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
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
|
/*
* Benchmark broadcast 1D
*/
kernel void bench_workgroup_broadcast_1D_int(global int *src,
global int *dst,
int reduce_loop,
uint wg_local_x,
uint wg_local_y)
{
uint offset = 0;
uint index = offset + get_global_id(0);
int val = src[index];
/* depending on generated ASM, volatile may be removed */
volatile int result;
for(; reduce_loop > 0; reduce_loop--){
result = work_group_broadcast(val,
wg_local_x);
}
dst[index] = result;
}
kernel void bench_workgroup_broadcast_1D_long(global long *src,
global long *dst,
int reduce_loop,
uint wg_local_x,
uint wg_local_y)
{
uint offset = 0;
uint index = offset + get_global_id(0);
long val = src[index];
/* depending on generated ASM, volatile may be removed */
volatile long result;
for(; reduce_loop > 0; reduce_loop--){
result = work_group_broadcast(val,
wg_local_x);
}
dst[index] = result;
}
/*
* Benchmark broadcast 2D
*/
kernel void bench_workgroup_broadcast_2D_int(global int *src,
global int *dst,
int reduce_loop,
uint wg_local_x,
uint wg_local_y)
{
uint lsize = get_local_size(0) * get_local_size(1);
uint offset = get_group_id(0) * lsize +
get_group_id(1) * get_num_groups(0) * lsize;
uint index = offset + get_local_id(0) +
get_local_id(1) * get_local_size(0);
int val = src[index];
/* depending on generated ASM, volatile may be removed */
int result;
for(; reduce_loop > 0; reduce_loop--){
result = work_group_broadcast(val,
wg_local_x,
wg_local_y);
}
dst[index] = result;
}
kernel void bench_workgroup_broadcast_2D_long(global long *src,
global long *dst,
int reduce_loop,
uint wg_local_x,
uint wg_local_y)
{
uint lsize = get_local_size(0) * get_local_size(1);
uint offset = get_group_id(0) * lsize +
get_group_id(1) * get_num_groups(0) * lsize;
uint index = offset + get_local_id(0) +
get_local_id(1) * get_local_size(0);
long val = src[index];
/* depending on generated ASM, volatile may be removed */
long result;
for(; reduce_loop > 0; reduce_loop--){
result = work_group_broadcast(val,
wg_local_x,
wg_local_y);
}
dst[index] = result;
}
/*
* Benchmark workgroup reduce add
*/
kernel void bench_workgroup_reduce_add_int(
global int *src,
global int *dst,
int reduce_loop)
{
int val;
int result;
for(; reduce_loop > 0; reduce_loop--){
val = src[get_global_id(0)];
result = work_group_reduce_add(val);
}
dst[get_global_id(0)] = result;
}
kernel void bench_workgroup_reduce_add_long(
global long *src,
global long *dst,
int reduce_loop)
{
long val;
long result;
for(; reduce_loop > 0; reduce_loop--){
val = src[get_global_id(0)];
result = work_group_reduce_add(val);
}
dst[get_global_id(0)] = result;
}
/*
* Benchmark workgroup reduce min
*/
kernel void bench_workgroup_reduce_min_int(
global int *src,
global int *dst,
int reduce_loop)
{
int val;
int result;
for(; reduce_loop > 0; reduce_loop--){
val = src[get_global_id(0)];
result = work_group_reduce_min(val);
}
dst[get_global_id(0)] = result;
}
kernel void bench_workgroup_reduce_min_long(
global long *src,
global long *dst,
int reduce_loop)
{
long val;
long result;
for(; reduce_loop > 0; reduce_loop--){
val = src[get_global_id(0)];
result = work_group_reduce_min(val);
}
dst[get_global_id(0)] = result;
}
/*
* Benchmark workgroup scan inclusive add
*/
kernel void bench_workgroup_scan_inclusive_add_int(
global int *src,
global int *dst,
int reduce_loop)
{
int val;
int result;
for(; reduce_loop > 0; reduce_loop--){
val = src[get_global_id(0)];
result = work_group_scan_inclusive_add(val);
}
dst[get_global_id(0)] = result;
}
kernel void bench_workgroup_scan_inclusive_add_long(
global long *src,
global long *dst,
int reduce_loop)
{
long val;
long result;
for(; reduce_loop > 0; reduce_loop--){
val = src[get_global_id(0)];
result = work_group_scan_inclusive_add(val);
}
dst[get_global_id(0)] = result;
}
/*
* Benchmark workgroup scan inclusive min
*/
kernel void bench_workgroup_scan_inclusive_min_int(
global int *src,
global int *dst,
int reduce_loop)
{
int val;
int result;
for(; reduce_loop > 0; reduce_loop--){
val = src[get_global_id(0)];
result = work_group_scan_inclusive_add(val);
}
dst[get_global_id(0)] = result;
}
kernel void bench_workgroup_scan_inclusive_min_long(
global long *src,
global long *dst,
int reduce_loop)
{
long val;
long result;
for(; reduce_loop > 0; reduce_loop--){
val = src[get_global_id(0)];
result = work_group_scan_inclusive_add(val);
}
dst[get_global_id(0)] = result;
}
|
