blob: 3bcd753dbbb15ee48954d6de798026a50c714b35 (
plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
|
/* { dg-do run } */
/* { dg-options "-O2" } */
/* { dg-additional-options "-msse2" { target sse2_runtime } } */
/* { dg-additional-options "-mavx" { target avx_runtime } } */
#include <stdlib.h>
#define EPS 0.00001
#define N 10000
#define M 1024
#pragma omp declare target
float Q[N][N];
#pragma omp declare simd uniform(i) linear(k) notinbranch
float Pfun (const int i, const int k)
{
return Q[i][k] * Q[k][i];
}
#pragma omp end declare target
void init ()
{
int i, j;
for (i = 0; i < N; i++)
for (j = 0; j < N; j++)
Q[i][j] = 0.001 * i * j;
}
float accum_ref ()
{
int i, k;
float tmp = 0.0;
for (i = 0; i < N; i++)
{
float tmp1 = 0.0;
for (k = 0; k < M; k++)
tmp1 += Pfun(i,k);
tmp += tmp1;
}
return tmp;
}
float accum ()
{
int i, k;
float tmp = 0.0;
#pragma omp target
#pragma omp parallel for reduction(+:tmp)
for (i = 0; i < N; i++)
{
float tmp1 = 0.0;
#pragma omp simd reduction(+:tmp1)
for (k = 0; k < M; k++)
tmp1 += Pfun(i,k);
tmp += tmp1;
}
return tmp;
}
void check (float a, float b)
{
float err = (b == 0.0) ? a : (a - b) / b;
if (((err > 0) ? err : -err) > EPS)
abort ();
}
int main ()
{
init ();
#pragma omp target update to(Q)
check (accum (), accum_ref ());
return 0;
}
|
