Repository navigation
Expand file tree
/
Copy pathlevel3.go
More file actions
855 lines (747 loc) · 31.5 KB
/
Copy pathlevel3.go
File metadata and controls
855 lines (747 loc) · 31.5 KB
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
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
package cublas
/*
#include <cublas_v2.h>
*/
import "C"
import (
"unsafe"
"github.com/gocnn/gocu/cudart"
"github.com/gocnn/gomat/blas"
)
func Sgemm(h *Handle, tA, tB Transpose, m, n, k int, alpha float32, a cudart.DevicePtr, lda int, b cudart.DevicePtr, ldb int, beta float32, c cudart.DevicePtr, ldc int) error {
if m < 0 {
panic(blas.ErrMLT0)
}
if n < 0 {
panic(blas.ErrMLT0)
}
if k < 0 {
panic(blas.ErrKLT0)
}
return Check(C.cublasSgemm(C.cublasHandle_t(h.h), C.cublasOperation_t(tA), C.cublasOperation_t(tB), C.int(m), C.int(n), C.int(k), (*C.float)(&alpha), (*C.float)(a), C.int(lda), (*C.float)(b), C.int(ldb), (*C.float)(&beta), (*C.float)(c), C.int(ldc)))
}
// Dgemm performs the matrix-matrix operation C = alpha*op(A)*op(B) + beta*C for double precision real matrices.
// op(X) is either X, X^T depending on the transpose parameter.
func Dgemm(h *Handle, tA, tB Transpose, m, n, k int, alpha float64, a cudart.DevicePtr, lda int, b cudart.DevicePtr, ldb int, beta float64, c cudart.DevicePtr, ldc int) error {
if m < 0 {
panic(blas.ErrMLT0)
}
if n < 0 {
panic(blas.ErrNLT0)
}
if k < 0 {
panic(blas.ErrKLT0)
}
if lda < 1 {
panic(blas.ErrShortA)
}
if ldb < 1 {
panic(blas.ErrShortB)
}
if ldc < 1 {
panic(blas.ErrShortC)
}
return Check(C.cublasDgemm(C.cublasHandle_t(h.h), C.cublasOperation_t(tA), C.cublasOperation_t(tB), C.int(m), C.int(n), C.int(k), (*C.double)(&alpha), (*C.double)(a), C.int(lda), (*C.double)(b), C.int(ldb), (*C.double)(&beta), (*C.double)(c), C.int(ldc)))
}
// Cgemm performs the matrix-matrix operation C = alpha*op(A)*op(B) + beta*C for single precision complex matrices.
// op(X) is either X, X^T, or X^H depending on the transpose parameter.
func Cgemm(h *Handle, tA, tB Transpose, m, n, k int, alpha complex64, a cudart.DevicePtr, lda int, b cudart.DevicePtr, ldb int, beta complex64, c cudart.DevicePtr, ldc int) error {
if m < 0 {
panic(blas.ErrMLT0)
}
if n < 0 {
panic(blas.ErrNLT0)
}
if k < 0 {
panic(blas.ErrKLT0)
}
if lda < 1 {
panic(blas.ErrShortA)
}
if ldb < 1 {
panic(blas.ErrShortB)
}
if ldc < 1 {
panic(blas.ErrShortC)
}
return Check(C.cublasCgemm(C.cublasHandle_t(h.h), C.cublasOperation_t(tA), C.cublasOperation_t(tB), C.int(m), C.int(n), C.int(k), (*C.cuComplex)(unsafe.Pointer(&alpha)), (*C.cuComplex)(a), C.int(lda), (*C.cuComplex)(b), C.int(ldb), (*C.cuComplex)(unsafe.Pointer(&beta)), (*C.cuComplex)(c), C.int(ldc)))
}
// Zgemm performs the matrix-matrix operation C = alpha*op(A)*op(B) + beta*C for double precision complex matrices.
// op(X) is either X, X^T, or X^H depending on the transpose parameter.
func Zgemm(h *Handle, tA, tB Transpose, m, n, k int, alpha complex128, a cudart.DevicePtr, lda int, b cudart.DevicePtr, ldb int, beta complex128, c cudart.DevicePtr, ldc int) error {
if m < 0 {
panic(blas.ErrMLT0)
}
if n < 0 {
panic(blas.ErrNLT0)
}
if k < 0 {
panic(blas.ErrKLT0)
}
if lda < 1 {
panic(blas.ErrShortA)
}
if ldb < 1 {
panic(blas.ErrShortB)
}
if ldc < 1 {
panic(blas.ErrShortC)
}
return Check(C.cublasZgemm(C.cublasHandle_t(h.h), C.cublasOperation_t(tA), C.cublasOperation_t(tB), C.int(m), C.int(n), C.int(k), (*C.cuDoubleComplex)(unsafe.Pointer(&alpha)), (*C.cuDoubleComplex)(a), C.int(lda), (*C.cuDoubleComplex)(b), C.int(ldb), (*C.cuDoubleComplex)(unsafe.Pointer(&beta)), (*C.cuDoubleComplex)(c), C.int(ldc)))
}
// Cgemm3m performs the matrix-matrix operation C = alpha*op(A)*op(B) + beta*C for single precision complex matrices
// using the 3M algorithm which reduces the number of real multiplications from 4 to 3 per complex multiplication.
func Cgemm3m(h *Handle, tA, tB Transpose, m, n, k int, alpha complex64, a cudart.DevicePtr, lda int, b cudart.DevicePtr, ldb int, beta complex64, c cudart.DevicePtr, ldc int) error {
if m < 0 {
panic(blas.ErrMLT0)
}
if n < 0 {
panic(blas.ErrNLT0)
}
if k < 0 {
panic(blas.ErrKLT0)
}
if lda < 1 {
panic(blas.ErrShortA)
}
if ldb < 1 {
panic(blas.ErrShortB)
}
if ldc < 1 {
panic(blas.ErrShortC)
}
return Check(C.cublasCgemm3m(C.cublasHandle_t(h.h), C.cublasOperation_t(tA), C.cublasOperation_t(tB), C.int(m), C.int(n), C.int(k), (*C.cuComplex)(unsafe.Pointer(&alpha)), (*C.cuComplex)(a), C.int(lda), (*C.cuComplex)(b), C.int(ldb), (*C.cuComplex)(unsafe.Pointer(&beta)), (*C.cuComplex)(c), C.int(ldc)))
}
// Zgemm3m performs the matrix-matrix operation C = alpha*op(A)*op(B) + beta*C for double precision complex matrices
// using the 3M algorithm which reduces the number of real multiplications from 4 to 3 per complex multiplication.
func Zgemm3m(h *Handle, tA, tB Transpose, m, n, k int, alpha complex128, a cudart.DevicePtr, lda int, b cudart.DevicePtr, ldb int, beta complex128, c cudart.DevicePtr, ldc int) error {
if m < 0 {
panic(blas.ErrMLT0)
}
if n < 0 {
panic(blas.ErrNLT0)
}
if k < 0 {
panic(blas.ErrKLT0)
}
if lda < 1 {
panic(blas.ErrShortA)
}
if ldb < 1 {
panic(blas.ErrShortB)
}
if ldc < 1 {
panic(blas.ErrShortC)
}
return Check(C.cublasZgemm3m(C.cublasHandle_t(h.h), C.cublasOperation_t(tA), C.cublasOperation_t(tB), C.int(m), C.int(n), C.int(k), (*C.cuDoubleComplex)(unsafe.Pointer(&alpha)), (*C.cuDoubleComplex)(a), C.int(lda), (*C.cuDoubleComplex)(b), C.int(ldb), (*C.cuDoubleComplex)(unsafe.Pointer(&beta)), (*C.cuDoubleComplex)(c), C.int(ldc)))
}
// Ssymm performs the symmetric matrix-matrix operation C = alpha*A*B + beta*C or C = alpha*B*A + beta*C for single precision real matrices.
// A is a symmetric matrix.
func Ssymm(h *Handle, side Side, uplo Uplo, m, n int, alpha float32, a cudart.DevicePtr, lda int, b cudart.DevicePtr, ldb int, beta float32, c cudart.DevicePtr, ldc int) error {
if m < 0 {
panic(blas.ErrMLT0)
}
if n < 0 {
panic(blas.ErrNLT0)
}
if lda < 1 {
panic(blas.ErrShortA)
}
if ldb < 1 {
panic(blas.ErrShortB)
}
if ldc < 1 {
panic(blas.ErrShortC)
}
return Check(C.cublasSsymm(C.cublasHandle_t(h.h), C.cublasSideMode_t(side), C.cublasFillMode_t(uplo), C.int(m), C.int(n), (*C.float)(&alpha), (*C.float)(a), C.int(lda), (*C.float)(b), C.int(ldb), (*C.float)(&beta), (*C.float)(c), C.int(ldc)))
}
// Dsymm performs the symmetric matrix-matrix operation C = alpha*A*B + beta*C or C = alpha*B*A + beta*C for double precision real matrices.
// A is a symmetric matrix.
func Dsymm(h *Handle, side Side, uplo Uplo, m, n int, alpha float64, a cudart.DevicePtr, lda int, b cudart.DevicePtr, ldb int, beta float64, c cudart.DevicePtr, ldc int) error {
if m < 0 {
panic(blas.ErrMLT0)
}
if n < 0 {
panic(blas.ErrNLT0)
}
if lda < 1 {
panic(blas.ErrShortA)
}
if ldb < 1 {
panic(blas.ErrShortB)
}
if ldc < 1 {
panic(blas.ErrShortC)
}
return Check(C.cublasDsymm(C.cublasHandle_t(h.h), C.cublasSideMode_t(side), C.cublasFillMode_t(uplo), C.int(m), C.int(n), (*C.double)(&alpha), (*C.double)(a), C.int(lda), (*C.double)(b), C.int(ldb), (*C.double)(&beta), (*C.double)(c), C.int(ldc)))
}
// Csymm performs the symmetric matrix-matrix operation C = alpha*A*B + beta*C or C = alpha*B*A + beta*C for single precision complex matrices.
// A is a symmetric matrix (not Hermitian).
func Csymm(h *Handle, side Side, uplo Uplo, m, n int, alpha complex64, a cudart.DevicePtr, lda int, b cudart.DevicePtr, ldb int, beta complex64, c cudart.DevicePtr, ldc int) error {
if m < 0 {
panic(blas.ErrMLT0)
}
if n < 0 {
panic(blas.ErrNLT0)
}
if lda < 1 {
panic(blas.ErrShortA)
}
if ldb < 1 {
panic(blas.ErrShortB)
}
if ldc < 1 {
panic(blas.ErrShortC)
}
return Check(C.cublasCsymm(C.cublasHandle_t(h.h), C.cublasSideMode_t(side), C.cublasFillMode_t(uplo), C.int(m), C.int(n), (*C.cuComplex)(unsafe.Pointer(&alpha)), (*C.cuComplex)(a), C.int(lda), (*C.cuComplex)(b), C.int(ldb), (*C.cuComplex)(unsafe.Pointer(&beta)), (*C.cuComplex)(c), C.int(ldc)))
}
// Zsymm performs the symmetric matrix-matrix operation C = alpha*A*B + beta*C or C = alpha*B*A + beta*C for double precision complex matrices.
// A is a symmetric matrix (not Hermitian).
func Zsymm(h *Handle, side Side, uplo Uplo, m, n int, alpha complex128, a cudart.DevicePtr, lda int, b cudart.DevicePtr, ldb int, beta complex128, c cudart.DevicePtr, ldc int) error {
if m < 0 {
panic(blas.ErrMLT0)
}
if n < 0 {
panic(blas.ErrNLT0)
}
if lda < 1 {
panic(blas.ErrShortA)
}
if ldb < 1 {
panic(blas.ErrShortB)
}
if ldc < 1 {
panic(blas.ErrShortC)
}
return Check(C.cublasZsymm(C.cublasHandle_t(h.h), C.cublasSideMode_t(side), C.cublasFillMode_t(uplo), C.int(m), C.int(n), (*C.cuDoubleComplex)(unsafe.Pointer(&alpha)), (*C.cuDoubleComplex)(a), C.int(lda), (*C.cuDoubleComplex)(b), C.int(ldb), (*C.cuDoubleComplex)(unsafe.Pointer(&beta)), (*C.cuDoubleComplex)(c), C.int(ldc)))
}
// Ssyrk performs the symmetric rank-k update C = alpha*A*A^T + beta*C or C = alpha*A^T*A + beta*C for single precision real matrices.
// C is a symmetric matrix.
func Ssyrk(h *Handle, uplo Uplo, trans Transpose, n, k int, alpha float32, a cudart.DevicePtr, lda int, beta float32, c cudart.DevicePtr, ldc int) error {
if n < 0 {
panic(blas.ErrNLT0)
}
if k < 0 {
panic(blas.ErrKLT0)
}
if lda < 1 {
panic(blas.ErrShortA)
}
if ldc < 1 {
panic(blas.ErrShortC)
}
return Check(C.cublasSsyrk(C.cublasHandle_t(h.h), C.cublasFillMode_t(uplo), C.cublasOperation_t(trans), C.int(n), C.int(k), (*C.float)(&alpha), (*C.float)(a), C.int(lda), (*C.float)(&beta), (*C.float)(c), C.int(ldc)))
}
// Dsyrk performs the symmetric rank-k update C = alpha*A*A^T + beta*C or C = alpha*A^T*A + beta*C for double precision real matrices.
// C is a symmetric matrix.
func Dsyrk(h *Handle, uplo Uplo, trans Transpose, n, k int, alpha float64, a cudart.DevicePtr, lda int, beta float64, c cudart.DevicePtr, ldc int) error {
if n < 0 {
panic(blas.ErrNLT0)
}
if k < 0 {
panic(blas.ErrKLT0)
}
if lda < 1 {
panic(blas.ErrShortA)
}
if ldc < 1 {
panic(blas.ErrShortC)
}
return Check(C.cublasDsyrk(C.cublasHandle_t(h.h), C.cublasFillMode_t(uplo), C.cublasOperation_t(trans), C.int(n), C.int(k), (*C.double)(&alpha), (*C.double)(a), C.int(lda), (*C.double)(&beta), (*C.double)(c), C.int(ldc)))
}
// Csyrk performs the symmetric rank-k update C = alpha*A*A^T + beta*C or C = alpha*A^T*A + beta*C for single precision complex matrices.
// C is a symmetric matrix (not Hermitian).
func Csyrk(h *Handle, uplo Uplo, trans Transpose, n, k int, alpha complex64, a cudart.DevicePtr, lda int, beta complex64, c cudart.DevicePtr, ldc int) error {
if n < 0 {
panic(blas.ErrNLT0)
}
if k < 0 {
panic(blas.ErrKLT0)
}
if lda < 1 {
panic(blas.ErrShortA)
}
if ldc < 1 {
panic(blas.ErrShortC)
}
return Check(C.cublasCsyrk(C.cublasHandle_t(h.h), C.cublasFillMode_t(uplo), C.cublasOperation_t(trans), C.int(n), C.int(k), (*C.cuComplex)(unsafe.Pointer(&alpha)), (*C.cuComplex)(a), C.int(lda), (*C.cuComplex)(unsafe.Pointer(&beta)), (*C.cuComplex)(c), C.int(ldc)))
}
// Zsyrk performs the symmetric rank-k update C = alpha*A*A^T + beta*C or C = alpha*A^T*A + beta*C for double precision complex matrices.
// C is a symmetric matrix (not Hermitian).
func Zsyrk(h *Handle, uplo Uplo, trans Transpose, n, k int, alpha complex128, a cudart.DevicePtr, lda int, beta complex128, c cudart.DevicePtr, ldc int) error {
if n < 0 {
panic(blas.ErrNLT0)
}
if k < 0 {
panic(blas.ErrKLT0)
}
if lda < 1 {
panic(blas.ErrShortA)
}
if ldc < 1 {
panic(blas.ErrShortC)
}
return Check(C.cublasZsyrk(C.cublasHandle_t(h.h), C.cublasFillMode_t(uplo), C.cublasOperation_t(trans), C.int(n), C.int(k), (*C.cuDoubleComplex)(unsafe.Pointer(&alpha)), (*C.cuDoubleComplex)(a), C.int(lda), (*C.cuDoubleComplex)(unsafe.Pointer(&beta)), (*C.cuDoubleComplex)(c), C.int(ldc)))
}
// Ssyr2k performs the symmetric rank-2k update C = alpha*(A*B^T + B*A^T) + beta*C or C = alpha*(A^T*B + B^T*A) + beta*C for single precision real matrices.
// C is a symmetric matrix.
func Ssyr2k(h *Handle, uplo Uplo, trans Transpose, n, k int, alpha float32, a cudart.DevicePtr, lda int, b cudart.DevicePtr, ldb int, beta float32, c cudart.DevicePtr, ldc int) error {
if n < 0 {
panic(blas.ErrNLT0)
}
if k < 0 {
panic(blas.ErrKLT0)
}
if lda < 1 {
panic(blas.ErrShortA)
}
if ldb < 1 {
panic(blas.ErrShortB)
}
if ldc < 1 {
panic(blas.ErrShortC)
}
return Check(C.cublasSsyr2k(C.cublasHandle_t(h.h), C.cublasFillMode_t(uplo), C.cublasOperation_t(trans), C.int(n), C.int(k), (*C.float)(&alpha), (*C.float)(a), C.int(lda), (*C.float)(b), C.int(ldb), (*C.float)(&beta), (*C.float)(c), C.int(ldc)))
}
// Dsyr2k performs the symmetric rank-2k update C = alpha*(A*B^T + B*A^T) + beta*C or C = alpha*(A^T*B + B^T*A) + beta*C for double precision real matrices.
// C is a symmetric matrix.
func Dsyr2k(h *Handle, uplo Uplo, trans Transpose, n, k int, alpha float64, a cudart.DevicePtr, lda int, b cudart.DevicePtr, ldb int, beta float64, c cudart.DevicePtr, ldc int) error {
if n < 0 {
panic(blas.ErrNLT0)
}
if k < 0 {
panic(blas.ErrKLT0)
}
if lda < 1 {
panic(blas.ErrShortA)
}
if ldb < 1 {
panic(blas.ErrShortB)
}
if ldc < 1 {
panic(blas.ErrShortC)
}
return Check(C.cublasDsyr2k(C.cublasHandle_t(h.h), C.cublasFillMode_t(uplo), C.cublasOperation_t(trans), C.int(n), C.int(k), (*C.double)(&alpha), (*C.double)(a), C.int(lda), (*C.double)(b), C.int(ldb), (*C.double)(&beta), (*C.double)(c), C.int(ldc)))
}
// Csyr2k performs the symmetric rank-2k update C = alpha*(A*B^T + B*A^T) + beta*C or C = alpha*(A^T*B + B^T*A) + beta*C for single precision complex matrices.
// C is a symmetric matrix (not Hermitian).
func Csyr2k(h *Handle, uplo Uplo, trans Transpose, n, k int, alpha complex64, a cudart.DevicePtr, lda int, b cudart.DevicePtr, ldb int, beta complex64, c cudart.DevicePtr, ldc int) error {
if n < 0 {
panic(blas.ErrNLT0)
}
if k < 0 {
panic(blas.ErrKLT0)
}
if lda < 1 {
panic(blas.ErrShortA)
}
if ldb < 1 {
panic(blas.ErrShortB)
}
if ldc < 1 {
panic(blas.ErrShortC)
}
return Check(C.cublasCsyr2k(C.cublasHandle_t(h.h), C.cublasFillMode_t(uplo), C.cublasOperation_t(trans), C.int(n), C.int(k), (*C.cuComplex)(unsafe.Pointer(&alpha)), (*C.cuComplex)(a), C.int(lda), (*C.cuComplex)(b), C.int(ldb), (*C.cuComplex)(unsafe.Pointer(&beta)), (*C.cuComplex)(c), C.int(ldc)))
}
// Zsyr2k performs the symmetric rank-2k update C = alpha*(A*B^T + B*A^T) + beta*C or C = alpha*(A^T*B + B^T*A) + beta*C for double precision complex matrices.
// C is a symmetric matrix (not Hermitian).
func Zsyr2k(h *Handle, uplo Uplo, trans Transpose, n, k int, alpha complex128, a cudart.DevicePtr, lda int, b cudart.DevicePtr, ldb int, beta complex128, c cudart.DevicePtr, ldc int) error {
if n < 0 {
panic(blas.ErrNLT0)
}
if k < 0 {
panic(blas.ErrKLT0)
}
if lda < 1 {
panic(blas.ErrShortA)
}
if ldb < 1 {
panic(blas.ErrShortB)
}
if ldc < 1 {
panic(blas.ErrShortC)
}
return Check(C.cublasZsyr2k(C.cublasHandle_t(h.h), C.cublasFillMode_t(uplo), C.cublasOperation_t(trans), C.int(n), C.int(k), (*C.cuDoubleComplex)(unsafe.Pointer(&alpha)), (*C.cuDoubleComplex)(a), C.int(lda), (*C.cuDoubleComplex)(b), C.int(ldb), (*C.cuDoubleComplex)(unsafe.Pointer(&beta)), (*C.cuDoubleComplex)(c), C.int(ldc)))
}
// Ssyrkx performs the symmetric rank-kx update C = alpha*A*B^T + beta*C or C = alpha*A^T*B + beta*C for single precision real matrices.
// This is a generalization of SYRK where A and B can be different matrices.
func Ssyrkx(h *Handle, uplo Uplo, trans Transpose, n, k int, alpha float32, a cudart.DevicePtr, lda int, b cudart.DevicePtr, ldb int, beta float32, c cudart.DevicePtr, ldc int) error {
if n < 0 {
panic(blas.ErrNLT0)
}
if k < 0 {
panic(blas.ErrKLT0)
}
if lda < 1 {
panic(blas.ErrShortA)
}
if ldb < 1 {
panic(blas.ErrShortB)
}
if ldc < 1 {
panic(blas.ErrShortC)
}
return Check(C.cublasSsyrkx(C.cublasHandle_t(h.h), C.cublasFillMode_t(uplo), C.cublasOperation_t(trans), C.int(n), C.int(k), (*C.float)(&alpha), (*C.float)(a), C.int(lda), (*C.float)(b), C.int(ldb), (*C.float)(&beta), (*C.float)(c), C.int(ldc)))
}
// Dsyrkx performs the symmetric rank-kx update C = alpha*A*B^T + beta*C or C = alpha*A^T*B + beta*C for double precision real matrices.
// This is a generalization of SYRK where A and B can be different matrices.
func Dsyrkx(h *Handle, uplo Uplo, trans Transpose, n, k int, alpha float64, a cudart.DevicePtr, lda int, b cudart.DevicePtr, ldb int, beta float64, c cudart.DevicePtr, ldc int) error {
if n < 0 {
panic(blas.ErrNLT0)
}
if k < 0 {
panic(blas.ErrKLT0)
}
if lda < 1 {
panic(blas.ErrShortA)
}
if ldb < 1 {
panic(blas.ErrShortB)
}
if ldc < 1 {
panic(blas.ErrShortC)
}
return Check(C.cublasDsyrkx(C.cublasHandle_t(h.h), C.cublasFillMode_t(uplo), C.cublasOperation_t(trans), C.int(n), C.int(k), (*C.double)(&alpha), (*C.double)(a), C.int(lda), (*C.double)(b), C.int(ldb), (*C.double)(&beta), (*C.double)(c), C.int(ldc)))
}
// Csyrkx performs the symmetric rank-kx update C = alpha*A*B^T + beta*C or C = alpha*A^T*B + beta*C for single precision complex matrices.
// This is a generalization of SYRK where A and B can be different matrices. C is a symmetric matrix (not Hermitian).
func Csyrkx(h *Handle, uplo Uplo, trans Transpose, n, k int, alpha complex64, a cudart.DevicePtr, lda int, b cudart.DevicePtr, ldb int, beta complex64, c cudart.DevicePtr, ldc int) error {
if n < 0 {
panic(blas.ErrNLT0)
}
if k < 0 {
panic(blas.ErrKLT0)
}
if lda < 1 {
panic(blas.ErrShortA)
}
if ldb < 1 {
panic(blas.ErrShortB)
}
if ldc < 1 {
panic(blas.ErrShortC)
}
return Check(C.cublasCsyrkx(C.cublasHandle_t(h.h), C.cublasFillMode_t(uplo), C.cublasOperation_t(trans), C.int(n), C.int(k), (*C.cuComplex)(unsafe.Pointer(&alpha)), (*C.cuComplex)(a), C.int(lda), (*C.cuComplex)(b), C.int(ldb), (*C.cuComplex)(unsafe.Pointer(&beta)), (*C.cuComplex)(c), C.int(ldc)))
}
// Zsyrkx performs the symmetric rank-kx update C = alpha*A*B^T + beta*C or C = alpha*A^T*B + beta*C for double precision complex matrices.
// This is a generalization of SYRK where A and B can be different matrices. C is a symmetric matrix (not Hermitian).
func Zsyrkx(h *Handle, uplo Uplo, trans Transpose, n, k int, alpha complex128, a cudart.DevicePtr, lda int, b cudart.DevicePtr, ldb int, beta complex128, c cudart.DevicePtr, ldc int) error {
if n < 0 {
panic(blas.ErrNLT0)
}
if k < 0 {
panic(blas.ErrKLT0)
}
if lda < 1 {
panic(blas.ErrShortA)
}
if ldb < 1 {
panic(blas.ErrShortB)
}
if ldc < 1 {
panic(blas.ErrShortC)
}
return Check(C.cublasZsyrkx(C.cublasHandle_t(h.h), C.cublasFillMode_t(uplo), C.cublasOperation_t(trans), C.int(n), C.int(k), (*C.cuDoubleComplex)(unsafe.Pointer(&alpha)), (*C.cuDoubleComplex)(a), C.int(lda), (*C.cuDoubleComplex)(b), C.int(ldb), (*C.cuDoubleComplex)(unsafe.Pointer(&beta)), (*C.cuDoubleComplex)(c), C.int(ldc)))
}
// Strmm performs the triangular matrix-matrix operation C = alpha*op(A)*B or C = alpha*B*op(A) for single precision real matrices.
// A is a triangular matrix.
func Strmm(h *Handle, side Side, uplo Uplo, trans Transpose, diag Diag, m, n int, alpha float32, a cudart.DevicePtr, lda int, b cudart.DevicePtr, ldb int, c cudart.DevicePtr, ldc int) error {
if m < 0 {
panic(blas.ErrMLT0)
}
if n < 0 {
panic(blas.ErrNLT0)
}
if lda < 1 {
panic(blas.ErrShortA)
}
if ldb < 1 {
panic(blas.ErrShortB)
}
if ldc < 1 {
panic(blas.ErrShortC)
}
return Check(C.cublasStrmm(C.cublasHandle_t(h.h), C.cublasSideMode_t(side), C.cublasFillMode_t(uplo), C.cublasOperation_t(trans), C.cublasDiagType_t(diag), C.int(m), C.int(n), (*C.float)(&alpha), (*C.float)(a), C.int(lda), (*C.float)(b), C.int(ldb), (*C.float)(c), C.int(ldc)))
}
// Dtrmm performs the triangular matrix-matrix operation C = alpha*op(A)*B or C = alpha*B*op(A) for double precision real matrices.
// A is a triangular matrix.
func Dtrmm(h *Handle, side Side, uplo Uplo, trans Transpose, diag Diag, m, n int, alpha float64, a cudart.DevicePtr, lda int, b cudart.DevicePtr, ldb int, c cudart.DevicePtr, ldc int) error {
if m < 0 {
panic(blas.ErrMLT0)
}
if n < 0 {
panic(blas.ErrNLT0)
}
if lda < 1 {
panic(blas.ErrShortA)
}
if ldb < 1 {
panic(blas.ErrShortB)
}
if ldc < 1 {
panic(blas.ErrShortC)
}
return Check(C.cublasDtrmm(C.cublasHandle_t(h.h), C.cublasSideMode_t(side), C.cublasFillMode_t(uplo), C.cublasOperation_t(trans), C.cublasDiagType_t(diag), C.int(m), C.int(n)
BD0
, (*C.double)(&alpha), (*C.double)(a), C.int(lda), (*C.double)(b), C.int(ldb), (*C.double)(c), C.int(ldc)))
}
// Ctrmm performs the triangular matrix-matrix operation C = alpha*op(A)*B or C = alpha*B*op(A) for single precision complex matrices.
// A is a triangular matrix.
func Ctrmm(h *Handle, side Side, uplo Uplo, trans Transpose, diag Diag, m, n int, alpha complex64, a cudart.DevicePtr, lda int, b cudart.DevicePtr, ldb int, c cudart.DevicePtr, ldc int) error {
if m < 0 {
panic(blas.ErrMLT0)
}
if n < 0 {
panic(blas.ErrNLT0)
}
if lda < 1 {
panic(blas.ErrShortA)
}
if ldb < 1 {
panic(blas.ErrShortB)
}
if ldc < 1 {
panic(blas.ErrShortC)
}
return Check(C.cublasCtrmm(C.cublasHandle_t(h.h), C.cublasSideMode_t(side), C.cublasFillMode_t(uplo), C.cublasOperation_t(trans), C.cublasDiagType_t(diag), C.int(m), C.int(n), (*C.cuComplex)(unsafe.Pointer(&alpha)), (*C.cuComplex)(a), C.int(lda), (*C.cuComplex)(b), C.int(ldb), (*C.cuComplex)(c), C.int(ldc)))
}
// Ztrmm performs the triangular matrix-matrix operation C = alpha*op(A)*B or C = alpha*B*op(A) for double precision complex matrices.
// A is a triangular matrix.
func Ztrmm(h *Handle, side Side, uplo Uplo, trans Transpose, diag Diag, m, n int, alpha complex128, a cudart.DevicePtr, lda int, b cudart.DevicePtr, ldb int, c cudart.DevicePtr, ldc int) error {
if m < 0 {
panic(blas.ErrMLT0)
}
if n < 0 {
panic(blas.ErrNLT0)
}
if lda < 1 {
panic(blas.ErrShortA)
}
if ldb < 1 {
panic(blas.ErrShortB)
}
if ldc < 1 {
panic(blas.ErrShortC)
}
return Check(C.cublasZtrmm(C.cublasHandle_t(h.h), C.cublasSideMode_t(side), C.cublasFillMode_t(uplo), C.cublasOperation_t(trans), C.cublasDiagType_t(diag), C.int(m), C.int(n), (*C.cuDoubleComplex)(unsafe.Pointer(&alpha)), (*C.cuDoubleComplex)(a), C.int(lda), (*C.cuDoubleComplex)(b), C.int(ldb), (*C.cuDoubleComplex)(c), C.int(ldc)))
}
// Strsm solves the triangular matrix equation op(A)*X = alpha*B or X*op(A) = alpha*B for single precision real matrices.
// A is a triangular matrix, and the solution X overwrites B.
func Strsm(h *Handle, side Side, uplo Uplo, trans Transpose, diag Diag, m, n int, alpha float32, a cudart.DevicePtr, lda int, b cudart.DevicePtr, ldb int) error {
if m < 0 {
panic(blas.ErrMLT0)
}
if n < 0 {
panic(blas.ErrNLT0)
}
if lda < 1 {
panic(blas.ErrShortA)
}
if ldb < 1 {
panic(blas.ErrShortB)
}
return Check(C.cublasStrsm(C.cublasHandle_t(h.h), C.cublasSideMode_t(side), C.cublasFillMode_t(uplo), C.cublasOperation_t(trans), C.cublasDiagType_t(diag), C.int(m), C.int(n), (*C.float)(&alpha), (*C.float)(a), C.int(lda), (*C.float)(b), C.int(ldb)))
}
// Dtrsm solves the triangular matrix equation op(A)*X = alpha*B or X*op(A) = alpha*B for double precision real matrices.
// A is a triangular matrix, and the solution X overwrites B.
func Dtrsm(h *Handle, side Side, uplo Uplo, trans Transpose, diag Diag, m, n int, alpha float64, a cudart.DevicePtr, lda int, b cudart.DevicePtr, ldb int) error {
if m < 0 {
panic(blas.ErrMLT0)
}
if n < 0 {
panic(blas.ErrNLT0)
}
if lda < 1 {
panic(blas.ErrShortA)
}
if ldb < 1 {
panic(blas.ErrShortB)
}
return Check(C.cublasDtrsm(C.cublasHandle_t(h.h), C.cublasSideMode_t(side), C.cublasFillMode_t(uplo), C.cublasOperation_t(trans), C.cublasDiagType_t(diag), C.int(m), C.int(n), (*C.double)(&alpha), (*C.double)(a), C.int(lda), (*C.double)(b), C.int(ldb)))
}
// Ctrsm solves the triangular matrix equation op(A)*X = alpha*B or X*op(A) = alpha*B for single precision complex matrices.
// A is a triangular matrix, and the solution X overwrites B.
func Ctrsm(h *Handle, side Side, uplo Uplo, trans Transpose, diag Diag, m, n int, alpha complex64, a cudart.DevicePtr, lda int, b cudart.DevicePtr, ldb int) error {
if m < 0 {
panic(blas.ErrMLT0)
}
if n < 0 {
panic(blas.ErrNLT0)
}
if lda < 1 {
panic(blas.ErrShortA)
}
if ldb < 1 {
panic(blas.ErrShortB)
}
return Check(C.cublasCtrsm(C.cublasHandle_t(h.h), C.cublasSideMode_t(side), C.cublasFillMode_t(uplo), C.cublasOperation_t(trans), C.cublasDiagType_t(diag), C.int(m), C.int(n), (*C.cuComplex)(unsafe.Pointer(&alpha)), (*C.cuComplex)(a), C.int(lda), (*C.cuComplex)(b), C.int(ldb)))
}
// Ztrsm solves the triangular matrix equation op(A)*X = alpha*B or X*op(A) = alpha*B for double precision complex matrices.
// A is a triangular matrix, and the solution X overwrites B.
func Ztrsm(h *Handle, side Side, uplo Uplo, trans Transpose, diag Diag, m, n int, alpha complex128, a cudart.DevicePtr, lda int, b cudart.DevicePtr, ldb int) error {
if m < 0 {
panic(blas.ErrMLT0)
}
if n < 0 {
panic(blas.ErrNLT0)
}
if lda < 1 {
panic(blas.ErrShortA)
}
if ldb < 1 {
panic(blas.ErrShortB)
}
return Check(C.cublasZtrsm(C.cublasHandle_t(h.h), C.cublasSideMode_t(side), C.cublasFillMode_t(uplo), C.cublasOperation_t(trans), C.cublasDiagType_t(diag), C.int(m), C.int(n), (*C.cuDoubleComplex)(unsafe.Pointer(&alpha)), (*C.cuDoubleComplex)(a), C.int(lda), (*C.cuDoubleComplex)(b), C.int(ldb)))
}
// Chemm performs the Hermitian matrix-matrix operation C = alpha*A*B + beta*C or C = alpha*B*A + beta*C for single precision complex matrices.
// A is a Hermitian matrix.
func Chemm(h *Handle, side Side, uplo Uplo, m, n int, alpha complex64, a cudart.DevicePtr, lda int, b cudart.DevicePtr, ldb int, beta complex64, c cudart.DevicePtr, ldc int) error {
if m < 0 {
panic(blas.ErrMLT0)
}
if n < 0 {
panic(blas.ErrNLT0)
}
if lda < 1 {
panic(blas.ErrShortA)
}
if ldb < 1 {
panic(blas.ErrShortB)
}
if ldc < 1 {
panic(blas.ErrShortC)
}
return Check(C.cublasChemm(C.cublasHandle_t(h.h), C.cublasSideMode_t(side), C.cublasFillMode_t(uplo), C.int(m), C.int(n), (*C.cuComplex)(unsafe.Pointer(&alpha)), (*C.cuComplex)(a), C.int(lda), (*C.cuComplex)(b), C.int(ldb), (*C.cuComplex)(unsafe.Pointer(&beta)), (*C.cuComplex)(c), C.int(ldc)))
}
// Zhemm performs the Hermitian matrix-matrix operation C = alpha*A*B + beta*C or C = alpha*B*A + beta*C for double precision complex matrices.
// A is a Hermitian matrix.
func Zhemm(h *Handle, side Side, uplo Uplo, m, n int, alpha complex128, a cudart.DevicePtr, lda int, b cudart.DevicePtr, ldb int, beta complex128, c cudart.DevicePtr, ldc int) error {
if m < 0 {
panic(blas.ErrMLT0)
}
if n < 0 {
panic(blas.ErrNLT0)
}
if lda < 1 {
panic(blas.ErrShortA)
}
if ldb < 1 {
panic(blas.ErrShortB)
}
if ldc < 1 {
panic(blas.ErrShortC)
}
return Check(C.cublasZhemm(C.cublasHandle_t(h.h), C.cublasSideMode_t(side), C.cublasFillMode_t(uplo), C.int(m), C.int(n), (*C.cuDoubleComplex)(unsafe.Pointer(&alpha)), (*C.cuDoubleComplex)(a), C.int(lda), (*C.cuDoubleComplex)(b), C.int(ldb), (*C.cuDoubleComplex)(unsafe.Pointer(&beta)), (*C.cuDoubleComplex)(c), C.int(ldc)))
}
// Cherk performs the Hermitian rank-k update C = alpha*A*A^H + beta*C or C = alpha*A^H*A + beta*C for single precision complex matrices.
// C is a Hermitian matrix. Note that alpha and beta are real scalars.
func Cherk(h *Handle, uplo Uplo, trans Transpose, n, k int, alpha float32, a cudart.DevicePtr, lda int, beta float32, c cudart.DevicePtr, ldc int) error {
if n < 0 {
panic(blas.ErrNLT0)
}
if k < 0 {
panic(blas.ErrKLT0)
}
if lda < 1 {
panic(blas.ErrShortA)
}
if ldc < 1 {
panic(blas.ErrShortC)
}
return Check(C.cublasCherk(C.cublasHandle_t(h.h), C.cublasFillMode_t(uplo), C.cublasOperation_t(trans), C.int(n), C.int(k), (*C.float)(&alpha), (*C.cuComplex)(a), C.int(lda), (*C.float)(&beta), (*C.cuComplex)(c), C.int(ldc)))
}
// Zherk performs the Hermitian rank-k update C = alpha*A*A^H + beta*C or C = alpha*A^H*A + beta*C for double precision complex matrices.
// C is a Hermitian matrix. Note that alpha and beta are real scalars.
func Zherk(h *Handle, uplo Uplo, trans Transpose, n, k int, alpha float64, a cudart.DevicePtr, lda int, beta float64, c cudart.DevicePtr, ldc int) error {
if n < 0 {
panic(blas.ErrNLT0)
}
if k < 0 {
panic(blas.ErrKLT0)
}
if lda < 1 {
panic(blas.ErrShortA)
}
if ldc < 1 {
panic(blas.ErrShortC)
}
return Check(C.cublasZherk(C.cublasHandle_t(h.h), C.cublasFillMode_t(uplo), C.cublasOperation_t(trans), C.int(n), C.int(k), (*C.double)(&alpha), (*C.cuDoubleComplex)(a), C.int(lda), (*C.double)(&beta), (*C.cuDoubleComplex)(c), C.int(ldc)))
}
// Cher2k performs the Hermitian rank-2k update C = alpha*(A*B^H + B*A^H) + beta*C or C = alpha*(A^H*B + B^H*A) + beta*C for single precision complex matrices.
// C is a Hermitian matrix. Note that beta is a real scalar.
func Cher2k(h *Handle, uplo Uplo, trans Transpose, n, k int, alpha complex64, a cudart.DevicePtr, lda int, b cudart.DevicePtr, ldb int, beta float32, c cudart.DevicePtr, ldc int) error {
if n < 0 {
panic(blas.ErrNLT0)
}
if k < 0 {
panic(blas.ErrKLT0)
}
if lda < 1 {
panic(blas.ErrShortA)
}
if ldb < 1 {
panic(blas.ErrShortB)
}
if ldc < 1 {
panic(blas.ErrShortC)
}
return Check(C.cublasCher2k(C.cublasHandle_t(h.h), C.cublasFillMode_t(uplo), C.cublasOperation_t(trans), C.int(n), C.int(k), (*C.cuComplex)(unsafe.Pointer(&alpha)), (*C.cuComplex)(a), C.int(lda), (*C.cuComplex)(b), C.int(ldb), (*C.float)(&beta), (*C.cuComplex)(c), C.int(ldc)))
}
// Zher2k performs the Hermitian rank-2k update C = alpha*(A*B^H + B*A^H) + beta*C or C = alpha*(A^H*B + B^H*A) + beta*C for double precision complex matrices.
// C is a Hermitian matrix. Note that beta is a real scalar.
func Zher2k(h *Handle, uplo Uplo, trans Transpose, n, k int, alpha complex128, a cudart.DevicePtr, lda int, b cudart.DevicePtr, ldb int, beta float64, c cudart.DevicePtr, ldc int) error {
if n < 0 {
panic(blas.ErrNLT0)
}
if k < 0 {
panic(blas.ErrKLT0)
}
if lda < 1 {
panic(blas.ErrShortA)
}
if ldb < 1 {
panic(blas.ErrShortB)
}
if ldc < 1 {
panic(blas.ErrShortC)
}
return Check(C.cublasZher2k(C.cublasHandle_t(h.h), C.cublasFillMode_t(uplo), C.cublasOperation_t(trans), C.int(n), C.int(k), (*C.cuDoubleComplex)(unsafe.Pointer(&alpha)), (*C.cuDoubleComplex)(a), C.int(lda), (*C.cuDoubleComplex)(b), C.int(ldb), (*C.double)(&beta), (*C.cuDoubleComplex)(c), C.int(ldc)))
}
// Cherkx performs the Hermitian rank-kx update C = alpha*A*B^H + beta*C or C = alpha*A^H*B + beta*C for single precision complex matrices.
// This is a generalization of HERK where A and B can be different matrices. Note that beta is a real scalar.
func Cherkx(h *Handle, uplo Uplo, trans Transpose, n, k int, alpha complex64, a cudart.DevicePtr, lda int, b cudart.DevicePtr, ldb int, beta float32, c cudart.DevicePtr, ldc int) error {
if n < 0 {
panic(blas.ErrNLT0)
}
if k < 0 {
panic(blas.ErrKLT0)
}
if lda < 1 {
panic(blas.ErrShortA)
}
if ldb < 1 {
panic(blas.ErrShortB)
}
if ldc < 1 {
panic(blas.ErrShortC)
}
return Check(C.cublasCherkx(C.cublasHandle_t(h.h), C.cublasFillMode_t(uplo), C.cublasOperation_t(trans), C.int(n), C.int(k), (*C.cuComplex)(unsafe.Pointer(&alpha)), (*C.cuComplex)(a), C.int(lda), (*C.cuComplex)(b), C.int(ldb), (*C.float)(&beta), (*C.cuComplex)(c), C.int(ldc)))
}
// Zherkx performs the Hermitian rank-kx update C = alpha*A*B^H + beta*C or C = alpha*A^H*B + beta*C for double precision complex matrices.
// This is a generalization of HERK where A and B can be different matrices. Note that beta is a real scalar.
func Zherkx(h *Handle, uplo Uplo, trans Transpose, n, k int, alpha complex128, a cudart.DevicePtr, lda int, b cudart.DevicePtr, ldb int, beta float64, c cudart.DevicePtr, ldc int) error {
if n < 0 {
panic(blas.ErrNLT0)
}
if k < 0 {
panic(blas.ErrKLT0)
}
if lda < 1 {
panic(blas.ErrShortA)
}
if ldb < 1 {
panic(blas.ErrShortB)
}
if ldc < 1 {
panic(blas.ErrShortC)
}
return Check(C.cublasZherkx(C.cublasHandle_t(h.h), C.cublasFillMode_t(uplo), C.cublasOperation_t(trans), C.int(n), C.int(k), (*C.cuDoubleComplex)(unsafe.Pointer(&alpha)), (*C.cuDoubleComplex)(a), C.int(lda), (*C.cuDoubleComplex)(b), C.int(ldb), (*C.double)(&beta), (*C.cuDoubleComplex)(c), C.int(ldc)))
}