[PATCH 5/6, OpenACC, libgomp] Async re-work, C/C++ testsuite changes
Chung-Lin Tang
chunglin_tang@mentor.com
Tue Sep 25 13:12:00 GMT 2018
These are the testsuite/libgomp.oacc-c-c++-common/* changes.
Thanks,
Chung-Lin
* testsuite/libgomp.oacc-c-c++-common/asyncwait-2.c: New testcase.
* testsuite/libgomp.oacc-c-c++-common/data-2-lib.c: Adjust testcase.
* testsuite/libgomp.oacc-c-c++-common/data-2.c: Likewise.
* testsuite/libgomp.oacc-c-c++-common/data-3.c: Likewise.
* testsuite/libgomp.oacc-c-c++-common/lib-71.c: Likewise.
* testsuite/libgomp.oacc-c-c++-common/lib-77.c: Likewise.
* testsuite/libgomp.oacc-c-c++-common/lib-79.c: Likewise.
* testsuite/libgomp.oacc-c-c++-common/lib-81.c: Likewise.
-------------- next part --------------
diff --git a/libgomp/testsuite/libgomp.oacc-c-c++-common/asyncwait-2.c b/libgomp/testsuite/libgomp.oacc-c-c++-common/asyncwait-2.c
new file mode 100644
index 0000000..9420540
--- /dev/null
+++ b/libgomp/testsuite/libgomp.oacc-c-c++-common/asyncwait-2.c
@@ -0,0 +1,904 @@
+/* { dg-do run { target openacc_nvidia_accel_selected } } */
+/* { dg-additional-options "-lcuda" } */
+
+#include <openacc.h>
+#include <stdlib.h>
+#include <cuda.h>
+
+#include <stdio.h>
+#include <time.h>
+#include <sys/time.h>
+
+int
+main (int argc, char **argv)
+{
+ CUresult r;
+ CUstream stream1;
+ int N = 128; //1024 * 1024;
+ float *a, *b, *c, *d, *e;
+ int i;
+ int nbytes;
+
+ srand (time (NULL));
+ int s = rand () % 100;
+
+ acc_init (acc_device_nvidia);
+
+ nbytes = N * sizeof (float);
+
+ a = (float *) malloc (nbytes);
+ b = (float *) malloc (nbytes);
+ c = (float *) malloc (nbytes);
+ d = (float *) malloc (nbytes);
+ e = (float *) malloc (nbytes);
+
+ for (i = 0; i < N; i++)
+ {
+ a[i] = 3.0;
+ b[i] = 0.0;
+ }
+
+ acc_set_default_async (s);
+
+#pragma acc data copy (a[0:N]) copy (b[0:N]) copyin (N)
+ {
+
+#pragma acc parallel async
+ {
+ int ii;
+
+ for (ii = 0; ii < N; ii++)
+ b[ii] = a[ii];
+ }
+
+#pragma acc wait
+
+ }
+
+ for (i = 0; i < N; i++)
+ {
+ if (a[i] != 3.0)
+ abort ();
+
+ if (b[i] != 3.0)
+ abort ();
+ }
+
+ for (i = 0; i < N; i++)
+ {
+ a[i] = 2.0;
+ b[i] = 0.0;
+ }
+
+#pragma acc data copy (a[0:N]) copy (b[0:N]) copyin (N)
+ {
+
+#pragma acc parallel async
+ {
+ int ii;
+
+ for (ii = 0; ii < N; ii++)
+ b[ii] = a[ii];
+ }
+
+#pragma acc wait (s)
+
+ }
+
+ for (i = 0; i < N; i++)
+ {
+ if (a[i] != 2.0)
+ abort ();
+
+ if (b[i] != 2.0)
+ abort ();
+ }
+
+ for (i = 0; i < N; i++)
+ {
+ a[i] = 3.0;
+ b[i] = 0.0;
+ c[i] = 0.0;
+ d[i] = 0.0;
+ }
+
+#pragma acc data copy (a[0:N]) copy (b[0:N]) copy (c[0:N]) copy (d[0:N]) copyin (N)
+ {
+
+#pragma acc parallel async
+ {
+ int ii;
+
+ for (ii = 0; ii < N; ii++)
+ b[ii] = (a[ii] * a[ii] * a[ii]) / a[ii];
+ }
+
+#pragma acc parallel async
+ {
+ int ii;
+
+ for (ii = 0; ii < N; ii++)
+ c[ii] = (a[ii] + a[ii] + a[ii] + a[ii]) / a[ii];
+ }
+
+
+#pragma acc parallel async
+ {
+ int ii;
+
+ for (ii = 0; ii < N; ii++)
+ d[ii] = ((a[ii] * a[ii] + a[ii]) / a[ii]) - a[ii];
+ }
+
+#pragma acc wait (s)
+
+ }
+
+ for (i = 0; i < N; i++)
+ {
+ if (a[i] != 3.0)
+ abort ();
+
+ if (b[i] != 9.0)
+ abort ();
+
+ if (c[i] != 4.0)
+ abort ();
+
+ if (d[i] != 1.0)
+ abort ();
+ }
+
+ for (i = 0; i < N; i++)
+ {
+ a[i] = 2.0;
+ b[i] = 0.0;
+ c[i] = 0.0;
+ d[i] = 0.0;
+ e[i] = 0.0;
+ }
+
+#pragma acc data copy (a[0:N], b[0:N], c[0:N], d[0:N], e[0:N]) copyin (N)
+ {
+
+#pragma acc parallel async
+ {
+ int ii;
+
+ for (ii = 0; ii < N; ii++)
+ b[ii] = (a[ii] * a[ii] * a[ii]) / a[ii];
+ }
+
+#pragma acc parallel async
+ {
+ int ii;
+
+ for (ii = 0; ii < N; ii++)
+ c[ii] = (a[ii] + a[ii] + a[ii] + a[ii]) / a[ii];
+ }
+
+#pragma acc parallel async
+ {
+ int ii;
+
+ for (ii = 0; ii < N; ii++)
+ d[ii] = ((a[ii] * a[ii] + a[ii]) / a[ii]) - a[ii];
+ }
+
+#pragma acc parallel wait (s) async (s)
+ {
+ int ii;
+
+ for (ii = 0; ii < N; ii++)
+ e[ii] = a[ii] + b[ii] + c[ii] + d[ii];
+ }
+
+#pragma acc wait (s)
+
+ }
+
+ for (i = 0; i < N; i++)
+ {
+ if (a[i] != 2.0)
+ abort ();
+
+ if (b[i] != 4.0)
+ abort ();
+
+ if (c[i] != 4.0)
+ abort ();
+
+ if (d[i] != 1.0)
+ abort ();
+
+ if (e[i] != 11.0)
+ abort ();
+ }
+
+
+ r = cuStreamCreate (&stream1, CU_STREAM_NON_BLOCKING);
+ if (r != CUDA_SUCCESS)
+ {
+ fprintf (stderr, "cuStreamCreate failed: %d\n", r);
+ abort ();
+ }
+
+ acc_set_cuda_stream (1, stream1);
+
+ for (i = 0; i < N; i++)
+ {
+ a[i] = 5.0;
+ b[i] = 0.0;
+ }
+
+#pragma acc data copy (a[0:N], b[0:N]) copyin (N)
+ {
+
+#pragma acc parallel async
+ {
+ int ii;
+
+ for (ii = 0; ii < N; ii++)
+ b[ii] = a[ii];
+ }
+
+#pragma acc wait (s)
+
+ }
+
+ for (i = 0; i < N; i++)
+ {
+ if (a[i] != 5.0)
+ abort ();
+
+ if (b[i] != 5.0)
+ abort ();
+ }
+
+ for (i = 0; i < N; i++)
+ {
+ a[i] = 7.0;
+ b[i] = 0.0;
+ c[i] = 0.0;
+ d[i] = 0.0;
+ }
+
+#pragma acc data copy (a[0:N]) copy (b[0:N]) copy (c[0:N]) copy (d[0:N]) copyin (N)
+ {
+
+#pragma acc parallel async
+ {
+ int ii;
+
+ for (ii = 0; ii < N; ii++)
+ b[ii] = (a[ii] * a[ii] * a[ii]) / a[ii];
+ }
+
+#pragma acc parallel async
+ {
+ int ii;
+
+ for (ii = 0; ii < N; ii++)
+ c[ii] = (a[ii] + a[ii] + a[ii] + a[ii]) / a[ii];
+ }
+
+#pragma acc parallel async
+ {
+ int ii;
+
+ for (ii = 0; ii < N; ii++)
+ d[ii] = ((a[ii] * a[ii] + a[ii]) / a[ii]) - a[ii];
+ }
+
+#pragma acc wait (s)
+
+ }
+
+ for (i = 0; i < N; i++)
+ {
+ if (a[i] != 7.0)
+ abort ();
+
+ if (b[i] != 49.0)
+ abort ();
+
+ if (c[i] != 4.0)
+ abort ();
+
+ if (d[i] != 1.0)
+ abort ();
+ }
+
+ for (i = 0; i < N; i++)
+ {
+ a[i] = 3.0;
+ b[i] = 0.0;
+ c[i] = 0.0;
+ d[i] = 0.0;
+ e[i] = 0.0;
+ }
+
+#pragma acc data copy (a[0:N], b[0:N], c[0:N], d[0:N], e[0:N]) copyin (N)
+ {
+
+#pragma acc parallel async
+ {
+ int ii;
+
+ for (ii = 0; ii < N; ii++)
+ b[ii] = (a[ii] * a[ii] * a[ii]) / a[ii];
+ }
+
+#pragma acc parallel async
+ {
+ int ii;
+
+ for (ii = 0; ii < N; ii++)
+ c[ii] = (a[ii] + a[ii] + a[ii] + a[ii]) / a[ii];
+ }
+
+#pragma acc parallel async
+ {
+ int ii;
+
+ for (ii = 0; ii < N; ii++)
+ d[ii] = ((a[ii] * a[ii] + a[ii]) / a[ii]) - a[ii];
+ }
+
+#pragma acc parallel wait (s) async (s)
+ {
+ int ii;
+
+ for (ii = 0; ii < N; ii++)
+ e[ii] = a[ii] + b[ii] + c[ii] + d[ii];
+ }
+
+#pragma acc wait (s)
+
+ }
+
+ for (i = 0; i < N; i++)
+ {
+ if (a[i] != 3.0)
+ abort ();
+
+ if (b[i] != 9.0)
+ abort ();
+
+ if (c[i] != 4.0)
+ abort ();
+
+ if (d[i] != 1.0)
+ abort ();
+
+ if (e[i] != 17.0)
+ abort ();
+ }
+
+ for (i = 0; i < N; i++)
+ {
+ a[i] = 4.0;
+ b[i] = 0.0;
+ c[i] = 0.0;
+ d[i] = 0.0;
+ e[i] = 0.0;
+ }
+
+#pragma acc data copyin (a[0:N], b[0:N], c[0:N]) copyin (N)
+ {
+
+#pragma acc parallel async
+ {
+ int ii;
+
+ for (ii = 0; ii < N; ii++)
+ b[ii] = (a[ii] * a[ii] * a[ii]) / a[ii];
+ }
+
+#pragma acc parallel async
+ {
+ int ii;
+
+ for (ii = 0; ii < N; ii++)
+ c[ii] = (a[ii] + a[ii] + a[ii] + a[ii]) / a[ii];
+ }
+
+#pragma acc update host (a[0:N], b[0:N], c[0:N]) wait (s)
+
+ }
+
+ for (i = 0; i < N; i++)
+ {
+ if (a[i] != 4.0)
+ abort ();
+
+ if (b[i] != 16.0)
+ abort ();
+
+ if (c[i] != 4.0)
+ abort ();
+ }
+
+
+ for (i = 0; i < N; i++)
+ {
+ a[i] = 5.0;
+ b[i] = 0.0;
+ c[i] = 0.0;
+ d[i] = 0.0;
+ e[i] = 0.0;
+ }
+
+#pragma acc data copyin (a[0:N], b[0:N], c[0:N]) copyin (N)
+ {
+
+#pragma acc parallel async
+ {
+ int ii;
+
+ for (ii = 0; ii < N; ii++)
+ b[ii] = (a[ii] * a[ii] * a[ii]) / a[ii];
+ }
+
+#pragma acc parallel async
+ {
+ int ii;
+
+ for (ii = 0; ii < N; ii++)
+ c[ii] = (a[ii] + a[ii] + a[ii] + a[ii]) / a[ii];
+ }
+
+#pragma acc update host (a[0:N], b[0:N], c[0:N]) async
+
+#pragma acc wait (s)
+
+ }
+
+ for (i = 0; i < N; i++)
+ {
+ if (a[i] != 5.0)
+ abort ();
+
+ if (b[i] != 25.0)
+ abort ();
+
+ if (c[i] != 4.0)
+ abort ();
+ }
+
+ for (i = 0; i < N; i++)
+ {
+ a[i] = 3.0;
+ b[i] = 0.0;
+ }
+
+#pragma acc data copy (a[0:N]) copy (b[0:N]) copyin (N)
+ {
+
+#pragma acc kernels async
+ {
+ int ii;
+
+ for (ii = 0; ii < N; ii++)
+ b[ii] = a[ii];
+ }
+
+#pragma acc wait
+
+ }
+
+ for (i = 0; i < N; i++)
+ {
+ if (a[i] != 3.0)
+ abort ();
+
+ if (b[i] != 3.0)
+ abort ();
+ }
+
+ for (i = 0; i < N; i++)
+ {
+ a[i] = 2.0;
+ b[i] = 0.0;
+ }
+
+#pragma acc data copy (a[0:N]) copy (b[0:N]) copyin (N)
+ {
+
+#pragma acc kernels async
+ {
+ int ii;
+
+ for (ii = 0; ii < N; ii++)
+ b[ii] = a[ii];
+ }
+
+#pragma acc wait (s)
+
+ }
+
+ for (i = 0; i < N; i++)
+ {
+ if (a[i] != 2.0)
+ abort ();
+
+ if (b[i] != 2.0)
+ abort ();
+ }
+
+ for (i = 0; i < N; i++)
+ {
+ a[i] = 3.0;
+ b[i] = 0.0;
+ c[i] = 0.0;
+ d[i] = 0.0;
+ }
+
+#pragma acc data copy (a[0:N]) copy (b[0:N]) copy (c[0:N]) copy (d[0:N]) copyin (N)
+ {
+
+#pragma acc kernels async
+ {
+ int ii;
+
+ for (ii = 0; ii < N; ii++)
+ b[ii] = (a[ii] * a[ii] * a[ii]) / a[ii];
+ }
+
+#pragma acc kernels async
+ {
+ int ii;
+
+ for (ii = 0; ii < N; ii++)
+ c[ii] = (a[ii] + a[ii] + a[ii] + a[ii]) / a[ii];
+ }
+
+
+#pragma acc kernels async
+ {
+ int ii;
+
+ for (ii = 0; ii < N; ii++)
+ d[ii] = ((a[ii] * a[ii] + a[ii]) / a[ii]) - a[ii];
+ }
+
+#pragma acc wait (s)
+
+ }
+
+ for (i = 0; i < N; i++)
+ {
+ if (a[i] != 3.0)
+ abort ();
+
+ if (b[i] != 9.0)
+ abort ();
+
+ if (c[i] != 4.0)
+ abort ();
+
+ if (d[i] != 1.0)
+ abort ();
+ }
+
+ for (i = 0; i < N; i++)
+ {
+ a[i] = 2.0;
+ b[i] = 0.0;
+ c[i] = 0.0;
+ d[i] = 0.0;
+ e[i] = 0.0;
+ }
+
+#pragma acc data copy (a[0:N], b[0:N], c[0:N], d[0:N], e[0:N]) copyin (N)
+ {
+
+#pragma acc kernels async
+ {
+ int ii;
+
+ for (ii = 0; ii < N; ii++)
+ b[ii] = (a[ii] * a[ii] * a[ii]) / a[ii];
+ }
+
+#pragma acc kernels async
+ {
+ int ii;
+
+ for (ii = 0; ii < N; ii++)
+ c[ii] = (a[ii] + a[ii] + a[ii] + a[ii]) / a[ii];
+ }
+
+#pragma acc kernels async
+ {
+ int ii;
+
+ for (ii = 0; ii < N; ii++)
+ d[ii] = ((a[ii] * a[ii] + a[ii]) / a[ii]) - a[ii];
+ }
+
+#pragma acc kernels wait (s) async (s)
+ {
+ int ii;
+
+ for (ii = 0; ii < N; ii++)
+ e[ii] = a[ii] + b[ii] + c[ii] + d[ii];
+ }
+
+#pragma acc wait (s)
+
+ }
+
+ for (i = 0; i < N; i++)
+ {
+ if (a[i] != 2.0)
+ abort ();
+
+ if (b[i] != 4.0)
+ abort ();
+
+ if (c[i] != 4.0)
+ abort ();
+
+ if (d[i] != 1.0)
+ abort ();
+
+ if (e[i] != 11.0)
+ abort ();
+ }
+
+
+ r = cuStreamCreate (&stream1, CU_STREAM_NON_BLOCKING);
+ if (r != CUDA_SUCCESS)
+ {
+ fprintf (stderr, "cuStreamCreate failed: %d\n", r);
+ abort ();
+ }
+
+ acc_set_cuda_stream (1, stream1);
+
+ for (i = 0; i < N; i++)
+ {
+ a[i] = 5.0;
+ b[i] = 0.0;
+ }
+
+#pragma acc data copy (a[0:N], b[0:N]) copyin (N)
+ {
+
+#pragma acc kernels async
+ {
+ int ii;
+
+ for (ii = 0; ii < N; ii++)
+ b[ii] = a[ii];
+ }
+
+#pragma acc wait (s)
+
+ }
+
+ for (i = 0; i < N; i++)
+ {
+ if (a[i] != 5.0)
+ abort ();
+
+ if (b[i] != 5.0)
+ abort ();
+ }
+
+ for (i = 0; i < N; i++)
+ {
+ a[i] = 7.0;
+ b[i] = 0.0;
+ c[i] = 0.0;
+ d[i] = 0.0;
+ }
+
+#pragma acc data copy (a[0:N]) copy (b[0:N]) copy (c[0:N]) copy (d[0:N]) copyin (N)
+ {
+
+#pragma acc kernels async
+ {
+ int ii;
+
+ for (ii = 0; ii < N; ii++)
+ b[ii] = (a[ii] * a[ii] * a[ii]) / a[ii];
+ }
+
+#pragma acc kernels async
+ {
+ int ii;
+
+ for (ii = 0; ii < N; ii++)
+ c[ii] = (a[ii] + a[ii] + a[ii] + a[ii]) / a[ii];
+ }
+
+#pragma acc kernels async
+ {
+ int ii;
+
+ for (ii = 0; ii < N; ii++)
+ d[ii] = ((a[ii] * a[ii] + a[ii]) / a[ii]) - a[ii];
+ }
+
+#pragma acc wait (s)
+
+ }
+
+ for (i = 0; i < N; i++)
+ {
+ if (a[i] != 7.0)
+ abort ();
+
+ if (b[i] != 49.0)
+ abort ();
+
+ if (c[i] != 4.0)
+ abort ();
+
+ if (d[i] != 1.0)
+ abort ();
+ }
+
+ for (i = 0; i < N; i++)
+ {
+ a[i] = 3.0;
+ b[i] = 0.0;
+ c[i] = 0.0;
+ d[i] = 0.0;
+ e[i] = 0.0;
+ }
+
+#pragma acc data copy (a[0:N], b[0:N], c[0:N], d[0:N], e[0:N]) copyin (N)
+ {
+
+#pragma acc kernels async
+ {
+ int ii;
+
+ for (ii = 0; ii < N; ii++)
+ b[ii] = (a[ii] * a[ii] * a[ii]) / a[ii];
+ }
+
+#pragma acc kernels async
+ {
+ int ii;
+
+ for (ii = 0; ii < N; ii++)
+ c[ii] = (a[ii] + a[ii] + a[ii] + a[ii]) / a[ii];
+ }
+
+#pragma acc kernels async
+ {
+ int ii;
+
+ for (ii = 0; ii < N; ii++)
+ d[ii] = ((a[ii] * a[ii] + a[ii]) / a[ii]) - a[ii];
+ }
+
+#pragma acc kernels wait (s) async (s)
+ {
+ int ii;
+
+ for (ii = 0; ii < N; ii++)
+ e[ii] = a[ii] + b[ii] + c[ii] + d[ii];
+ }
+
+#pragma acc wait (s)
+
+ }
+
+ for (i = 0; i < N; i++)
+ {
+ if (a[i] != 3.0)
+ abort ();
+
+ if (b[i] != 9.0)
+ abort ();
+
+ if (c[i] != 4.0)
+ abort ();
+
+ if (d[i] != 1.0)
+ abort ();
+
+ if (e[i] != 17.0)
+ abort ();
+ }
+
+ for (i = 0; i < N; i++)
+ {
+ a[i] = 4.0;
+ b[i] = 0.0;
+ c[i] = 0.0;
+ d[i] = 0.0;
+ e[i] = 0.0;
+ }
+
+#pragma acc data copyin (a[0:N], b[0:N], c[0:N]) copyin (N)
+ {
+
+#pragma acc kernels async
+ {
+ int ii;
+
+ for (ii = 0; ii < N; ii++)
+ b[ii] = (a[ii] * a[ii] * a[ii]) / a[ii];
+ }
+
+#pragma acc kernels async
+ {
+ int ii;
+
+ for (ii = 0; ii < N; ii++)
+ c[ii] = (a[ii] + a[ii] + a[ii] + a[ii]) / a[ii];
+ }
+
+#pragma acc update host (a[0:N], b[0:N], c[0:N]) wait (s)
+
+ }
+
+ for (i = 0; i < N; i++)
+ {
+ if (a[i] != 4.0)
+ abort ();
+
+ if (b[i] != 16.0)
+ abort ();
+
+ if (c[i] != 4.0)
+ abort ();
+ }
+
+
+ for (i = 0; i < N; i++)
+ {
+ a[i] = 5.0;
+ b[i] = 0.0;
+ c[i] = 0.0;
+ d[i] = 0.0;
+ e[i] = 0.0;
+ }
+
+#pragma acc data copyin (a[0:N], b[0:N], c[0:N]) copyin (N)
+ {
+
+#pragma acc kernels async
+ {
+ int ii;
+
+ for (ii = 0; ii < N; ii++)
+ b[ii] = (a[ii] * a[ii] * a[ii]) / a[ii];
+ }
+
+#pragma acc kernels async
+ {
+ int ii;
+
+ for (ii = 0; ii < N; ii++)
+ c[ii] = (a[ii] + a[ii] + a[ii] + a[ii]) / a[ii];
+ }
+
+#pragma acc update host (a[0:N], b[0:N], c[0:N]) async
+
+#pragma acc wait (s)
+
+ }
+
+ for (i = 0; i < N; i++)
+ {
+ if (a[i] != 5.0)
+ abort ();
+
+ if (b[i] != 25.0)
+ abort ();
+
+ if (c[i] != 4.0)
+ abort ();
+ }
+
+ acc_shutdown (acc_device_nvidia);
+
+ return 0;
+}
diff --git a/libgomp/testsuite/libgomp.oacc-c-c++-common/data-2-lib.c b/libgomp/testsuite/libgomp.oacc-c-c++-common/data-2-lib.c
index 2ddfa7d..f553d3d 100644
--- a/libgomp/testsuite/libgomp.oacc-c-c++-common/data-2-lib.c
+++ b/libgomp/testsuite/libgomp.oacc-c-c++-common/data-2-lib.c
@@ -153,7 +153,7 @@ main (int argc, char **argv)
d[ii] = ((a[ii] * a[ii] + a[ii]) / a[ii]) - a[ii];
#pragma acc parallel present (a[0:N], b[0:N], c[0:N], d[0:N], e[0:N], N) \
- async (4)
+ wait (1, 2, 3) async (4)
for (int ii = 0; ii < N; ii++)
e[ii] = a[ii] + b[ii] + c[ii] + d[ii];
diff --git a/libgomp/testsuite/libgomp.oacc-c-c++-common/data-2.c b/libgomp/testsuite/libgomp.oacc-c-c++-common/data-2.c
index 0c6abe6..81d623a 100644
--- a/libgomp/testsuite/libgomp.oacc-c-c++-common/data-2.c
+++ b/libgomp/testsuite/libgomp.oacc-c-c++-common/data-2.c
@@ -162,7 +162,7 @@ main (int argc, char **argv)
d[ii] = ((a[ii] * a[ii] + a[ii]) / a[ii]) - a[ii];
#pragma acc parallel present (a[0:N], b[0:N], c[0:N], d[0:N], e[0:N]) \
- wait (1) async (4)
+ wait (1, 2, 3) async (4)
for (int ii = 0; ii < N; ii++)
e[ii] = a[ii] + b[ii] + c[ii] + d[ii];
diff --git a/libgomp/testsuite/libgomp.oacc-c-c++-common/data-3.c b/libgomp/testsuite/libgomp.oacc-c-c++-common/data-3.c
index 0bf706a..5ec50b8 100644
--- a/libgomp/testsuite/libgomp.oacc-c-c++-common/data-3.c
+++ b/libgomp/testsuite/libgomp.oacc-c-c++-common/data-3.c
@@ -138,7 +138,7 @@ main (int argc, char **argv)
d[ii] = ((a[ii] * a[ii] + a[ii]) / a[ii]) - a[ii];
#pragma acc parallel present (a[0:N], b[0:N], c[0:N], d[0:N], e[0:N]) \
- wait (1,5) async (4)
+ wait (1, 2, 3, 5) async (4)
for (int ii = 0; ii < N; ii++)
e[ii] = a[ii] + b[ii] + c[ii] + d[ii];
diff --git a/libgomp/testsuite/libgomp.oacc-c-c++-common/lib-71.c b/libgomp/testsuite/libgomp.oacc-c-c++-common/lib-71.c
index c85e824..6afe2a0 100644
--- a/libgomp/testsuite/libgomp.oacc-c-c++-common/lib-71.c
+++ b/libgomp/testsuite/libgomp.oacc-c-c++-common/lib-71.c
@@ -92,16 +92,22 @@ main (int argc, char **argv)
abort ();
}
- fprintf (stderr, "CheCKpOInT\n");
- if (acc_async_test (1) != 0)
+ if (acc_async_test (0) != 0)
{
fprintf (stderr, "asynchronous operation not running\n");
abort ();
}
+ /* Test unseen async number. */
+ if (acc_async_test (1) != 1)
+ {
+ fprintf (stderr, "acc_async_test failed on unseen number\n");
+ abort ();
+ }
+
sleep ((int) (dtime / 1000.0f) + 1);
- if (acc_async_test (1) != 1)
+ if (acc_async_test (0) != 1)
{
fprintf (stderr, "found asynchronous operation still running\n");
abort ();
@@ -116,7 +122,3 @@ main (int argc, char **argv)
return 0;
}
-
-/* { dg-output "CheCKpOInT(\n|\r\n|\r).*" } */
-/* { dg-output "unknown async \[0-9\]+" } */
-/* { dg-shouldfail "" } */
diff --git a/libgomp/testsuite/libgomp.oacc-c-c++-common/lib-77.c b/libgomp/testsuite/libgomp.oacc-c-c++-common/lib-77.c
index f4f196d..2821f88 100644
--- a/libgomp/testsuite/libgomp.oacc-c-c++-common/lib-77.c
+++ b/libgomp/testsuite/libgomp.oacc-c-c++-common/lib-77.c
@@ -111,7 +111,7 @@ main (int argc, char **argv)
start_timer (0);
- acc_wait (1);
+ acc_wait (0);
atime = stop_timer (0);
@@ -132,7 +132,3 @@ main (int argc, char **argv)
return 0;
}
-
-/* { dg-output "CheCKpOInT(\n|\r\n|\r).*" } */
-/* { dg-output "unknown async \[0-9\]+" } */
-/* { dg-shouldfail "" } */
diff --git a/libgomp/testsuite/libgomp.oacc-c-c++-common/lib-79.c b/libgomp/testsuite/libgomp.oacc-c-c++-common/lib-79.c
index ef3df13..b22af26 100644
--- a/libgomp/testsuite/libgomp.oacc-c-c++-common/lib-79.c
+++ b/libgomp/testsuite/libgomp.oacc-c-c++-common/lib-79.c
@@ -114,6 +114,7 @@ main (int argc, char **argv)
for (i = 0; i < N; i++)
{
+ stream = (CUstream) acc_get_cuda_stream (i & 1);
r = cuLaunchKernel (delay, 1, 1, 1, 1, 1, 1, 0, stream, kargs, 0);
if (r != CUDA_SUCCESS)
{
@@ -122,11 +123,11 @@ main (int argc, char **argv)
}
}
- acc_wait_async (0, 1);
-
if (acc_async_test (0) != 0)
abort ();
+ acc_wait_async (0, 1);
+
if (acc_async_test (1) != 0)
abort ();
diff --git a/libgomp/testsuite/libgomp.oacc-c-c++-common/lib-81.c b/libgomp/testsuite/libgomp.oacc-c-c++-common/lib-81.c
index d5f18f0..30a4b57 100644
--- a/libgomp/testsuite/libgomp.oacc-c-c++-common/lib-81.c
+++ b/libgomp/testsuite/libgomp.oacc-c-c++-common/lib-81.c
@@ -133,7 +133,7 @@ main (int argc, char **argv)
for (i = 0; i <= N; i++)
{
- if (acc_async_test (i) != 0)
+ if (acc_async_test (i) == 0)
abort ();
}
More information about the Gcc-patches
mailing list