|
@@ -0,0 +1,126 @@
|
|
|
+/* StarPU --- Runtime system for heterogeneous multicore architectures.
|
|
|
+ *
|
|
|
+ * Copyright (C) 2012 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.
|
|
|
+ */
|
|
|
+#ifndef STARPU_GCC_PLUGIN
|
|
|
+# error must be compiled with the StarPU GCC plug-in
|
|
|
+#endif
|
|
|
+
|
|
|
+/*
|
|
|
+ * This is a very simple example that is primarily meant to test the
|
|
|
+ * features offered by the StarPU GCC plug-in.
|
|
|
+ *
|
|
|
+ * Currently tested features :
|
|
|
+ * - multi-implementations
|
|
|
+ * - CPU codelet
|
|
|
+ *
|
|
|
+ * Features to test in a near future :
|
|
|
+ * - CUDA
|
|
|
+ * - OpenCL
|
|
|
+ * - Filters
|
|
|
+ */
|
|
|
+#include <stdio.h>
|
|
|
+#include <stdlib.h>
|
|
|
+
|
|
|
+#ifdef __SSE__
|
|
|
+#include <xmmintrin.h>
|
|
|
+#endif
|
|
|
+
|
|
|
+#define NX 16
|
|
|
+#define FACTOR 3.14
|
|
|
+
|
|
|
+static float vector[NX];
|
|
|
+
|
|
|
+static void vector_scal(float *v, size_t n, float factor)
|
|
|
+__attribute__ ((task));
|
|
|
+
|
|
|
+static void vector_scal_cpu(float *v, size_t n, float factor)
|
|
|
+__attribute__ ((task_implementation ("cpu", vector_scal)));
|
|
|
+
|
|
|
+#ifdef __SSE__
|
|
|
+static void vector_scal_sse(float *v, size_t n, float factor)
|
|
|
+__attribute__ ((task_implementation ("cpu", vector_scal)));
|
|
|
+#endif /* !__SSE__ */
|
|
|
+
|
|
|
+static void
|
|
|
+vector_scal_cpu(float *v, size_t n, float factor)
|
|
|
+{
|
|
|
+ int i;
|
|
|
+ for (i = 0; i < n; i++)
|
|
|
+ v[i] *= factor;
|
|
|
+}
|
|
|
+
|
|
|
+#ifdef __SSE__
|
|
|
+static void
|
|
|
+vector_scal_sse(float *vector, size_t n, float factor)
|
|
|
+{
|
|
|
+ unsigned int n_iterations = n/4;
|
|
|
+
|
|
|
+ __m128 *VECTOR = (__m128*) vector;
|
|
|
+ __m128 _FACTOR __attribute__((aligned(16)));
|
|
|
+ _FACTOR = _mm_set1_ps(factor);
|
|
|
+
|
|
|
+ unsigned int i;
|
|
|
+ for (i = 0; i < n_iterations; i++)
|
|
|
+ VECTOR[i] = _mm_mul_ps(_FACTOR, VECTOR[i]);
|
|
|
+
|
|
|
+ unsigned int remainder = n%4;
|
|
|
+ if (remainder != 0)
|
|
|
+ {
|
|
|
+ unsigned int start = 4 * n_iterations;
|
|
|
+ for (i = start; i < start+remainder; ++i)
|
|
|
+ {
|
|
|
+ vector[i] = factor * vector[i];
|
|
|
+ }
|
|
|
+ }
|
|
|
+}
|
|
|
+#endif /* !__SSE__ */
|
|
|
+
|
|
|
+static void
|
|
|
+init_data(void)
|
|
|
+{
|
|
|
+ int i;
|
|
|
+ for (i = 0; i < NX; i++)
|
|
|
+ vector[i] = (float) i;
|
|
|
+}
|
|
|
+
|
|
|
+#define EPSILON 1e-3
|
|
|
+static int
|
|
|
+check(void)
|
|
|
+{
|
|
|
+ int i;
|
|
|
+ for (i = 0; i < NX; i++)
|
|
|
+ if (vector[i] - i*FACTOR > EPSILON)
|
|
|
+ return 1;
|
|
|
+
|
|
|
+ return 0;
|
|
|
+}
|
|
|
+
|
|
|
+int
|
|
|
+main(void)
|
|
|
+{
|
|
|
+#pragma starpu initialize
|
|
|
+
|
|
|
+ init_data();
|
|
|
+
|
|
|
+#pragma starpu register &vector NX
|
|
|
+
|
|
|
+ vector_scal(vector, NX, FACTOR);
|
|
|
+
|
|
|
+#pragma starpu wait
|
|
|
+#pragma starpu unregister &vector
|
|
|
+#pragma starpu shutdown
|
|
|
+
|
|
|
+ return check()?EXIT_FAILURE:EXIT_SUCCESS;
|
|
|
+}
|