float *y, int n)
+{
+ int i, s = -1;
+ for ( i=0; i<N; i++ )
+ {
+ b[i] = i*i*s;
+ y[i] = b[i] * 0.1f;
+ s = -s;
+ }
+}
+
+#pragma omp declare simd linear(p:1) notinbranch
+int foo(int *p){
+ *p = *p + 10;
+ return *p;
+}
+
+int myaddint(int *a, int *b, int n)
+{
+#pragma omp simd
+ for (int i=0; i<n; i++){
+ a[i] = foo(&b[i]); /* foo is not called under a condition */
+ }
+ return a[n-1];
+}
+
+int myaddint_ref(int *a, int *b, int n)
+{
+ for (int i=0; i<n; i++){
+ a[i] = foo(&b[i]);
+ }
+ return a[n-1];
+}
+
+#pragma omp declare simd linear(p:1) inbranch
+float goo(float *p){
+ *p = *p + 18.5f;
+ return *p;
+}
+
+int myaddfloat(float *x, float *y, int n)
+{
+#pragma omp simd
+ for (int i=0; i<n; i++){
+ x[i] = (x[i] > y[i]) ? goo(&y[i]) : y[i];
+ /* goo is called under the condition (or within a branch) */
+ }
+ return x[n-1];
+}
+
+int myaddfloat_ref(float *x, float *y, int n)
+{
+ for (int i=0; i<n; i++){
+ x[i] = (x[i] > y[i]) ? goo(&y[i]) : y[i];
+ }
+ return x[n-1];
+}
+
+void check_addint (int *a, int *b)
+{
+ int i;
+ for (i = 0; i < N; i++)
+ if (a[i] != b[i])
+ abort ();
+}
+
+void check_addfloat (float *a, float *b)
+{
+ int i;
+ for (i = 0; i < N; i++)
+ if (a[i] - b[i] > EPS || b[i] - a[i] > EPS)
+ abort ();
+}
+
+int main ()
+{
+ int i;
+ int a[N], a_ref[N], b[N];
+ float x[N], x_ref[N], y[N];
+
+ init(a, x, N);
+ init(b, y, N);
+ myaddint(a, b, N);
+ myaddfloat(x, y, N);
+
+ init(a_ref, x_ref, N);
+ init(b, y, N);
+ myaddint_ref(a_ref, b, N);
+ myaddfloat_ref(x_ref, y, N);
+
+ check_addint(a, a_ref);
+ check_addfloat(x, x_ref);
+
+ return 0;
+}
--- /dev/null
+/* { dg-do run { target vect_simd_clones } } */
+/* { dg-additional-options "-msse2" { target sse2_runtime } } */
+/* { dg-additional-options "-mavx" { target avx_runtime } } */
+
+#include <stdio.h>
+#include <stdlib.h>
+
+#define N 45
+int a[N], a_ref[N], b[N];
+
+#pragma omp declare simd inbranch
+int fib( int n )
+{
+ if (n <= 2)
+ return n;
+ else {
+ return fib(n-1) + fib(n-2);
+ }
+}
+
+int main(void)
+{
+ int i;
+
+#pragma omp simd
+ for (i=0; i < N; i++)
+ b[i] = i;
+
+#pragma omp simd
+ for (i=0; i < N; i++)
+ a[i] = fib(b[i]);
+
+ for (i=0; i < N; i++)
+ a_ref[i] = fib(b[i]);
+
+ for (i=0; i < N; i++)
+ if (a[i] != a_ref[i])
+ abort ();
+
+ return 0;
+}
--- /dev/null
+/* { dg-do run } */
+/* { dg-additional-options "-msse2" { target sse2_runtime } } */
+/* { dg-additional-options "-mavx" { target avx_runtime } } */
+
+#include <stdlib.h>
+#include <math.h>
+
+int P[1000];
+float A[1000];
+
+float do_work(float *arr)
+{
+ float pri;
+
+#pragma omp simd lastprivate(pri)
+ for (int i = 0; i < 999; ++i)
+ {
+ int j = P[i];
+
+ pri = 0.5f;
+ if (j % 2 == 0)
+ {
+ pri = A[j+1] + arr[i];
+ }
+ A[j] = pri * 1.5f;
+ pri = pri + A[j];
+ }
+
+ return pri;
+}
+
+int main(void)
+{
+ float pri, arr[1000];
+
+ for (int i = 0; i < 1000; ++i)
+ {
+ P[i] = i;
+ A[i] = i * 1.5f;
+ arr[i] = i * 1.8f;
+ }
+
+ pri = do_work(&arr[0]);
+
+ if (pri != 8237.25)
+ abort ();
+
+ return 0;
+}
--- /dev/null
+/* { dg-do run } */
+
+#include <stdlib.h>
+int main()
+{
+ int x = 1;
+ #pragma omp parallel
+ #pragma omp single
+ {
+ #pragma omp task shared(x) depend(out: x)
+ x = 2;
+ #pragma omp task shared(x) depend(in: x)
+ if (x != 2)
+ abort ();
+ }
+ return 0;
+}
--- /dev/null
+/* { dg-do run } */
+
+#include <stdlib.h>
+int main()
+{
+ int x = 1;
+ #pragma omp parallel
+ #pragma omp single
+ {
+ #pragma omp task shared(x) depend(in: x)
+ if (x != 1)
+ abort ();
+ #pragma omp task shared(x) depend(out: x)
+ x = 2;
+ }
+ return 0;
+}
--- /dev/null
+/* { dg-do run } */
+
+#include <stdlib.h>
+
+int main()
+{
+ int x = 0;
+ #pragma omp parallel
+ #pragma omp single
+ {
+ #pragma omp task shared(x) depend(out: x)
+ x = 1;
+ #pragma omp task shared(x) depend(out: x)
+ x = 2;
+ #pragma omp taskwait
+ if (x != 1 && x != 2)
+ abort ();
+ }
+ return 0;
+}
--- /dev/null
+/* { dg-do run } */
+
+#include <stdlib.h>
+int main()
+{
+ int x = 1;
+ #pragma omp parallel
+ #pragma omp single
+ {
+ #pragma omp task shared(x) depend(out: x)
+ x = 2;
+ #pragma omp task shared(x) depend(in: x)
+ if (x != 2)
+ abort ();
+ #pragma omp task shared(x) depend(in: x)
+ if (x != 2)
+ abort ();
+ }
+ return 0;
+}
--- /dev/null
+/* { dg-do run } */
+
+#define N 128
+#define BS 16
+#define EPS 0.000001
+
+#include <stdlib.h>
+
+void matmul_depend (float A[N][N], float B[N][N], float C[N][N])
+{
+ int i, j, k, ii, jj, kk;
+ for (i = 0; i < N; i+=BS)
+ for (j = 0; j < N; j+=BS)
+ for (k = 0; k < N; k+=BS)
+// Note 1: i, j, k, A, B, C are firstprivate by default
+// Note 2: A, B and C are just pointers
+#pragma omp task private(ii, jj, kk) \
+ depend ( in: A[i:BS][k:BS], B[k:BS][j:BS] ) \
+ depend ( inout: C[i:BS][j:BS] )
+ for (ii = i; ii < i+BS; ii++ )
+ for (jj = j; jj < j+BS; jj++ )
+ for (kk = k; kk < k+BS; kk++ )
+ C[ii][jj] = C[ii][jj] + A[ii][kk] * B[kk][jj];
+}
+
+void matmul_ref (float A[N][N], float B[N][N], float C[N][N])
+{
+ int i, j, k;
+
+ for (i = 0; i < N; i++)
+ for (j = 0; j < N; j++)
+ for (k = 0; k < N; k++)
+ C[i][j] += A[i][k] * B[k][j];
+}
+
+void init (float A[N][N], float B[N][N])
+{
+ int i, j, s = -1;
+ for (i = 0; i < N; i++)
+ for (j = 0; j < N; j++)
+ {
+ A[i][j] = i * j * s;
+ B[i][j] = i + j;
+ s = -s;
+ }
+}
+
+void init_zero (float A[N][N], float B[N][N])
+{
+ int i, j, s = -1;
+ for (i = 0; i < N; i++)
+ for (j = 0; j < N; j++)
+ {
+ A[i][j] = 0;
+ B[i][j] = 0;
+ }
+}
+
+void check (float A[N][N], float B[N][N])
+{
+ int i, j;
+ for (i = 0; i < N; i++)
+ for (j = 0; j < N; j++)
+ if (A[i][j] - B[i][j] > EPS || B[i][j] - A[i][j] > EPS)
+ abort ();
+}
+
+int main ()
+{
+ float A[N][N], B[N][N], C[N][N], C_ref[N][N];
+
+ init (A, B);
+ init_zero (C, C_ref);
+
+ matmul_depend (A, B, C);
+ matmul_ref (A, B, C_ref);
+
+ check (C, C_ref);
+
+ return 0;
+}
!$omp end target
!$omp end task
!$omp end target data
+
+ !$omp taskwait
call check (p, N)
end subroutine
--- /dev/null
+! { dg-do run }
+! { dg-additional-options "-msse2" { target sse2_runtime } }
+! { dg-additional-options "-mavx" { target avx_runtime } }
+
+module SIMD1_mod
+contains
+ subroutine init (a, a_ref, b, c, n, ioff_ptr)
+ double precision :: a(*), a_ref(*), b(*), c(*)
+ integer :: n, i, s
+ integer, pointer :: ioff_ptr
+
+ s = -1
+ do i = 1, n
+ a(i) = i * i * s
+ a_ref(i) = a(i)
+ b(i) = i + i
+ end do
+
+ do i = 1, n+ioff_ptr
+ c(i) = i * 3
+ end do
+
+ end subroutine
+
+ subroutine check (a, b, n)
+ integer :: i, n
+ double precision, parameter :: EPS = 0.0000000000001
+ double precision :: diff, a(*), b(*)
+ do i = 1, n
+ diff = a(i) - b(i)
+ if (diff > EPS .or. -diff > EPS) call abort
+ end do
+ end subroutine
+
+ subroutine star(a, a_ref, b, c, n, ioff_ptr)
+ double precision :: a(*), a_ref(*), b(*), c(*)
+ integer :: n, i
+ integer, pointer :: ioff_ptr
+
+ call init (a, a_ref, b, c, n, ioff_ptr)
+
+ !$omp simd
+ do i = 1,n
+ a(i) = a(i) * b(i) * c(i+ioff_ptr)
+ end do
+
+ do i = 1,n
+ a_ref(i) = a_ref(i) * b(i) * c(i+ioff_ptr)
+ end do
+
+ call check (a, a_ref, n)
+
+ end subroutine
+end module
+
+program SIMD1
+ use SIMD1_mod, only : star
+ double precision :: a(128), a_ref(128), b(128), c(144)
+ integer, pointer:: ioff_ptr
+ integer, target:: offset
+
+ offset = 16
+ ioff_ptr => offset
+
+ call star (a, a_ref, b, c, 128, ioff_ptr)
+
+end program
--- /dev/null
+! { dg-do run { target vect_simd_clones } }
+! { dg-additional-options "-msse2" { target sse2_runtime } }
+! { dg-additional-options "-mavx" { target avx_runtime } }
+
+module SIMD2_mod
+contains
+ function add1(a,b,fact) result(c)
+ !$omp declare simd(add1) uniform(fact)
+ double precision :: a,b,fact, c
+ c = a + b + fact
+ end function
+
+ function add2(a,b,i, fact) result(c)
+ !$omp declare simd(add2) uniform(a,b,fact) linear(i:1)
+ integer, value :: i
+ double precision, dimension(:) :: a, b
+ double precision :: fact, c
+ c = a(i) + b(i) + fact
+ end function
+
+ subroutine work(a, b, n )
+ implicit none
+ double precision :: a(n),b(n), tmp
+ integer :: n, i
+
+ !$omp simd private(tmp)
+ do i = 1,n
+ tmp = add1(a(i), b(i), 1.0d0)
+ a(i) = add2(a, b, i, 1.0d0) + tmp
+ a(i) = a(i) + b(i) + 1.0d0
+ end do
+ end subroutine
+
+ subroutine work_ref(a, b, n )
+ implicit none
+ double precision :: a(n),b(n), tmp
+ integer :: n, i
+
+ do i = 1,n
+ tmp = add1(a(i), b(i), 1.0d0)
+ a(i) = add2(a, b, i, 1.0d0) + tmp
+ a(i) = a(i) + b(i) + 1.0d0
+ end do
+ end subroutine
+
+ subroutine check (a, b, n)
+ integer :: i, n
+ double precision, parameter :: EPS = 0.0000000000001
+ double precision :: diff, a(*), b(*)
+ do i = 1, n
+ diff = a(i) - b(i)
+ if (diff > EPS .or. -diff > EPS) call abort
+ end do
+ end subroutine
+end module
+
+program main
+ use SIMD2_mod
+ integer, parameter :: N=32
+ integer :: i
+ double precision :: a(N), b(N), a_ref(N)
+ do i = 1,N
+ a(i) = i-1
+ a_ref(i) = a(i)
+ b(i) = N-(i-1)
+ end do
+
+ call work(a, b, N )
+ call work_ref(a_ref, b, N )
+
+ call check(a, a_ref, N )
+end program
--- /dev/null
+! { dg-do run }
+! { dg-additional-options "-msse2" { target sse2_runtime } }
+! { dg-additional-options "-mavx" { target avx_runtime } }
+
+module SIMD3_mod
+contains
+ subroutine work( a, b, n, sum )
+ implicit none
+ integer :: i, n
+ double precision :: a(n), b(n), sum, tmp
+
+ sum = 0.0d0
+ call init(a, b, n)
+ !$omp simd private(tmp) reduction(+:sum)
+ do i = 1,n
+ tmp = a(i) + b(i)
+ sum = sum + tmp
+ end do
+
+ end subroutine work
+
+ subroutine work_ref( a, b, n, sum )
+ implicit none
+ integer :: i, n
+ double precision :: a(n), b(n), sum, tmp
+
+ sum = 0.0d0
+ call init(a, b, n)
+ do i = 1,n
+ tmp = a(i) + b(i)
+ sum = sum + tmp
+ end do
+
+ end subroutine work_ref
+
+ subroutine init (a, b, n)
+ double precision :: a(*), b(*)
+ integer :: n, i, s
+
+ s = -1
+ do i = 1, n
+ a(i) = i * i * s
+ b(i) = i + i
+ s = -s
+ end do
+
+ end subroutine
+end module
+
+program SIMD3
+ use SIMD3_mod
+ double precision :: a(128), b(128), sum, sum_ref
+
+ call work(a, b, 128, sum)
+ call work_ref(a, b, 128, sum_ref)
+
+ if (sum .ne. sum_ref) call abort
+
+end program
--- /dev/null
+! { dg-do run }
+! { dg-additional-options "-msse2" { target sse2_runtime } }
+! { dg-additional-options "-mavx" { target avx_runtime } }
+
+module SIMD4_mod
+contains
+ subroutine work( b, n, m )
+ implicit none
+ real :: b(n)
+ integer :: i,n,m
+
+ call init(b, n)
+
+ !$omp simd safelen(16)
+ do i = m+1, n
+ b(i) = b(i-m) - 1.0
+ end do
+ end subroutine work
+
+ subroutine work_ref( b, n, m )
+ implicit none
+ real :: b(n)
+ integer :: i,n,m
+
+ call init(b, n)
+
+ do i = m+1, n
+ b(i) = b(i-m) - 1.0
+ end do
+ end subroutine work_ref
+
+ subroutine init (b, n)
+ real :: b(*)
+ integer :: n, i, s
+
+ s = -1
+ do i = 1, n
+ b(i) = i * i * s
+ s = -s
+ end do
+
+ end subroutine
+
+ subroutine check (a, b, n)
+ integer :: i, n
+ real, parameter :: EPS = 0.000001
+ real :: diff, a(*), b(*)
+ do i = 1, n
+ diff = a(i) - b(i)
+ if (diff > EPS .or. -diff > EPS) call abort
+ end do
+ end subroutine
+
+end module
+
+program SIMD4
+ use SIMD4_mod
+ real :: b(128), b_ref(128)
+
+ call work(b, 128, 32)
+ call work_ref(b_ref, 128, 32)
+
+ call check(b, b_ref, 128)
+end program
--- /dev/null
+! { dg-do run }
+! { dg-additional-options "-msse2" { target sse2_runtime } }
+! { dg-additional-options "-mavx" { target avx_runtime } }
+
+module SIMD5_mod
+contains
+ subroutine work( a, b, c, n )
+ implicit none
+ integer :: i,j,n
+ double precision :: a(n,n), b(n,n), c(n,n), tmp
+
+ !$omp do simd collapse(2) private(tmp)
+ do j = 1,n
+ do i = 1,n
+ tmp = a(i,j) + b(i,j)
+ c(i,j) = tmp
+ end do
+ end do
+
+ end subroutine work
+
+ subroutine work_ref( a, b, c, n )
+ implicit none
+ integer :: i,j,n
+ double precision :: a(n,n), b(n,n), c(n,n), tmp
+
+ do j = 1,n
+ do i = 1,n
+ tmp = a(i,j) + b(i,j)
+ c(i,j) = tmp
+ end do
+ end do
+
+ end subroutine work_ref
+
+ subroutine init (a, b, n)
+ integer :: i,j,n,s
+ double precision :: a(n,n), b(n,n)
+
+ s = -1
+
+ do j = 1,n
+ do i = 1,n
+ a(i,j) = i*j*s
+ b(i,j) = i+j
+ s = -s
+ end do
+ end do
+
+ end subroutine
+
+ subroutine check (a, b, n)
+ integer :: i, j, n
+ double precision, parameter :: EPS = 0.0000000000000001
+ double precision :: diff, a(n,n), b(n,n)
+ do j = 1, n
+ do i = 1, n
+ diff = a(i,j) - b(i,j)
+ if (diff > EPS .or. -diff > EPS) call abort
+ end do
+ end do
+ end subroutine
+
+end module
+
+program SIMD5
+ use SIMD5_mod
+ double precision, dimension(32, 32) :: a, b, c, c_ref
+
+ call init(a, b, 32)
+
+ call work(a, b, c, 32)
+ call work_ref(a, b, c_ref, 32)
+
+ call check(c, c_ref, 32)
+end program
--- /dev/null
+! { dg-do run { target vect_simd_clones } }
+! { dg-additional-options "-msse2" { target sse2_runtime } }
+! { dg-additional-options "-mavx" { target avx_runtime } }
+
+module SIMD6_mod
+contains
+ function foo(p) result(r)
+ !$omp declare simd(foo) notinbranch
+ integer :: p, r
+ p = p + 10
+ r = p
+ end function foo
+
+ function myaddint(a, b, n) result(r)
+ implicit none
+ integer :: a(*), b(*), n, r
+ integer :: i
+
+ !$omp simd
+ do i=1, n
+ a(i) = foo(b(i)) ! foo is not called under a condition
+ end do
+ r = a(n)
+
+ end function myaddint
+
+ function myaddint_ref(a, b, n) result(r)
+ implicit none
+ integer :: a(*), b(*), n, r
+ integer :: i
+
+ do i=1, n
+ a(i) = foo(b(i))
+ end do
+ r = a(n)
+
+ end function myaddint_ref
+
+ function goo(p) result(r)
+ !$omp declare simd(goo) inbranch
+ real :: p, r
+ p = p + 18.5
+ r = p
+ end function goo
+
+ function myaddfloat(x, y, n) result(r)
+ implicit none
+ real :: x(*), y(*), r
+ integer :: n
+ integer :: i
+
+ !$omp simd
+ do i=1, n
+ if (x(i) > y(i)) then
+ x(i) = goo(y(i))
+ ! goo is called under the condition (or within a branch)
+ else
+ x(i) = y(i)
+ endif
+ end do
+
+ r = x(n)
+ end function myaddfloat
+
+ function myaddfloat_ref(x, y, n) result(r)
+ implicit none
+ real :: x(*), y(*), r
+ integer :: n
+ integer :: i
+
+ do i=1, n
+ if (x(i) > y(i)) then
+ x(i) = goo(y(i))
+ else
+ x(i) = y(i)
+ endif
+ end do
+
+ r = x(n)
+ end function myaddfloat_ref
+
+ subroutine init (b, y, n)
+ integer :: b(128)
+ real :: y(128)
+
+ s = -1
+ do i = 1, n
+ b(i) = i*i*s
+ y(i) = i*i*s
+ s = -s
+ end do
+
+ end subroutine
+
+ subroutine init2 (b, y, n)
+ integer :: b(128)
+ real :: y(128)
+
+ do i = 1, n
+ b(i) = i
+ y(i) = i
+ end do
+
+ end subroutine
+
+ subroutine checkfloat (a, b, n)
+ integer :: i, n
+ real, parameter :: EPS = 0.000001
+ real :: diff, a(*), b(*)
+ do i = 1, n
+ diff = a(i) - b(i)
+ if (diff > EPS .or. -diff > EPS) call abort
+ end do
+ end subroutine
+
+ subroutine checkint (a, b, n)
+ integer :: i, n, a(*), b(*)
+ do i = 1, n
+ if (a(i) .ne. b(i)) call abort
+ end do
+ end subroutine
+
+ subroutine test ()
+ integer :: a(128), a_ref(128), b(128), ri, ri_ref
+ real :: x(128), x_ref(128), y(128), rf, rf_ref
+
+ call init2(a, x, 128)
+ call init2(a_ref, x_ref, 128)
+
+ call init(b, y, 128)
+
+ ri = myaddint (a, b, 128)
+ rf = myaddfloat (x, y, 128)
+
+ call init(b, y, 128)
+
+ ri_ref = myaddint_ref (a_ref, b, 128)
+ rf_ref = myaddfloat_ref (x_ref, y, 128)
+
+ call checkint (a, a_ref, 128)
+ call checkfloat (x, x_ref, 128)
+ end subroutine
+
+end module
+
+program SIMD6
+ use SIMD6_mod, only: test
+
+ call test ()
+
+end program
--- /dev/null
+! { dg-do run { target vect_simd_clones } }
+! { dg-additional-options "-msse2" { target sse2_runtime } }
+! { dg-additional-options "-mavx" { target avx_runtime } }
+
+program fibonacci
+ implicit none
+ integer,parameter :: N=45
+ integer :: a(0:N-1), b(0:N-1)
+ integer :: a_ref(0:N-1), b_ref(0:N-1)
+ integer :: i
+ integer, external :: fib
+
+ !$omp simd
+ do i = 0,N-1
+ b(i) = i
+ end do
+
+ do i = 0,N-1
+ b_ref(i) = i
+ end do
+
+ !$omp simd
+ do i=0,N-1
+ a(i) = fib(b(i))
+ end do
+
+ do i=0,N-1
+ a_ref(i) = fib(b_ref(i))
+ end do
+
+ do i = 0, N-1
+ if (a(i) .ne. a_ref(i)) call abort ()
+ end do
+
+ if (a(44) .ne. 1134903170) call abort()
+
+end program
+
+recursive function fib(n) result(r)
+!$omp declare simd(fib) inbranch
+ integer :: n, r
+
+ if (n <= 2) then
+ r = n
+ else
+ r = fib(n-1) + fib(n-2)
+ endif
+
+end function fib
--- /dev/null
+! { dg-do run }
+! { dg-additional-options "-msse2" { target sse2_runtime } }
+! { dg-additional-options "-mavx" { target avx_runtime } }
+
+module work
+
+integer :: P(1000)
+real :: A(1000)
+
+contains
+function do_work(arr) result(pri)
+ implicit none
+ real, dimension(*) :: arr
+
+ real :: pri
+ integer :: i, j
+
+ !$omp simd private(j) lastprivate(pri)
+ do i = 1, 999
+ j = P(i)
+
+ pri = 0.5
+ if (mod(j-1, 2) == 0) then
+ pri = A(j+1) + arr(i)
+ endif
+ A(j) = pri * 1.5
+ pri = pri + A(j)
+ end do
+
+end function do_work
+
+end module work
+
+program simd_8f
+ use work
+ implicit none
+ real :: pri, arr(1000)
+ integer :: i
+
+ do i = 1, 1000
+ P(i) = i
+ A(i) = (i-1) * 1.5
+ arr(i) = (i-1) * 1.8
+ end do
+ pri = do_work(arr)
+ if (pri .ne. 8237.25) call abort ()
+
+end program
--- /dev/null
+! { dg-do run }
+
+program example
+ integer :: x
+ x = 1
+ !$omp parallel
+ !$omp single
+ !$omp task shared(x) depend(out: x)
+ x = 2
+ !$omp end task
+ !$omp task shared(x) depend(in: x)
+ if (x .ne. 2) call abort ()
+ !$omp end task
+ !$omp end single
+ !$omp end parallel
+end program
--- /dev/null
+! { dg-do run }
+
+program example
+ integer :: x
+ x = 1
+ !$omp parallel
+ !$omp single
+ !$omp task shared(x) depend(in: x)
+ if (x .ne. 1) call abort ()
+ !$omp end task
+ !$omp task shared(x) depend(out: x)
+ x = 2
+ !$omp end task
+ !$omp end single
+ !$omp end parallel
+end program
--- /dev/null
+! { dg-do run }
+
+program example
+ integer :: x
+ x = 0
+ !$omp parallel
+ !$omp single
+ !$omp task shared(x) depend(out: x)
+ x = 1
+ !$omp end task
+ !$omp task shared(x) depend(out: x)
+ x = 2
+ !$omp end task
+ !$omp taskwait
+ if ((x .ne. 1) .and. (x .ne. 2)) call abort()
+ !$omp end single
+ !$omp end parallel
+end program
--- /dev/null
+! { dg-do run }
+
+program example
+ integer :: x
+ x = 1
+ !$omp parallel
+ !$omp single
+ !$omp task shared(x) depend(out: x)
+ x = 2
+ !$omp end task
+ !$omp task shared(x) depend(in: x)
+ if (x .ne. 2) call abort ()
+ !$omp end task
+ !$omp task shared(x) depend(in: x)
+ if (x .ne. 2) call abort ()
+ !$omp end task
+ !$omp end single
+ !$omp end parallel
+end program
--- /dev/null
+! { dg-do run }
+
+module task_dep5_mod
+contains
+ subroutine matmul_depend (N, BS, A, B, C)
+ implicit none
+ integer :: N, BS, BM
+ real, dimension(N, N) :: A, B, C
+ integer :: i, j, k, ii, jj, kk
+ BM = BS - 1
+ do i = 1, N, BS
+ do j = 1, N, BS
+ do k = 1, N, BS
+ !$omp task shared(A,B,C) private(ii,jj,kk) & ! I,J,K are firstprivate by default
+ !$omp depend ( in: A(i:i+BM, k:k+BM), B(k:k+BM, j:j+BM) ) &
+ !$omp depend ( inout: C(i:i+BM, j:j+BM) )
+ do ii = i, i+BM
+ do jj = j, j+BM
+ do kk = k, k+BM
+ C(jj,ii) = C(jj,ii) + A(kk,ii) * B(jj,kk)
+ end do
+ end do
+ end do
+ !$omp end task
+ end do
+ end do
+ end do
+ end subroutine
+
+ subroutine matmul_ref (N, A, B, C)
+ implicit none
+ integer :: N
+ real, dimension(N, N) :: A, B, C
+ integer :: i, j, k
+ do i = 1, N
+ do j = 1, N
+ do k = 1, N
+ C(j,i) = C(j,i) + A(k,i) * B(j,k)
+ end do
+ end do
+ end do
+ end subroutine
+
+ subroutine check (N, A, B)
+ integer :: N
+ integer :: i, j
+ integer, parameter :: EPS = 0.000001
+ real, dimension(N,N) :: A, B
+ real :: diff
+ do i = 1, N
+ do j = 1, N
+ diff = A(i, j) - B(i, j)
+ if (diff > EPS .or. -diff > EPS) then
+ call abort ()
+ end if
+ end do
+ end do
+ end subroutine
+
+ subroutine init (N, A, B)
+ integer :: N
+ integer :: i, j, s
+ real, dimension(N,N) :: A, B
+ s = -1
+ do i = 1, N
+ do j = 1, N
+ A(i, j) = i*j*s
+ B(i, j) = i+j
+ s = -s
+ end do
+ end do
+ end subroutine
+
+ subroutine zero_init (N, A, B)
+ integer :: N
+ integer :: i, j
+ real, dimension(N,N) :: A, B
+ do i = 1, N
+ do j = 1, N
+ A(i, j) = 0
+ B(i, j) = 0
+ end do
+ end do
+ end subroutine
+
+end module
+
+program main
+ use task_dep5_mod
+ real, dimension(32, 32) :: A, B, C, C_ref
+
+ call init (32, A, B)
+ call zero_init (32, C, C_ref)
+
+ call matmul_depend(32, 4, A, B, C)
+ call matmul_ref(32, A, B, C_ref)
+
+ call check (32, C, C_ref)
+
+end program
git://gcc.gnu.org / gcc.git / commitdiff