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
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
|
/******************************************************************************
* Copyright (c) Intel Corporation - All rights reserved. *
* This file is part of the LIBXSMM library. *
* *
* For information on the license, see the LICENSE file. *
* Further information: https://github.com/hfp/libxsmm/ *
* SPDX-License-Identifier: BSD-3-Clause *
******************************************************************************/
/* Hans Pabst (Intel Corp.)
******************************************************************************/
#include <libxsmm_source.h>
#if defined(LIBXSMM_OFFLOAD_TARGET)
# pragma offload_attribute(push,target(LIBXSMM_OFFLOAD_TARGET))
#endif
#include <algorithm>
#include <stdexcept>
#include <cstdlib>
#include <cstring>
#include <cassert>
#include <cstdio>
#include <vector>
#include <cmath>
#if defined(__MKL) || defined(MKL_DIRECT_CALL_SEQ) || defined(MKL_DIRECT_CALL)
# include <mkl_service.h>
# include <mkl.h>
#endif
#if defined(_OPENMP)
# include <omp.h>
#endif
#if defined(LIBXSMM_OFFLOAD_TARGET)
# pragma offload_attribute(pop)
#endif
#if 0 /* enable padding on a per-matrix basis */
# define PAD(TYPE, VALUE) (LIBXSMM_UP2((VALUE) * sizeof(TYPE), LIBXSMM_ALIGNMENT) / sizeof(TYPE))
#else
# define PAD(TYPE, VALUE) (VALUE)
#endif
#if !defined(RANDOMIZED) && 0
# define RANDOMIZED
#endif
#if !defined(ITYPE)
# define ITYPE double
#endif
#if !defined(OTYPE)
# define OTYPE ITYPE
#endif
#if (LIBXSMM_EQUAL(ITYPE, float) || LIBXSMM_EQUAL(ITYPE, double)) \
&& !defined(MKL_DIRECT_CALL_SEQ) && !defined(MKL_DIRECT_CALL)
LIBXSMM_BLAS_SYMBOL_DECL(ITYPE, gemm)
#endif
int main(int argc, char* argv[])
{
int result = EXIT_SUCCESS;
try {
#if defined(__BLAS) && (0 == __BLAS)
LIBXSMM_UNUSED(argc); LIBXSMM_UNUSED(argv);
throw "LAPACK/BLAS library must be available for this sample code!";
#else /* BLAS available */
const libxsmm_blasint benchmark = (1 < argc ? std::atoi(argv[1]) : 0);
LIBXSMM_BLAS_CONST libxsmm_blasint m = (2 < argc ? std::atoi(argv[2]) : 23);
LIBXSMM_BLAS_CONST libxsmm_blasint k = (4 < argc ? std::atoi(argv[4]) : m);
LIBXSMM_BLAS_CONST libxsmm_blasint n = (3 < argc ? std::atoi(argv[3]) : k);
const libxsmm_blasint q = (5 < argc ? std::atoi(argv[5]) : 0/*auto*/);
const libxsmm_blasint nrepeat = (6 < argc ? std::atoi(argv[6]) : (0 >= q ? 13 : 1));
LIBXSMM_BLAS_CONST libxsmm_blasint lda = m, ldb = k, ldc = m;
LIBXSMM_BLAS_CONST char transa = 'N', transb = 'N';
LIBXSMM_BLAS_CONST OTYPE alpha = 1, beta = 1;
const libxsmm_blasint asize = PAD(ITYPE, lda * k), bsize = PAD(ITYPE, ldb * n), csize = PAD(OTYPE, ldc * n);
const libxsmm_blasint max_size = ((2ULL << 30/*2 GB*/) / ((static_cast<size_t>(asize) + bsize) * sizeof(ITYPE) + csize * sizeof(OTYPE)));
const libxsmm_blasint s = LIBXSMM_MIN(0 < q ? q : max_size, max_size);
const libxsmm_blasint aspace = LIBXSMM_ALIGNMENT / sizeof(ITYPE);
const size_t bwsize = (static_cast<size_t>(asize)/*load*/ + static_cast<size_t>(bsize)/*load*/) * sizeof(ITYPE)
+ (sizeof(OTYPE) * static_cast<size_t>(csize) * 2/*RFO*/);
const double gflops = 2E-9 * s * m * n * k;
#if LIBXSMM_TYPEINFO(ITYPE, FP)
const char ops[] = "FLOPS";
const double scale = 1.0 / s;
#else
const char ops[] = "OPS";
const double scale = 1;
#endif
#if defined(_OPENMP) && !defined(_DEBUG)
const char *const env_check = getenv("CHECK");
const int check = (NULL == env_check ? 0 : atoi(env_check));
#elif (defined(__MKL) || defined(MKL_DIRECT_CALL_SEQ) || defined(MKL_DIRECT_CALL)) && (LIBXSMM_VERSION2(11, 3) <= INTEL_MKL_VERSION)
/*const*/ int check = 1;
#endif
#if defined(LIBXSMM_OFFLOAD_TARGET)
# pragma offload target(LIBXSMM_OFFLOAD_TARGET)
#endif
{
#if defined(_OPENMP)
const libxsmm_blasint chunksize = s / omp_get_max_threads();
#endif
struct raii { // avoid std::vector (first-touch init. causes NUMA issue)
ITYPE *a, *b;
OTYPE *c, *d;
size_t m_size, m_shuffle;
raii(libxsmm_blasint asize_, libxsmm_blasint bsize_, libxsmm_blasint csize_, libxsmm_blasint size_)
: a(new ITYPE[static_cast<size_t>(asize_)]), b(new ITYPE[static_cast<size_t>(bsize_)])
, c(new OTYPE[static_cast<size_t>(csize_)]), d(new OTYPE[static_cast<size_t>(csize_)])
, m_size(static_cast<size_t>(size_)), m_shuffle(libxsmm_shuffle(static_cast<unsigned int>(size_)))
{}
~raii() { delete[] a; delete[] b; delete[] c; delete[] d; }
#if defined(RANDOMIZED)
libxsmm_blasint shuffle(libxsmm_blasint i) const { return (i * m_shuffle) % m_size; }
#else
libxsmm_blasint shuffle(libxsmm_blasint i) const { return i; }
#endif
} helper(s * asize + aspace - 1, s * bsize + aspace - 1, s * csize + aspace - 1, s);
ITYPE *const a = LIBXSMM_ALIGN(helper.a, LIBXSMM_ALIGNMENT);
ITYPE *const b = LIBXSMM_ALIGN(helper.b, LIBXSMM_ALIGNMENT);
OTYPE *const c = LIBXSMM_ALIGN(helper.c, LIBXSMM_ALIGNMENT);
OTYPE *const d = LIBXSMM_ALIGN(helper.d, LIBXSMM_ALIGNMENT);
#if defined(_OPENMP)
const int nthreads = omp_get_max_threads();
# pragma omp parallel for num_threads(nthreads) schedule(static)
#endif
for (libxsmm_blasint i = 0; i < s; ++i) {
LIBXSMM_MATINIT(ITYPE, 42 + helper.shuffle(i), a + static_cast<size_t>(asize) * helper.shuffle(i), m, k, lda, scale);
LIBXSMM_MATINIT(ITYPE, 24 + helper.shuffle(i), b + static_cast<size_t>(bsize) * helper.shuffle(i), k, n, ldb, scale);
LIBXSMM_MATINIT(OTYPE, 22 + i, c + static_cast<size_t>(csize) * i, m, n, ldc, scale);
LIBXSMM_MATINIT(OTYPE, 22 + i, d + static_cast<size_t>(csize) * i, m, n, ldc, scale);
}
#if defined(MKL_ENABLE_AVX512)
mkl_enable_instructions(MKL_ENABLE_AVX512);
#endif
// initialize LIBXSMM
libxsmm_init();
fprintf(stdout, "m=%lli n=%lli k=%lli size=%lli memory=%.1f MB (input=%s output=%s)\n\n",
static_cast<long long>(m), static_cast<long long>(n), static_cast<long long>(k), static_cast<long long>(s),
1.0 * (s * ((static_cast<size_t>(asize) + bsize) * sizeof(ITYPE) + csize * sizeof(OTYPE))) / (1ULL << 20),
LIBXSMM_TYPENAME(ITYPE), LIBXSMM_TYPENAME(OTYPE));
// LAPACK/BLAS3 (warm-up BLAS Library)
#if defined(_OPENMP)
# pragma omp parallel for schedule(static)
#endif
for (libxsmm_blasint i = 0; i < s; ++i) {
LIBXSMM_GEMM_SYMBOL(ITYPE)(&transa, &transb, &m, &n, &k,
&alpha, a + static_cast<size_t>(asize) * helper.shuffle(i), &lda, b + static_cast<size_t>(bsize) * helper.shuffle(i), &ldb,
&beta, c + static_cast<size_t>(csize) * i, &ldc);
}
#if (defined(__MKL) || defined(MKL_DIRECT_CALL_SEQ) || defined(MKL_DIRECT_CALL)) && (LIBXSMM_VERSION2(11, 3) <= INTEL_MKL_VERSION)
std::vector<const ITYPE*> va_array(static_cast<size_t>(s)), vb_array(static_cast<size_t>(s));
std::vector<OTYPE*> vc_array(static_cast<size_t>(s));
const ITYPE* *const a_array = &va_array[0];
const ITYPE* *const b_array = &vb_array[0];
OTYPE* *const c_array = &vc_array[0];
const libxsmm_blasint group_count = 1;
for (libxsmm_blasint i = 0; i < s; ++i) { // setup batched (A,B,C)
a_array[i] = a + static_cast<size_t>(asize) * helper.shuffle(i);
b_array[i] = b + static_cast<size_t>(bsize) * helper.shuffle(i);
c_array[i] = d + static_cast<size_t>(csize) * i;
}
// additional warm-up (also to eventually match the Gold result)
LIBXSMM_TPREFIX(ITYPE,gemm_batch)(&transa, &transb, &m, &n, &k,
&alpha, &a_array[0], &lda, &b_array[0], &ldb,
&beta, &c_array[0], &ldc, &group_count, &s);
#endif
switch (benchmark) {
case 0: { // batched
fprintf(stdout, "Batched (A,B,C)...\n");
const unsigned long long start = libxsmm_timer_tick();
for (libxsmm_blasint r = 0; r < nrepeat; ++r) {
#if defined(_OPENMP)
# pragma omp parallel for num_threads(nthreads) schedule(static)
#endif
for (libxsmm_blasint i = 0; i < s; ++i) {
LIBXSMM_GEMM_SYMBOL(ITYPE)(&transa, &transb, &m, &n, &k,
&alpha, a + static_cast<size_t>(asize) * helper.shuffle(i), &lda, b + static_cast<size_t>(bsize) * helper.shuffle(i), &ldb,
&beta, c + static_cast<size_t>(csize) * i, &ldc);
}
}
const unsigned long long ncycles = libxsmm_timer_ncycles(start, libxsmm_timer_tick());
const double duration = libxsmm_timer_duration(0, ncycles) / nrepeat;
if (0 < duration && 0 != ncycles) {
fprintf(stdout, "\tpseudo-perf.: %.1f %s/cycle\n", (2.0 * k - 1.0) * (static_cast<double>(s) * m * n) / ncycles, ops);
fprintf(stdout, "\tperformance: %.1f G%s/s\n", gflops / duration, ops);
fprintf(stdout, "\tbandwidth: %.1f GB/s\n", s * bwsize / (duration * (1ULL << 30)));
}
fprintf(stdout, "\tduration: %.0f ms\n", 1000.0 * duration);
} /* fallthrough */
#if (defined(__MKL) || defined(MKL_DIRECT_CALL_SEQ) || defined(MKL_DIRECT_CALL)) && (LIBXSMM_VERSION2(11, 3) <= INTEL_MKL_VERSION)
case 1: { // batched indirect
fprintf(stdout, "Indirect (A,B,C)...\n");
const unsigned long long start = libxsmm_timer_tick();
for (libxsmm_blasint r = 0; r < nrepeat; ++r) {
LIBXSMM_TPREFIX(ITYPE,gemm_batch)(&transa, &transb, &m, &n, &k,
&alpha, &a_array[0], &lda, &b_array[0], &ldb,
&beta, &c_array[0], &ldc, &group_count, &s);
}
const unsigned long long ncycles = libxsmm_timer_ncycles(start, libxsmm_timer_tick());
const double duration = libxsmm_timer_duration(0, ncycles) / nrepeat;
if (0 < duration && 0 != ncycles) {
fprintf(stdout, "\tpseudo-perf.: %.1f %s/cycle\n", (2.0 * k - 1.0) * (static_cast<double>(s) * m * n) / ncycles, ops);
fprintf(stdout, "\tperformance: %.1f G%s/s\n", gflops / duration, ops);
fprintf(stdout, "\tbandwidth: %.1f GB/s\n", s * bwsize / (duration * (1ULL << 30)));
}
fprintf(stdout, "\tduration: %.0f ms\n", 1000.0 * duration);
if (0 == (benchmark & 1) && 0 != check) { /* Gold result is available */
libxsmm_matdiff_info diff;
libxsmm_matdiff_clear(&diff);
for (libxsmm_blasint h = 0; h < s; ++h) {
const OTYPE *const u = c + static_cast<size_t>(csize) * h, *const v = c_array[h];
libxsmm_matdiff_info dv;
result = libxsmm_matdiff(&dv, LIBXSMM_DATATYPE(OTYPE), m, n, u, v, &ldc, &ldc);
if (EXIT_SUCCESS == result) libxsmm_matdiff_reduce(&diff, &dv);
}
fprintf(stdout, "\tdiff: L2abs=%f Linf=%f\n", diff.l2_abs, diff.linf_abs);
if (check < diff.l2_rel) {
fprintf(stderr, "FAILED.\n");
result = EXIT_FAILURE;
}
}
}
#endif
break;
case 2: { // streaming A and C
fprintf(stdout, "Streamed (A,C)...\n");
const unsigned long long start = libxsmm_timer_tick();
for (libxsmm_blasint r = 0; r < nrepeat; ++r) {
#if defined(_OPENMP)
# pragma omp parallel for num_threads(nthreads) schedule(static)
#endif
for (libxsmm_blasint i = 0; i < s; ++i) {
LIBXSMM_GEMM_SYMBOL(ITYPE)(&transa, &transb, &m, &n, &k,
&alpha, a + static_cast<size_t>(asize) * helper.shuffle(i), &lda, b, &ldb,
&beta, c + static_cast<size_t>(csize) * i, &ldc);
}
}
const unsigned long long ncycles = libxsmm_timer_ncycles(start, libxsmm_timer_tick());
const double duration = libxsmm_timer_duration(0, ncycles) / nrepeat;
if (0 < duration && 0 != ncycles) {
fprintf(stdout, "\tpseudo-perf.: %.1f %s/cycle\n", (2.0 * k - 1.0) * (static_cast<double>(s) * m * n) / ncycles, ops);
fprintf(stdout, "\tperformance: %.1f G%s/s\n", gflops / duration, ops);
fprintf(stdout, "\tbandwidth: %.1f GB/s\n", s * (bwsize - bsize * sizeof(ITYPE)) / (duration * (1ULL << 30)));
}
fprintf(stdout, "\tduration: %.0f ms\n", 1000.0 * duration);
} /* fallthrough */
#if (defined(__MKL) || defined(MKL_DIRECT_CALL_SEQ) || defined(MKL_DIRECT_CALL)) && (LIBXSMM_VERSION2(11, 3) <= INTEL_MKL_VERSION)
case 3: { // indirect A and C
fprintf(stdout, "Indirect (A,C)...\n");
for (libxsmm_blasint i = 0; i < s; ++i) {
a_array[i] = a + static_cast<size_t>(asize) * helper.shuffle(i);
b_array[i] = b;
c_array[i] = d + static_cast<size_t>(csize) * i;
}
const unsigned long long start = libxsmm_timer_tick();
for (libxsmm_blasint r = 0; r < nrepeat; ++r) {
LIBXSMM_TPREFIX(ITYPE, gemm_batch)(&transa, &transb, &m, &n, &k,
&alpha, &a_array[0], &lda, &b_array[0], &ldb,
&beta, &c_array[0], &ldc, &group_count, &s);
}
const unsigned long long ncycles = libxsmm_timer_ncycles(start, libxsmm_timer_tick());
const double duration = libxsmm_timer_duration(0, ncycles) / nrepeat;
if (0 < duration && 0 != ncycles) {
fprintf(stdout, "\tpseudo-perf.: %.1f %s/cycle\n", (2.0 * k - 1.0) * (static_cast<double>(s) * m * n) / ncycles, ops);
fprintf(stdout, "\tperformance: %.1f G%s/s\n", gflops / duration, ops);
fprintf(stdout, "\tbandwidth: %.1f GB/s\n", s * (bwsize - bsize * sizeof(ITYPE)) / (duration * (1ULL << 30)));
}
fprintf(stdout, "\tduration: %.0f ms\n", 1000.0 * duration);
if (0 == (benchmark & 1) && 0 != check) { /* Gold result is available */
libxsmm_matdiff_info diff;
libxsmm_matdiff_clear(&diff);
for (libxsmm_blasint h = 0; h < s; ++h) {
const OTYPE *const u = c + static_cast<size_t>(csize) * h, *const v = c_array[h];
libxsmm_matdiff_info dv;
result = libxsmm_matdiff(&dv, LIBXSMM_DATATYPE(OTYPE), m, n, u, v, &ldc, &ldc);
if (EXIT_SUCCESS == result) libxsmm_matdiff_reduce(&diff, &dv);
}
fprintf(stdout, "\tdiff: L2abs=%f Linf=%f\n", diff.l2_abs, diff.linf_abs);
if (check < diff.l2_rel) {
fprintf(stderr, "FAILED.\n");
result = EXIT_FAILURE;
}
}
}
#endif
break;
case 4: { // streaming B and C
fprintf(stdout, "Streamed (B,C)...\n");
const unsigned long long start = libxsmm_timer_tick();
for (libxsmm_blasint r = 0; r < nrepeat; ++r) {
#if defined(_OPENMP)
# pragma omp parallel for num_threads(nthreads) schedule(static)
#endif
for (libxsmm_blasint i = 0; i < s; ++i) {
LIBXSMM_GEMM_SYMBOL(ITYPE)(&transa, &transb, &m, &n, &k,
&alpha, a, &lda, b + static_cast<size_t>(bsize) * helper.shuffle(i), &ldb,
&beta, c + static_cast<size_t>(csize) * i, &ldc);
}
}
const unsigned long long ncycles = libxsmm_timer_ncycles(start, libxsmm_timer_tick());
const double duration = libxsmm_timer_duration(0, ncycles) / nrepeat;
if (0 < duration && 0 != ncycles) {
fprintf(stdout, "\tpseudo-perf.: %.1f %s/cycle\n", (2.0 * k - 1.0) * (static_cast<double>(s) * m * n) / ncycles, ops);
fprintf(stdout, "\tperformance: %.1f G%s/s\n", gflops / duration, ops);
fprintf(stdout, "\tbandwidth: %.1f GB/s\n", s * (bwsize - asize * sizeof(ITYPE)) / (duration * (1ULL << 30)));
}
fprintf(stdout, "\tduration: %.0f ms\n", 1000.0 * duration);
} /* fallthrough */
#if (defined(__MKL) || defined(MKL_DIRECT_CALL_SEQ) || defined(MKL_DIRECT_CALL)) && (LIBXSMM_VERSION2(11, 3) <= INTEL_MKL_VERSION)
case 5: { // indirect B and C
fprintf(stdout, "Indirect (B,C)...\n");
for (libxsmm_blasint i = 0; i < s; ++i) {
a_array[i] = a;
b_array[i] = b + static_cast<size_t>(bsize) * helper.shuffle(i);
c_array[i] = d + static_cast<size_t>(csize) * i;
}
const unsigned long long start = libxsmm_timer_tick();
for (libxsmm_blasint r = 0; r < nrepeat; ++r) {
LIBXSMM_TPREFIX(ITYPE, gemm_batch)(&transa, &transb, &m, &n, &k,
&alpha, &a_array[0], &lda, &b_array[0], &ldb,
&beta, &c_array[0], &ldc, &group_count, &s);
}
const unsigned long long ncycles = libxsmm_timer_ncycles(start, libxsmm_timer_tick());
const double duration = libxsmm_timer_duration(0, ncycles) / nrepeat;
if (0 < duration && 0 != ncycles) {
fprintf(stdout, "\tpseudo-perf.: %.1f %s/cycle\n", (2.0 * k - 1.0) * (static_cast<double>(s) * m * n) / ncycles, ops);
fprintf(stdout, "\tperformance: %.1f G%s/s\n", gflops / duration, ops);
fprintf(stdout, "\tbandwidth: %.1f GB/s\n", s * (bwsize - asize * sizeof(ITYPE)) / (duration * (1ULL << 30)));
}
fprintf(stdout, "\tduration: %.0f ms\n", 1000.0 * duration);
if (0 == (benchmark & 1) && 0 != check) { /* Gold result is available */
libxsmm_matdiff_info diff;
libxsmm_matdiff_clear(&diff);
for (libxsmm_blasint h = 0; h < s; ++h) {
const OTYPE *const u = c + static_cast<size_t>(csize) * h, *const v = c_array[h];
libxsmm_matdiff_info dv;
result = libxsmm_matdiff(&dv, LIBXSMM_DATATYPE(OTYPE), m, n, u, v, &ldc, &ldc);
if (EXIT_SUCCESS == result) libxsmm_matdiff_reduce(&diff, &dv);
}
fprintf(stdout, "\tdiff: L2abs=%f Linf=%f\n", diff.l2_abs, diff.linf_abs);
if (check < diff.l2_rel) {
fprintf(stderr, "FAILED.\n");
result = EXIT_FAILURE;
}
}
}
#endif
break;
case 6: { // streaming A and B
fprintf(stdout, "Streamed (A,B)...\n");
const unsigned long long start = libxsmm_timer_tick();
for (libxsmm_blasint r = 0; r < nrepeat; ++r) {
#if defined(_OPENMP)
# pragma omp parallel for num_threads(0 == check ? nthreads : 1) schedule(static)
#endif
for (libxsmm_blasint i = 0; i < s; ++i) {
libxsmm_blasint j = 0;
#if defined(_OPENMP) /* attempt to write to disjunct cachelines */
if (0 == check) j = omp_get_thread_num() * chunksize * csize;
#endif
LIBXSMM_GEMM_SYMBOL(ITYPE)(&transa, &transb, &m, &n, &k,
&alpha, a + static_cast<size_t>(asize) * helper.shuffle(i), &lda, b + static_cast<size_t>(bsize) * helper.shuffle(i), &ldb,
&beta, c + j, &ldc);
}
}
const unsigned long long ncycles = libxsmm_timer_ncycles(start, libxsmm_timer_tick());
const double duration = libxsmm_timer_duration(0, ncycles) / nrepeat;
if (0 < duration && 0 != ncycles) {
fprintf(stdout, "\tpseudo-perf.: %.1f %s/cycle\n", (2.0 * k - 1.0) * (static_cast<double>(s) * m * n) / ncycles, ops);
fprintf(stdout, "\tperformance: %.1f G%s/s\n", gflops / duration, ops);
fprintf(stdout, "\tbandwidth: %.1f GB/s\n", s * (bwsize - sizeof(OTYPE) * csize * 2) / (duration * (1ULL << 30)));
}
fprintf(stdout, "\tduration: %.0f ms\n", 1000.0 * duration);
} /* fallthrough */
#if (defined(__MKL) || defined(MKL_DIRECT_CALL_SEQ) || defined(MKL_DIRECT_CALL)) && (LIBXSMM_VERSION2(11, 3) <= INTEL_MKL_VERSION)
case 7: { // indirect A and B
fprintf(stdout, "Indirect (A,B)...\n");
#if defined(_OPENMP)
# pragma omp parallel for num_threads(0 == check ? nthreads : 1) schedule(static)
#endif
for (libxsmm_blasint i = 0; i < s; ++i) {
a_array[i] = a + static_cast<size_t>(asize) * helper.shuffle(i);
b_array[i] = b + static_cast<size_t>(bsize) * helper.shuffle(i);
#if defined(_OPENMP) /* attempt to write to disjunct cachelines */
if (0 == check) {
c_array[i] = d + static_cast<size_t>(csize) * chunksize * omp_get_thread_num();
}
else
#endif
c_array[i] = d;
}
#if defined(_OPENMP)
omp_set_num_threads(0 == check ? nthreads : 1);
#endif
const unsigned long long start = libxsmm_timer_tick();
for (libxsmm_blasint r = 0; r < nrepeat; ++r) {
LIBXSMM_TPREFIX(ITYPE, gemm_batch)(&transa, &transb, &m, &n, &k,
&alpha, &a_array[0], &lda, &b_array[0], &ldb,
&beta, &c_array[0], &ldc, &group_count, &s);
}
const unsigned long long ncycles = libxsmm_timer_ncycles(start, libxsmm_timer_tick());
const double duration = libxsmm_timer_duration(0, ncycles) / nrepeat;
if (0 < duration && 0 != ncycles) {
fprintf(stdout, "\tpseudo-perf.: %.1f %s/cycle\n", (2.0 * k - 1.0) * (static_cast<double>(s) * m * n) / ncycles, ops);
fprintf(stdout, "\tperformance: %.1f G%s/s\n", gflops / duration, ops);
fprintf(stdout, "\tbandwidth: %.1f GB/s\n", s * (bwsize - sizeof(OTYPE) * csize * 2) / (duration * (1ULL << 30)));
}
fprintf(stdout, "\tduration: %.0f ms\n", 1000.0 * duration);
if (0 == (benchmark & 1) && 0 != check) { /* Gold result is available */
libxsmm_matdiff_info diff;
result = libxsmm_matdiff(&diff, LIBXSMM_DATATYPE(OTYPE), m, n, c, d, &ldc, &ldc);
fprintf(stdout, "\tdiff: L2abs=%f Linf=%f\n", diff.l2_abs, diff.linf_abs);
if (check < diff.l2_rel) {
fprintf(stderr, "FAILED.\n");
result = EXIT_FAILURE;
}
}
}
#endif
break;
case 8: { // cached
fprintf(stdout, "Cached...\n");
const unsigned long long start = libxsmm_timer_tick();
for (libxsmm_blasint r = 0; r < nrepeat; ++r) {
#if defined(_OPENMP)
# pragma omp parallel for num_threads(0 == check ? nthreads : 1) schedule(static)
#endif
for (libxsmm_blasint i = 0; i < s; ++i) {
libxsmm_blasint j = 0;
#if defined(_OPENMP) /* attempt to write to disjunct cachelines */
if (0 == check) j = omp_get_thread_num() * chunksize * csize;
#endif
LIBXSMM_GEMM_SYMBOL(ITYPE)(&transa, &transb, &m, &n, &k,
&alpha, a, &lda, b, &ldb, &beta, c + j, &ldc);
}
}
const unsigned long long ncycles = libxsmm_timer_ncycles(start, libxsmm_timer_tick());
const double duration = libxsmm_timer_duration(0, ncycles) / nrepeat;
if (0 < duration && 0 != ncycles) {
fprintf(stdout, "\tpseudo-perf.: %.1f %s/cycle\n", (2.0 * k - 1.0) * (static_cast<double>(s) * m * n) / ncycles, ops);
fprintf(stdout, "\tperformance: %.1f G%s/s\n", gflops / duration, ops);
}
fprintf(stdout, "\tduration: %.0f ms\n", 1000.0 * duration);
} /* fallthrough */
#if (defined(__MKL) || defined(MKL_DIRECT_CALL_SEQ) || defined(MKL_DIRECT_CALL)) && (LIBXSMM_VERSION2(11, 3) <= INTEL_MKL_VERSION)
case 9: { // indirect cached
fprintf(stdout, "Indirect cached...\n");
#if defined(_OPENMP)
# pragma omp parallel for num_threads(0 == check ? nthreads : 1) schedule(static)
#endif
for (libxsmm_blasint i = 0; i < s; ++i) {
a_array[i] = a; b_array[i] = b;
#if defined(_OPENMP) /* attempt to write to disjunct cachelines */
if (0 == check) {
c_array[i] = d + static_cast<size_t>(csize) * chunksize * omp_get_thread_num();
}
else
#endif
c_array[i] = d;
}
#if defined(_OPENMP)
omp_set_num_threads(0 == check ? nthreads : 1);
#endif
const unsigned long long start = libxsmm_timer_tick();
for (libxsmm_blasint r = 0; r < nrepeat; ++r) {
LIBXSMM_TPREFIX(ITYPE, gemm_batch)(&transa, &transb, &m, &n, &k,
&alpha, &a_array[0], &lda, &b_array[0], &ldb,
&beta, &c_array[0], &ldc, &group_count, &s);
}
const unsigned long long ncycles = libxsmm_timer_ncycles(start, libxsmm_timer_tick());
const double duration = libxsmm_timer_duration(0, ncycles) / nrepeat;
if (0 < duration && 0 != ncycles) {
fprintf(stdout, "\tpseudo-perf.: %.1f %s/cycle\n", (2.0 * k - 1.0) * (static_cast<double>(s) * m * n) / ncycles, ops);
fprintf(stdout, "\tperformance: %.1f G%s/s\n", gflops / duration, ops);
}
fprintf(stdout, "\tduration: %.0f ms\n", 1000.0 * duration);
if (0 == (benchmark & 1) && 0 != check) { /* Gold result is available */
libxsmm_matdiff_info diff;
result = libxsmm_matdiff(&diff, LIBXSMM_DATATYPE(OTYPE), m, n, c, d, &ldc, &ldc);
fprintf(stdout, "\tdiff: L2abs=%f Linf=%f\n", diff.l2_abs, diff.linf_abs);
if (check < diff.l2_rel) {
fprintf(stderr, "FAILED.\n");
result = EXIT_FAILURE;
}
}
}
#endif
break;
default: throw "invalid case selected!";
} /*switch*/
// finalize LIBXSMM
libxsmm_finalize();
fprintf(stdout, "Finished\n");
}
#endif
}
catch(const std::exception& e) {
fprintf(stderr, "Error: %s\n", e.what());
result = EXIT_FAILURE;
}
catch(const char* message) {
fprintf(stderr, "Error: %s\n", message);
result = EXIT_FAILURE;
}
catch(...) {
fprintf(stderr, "Error: unknown exception caught!\n");
result = EXIT_FAILURE;
}
return result;
}
|
