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
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
|
/* Tests of reduction on loop directive. */
/* { dg-additional-options "-Wopenacc-parallelism" } for testing/documenting
aspects of that functionality. */
#include <assert.h>
/* Test of reduction on loop directive (gangs, non-private reduction
variable). */
void g_np_1()
{
int i, arr[1024], res = 0, hres = 0;
for (i = 0; i < 1024; i++)
arr[i] = i;
#pragma acc parallel num_gangs(32) num_workers(32) vector_length(32)
/* { dg-warning "region is worker partitioned but does not contain worker partitioned code" "" { target *-*-* } .-1 } */
/* { dg-warning "region is vector partitioned but does not contain vector partitioned code" "" { target *-*-* } .-2 } */
{
#pragma acc loop gang reduction(+:res)
for (i = 0; i < 1024; i++)
res += arr[i];
}
for (i = 0; i < 1024; i++)
hres += arr[i];
assert (res == hres);
res = hres = 1;
#pragma acc parallel num_gangs(32) num_workers(32) vector_length(32)
/* { dg-warning "region is worker partitioned but does not contain worker partitioned code" "" { target *-*-* } .-1 } */
/* { dg-warning "region is vector partitioned but does not contain vector partitioned code" "" { target *-*-* } .-2 } */
{
#pragma acc loop gang reduction(*:res)
for (i = 0; i < 12; i++)
res *= arr[i];
}
for (i = 0; i < 12; i++)
hres *= arr[i];
assert (res == hres);
}
/* Test of reduction on loop directive (gangs and vectors, non-private
reduction variable). */
void gv_np_1()
{
int i, arr[1024], res = 0, hres = 0;
for (i = 0; i < 1024; i++)
arr[i] = i;
#pragma acc parallel num_gangs(32) num_workers(32) vector_length(32)
/* { dg-warning "region is worker partitioned but does not contain worker partitioned code" "" { target *-*-* } .-1 } */
{
#pragma acc loop gang vector reduction(+:res)
for (i = 0; i < 1024; i++)
res += arr[i];
}
for (i = 0; i < 1024; i++)
hres += arr[i];
assert (res == hres);
}
/* Test of reduction on loop directive (gangs and workers, non-private
reduction variable). */
void gw_np_1()
{
int i, arr[1024], res = 0, hres = 0;
for (i = 0; i < 1024; i++)
arr[i] = i;
#pragma acc parallel num_gangs(32) num_workers(32) vector_length(32)
/* { dg-warning "region is vector partitioned but does not contain vector partitioned code" "" { target *-*-* } .-1 } */
{
#pragma acc loop gang worker reduction(+:res)
for (i = 0; i < 1024; i++)
res += arr[i];
}
for (i = 0; i < 1024; i++)
hres += arr[i];
assert (res == hres);
}
/* Test of reduction on loop directive (gangs, workers and vectors, non-private
reduction variable). */
void gwv_np_1()
{
int i, arr[1024], res = 0, hres = 0;
for (i = 0; i < 1024; i++)
arr[i] = i;
#pragma acc parallel num_gangs(32) num_workers(32) vector_length(32)
{
#pragma acc loop gang worker vector reduction(+:res)
for (i = 0; i < 1024; i++)
res += arr[i];
}
for (i = 0; i < 1024; i++)
hres += arr[i];
assert (res == hres);
}
/* Test of reduction on loop directive (gangs, workers and vectors, non-private
reduction variable: separate gang and worker/vector loops). */
void gwv_np_2()
{
int i, j, arr[32768], res = 0, hres = 0;
for (i = 0; i < 32768; i++)
arr[i] = i;
#pragma acc parallel num_gangs(32) num_workers(32) vector_length(32)
{
#pragma acc loop gang reduction(+:res)
for (j = 0; j < 32; j++)
{
#pragma acc loop worker vector reduction(+:res)
for (i = 0; i < 1024; i++)
res += arr[j * 1024 + i];
}
/* "res" is non-private, and is not available until after the parallel
region. */
}
for (i = 0; i < 32768; i++)
hres += arr[i];
assert (res == hres);
}
/* Test of reduction on loop directive (gangs, workers and vectors, non-private
reduction variable: separate gang and worker/vector loops). */
void gwv_np_3()
{
int i, j;
double arr[32768], res = 0, hres = 0;
for (i = 0; i < 32768; i++)
arr[i] = i;
#pragma acc parallel num_gangs(32) num_workers(32) vector_length(32) \
copyin(arr)
{
#pragma acc loop gang reduction(+:res)
for (j = 0; j < 32; j++)
{
#pragma acc loop worker vector reduction(+:res)
for (i = 0; i < 1024; i++)
res += arr[j * 1024 + i];
}
}
for (i = 0; i < 32768; i++)
hres += arr[i];
assert (res == hres);
}
#if ACC_DEVICE_TYPE_nvidia
/* To avoid 'libgomp: The Nvidia accelerator has insufficient resources'. */
#define NUM_WORKERS 28
#else
#define NUM_WORKERS 32
#endif
/* Test of reduction on loop directive (gangs, workers and vectors, multiple
non-private reduction variables, float type). */
void gwv_np_4()
{
int i, j;
float arr[32768];
float res = 0, mres = 0, hres = 0, hmres = 0;
for (i = 0; i < 32768; i++)
arr[i] = i % (32768 / 64);
#pragma acc parallel num_gangs(32) num_workers(NUM_WORKERS) vector_length(32)
{
#pragma acc loop gang reduction(+:res) reduction(max:mres)
for (j = 0; j < 32; j++)
{
#pragma acc loop worker vector reduction(+:res) reduction(max:mres)
for (i = 0; i < 1024; i++)
{
res += arr[j * 1024 + i];
if (arr[j * 1024 + i] > mres)
mres = arr[j * 1024 + i];
}
#pragma acc loop worker vector reduction(+:res) reduction(max:mres)
for (i = 0; i < 1024; i++)
{
res += arr[j * 1024 + (1023 - i)];
if (arr[j * 1024 + (1023 - i)] > mres)
mres = arr[j * 1024 + (1023 - i)];
}
}
}
for (j = 0; j < 32; j++)
for (i = 0; i < 1024; i++)
{
hres += arr[j * 1024 + i];
hres += arr[j * 1024 + (1023 - i)];
if (arr[j * 1024 + i] > hmres)
hmres = arr[j * 1024 + i];
if (arr[j * 1024 + (1023 - i)] > hmres)
hmres = arr[j * 1024 + (1023 - i)];
}
assert (hres <= 16777216);
assert (res == hres);
assert (hmres <= 16777216);
assert (mres == hmres);
}
#undef NUM_WORKERS
/* Test of reduction on loop directive (vectors, private reduction
variable). */
void v_p_1()
{
int i, j, arr[1024], out[32], res = 0, hres = 0;
for (i = 0; i < 1024; i++)
arr[i] = i;
#pragma acc parallel num_gangs(32) num_workers(32) vector_length(32) \
private(res) copyout(out)
/* { dg-warning "region is worker partitioned but does not contain worker partitioned code" "" { target *-*-* } .-2 } */
{
#pragma acc loop gang
for (j = 0; j < 32; j++)
{
res = 0;
#pragma acc loop vector reduction(+:res)
for (i = 0; i < 32; i++)
res += arr[j * 32 + i];
out[j] = res;
}
}
for (j = 0; j < 32; j++)
{
hres = 0;
for (i = 0; i < 32; i++)
hres += arr[j * 32 + i];
assert (out[j] == hres);
}
}
/* Test of reduction on loop directive (vector reduction in
gang-partitioned/worker-partitioned mode, private reduction variable). */
void v_p_2()
{
int i, j, k;
double ina[1024], inb[1024], out[1024], acc;
for (j = 0; j < 32; j++)
for (i = 0; i < 32; i++)
{
ina[j * 32 + i] = (i == j) ? 2.0 : 0.0;
inb[j * 32 + i] = (double) (i + j);
}
#pragma acc parallel num_gangs(32) num_workers(32) vector_length(32) \
private(acc) copyin(ina, inb) copyout(out)
{
#pragma acc loop gang worker
for (k = 0; k < 32; k++)
for (j = 0; j < 32; j++)
{
acc = 0;
#pragma acc loop vector reduction(+:acc)
for (i = 0; i < 32; i++)
acc += ina[k * 32 + i] * inb[i * 32 + j];
out[k * 32 + j] = acc;
}
}
for (j = 0; j < 32; j++)
for (i = 0; i < 32; i++)
assert (out[j * 32 + i] == (i + j) * 2);
}
/* Test of reduction on loop directive (workers, private reduction
variable). */
void w_p_1()
{
int i, j, arr[1024], out[32], res = 0, hres = 0;
for (i = 0; i < 1024; i++)
arr[i] = i;
#pragma acc parallel num_gangs(32) num_workers(32) vector_length(32) \
private(res) copyout(out)
/* { dg-warning "region is vector partitioned but does not contain vector partitioned code" "" { target *-*-* } .-2 } */
{
#pragma acc loop gang
for (j = 0; j < 32; j++)
{
res = 0;
#pragma acc loop worker reduction(+:res)
for (i = 0; i < 32; i++)
res += arr[j * 32 + i];
out[j] = res;
}
}
for (j = 0; j < 32; j++)
{
hres = 0;
for (i = 0; i < 32; i++)
hres += arr[j * 32 + i];
assert (out[j] == hres);
}
}
/* Test of reduction on loop directive (workers and vectors, private reduction
variable). */
void wv_p_1()
{
int i, j, arr[1024], out[32], res = 0, hres = 0;
for (i = 0; i < 1024; i++)
arr[i] = i;
#pragma acc parallel num_gangs(32) num_workers(32) vector_length(32) \
private(res) copyout(out)
{
#pragma acc loop gang
for (j = 0; j < 32; j++)
{
res = 0;
#pragma acc loop worker vector reduction(+:res)
for (i = 0; i < 32; i++)
res += arr[j * 32 + i];
out[j] = res;
}
}
for (j = 0; j < 32; j++)
{
hres = 0;
for (i = 0; i < 32; i++)
hres += arr[j * 32 + i];
assert (out[j] == hres);
}
}
/* Test of reduction on loop directive (workers and vectors, private reduction
variable). */
void wv_p_2()
{
int i, j, arr[32768], out[32], res = 0, hres = 0;
for (i = 0; i < 32768; i++)
arr[i] = i;
#pragma acc parallel num_gangs(32) num_workers(32) vector_length(32) \
private(res) copyout(out)
{
#pragma acc loop gang
for (j = 0; j < 32; j++)
{
res = j;
#pragma acc loop worker reduction(+:res)
for (i = 0; i < 1024; i++)
res += arr[j * 1024 + i];
#pragma acc loop vector reduction(+:res)
for (i = 1023; i >= 0; i--)
res += arr[j * 1024 + i];
out[j] = res;
}
}
for (j = 0; j < 32; j++)
{
hres = j;
for (i = 0; i < 1024; i++)
hres += arr[j * 1024 + i] * 2;
assert (out[j] == hres);
}
}
/* Test of reduction on loop directive (workers and vectors, private reduction
variable: gang-redundant mode). */
void wv_p_3()
{
int i, arr[1024], out[32], res = 0, hres = 0;
for (i = 0; i < 1024; i++)
arr[i] = i ^ 33;
#pragma acc parallel num_gangs(32) num_workers(32) vector_length(32) \
private(res) copyin(arr) copyout(out)
{
/* Private variables aren't initialized by default in openacc. */
res = 0;
/* "res" should be available at the end of the following loop (and should
have the same value redundantly in each gang). */
#pragma acc loop worker vector reduction(+:res)
for (i = 0; i < 1024; i++)
res += arr[i];
#pragma acc loop gang (static: 1)
for (i = 0; i < 32; i++)
out[i] = res;
}
for (i = 0; i < 1024; i++)
hres += arr[i];
for (i = 0; i < 32; i++)
assert (out[i] == hres);
}
int main()
{
g_np_1();
gv_np_1();
gw_np_1();
gwv_np_1();
gwv_np_2();
gwv_np_3();
gwv_np_4();
v_p_1();
v_p_2();
w_p_1();
wv_p_1();
wv_p_2();
wv_p_3();
return 0;
}
|
