-
Notifications
You must be signed in to change notification settings - Fork 0
/
ca_main.cu
204 lines (189 loc) · 5.77 KB
/
ca_main.cu
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
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
#include "config.h"
#include "load_rle.h"
#include "render.h"
#include <stdio.h>
#include <time.h>
#include <unistd.h>
static cell *cuda_buffer1 = NULL, *cuda_buffer2 = NULL;
cell* host_buffer = NULL;
__host__ __device__ inline size_t idx(ssize_t X, ssize_t Y)
{
if (Y >= 0 && X >= 0 && ((Y * CANVAS_SIZE_X + X) < NUM_CELLS - 1)) {
return (Y * CANVAS_SIZE_X + X);
} else {
return NUM_CELLS - 1;
}
}
//#define idx(X, Y) \
//(((int)Y >= 0 && (int)X >= 0 && ((Y * CANVAS_SIZE_X + X) < NUM_CELLS - 1)) ? (Y * CANVAS_SIZE_X + X) : NUM_CELLS - 1)
__host__ __device__ inline size_t xdi_x(size_t i)
{
return i % CANVAS_SIZE_X;
}
__host__ __device__ inline size_t xdi_y(size_t i)
{
return i / CANVAS_SIZE_X;
}
void init()
{
debug_print("Using %lu dimensional canvas of size %zux%zu with %d bit colors\n", NDIM, CANVAS_SIZE_X, CANVAS_SIZE_Y, N_COLOR_BIT);
debug_print("Using two buffer each of size %lu mb, or %lu cells\n", BUFFER_SIZE / 1024 / 1024, BUFFER_SIZE / sizeof(cell));
#ifdef USE_CUDA
if (cuda_buffer1) {
cudaFree(cuda_buffer1);
cuda_buffer1 = NULL;
}
if (cuda_buffer2) {
cudaFree(cuda_buffer2);
cuda_buffer2 = NULL;
}
cudaCalloc(&cuda_buffer1, BUFFER_SIZE, 1);
cudaCalloc(&cuda_buffer2, BUFFER_SIZE, 1);
#else
if (cuda_buffer1) {
free(cuda_buffer1);
cuda_buffer1 = NULL;
}
if (cuda_buffer2) {
free(cuda_buffer2);
cuda_buffer2 = NULL;
}
cuda_buffer1 = (cell*)calloc(BUFFER_SIZE, 1);
cuda_buffer2 = (cell*)calloc(BUFFER_SIZE, 1);
#endif
if (host_buffer) {
free(host_buffer);
host_buffer = NULL;
}
host_buffer = (cell*)calloc(1, BUFFER_SIZE);
debug_print("Initialization done!\n");
}
__host__ __device__ inline unsigned int update_cell_bin_2d(cell& center, cell& c1, cell& c2, cell& c3, cell& c4, cell& c5, cell& c6, cell& c7, cell& c8)
{
unsigned int sum = c1.x
+ c2.x
+ c3.x
+ c4.x
+ c5.x
+ c6.x
+ c7.x
+ c8.x;
if (sum < 2 || sum > 3) {
return 0;
} else if ((center.x && (sum == 2 || sum == 3)) || (!center.x) && sum == 3) {
return 1;
}
return 0;
}
__global__ void update_cell_bin_2d_CUDA(cell* dest, cell* origin)
{
size_t i = blockIdx.x * blockDim.x + threadIdx.x;
if (i < CANVAS_SIZE_X * CANVAS_SIZE_Y) {
size_t x = xdi_x(i), y = xdi_y(i);
dest[i].x = update_cell_bin_2d(origin[i],
origin[idx(x - 1, y - 1)],
origin[idx(x + 1, y + 1)],
origin[idx(x, y - 1)],
origin[idx(x - 1, y)],
origin[idx(x + 1, y)],
origin[idx(x, y + 1)],
origin[idx(x - 1, y + 1)],
origin[idx(x + 1, y - 1)]);
}
}
#ifndef USE_CUDA
inline
#endif
void
update(cell* dest, cell* origin)
{
#ifdef USE_CUDA
update_cell_bin_2d_CUDA<<<ceil(CANVAS_SIZE_X * CANVAS_SIZE_Y / BLOCK_SIZE), BLOCK_SIZE>>>(dest, origin);
return;
#else
for (ssize_t x = 0; x < CANVAS_SIZE_X; x++) {
for (ssize_t y = 0; y < CANVAS_SIZE_Y; y++) {
size_t i = idx(x, y);
dest[i].x = update_cell_bin_2d(origin[i],
origin[idx(x - 1, y - 1)],
origin[idx(x + 1, y + 1)],
origin[idx(x, y - 1)],
origin[idx(x - 1, y)],
origin[idx(x + 1, y)],
origin[idx(x, y + 1)],
origin[idx(x - 1, y + 1)],
origin[idx(x + 1, y - 1)]);
}
}
#endif
}
inline void copy_buffer_to_host(cell* dst, cell* src)
{
#ifndef USE_CUDA
memcpy(dst, src, BUFFER_SIZE);
#else
cudaMemcpy(dst, src, BUFFER_SIZE, cudaMemcpyDeviceToHost);
#endif
}
inline void copy_buffer_to_device(cell* dst, cell* src)
{
#ifndef USE_CUDA
memcpy(dst, src, BUFFER_SIZE);
#else
cudaMemcpy(dst, src, BUFFER_SIZE, cudaMemcpyHostToDevice);
#endif
}
inline void print_buffer(cell* src)
{
#ifdef DO_TERM_DISPLAY
fprintf(stdout, "---------------------Iteration-------------------------\n");
for (ssize_t x = 0; x < CANVAS_SIZE_X; x++) {
for (ssize_t y = 0; y < CANVAS_SIZE_Y; y++) {
//debug_print("(%zu, %zu) -> %zu\n", x, y, idx(x, y));
fprintf(stdout, src[idx(x, y)].x ? " " : "█");
}
fprintf(stdout, "|\n");
}
#endif
}
int main(int argc, char** argv)
{
debug_print("-------------CA Running-----------\n");
init();
//for (int i = CANVAS_SIZE_X * 3 / 7; i < CANVAS_SIZE_X * 4 / 7; i++) {
//host_buffer[idx(i, i)].x = 1;
//host_buffer[idx(i, i + 1)].x = 1;
//host_buffer[idx(i, i - 1)].x = 1;
//host_buffer[idx(i, i - 5)].x = 1;
//host_buffer[idx(i - 1, i - 6)].x = 1;
//host_buffer[idx(i, i - 6)].x = 1;
//}
load_rle_file(host_buffer, argv[1] == NULL ? "./saves/glider.rle" : argv[1], .5, .5);
//load_rle_file(host_buffer, argv[1] == NULL ? "./saves/glider.rle" : argv[1], .3, .85);
//load_rle_file(host_buffer, "./saves/Gosper glider gun.rle", .5, .5);
//load_rle_file(host_buffer, "./saves/rats_synth.rle", .5, .5);
//load_rle_file(host_buffer, "./saves/fullutm.rle", 0, 0);
print_buffer(host_buffer);
copy_buffer_to_device(cuda_buffer1, host_buffer);
render_init();
int iteration = 80000;
#ifdef DO_TERM_DISPLAY
int delay = 30000;
#else
//int delay = 10000;
int delay = 0;
#endif
//while (0 == display_ready)
//; //wait for display
debug_print("Display is now ready\n");
for (int i = 0; i < iteration / 2; i++) {
update(cuda_buffer2, cuda_buffer1);
copy_buffer_to_host(host_buffer, cuda_buffer2);
print_buffer(host_buffer);
usleep(delay);
update(cuda_buffer1, cuda_buffer2);
copy_buffer_to_host(host_buffer, cuda_buffer1);
print_buffer(host_buffer);
usleep(delay);
}
}