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
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
|
/*
-- MAGMA (version 2.9.0) --
Univ. of Tennessee, Knoxville
Univ. of California, Berkeley
Univ. of Colorado, Denver
@date January 2025
@author Mark Gates
*/
#include "magma_internal.h"
#include "error.h"
#include <cuda_runtime.h>
#if defined(MAGMA_HAVE_CUDA) || defined(MAGMA_HAVE_HIP)
// Generic, type-independent routines to copy data.
// Type-safe versions which avoid the user needing sizeof(...) are in headers;
// see magma_{s,d,c,z,i,index_}{set,get,copy}{matrix,vector}
/***************************************************************************//**
@fn magma_setvector( n, elemSize, hx_src, incx, dy_dst, incy, queue )
Copy vector hx_src on CPU host to dy_dst on GPU device.
Elements may be arbitrary size.
Type-safe versions set elemSize appropriately.
This version synchronizes the queue after the transfer.
See magma_setvector_async() for an asynchronous version.
@param[in]
n Number of elements in vector.
@param[in]
elemSize Size of each element, e.g., sizeof(double).
@param[in]
hx_src Source array of dimension (1 + (n-1))*incx, on CPU host.
@param[in]
incx Increment between elements of hx_src. incx > 0.
@param[out]
dy_dst Destination array of dimension (1 + (n-1))*incy, on GPU device.
@param[in]
incy Increment between elements of dy_dst. incy > 0.
@param[in]
queue Queue to execute in.
@ingroup magma_setvector
*******************************************************************************/
extern "C" void
magma_setvector_internal(
magma_int_t n, magma_int_t elemSize,
void const* hx_src, magma_int_t incx,
magma_ptr dy_dst, magma_int_t incy,
magma_queue_t queue,
const char* func, const char* file, int line )
{
cudaStream_t stream = NULL;
if ( queue != NULL ) {
stream = queue->cuda_stream();
}
cublasStatus_t status;
status = cublasSetVectorAsync(
int(n), int(elemSize),
hx_src, int(incx),
dy_dst, int(incy), stream );
if ( queue != NULL )
cudaStreamSynchronize( stream );
check_xerror( status, func, file, line );
MAGMA_UNUSED( status );
}
/***************************************************************************//**
@fn magma_setvector_async( n, elemSize, hx_src, incx, dy_dst, incy, queue )
Copy vector hx_src on CPU host to dy_dst on GPU device.
Elements may be arbitrary size.
Type-safe versions set elemSize appropriately.
This version is asynchronous: it may return before the transfer finishes,
if hx_src is pinned CPU memory.
See magma_setvector() for a synchronous version.
@param[in]
n Number of elements in vector.
@param[in]
elemSize Size of each element, e.g., sizeof(double).
@param[in]
hx_src Source array of dimension (1 + (n-1))*incx, on CPU host.
@param[in]
incx Increment between elements of hx_src. incx > 0.
@param[out]
dy_dst Destination array of dimension (1 + (n-1))*incy, on GPU device.
@param[in]
incy Increment between elements of dy_dst. incy > 0.
@param[in]
queue Queue to execute in.
@ingroup magma_setvector
*******************************************************************************/
extern "C" void
magma_setvector_async_internal(
magma_int_t n, magma_int_t elemSize,
void const* hx_src, magma_int_t incx,
magma_ptr dy_dst, magma_int_t incy,
magma_queue_t queue,
const char* func, const char* file, int line )
{
// for backwards compatability, accepts NULL queue to mean NULL stream.
cudaStream_t stream = NULL;
if ( queue != NULL ) {
stream = queue->cuda_stream();
}
else {
fprintf( stderr, "Warning: %s got NULL queue\n", __func__ );
}
cublasStatus_t status;
status = cublasSetVectorAsync(
int(n), int(elemSize),
hx_src, int(incx),
dy_dst, int(incy), stream );
check_xerror( status, func, file, line );
MAGMA_UNUSED( status );
}
/***************************************************************************//**
@fn magma_getvector( n, elemSize, dx_src, incx, hy_dst, incy, queue )
Copy vector dx_src on GPU device to hy_dst on CPU host.
Elements may be arbitrary size.
Type-safe versions set elemSize appropriately.
This version synchronizes the queue after the transfer.
See magma_getvector_async() for an asynchronous version.
@param[in]
n Number of elements in vector.
@param[in]
elemSize Size of each element, e.g., sizeof(double).
@param[in]
dx_src Source array of dimension (1 + (n-1))*incx, on GPU device.
@param[in]
incx Increment between elements of hx_src. incx > 0.
@param[out]
hy_dst Destination array of dimension (1 + (n-1))*incy, on CPU host.
@param[in]
incy Increment between elements of dy_dst. incy > 0.
@param[in]
queue Queue to execute in.
@ingroup magma_getvector
*******************************************************************************/
extern "C" void
magma_getvector_internal(
magma_int_t n, magma_int_t elemSize,
magma_const_ptr dx_src, magma_int_t incx,
void* hy_dst, magma_int_t incy,
magma_queue_t queue,
const char* func, const char* file, int line )
{
cudaStream_t stream = NULL;
if ( queue != NULL ) {
stream = queue->cuda_stream();
}
cublasStatus_t status;
status = cublasGetVectorAsync(
int(n), int(elemSize),
dx_src, int(incx),
hy_dst, int(incy), stream );
if ( queue != NULL )
cudaStreamSynchronize( stream );
check_xerror( status, func, file, line );
MAGMA_UNUSED( status );
}
/***************************************************************************//**
@fn magma_getvector_async( n, elemSize, dx_src, incx, hy_dst, incy, queue )
Copy vector dx_src on GPU device to hy_dst on CPU host.
Elements may be arbitrary size.
Type-safe versions set elemSize appropriately.
This version is asynchronous: it may return before the transfer finishes,
if hy_dst is pinned CPU memory.
See magma_getvector() for a synchronous version.
@param[in]
n Number of elements in vector.
@param[in]
elemSize Size of each element, e.g., sizeof(double).
@param[in]
dx_src Source array of dimension (1 + (n-1))*incx, on GPU device.
@param[in]
incx Increment between elements of hx_src. incx > 0.
@param[out]
hy_dst Destination array of dimension (1 + (n-1))*incy, on CPU host.
@param[in]
incy Increment between elements of dy_dst. incy > 0.
@param[in]
queue Queue to execute in.
@ingroup magma_getvector
*******************************************************************************/
extern "C" void
magma_getvector_async_internal(
magma_int_t n, magma_int_t elemSize,
magma_const_ptr dx_src, magma_int_t incx,
void* hy_dst, magma_int_t incy,
magma_queue_t queue,
const char* func, const char* file, int line )
{
// for backwards compatability, accepts NULL queue to mean NULL stream.
cudaStream_t stream = NULL;
if ( queue != NULL ) {
stream = queue->cuda_stream();
}
else {
fprintf( stderr, "Warning: %s got NULL queue\n", __func__ );
}
cublasStatus_t status;
status = cublasGetVectorAsync(
int(n), int(elemSize),
dx_src, int(incx),
hy_dst, int(incy), stream );
check_xerror( status, func, file, line );
MAGMA_UNUSED( status );
}
/***************************************************************************//**
@fn magma_copyvector( n, elemSize, dx_src, incx, dy_dst, incy, queue )
Copy vector dx_src on GPU device to dy_dst on GPU device.
Elements may be arbitrary size.
Type-safe versions set elemSize appropriately.
With CUDA unified addressing, dx and dy can be on different GPUs.
This version synchronizes the queue after the transfer.
See magma_copyvector_async() for an asynchronous version.
@param[in]
n Number of elements in vector.
@param[in]
elemSize Size of each element, e.g., sizeof(double).
@param[in]
dx_src Source array of dimension (1 + (n-1))*incx, on GPU device.
@param[in]
incx Increment between elements of hx_src. incx > 0.
@param[out]
dy_dst Destination array of dimension (1 + (n-1))*incy, on GPU device.
@param[in]
incy Increment between elements of dy_dst. incy > 0.
@param[in]
queue Queue to execute in.
@ingroup magma_copyvector
*******************************************************************************/
// TODO compare performance with cublasZcopy BLAS function.
// But this implementation can handle any element size, not just [sdcz] precisions.
extern "C" void
magma_copyvector_internal(
magma_int_t n, magma_int_t elemSize,
magma_const_ptr dx_src, magma_int_t incx,
magma_ptr dy_dst, magma_int_t incy,
magma_queue_t queue,
const char* func, const char* file, int line )
{
cudaStream_t stream = NULL;
if ( queue != NULL ) {
stream = queue->cuda_stream();
}
if ( incx == 1 && incy == 1 ) {
cudaError_t status;
status = cudaMemcpyAsync(
dy_dst,
dx_src,
int(n*elemSize), cudaMemcpyDeviceToDevice, stream );
if ( queue != NULL )
cudaStreamSynchronize( stream );
check_xerror( status, func, file, line );
MAGMA_UNUSED( status );
}
else {
magma_copymatrix_internal(
1, n, elemSize, dx_src, incx, dy_dst, incy, queue, func, file, line );
}
}
/***************************************************************************//**
@fn magma_copyvector_async( n, elemSize, dx_src, incx, dy_dst, incy, queue )
Copy vector dx_src on GPU device to dy_dst on GPU device.
Elements may be arbitrary size.
Type-safe versions set elemSize appropriately.
With CUDA unified addressing, dx and dy can be on different GPUs.
This version is asynchronous: it may return before the transfer finishes.
See magma_copyvector() for a synchronous version.
@param[in]
n Number of elements in vector.
@param[in]
elemSize Size of each element, e.g., sizeof(double).
@param[in]
dx_src Source array of dimension (1 + (n-1))*incx, on GPU device.
@param[in]
incx Increment between elements of hx_src. incx > 0.
@param[out]
dy_dst Destination array of dimension (1 + (n-1))*incy, on GPU device.
@param[in]
incy Increment between elements of dy_dst. incy > 0.
@param[in]
queue Queue to execute in.
@ingroup magma_copyvector
*******************************************************************************/
extern "C" void
magma_copyvector_async_internal(
magma_int_t n, magma_int_t elemSize,
magma_const_ptr dx_src, magma_int_t incx,
magma_ptr dy_dst, magma_int_t incy,
magma_queue_t queue,
const char* func, const char* file, int line )
{
// for backwards compatability, accepts NULL queue to mean NULL stream.
cudaStream_t stream = NULL;
if ( queue != NULL ) {
stream = queue->cuda_stream();
}
else {
fprintf( stderr, "Warning: %s got NULL queue\n", __func__ );
}
if ( incx == 1 && incy == 1 ) {
cudaError_t status;
status = cudaMemcpyAsync(
dy_dst,
dx_src,
int(n*elemSize), cudaMemcpyDeviceToDevice, stream );
check_xerror( status, func, file, line );
MAGMA_UNUSED( status );
}
else {
magma_copymatrix_async_internal(
1, n, elemSize, dx_src, incx, dy_dst, incy, queue, func, file, line );
}
}
/***************************************************************************//**
@fn magma_setmatrix( m, n, elemSize, hA_src, lda, dB_dst, lddb, queue )
Copy all or part of matrix hA_src on CPU host to dB_dst on GPU device.
Elements may be arbitrary size.
Type-safe versions set elemSize appropriately.
This version synchronizes the queue after the transfer.
See magma_setmatrix_async() for an asynchronous version.
@param[in]
m Number of rows of matrix A. m >= 0.
@param[in]
n Number of columns of matrix A. n >= 0.
@param[in]
elemSize Size of each element, e.g., sizeof(double).
@param[in]
hA_src Source array of dimension (lda,n), on CPU host.
@param[in]
lda Leading dimension of matrix A. lda >= m.
@param[out]
dB_dst Destination array of dimension (lddb,n), on GPU device.
@param[in]
lddb Leading dimension of matrix B. lddb >= m.
@param[in]
queue Queue to execute in.
@ingroup magma_setmatrix
*******************************************************************************/
extern "C" void
magma_setmatrix_internal(
magma_int_t m, magma_int_t n, magma_int_t elemSize,
void const* hA_src, magma_int_t lda,
magma_ptr dB_dst, magma_int_t lddb,
magma_queue_t queue,
const char* func, const char* file, int line )
{
cudaStream_t stream = NULL;
if ( queue != NULL ) {
stream = queue->cuda_stream();
}
cublasStatus_t status;
status = cublasSetMatrixAsync(
int(m), int(n), int(elemSize),
hA_src, int(lda),
dB_dst, int(lddb), stream );
if ( queue != NULL )
cudaStreamSynchronize( stream );
check_xerror( status, func, file, line );
MAGMA_UNUSED( status );
}
/***************************************************************************//**
@fn magma_setmatrix_async( m, n, elemSize, hA_src, lda, dB_dst, lddb, queue )
Copy all or part of matrix hA_src on CPU host to dB_dst on GPU device.
Elements may be arbitrary size.
Type-safe versions set elemSize appropriately.
This version is asynchronous: it may return before the transfer finishes,
if hA_src is pinned CPU memory.
See magma_setmatrix() for a synchronous version.
@param[in]
m Number of rows of matrix A. m >= 0.
@param[in]
n Number of columns of matrix A. n >= 0.
@param[in]
elemSize Size of each element, e.g., sizeof(double).
@param[in]
hA_src Source array of dimension (lda,n), on CPU host.
@param[in]
lda Leading dimension of matrix A. lda >= m.
@param[out]
dB_dst Destination array of dimension (lddb,n), on GPU device.
@param[in]
lddb Leading dimension of matrix B. lddb >= m.
@param[in]
queue Queue to execute in.
@ingroup magma_setmatrix
*******************************************************************************/
extern "C" void
magma_setmatrix_async_internal(
magma_int_t m, magma_int_t n, magma_int_t elemSize,
void const* hA_src, magma_int_t lda,
magma_ptr dB_dst, magma_int_t lddb,
magma_queue_t queue,
const char* func, const char* file, int line )
{
// for backwards compatability, accepts NULL queue to mean NULL stream.
cudaStream_t stream = NULL;
if ( queue != NULL ) {
stream = queue->cuda_stream();
}
else {
fprintf( stderr, "Warning: %s got NULL queue\n", __func__ );
}
cublasStatus_t status;
status = cublasSetMatrixAsync(
int(m), int(n), int(elemSize),
hA_src, int(lda),
dB_dst, int(lddb), stream );
check_xerror( status, func, file, line );
MAGMA_UNUSED( status );
}
/***************************************************************************//**
@fn magma_getmatrix( m, n, elemSize, dA_src, ldda, hB_dst, ldb, queue )
Copy all or part of matrix dA_src on GPU device to hB_dst on CPU host.
Elements may be arbitrary size.
Type-safe versions set elemSize appropriately.
This version synchronizes the queue after the transfer.
See magma_getmatrix_async() for an asynchronous version.
@param[in]
m Number of rows of matrix A. m >= 0.
@param[in]
n Number of columns of matrix A. n >= 0.
@param[in]
elemSize Size of each element, e.g., sizeof(double).
@param[in]
dA_src Source array of dimension (ldda,n), on GPU device.
@param[in]
ldda Leading dimension of matrix A. ldda >= m.
@param[out]
hB_dst Destination array of dimension (ldb,n), on CPU host.
@param[in]
ldb Leading dimension of matrix B. ldb >= m.
@param[in]
queue Queue to execute in.
@ingroup magma_getmatrix
*******************************************************************************/
extern "C" void
magma_getmatrix_internal(
magma_int_t m, magma_int_t n, magma_int_t elemSize,
magma_const_ptr dA_src, magma_int_t ldda,
void* hB_dst, magma_int_t ldb,
magma_queue_t queue,
const char* func, const char* file, int line )
{
cudaStream_t stream = NULL;
if ( queue != NULL ) {
stream = queue->cuda_stream();
}
cublasStatus_t status;
status = cublasGetMatrixAsync(
int(m), int(n), int(elemSize),
dA_src, int(ldda),
hB_dst, int(ldb), stream );
if ( queue != NULL )
cudaStreamSynchronize( stream );
check_xerror( status, func, file, line );
MAGMA_UNUSED( status );
}
/***************************************************************************//**
@fn magma_getmatrix_async( m, n, elemSize, dA_src, ldda, hB_dst, ldb, queue )
Copy all or part of matrix dA_src on GPU device to hB_dst on CPU host.
Elements may be arbitrary size.
Type-safe versions set elemSize appropriately.
This version is asynchronous: it may return before the transfer finishes,
if hB_dst is pinned CPU memory.
See magma_getmatrix() for a synchronous version.
@param[in]
m Number of rows of matrix A. m >= 0.
@param[in]
n Number of columns of matrix A. n >= 0.
@param[in]
elemSize Size of each element, e.g., sizeof(double).
@param[in]
dA_src Source array of dimension (ldda,n), on GPU device.
@param[in]
ldda Leading dimension of matrix A. ldda >= m.
@param[out]
hB_dst Destination array of dimension (ldb,n), on CPU host.
@param[in]
ldb Leading dimension of matrix B. ldb >= m.
@param[in]
queue Queue to execute in.
@ingroup magma_getmatrix
*******************************************************************************/
extern "C" void
magma_getmatrix_async_internal(
magma_int_t m, magma_int_t n, magma_int_t elemSize,
magma_const_ptr dA_src, magma_int_t ldda,
void* hB_dst, magma_int_t ldb,
magma_queue_t queue,
const char* func, const char* file, int line )
{
cudaStream_t stream = NULL;
if ( queue != NULL ) {
stream = queue->cuda_stream();
}
else {
fprintf( stderr, "Warning: %s got NULL queue\n", __func__ );
}
cublasStatus_t status;
status = cublasGetMatrixAsync(
int(m), int(n), int(elemSize),
dA_src, int(ldda),
hB_dst, int(ldb), stream );
check_xerror( status, func, file, line );
MAGMA_UNUSED( status );
}
/***************************************************************************//**
@fn magma_copymatrix( m, n, elemSize, dA_src, ldda, dB_dst, lddb, queue )
Copy all or part of matrix dA_src on GPU device to dB_dst on GPU device.
Elements may be arbitrary size.
Type-safe versions set elemSize appropriately.
With CUDA unified addressing, dA and dB can be on different GPUs.
This version synchronizes the queue after the transfer.
See magma_copymatrix_async() for an asynchronous version.
@param[in]
m Number of rows of matrix A. m >= 0.
@param[in]
n Number of columns of matrix A. n >= 0.
@param[in]
elemSize Size of each element, e.g., sizeof(double).
@param[in]
dA_src Source array of dimension (ldda,n).
@param[in]
ldda Leading dimension of matrix A. ldda >= m.
@param[out]
dB_dst Destination array of dimension (lddb,n), on GPU device.
@param[in]
lddb Leading dimension of matrix B. lddb >= m.
@param[in]
queue Queue to execute in.
@ingroup magma_copymatrix
*******************************************************************************/
extern "C" void
magma_copymatrix_internal(
magma_int_t m, magma_int_t n, magma_int_t elemSize,
magma_const_ptr dA_src, magma_int_t ldda,
magma_ptr dB_dst, magma_int_t lddb,
magma_queue_t queue,
const char* func, const char* file, int line )
{
cudaStream_t stream = NULL;
if ( queue != NULL ) {
stream = queue->cuda_stream();
}
cudaError_t status;
status = cudaMemcpy2DAsync(
dB_dst, int(lddb*elemSize),
dA_src, int(ldda*elemSize),
int(m*elemSize), int(n), cudaMemcpyDeviceToDevice, stream );
if ( queue != NULL )
cudaStreamSynchronize( stream );
check_xerror( status, func, file, line );
MAGMA_UNUSED( status );
}
/***************************************************************************//**
@fn magma_copymatrix_async( m, n, elemSize, dA_src, ldda, dB_dst, lddb, queue )
Copy all or part of matrix dA_src on GPU device to dB_dst on GPU device.
Elements may be arbitrary size.
Type-safe versions set elemSize appropriately.
With CUDA unified addressing, dA and dB can be on different GPUs.
This version is asynchronous: it may return before the transfer finishes.
See magma_copyvector() for a synchronous version.
@param[in]
m Number of rows of matrix A. m >= 0.
@param[in]
n Number of columns of matrix A. n >= 0.
@param[in]
elemSize Size of each element, e.g., sizeof(double).
@param[in]
dA_src Source array of dimension (ldda,n), on GPU device.
@param[in]
ldda Leading dimension of matrix A. ldda >= m.
@param[out]
dB_dst Destination array of dimension (lddb,n), on GPU device.
@param[in]
lddb Leading dimension of matrix B. lddb >= m.
@param[in]
queue Queue to execute in.
@ingroup magma_copymatrix
*******************************************************************************/
extern "C" void
magma_copymatrix_async_internal(
magma_int_t m, magma_int_t n, magma_int_t elemSize,
magma_const_ptr dA_src, magma_int_t ldda,
magma_ptr dB_dst, magma_int_t lddb,
magma_queue_t queue,
const char* func, const char* file, int line )
{
// for backwards compatability, accepts NULL queue to mean NULL stream.
cudaStream_t stream = NULL;
if ( queue != NULL ) {
stream = queue->cuda_stream();
}
else {
fprintf( stderr, "Warning: %s got NULL queue\n", __func__ );
}
cudaError_t status;
status = cudaMemcpy2DAsync(
dB_dst, int(lddb*elemSize),
dA_src, int(ldda*elemSize),
int(m*elemSize), int(n), cudaMemcpyDeviceToDevice, stream );
check_xerror( status, func, file, line );
MAGMA_UNUSED( status );
}
#endif // MAGMA_HAVE_CUDA
|
