|
|
@@ -0,0 +1,105 @@
|
|
|
+/* StarPU --- Runtime system for heterogeneous multicore architectures.
|
|
|
+ *
|
|
|
+ * Copyright (C) 2014 Inria
|
|
|
+ *
|
|
|
+ * StarPU is free software; you can redistribute it and/or modify
|
|
|
+ * it under the terms of the GNU Lesser General Public License as published by
|
|
|
+ * the Free Software Foundation; either version 2.1 of the License, or (at
|
|
|
+ * your option) any later version.
|
|
|
+ *
|
|
|
+ * StarPU is distributed in the hope that it will be useful, but
|
|
|
+ * WITHOUT ANY WARRANTY; without even the implied warranty of
|
|
|
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
|
|
|
+ *
|
|
|
+ * See the GNU Lesser General Public License in COPYING.LGPL for more details.
|
|
|
+ */
|
|
|
+
|
|
|
+#include <starpu.h>
|
|
|
+#include <stdlib.h>
|
|
|
+#include <stdio.h>
|
|
|
+#include <unistd.h>
|
|
|
+#ifdef STARPU_USE_CUDA
|
|
|
+#include <starpu_cuda.h>
|
|
|
+#endif
|
|
|
+
|
|
|
+static unsigned cnt;
|
|
|
+
|
|
|
+static void cpu_memcpy(void *descr[], void *cl_arg)
|
|
|
+{
|
|
|
+ double *res = (double *)STARPU_VECTOR_GET_PTR(descr[0]);
|
|
|
+ double *dst = (double *)STARPU_VECTOR_GET_PTR(descr[1]);
|
|
|
+ int n = (int) STARPU_VECTOR_GET_NX(descr[0]);
|
|
|
+ int me = (uintptr_t)cl_arg;
|
|
|
+
|
|
|
+ if (me == 0)
|
|
|
+ {
|
|
|
+ sleep(1);
|
|
|
+ STARPU_ASSERT(STARPU_ATOMIC_ADD(&cnt,1) == 1);
|
|
|
+ }
|
|
|
+ else
|
|
|
+ STARPU_ASSERT(STARPU_ATOMIC_ADD(&cnt,1) != 1);
|
|
|
+
|
|
|
+ memcpy(dst, res, n*sizeof(double));
|
|
|
+}
|
|
|
+
|
|
|
+
|
|
|
+static struct starpu_codelet my_cl = {
|
|
|
+ .where = STARPU_CPU,
|
|
|
+ .cpu_funcs = {cpu_memcpy, NULL},
|
|
|
+ .nbuffers = STARPU_VARIABLE_NBUFFERS
|
|
|
+};
|
|
|
+
|
|
|
+int
|
|
|
+main()
|
|
|
+{
|
|
|
+ double *res, *a;
|
|
|
+ unsigned n=100000, i;
|
|
|
+ starpu_data_handle_t res_handle, a_handle;
|
|
|
+ unsigned nb_tasks = 10, worker;
|
|
|
+
|
|
|
+ if (starpu_init(NULL))
|
|
|
+ exit(-1);
|
|
|
+
|
|
|
+ starpu_malloc((void**)&res, n*sizeof(double));
|
|
|
+ starpu_malloc((void**)&a, n*sizeof(double));
|
|
|
+
|
|
|
+ for(i=0; i < n; i++)
|
|
|
+ res[i] = a[i] = 1.0;
|
|
|
+
|
|
|
+ starpu_vector_data_register(&res_handle, 0, (uintptr_t)res, (uint32_t)n, sizeof(double));
|
|
|
+ starpu_vector_data_register(&a_handle, 0, (uintptr_t)a, (uint32_t)n, sizeof(double));
|
|
|
+
|
|
|
+ for (i = 0; i < nb_tasks; i++) {
|
|
|
+ struct starpu_task *task = starpu_task_create();
|
|
|
+ task->cl=&my_cl;
|
|
|
+ task->nbuffers = 2;
|
|
|
+ task->handles[0] = res_handle;
|
|
|
+
|
|
|
+ if (i == 0)
|
|
|
+ task->modes[0] = STARPU_RW;
|
|
|
+ else
|
|
|
+ task->modes[0] = STARPU_RW | STARPU_COMMUTE;
|
|
|
+
|
|
|
+ task->handles[1] = a_handle;
|
|
|
+ task->modes[1] = STARPU_R;
|
|
|
+ task->cl_arg = (void*)(uintptr_t)i;
|
|
|
+
|
|
|
+ if (starpu_task_submit(task)) {
|
|
|
+ fprintf(stderr,"taks submmition failed!");
|
|
|
+ exit(EXIT_FAILURE);
|
|
|
+ }
|
|
|
+ }
|
|
|
+
|
|
|
+ starpu_task_wait_for_all ();
|
|
|
+
|
|
|
+ starpu_data_unregister(res_handle);
|
|
|
+ starpu_data_unregister(a_handle);
|
|
|
+
|
|
|
+ starpu_free(res);
|
|
|
+ starpu_free(a);
|
|
|
+
|
|
|
+ starpu_shutdown();
|
|
|
+ return EXIT_SUCCESS;
|
|
|
+}
|
|
|
+
|
|
|
+
|
