-
Notifications
You must be signed in to change notification settings - Fork 28
/
Copy pathDRB056-jacobi2d-tile-no.c
173 lines (167 loc) · 13.5 KB
/
DRB056-jacobi2d-tile-no.c
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
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
/**
* jacobi-2d-imper.c: This file is part of the PolyBench/C 3.2 test suite.
* Jacobi with array copying, no reduction. with tiling and nested SIMD.
*
* Contact: Louis-Noel Pouchet <[email protected]>
* Web address: http://polybench.sourceforge.net
* License: /LICENSE.OSU.txt
*/
#include <stdio.h>
#include <unistd.h>
#include <string.h>
#include <math.h>
/* Include polybench common header. */
#include "polybench/polybench.h"
/* Include benchmark-specific header. */
/* Default data type is double, default size is 20x1000. */
#include "polybench/jacobi-2d-imper.h"
/* Array initialization. */
static void init_array(int n,double A[500 + 0][500 + 0],double B[500 + 0][500 + 0])
{
//int i;
//int j;
{
int c1;
int c2;
int c4;
int c3;
if (n >= 1) {
#pragma omp parallel for private(c3, c4, c2)
for (c1 = 0; c1 <= (((n + -1) * 16 < 0?((16 < 0?-((-(n + -1) + 16 + 1) / 16) : -((-(n + -1) + 16 - 1) / 16))) : (n + -1) / 16)); c1++) {
for (c2 = 0; c2 <= (((n + -1) * 16 < 0?((16 < 0?-((-(n + -1) + 16 + 1) / 16) : -((-(n + -1) + 16 - 1) / 16))) : (n + -1) / 16)); c2++) {
for (c3 = 16 * c2; c3 <= ((16 * c2 + 15 < n + -1?16 * c2 + 15 : n + -1)); c3++) {
#pragma omp simd
for (c4 = 16 * c1; c4 <= ((16 * c1 + 15 < n + -1?16 * c1 + 15 : n + -1)); c4++) {
A[c4][c3] = (((double )c4) * (c3 + 2) + 2) / n;
B[c4][c3] = (((double )c4) * (c3 + 3) + 3) / n;
}
}
}
}
}
}
}
/* DCE code. Must scan the entire live-out data.
Can be used also to check the correctness of the output. */
static void print_array(int n,double A[500 + 0][500 + 0])
{
int i;
int j;
for (i = 0; i < n; i++)
for (j = 0; j < n; j++) {
fprintf(stderr,"%0.2lf ",A[i][j]);
if ((i * n + j) % 20 == 0)
fprintf(stderr,"\n");
}
fprintf(stderr,"\n");
}
/* Main computational kernel. The whole function will be timed,
including the call and return. */
static void kernel_jacobi_2d_imper(int tsteps,int n,double A[500 + 0][500 + 0],double B[500 + 0][500 + 0])
{
//int t;
//int i;
//int j;
//#pragma scop
{
int c0;
int c1;
int c3;
int c2;
int c4;
int c5;
if (n >= 3 && tsteps >= 1) {
for (c0 = 0; c0 <= (((n + 3 * tsteps + -4) * 16 < 0?((16 < 0?-((-(n + 3 * tsteps + -4) + 16 + 1) / 16) : -((-(n + 3 * tsteps + -4) + 16 - 1) / 16))) : (n + 3 * tsteps + -4) / 16)); c0++) {
#pragma omp parallel for private(c5, c4, c2, c3)
for (c1 = (((2 * c0 * 3 < 0?-(-(2 * c0) / 3) : ((3 < 0?(-(2 * c0) + - 3 - 1) / - 3 : (2 * c0 + 3 - 1) / 3)))) > (((16 * c0 + -1 * tsteps + 1) * 16 < 0?-(-(16 * c0 + -1 * tsteps + 1) / 16) : ((16 < 0?(-(16 * c0 + -1 * tsteps + 1) + - 16 - 1) / - 16 : (16 * c0 + -1 * tsteps + 1 + 16 - 1) / 16))))?((2 * c0 * 3 < 0?-(-(2 * c0) / 3) : ((3 < 0?(-(2 * c0) + - 3 - 1) / - 3 : (2 * c0 + 3 - 1) / 3)))) : (((16 * c0 + -1 * tsteps + 1) * 16 < 0?-(-(16 * c0 + -1 * tsteps + 1) / 16) : ((16 < 0?(-(16 * c0 + -1 * tsteps + 1) + - 16 - 1) / - 16 : (16 * c0 + -1 * tsteps + 1 + 16 - 1) / 16))))); c1 <= (((((((n + 2 * tsteps + -3) * 16 < 0?((16 < 0?-((-(n + 2 * tsteps + -3) + 16 + 1) / 16) : -((-(n + 2 * tsteps + -3) + 16 - 1) / 16))) : (n + 2 * tsteps + -3) / 16)) < (((32 * c0 + n + 29) * 48 < 0?((48 < 0?-((-(32 * c0 + n + 29) + 48 + 1) / 48) : -((-(32 * c0 + n + 29) + 48 - 1) / 48))) : (32 * c0 + n + 29) / 48))?(((n + 2 * tsteps + -3) * 16 < 0?((16 < 0?-((-(n + 2 * tsteps + -3) + 16 + 1) / 16) : -((-(n + 2 * tsteps + -3) + 16 - 1) / 16))) : (n + 2 * tsteps + -3) / 16)) : (((32 * c0 + n + 29) * 48 < 0?((48 < 0?-((-(32 * c0 + n + 29) + 48 + 1) / 48) : -((-(32 * c0 + n + 29) + 48 - 1) / 48))) : (32 * c0 + n + 29) / 48)))) < c0?(((((n + 2 * tsteps + -3) * 16 < 0?((16 < 0?-((-(n + 2 * tsteps + -3) + 16 + 1) / 16) : -((-(n + 2 * tsteps + -3) + 16 - 1) / 16))) : (n + 2 * tsteps + -3) / 16)) < (((32 * c0 + n + 29) * 48 < 0?((48 < 0?-((-(32 * c0 + n + 29) + 48 + 1) / 48) : -((-(32 * c0 + n + 29) + 48 - 1) / 48))) : (32 * c0 + n + 29) / 48))?(((n + 2 * tsteps + -3) * 16 < 0?((16 < 0?-((-(n + 2 * tsteps + -3) + 16 + 1) / 16) : -((-(n + 2 * tsteps + -3) + 16 - 1) / 16))) : (n + 2 * tsteps + -3) / 16)) : (((32 * c0 + n + 29) * 48 < 0?((48 < 0?-((-(32 * c0 + n + 29) + 48 + 1) / 48) : -((-(32 * c0 + n + 29) + 48 - 1) / 48))) : (32 * c0 + n + 29) / 48)))) : c0)); c1++) {
for (c2 = ((((16 * c1 + -1 * n + -12) * 16 < 0?-(-(16 * c1 + -1 * n + -12) / 16) : ((16 < 0?(-(16 * c1 + -1 * n + -12) + - 16 - 1) / - 16 : (16 * c1 + -1 * n + -12 + 16 - 1) / 16)))) > 2 * c0 + -2 * c1?(((16 * c1 + -1 * n + -12) * 16 < 0?-(-(16 * c1 + -1 * n + -12) / 16) : ((16 < 0?(-(16 * c1 + -1 * n + -12) + - 16 - 1) / - 16 : (16 * c1 + -1 * n + -12 + 16 - 1) / 16)))) : 2 * c0 + -2 * c1); c2 <= (((((((16 * c1 + n + 12) * 16 < 0?((16 < 0?-((-(16 * c1 + n + 12) + 16 + 1) / 16) : -((-(16 * c1 + n + 12) + 16 - 1) / 16))) : (16 * c1 + n + 12) / 16)) < (((n + 2 * tsteps + -3) * 16 < 0?((16 < 0?-((-(n + 2 * tsteps + -3) + 16 + 1) / 16) : -((-(n + 2 * tsteps + -3) + 16 - 1) / 16))) : (n + 2 * tsteps + -3) / 16))?(((16 * c1 + n + 12) * 16 < 0?((16 < 0?-((-(16 * c1 + n + 12) + 16 + 1) / 16) : -((-(16 * c1 + n + 12) + 16 - 1) / 16))) : (16 * c1 + n + 12) / 16)) : (((n + 2 * tsteps + -3) * 16 < 0?((16 < 0?-((-(n + 2 * tsteps + -3) + 16 + 1) / 16) : -((-(n + 2 * tsteps + -3) + 16 - 1) / 16))) : (n + 2 * tsteps + -3) / 16)))) < (((32 * c0 + -32 * c1 + n + 29) * 16 < 0?((16 < 0?-((-(32 * c0 + -32 * c1 + n + 29) + 16 + 1) / 16) : -((-(32 * c0 + -32 * c1 + n + 29) + 16 - 1) / 16))) : (32 * c0 + -32 * c1 + n + 29) / 16))?(((((16 * c1 + n + 12) * 16 < 0?((16 < 0?-((-(16 * c1 + n + 12) + 16 + 1) / 16) : -((-(16 * c1 + n + 12) + 16 - 1) / 16))) : (16 * c1 + n + 12) / 16)) < (((n + 2 * tsteps + -3) * 16 < 0?((16 < 0?-((-(n + 2 * tsteps + -3) + 16 + 1) / 16) : -((-(n + 2 * tsteps + -3) + 16 - 1) / 16))) : (n + 2 * tsteps + -3) / 16))?(((16 * c1 + n + 12) * 16 < 0?((16 < 0?-((-(16 * c1 + n + 12) + 16 + 1) / 16) : -((-(16 * c1 + n + 12) + 16 - 1) / 16))) : (16 * c1 + n + 12) / 16)) : (((n + 2 * tsteps + -3) * 16 < 0?((16 < 0?-((-(n + 2 * tsteps + -3) + 16 + 1) / 16) : -((-(n + 2 * tsteps + -3) + 16 - 1) / 16))) : (n + 2 * tsteps + -3) / 16)))) : (((32 * c0 + -32 * c1 + n + 29) * 16 < 0?((16 < 0?-((-(32 * c0 + -32 * c1 + n + 29) + 16 + 1) / 16) : -((-(32 * c0 + -32 * c1 + n + 29) + 16 - 1) / 16))) : (32 * c0 + -32 * c1 + n + 29) / 16)))); c2++) {
if (c0 <= (((32 * c1 + 16 * c2 + -1 * n + 1) * 32 < 0?((32 < 0?-((-(32 * c1 + 16 * c2 + -1 * n + 1) + 32 + 1) / 32) : -((-(32 * c1 + 16 * c2 + -1 * n + 1) + 32 - 1) / 32))) : (32 * c1 + 16 * c2 + -1 * n + 1) / 32)) && c1 <= c2 + -1) {
if ((n + 1) % 2 == 0) {
for (c4 = (16 * c1 > 16 * c2 + -1 * n + 3?16 * c1 : 16 * c2 + -1 * n + 3); c4 <= 16 * c1 + 15; c4++) {
A[-16 * c2 + c4 + n + -2][n + -2] = B[-16 * c2 + c4 + n + -2][n + -2];
}
}
}
if (c0 <= (((48 * c1 + -1 * n + 1) * 32 < 0?((32 < 0?-((-(48 * c1 + -1 * n + 1) + 32 + 1) / 32) : -((-(48 * c1 + -1 * n + 1) + 32 - 1) / 32))) : (48 * c1 + -1 * n + 1) / 32)) && c1 >= c2) {
if ((n + 1) % 2 == 0) {
for (c5 = (16 * c2 > 16 * c1 + -1 * n + 3?16 * c2 : 16 * c1 + -1 * n + 3); c5 <= ((16 * c1 < 16 * c2 + 15?16 * c1 : 16 * c2 + 15)); c5++) {
A[n + -2][-16 * c1 + c5 + n + -2] = B[n + -2][-16 * c1 + c5 + n + -2];
}
}
}
for (c3 = ((((((16 * c1 + -1 * n + 2) * 2 < 0?-(-(16 * c1 + -1 * n + 2) / 2) : ((2 < 0?(-(16 * c1 + -1 * n + 2) + - 2 - 1) / - 2 : (16 * c1 + -1 * n + 2 + 2 - 1) / 2)))) > (((16 * c2 + -1 * n + 2) * 2 < 0?-(-(16 * c2 + -1 * n + 2) / 2) : ((2 < 0?(-(16 * c2 + -1 * n + 2) + - 2 - 1) / - 2 : (16 * c2 + -1 * n + 2 + 2 - 1) / 2))))?(((16 * c1 + -1 * n + 2) * 2 < 0?-(-(16 * c1 + -1 * n + 2) / 2) : ((2 < 0?(-(16 * c1 + -1 * n + 2) + - 2 - 1) / - 2 : (16 * c1 + -1 * n + 2 + 2 - 1) / 2)))) : (((16 * c2 + -1 * n + 2) * 2 < 0?-(-(16 * c2 + -1 * n + 2) / 2) : ((2 < 0?(-(16 * c2 + -1 * n + 2) + - 2 - 1) / - 2 : (16 * c2 + -1 * n + 2 + 2 - 1) / 2)))))) > 16 * c0 + -16 * c1?(((((16 * c1 + -1 * n + 2) * 2 < 0?-(-(16 * c1 + -1 * n + 2) / 2) : ((2 < 0?(-(16 * c1 + -1 * n + 2) + - 2 - 1) / - 2 : (16 * c1 + -1 * n + 2 + 2 - 1) / 2)))) > (((16 * c2 + -1 * n + 2) * 2 < 0?-(-(16 * c2 + -1 * n + 2) / 2) : ((2 < 0?(-(16 * c2 + -1 * n + 2) + - 2 - 1) / - 2 : (16 * c2 + -1 * n + 2 + 2 - 1) / 2))))?(((16 * c1 + -1 * n + 2) * 2 < 0?-(-(16 * c1 + -1 * n + 2) / 2) : ((2 < 0?(-(16 * c1 + -1 * n + 2) + - 2 - 1) / - 2 : (16 * c1 + -1 * n + 2 + 2 - 1) / 2)))) : (((16 * c2 + -1 * n + 2) * 2 < 0?-(-(16 * c2 + -1 * n + 2) / 2) : ((2 < 0?(-(16 * c2 + -1 * n + 2) + - 2 - 1) / - 2 : (16 * c2 + -1 * n + 2 + 2 - 1) / 2)))))) : 16 * c0 + -16 * c1); c3 <= ((((((8 * c1 + 6 < 8 * c2 + 6?8 * c1 + 6 : 8 * c2 + 6)) < tsteps + -1?((8 * c1 + 6 < 8 * c2 + 6?8 * c1 + 6 : 8 * c2 + 6)) : tsteps + -1)) < 16 * c0 + -16 * c1 + 15?((((8 * c1 + 6 < 8 * c2 + 6?8 * c1 + 6 : 8 * c2 + 6)) < tsteps + -1?((8 * c1 + 6 < 8 * c2 + 6?8 * c1 + 6 : 8 * c2 + 6)) : tsteps + -1)) : 16 * c0 + -16 * c1 + 15)); c3++) {
if (c1 <= ((c3 * 8 < 0?((8 < 0?-((-c3 + 8 + 1) / 8) : -((-c3 + 8 - 1) / 8))) : c3 / 8))) {
for (c5 = (16 * c2 > 2 * c3 + 1?16 * c2 : 2 * c3 + 1); c5 <= ((16 * c2 + 15 < 2 * c3 + n + -2?16 * c2 + 15 : 2 * c3 + n + -2)); c5++) {
B[1][-2 * c3 + c5] = 0.2 * (A[1][-2 * c3 + c5] + A[1][-2 * c3 + c5 - 1] + A[1][1 + (-2 * c3 + c5)] + A[1 + 1][-2 * c3 + c5] + A[1 - 1][-2 * c3 + c5]);
}
}
for (c4 = (16 * c1 > 2 * c3 + 2?16 * c1 : 2 * c3 + 2); c4 <= ((16 * c1 + 15 < 2 * c3 + n + -2?16 * c1 + 15 : 2 * c3 + n + -2)); c4++) {
if (c2 <= ((c3 * 8 < 0?((8 < 0?-((-c3 + 8 + 1) / 8) : -((-c3 + 8 - 1) / 8))) : c3 / 8))) {
B[-2 * c3 + c4][1] = 0.2 * (A[-2 * c3 + c4][1] + A[-2 * c3 + c4][1 - 1] + A[-2 * c3 + c4][1 + 1] + A[1 + (-2 * c3 + c4)][1] + A[-2 * c3 + c4 - 1][1]);
}
for (c5 = (16 * c2 > 2 * c3 + 2?16 * c2 : 2 * c3 + 2); c5 <= ((16 * c2 + 15 < 2 * c3 + n + -2?16 * c2 + 15 : 2 * c3 + n + -2)); c5++) {
B[-2 * c3 + c4][-2 * c3 + c5] = 0.2 * (A[-2 * c3 + c4][-2 * c3 + c5] + A[-2 * c3 + c4][-2 * c3 + c5 - 1] + A[-2 * c3 + c4][1 + (-2 * c3 + c5)] + A[1 + (-2 * c3 + c4)][-2 * c3 + c5] + A[-2 * c3 + c4 - 1][-2 * c3 + c5]);
A[-2 * c3 + c4 + -1][-2 * c3 + c5 + -1] = B[-2 * c3 + c4 + -1][-2 * c3 + c5 + -1];
}
if (c2 >= (((2 * c3 + n + -16) * 16 < 0?-(-(2 * c3 + n + -16) / 16) : ((16 < 0?(-(2 * c3 + n + -16) + - 16 - 1) / - 16 : (2 * c3 + n + -16 + 16 - 1) / 16))))) {
A[-2 * c3 + c4 + -1][n + -2] = B[-2 * c3 + c4 + -1][n + -2];
}
}
if (c1 >= (((2 * c3 + n + -16) * 16 < 0?-(-(2 * c3 + n + -16) / 16) : ((16 < 0?(-(2 * c3 + n + -16) + - 16 - 1) / - 16 : (2 * c3 + n + -16 + 16 - 1) / 16))))) {
for (c5 = (16 * c2 > 2 * c3 + 2?16 * c2 : 2 * c3 + 2); c5 <= ((16 * c2 + 15 < 2 * c3 + n + -1?16 * c2 + 15 : 2 * c3 + n + -1)); c5++) {
A[n + -2][-2 * c3 + c5 + -1] = B[n + -2][-2 * c3 + c5 + -1];
}
}
}
if (c0 >= (((2 * c1 + c2 + -1) * 2 < 0?-(-(2 * c1 + c2 + -1) / 2) : ((2 < 0?(-(2 * c1 + c2 + -1) + - 2 - 1) / - 2 : (2 * c1 + c2 + -1 + 2 - 1) / 2)))) && c1 >= c2 + 1 && c2 <= (((tsteps + -8) * 8 < 0?((8 < 0?-((-(tsteps + -8) + 8 + 1) / 8) : -((-(tsteps + -8) + 8 - 1) / 8))) : (tsteps + -8) / 8))) {
for (c4 = 16 * c1; c4 <= ((16 * c1 + 15 < 16 * c2 + n + 12?16 * c1 + 15 : 16 * c2 + n + 12)); c4++) {
B[-16 * c2 + c4 + -14][1] = 0.2 * (A[-16 * c2 + c4 + -14][1] + A[-16 * c2 + c4 + -14][1 - 1] + A[-16 * c2 + c4 + -14][1 + 1] + A[1 + (-16 * c2 + c4 + -14)][1] + A[-16 * c2 + c4 + -14 - 1][1]);
}
}
if (c0 >= (((3 * c1 + -1) * 2 < 0?-(-(3 * c1 + -1) / 2) : ((2 < 0?(-(3 * c1 + -1) + - 2 - 1) / - 2 : (3 * c1 + -1 + 2 - 1) / 2)))) && c1 <= (((((tsteps + -8) * 8 < 0?((8 < 0?-((-(tsteps + -8) + 8 + 1) / 8) : -((-(tsteps + -8) + 8 - 1) / 8))) : (tsteps + -8) / 8)) < c2?(((tsteps + -8) * 8 < 0?((8 < 0?-((-(tsteps + -8) + 8 + 1) / 8) : -((-(tsteps + -8) + 8 - 1) / 8))) : (tsteps + -8) / 8)) : c2))) {
for (c5 = (16 * c2 > 16 * c1 + 15?16 * c2 : 16 * c1 + 15); c5 <= ((16 * c2 + 15 < 16 * c1 + n + 12?16 * c2 + 15 : 16 * c1 + n + 12)); c5++) {
B[1][-16 * c1 + c5 + -14] = 0.2 * (A[1][-16 * c1 + c5 + -14] + A[1][-16 * c1 + c5 + -14 - 1] + A[1][1 + (-16 * c1 + c5 + -14)] + A[1 + 1][-16 * c1 + c5 + -14] + A[1 - 1][-16 * c1 + c5 + -14]);
}
}
}
}
}
}
}
//#pragma endscop
}
int main(int argc,char **argv)
{
/* Retrieve problem size. */
int n = 500;
int tsteps = 10;
/* Variable declaration/allocation. */
double (*A)[500 + 0][500 + 0];
A = ((double (*)[500 + 0][500 + 0])(polybench_alloc_data(((500 + 0) * (500 + 0)),(sizeof(double )))));
;
double (*B)[500 + 0][500 + 0];
B = ((double (*)[500 + 0][500 + 0])(polybench_alloc_data(((500 + 0) * (500 + 0)),(sizeof(double )))));
;
/* Initialize array(s). */
init_array(n, *A, *B);
/* Start timer. */
polybench_timer_start();
;
/* Run kernel. */
kernel_jacobi_2d_imper(tsteps,n, *A, *B);
/* Stop and print timer. */
polybench_timer_stop();
;
polybench_timer_print();
;
/* Prevent dead-code elimination. All live-out data must be printed
by the function call in argument. */
if (argc > 42 && !strcmp(argv[0],""))
print_array(n, *A);
/* Be clean. */
free(((void *)A));
;
free(((void *)B));
;
return 0;
}