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
|
/* ---------------------------------------------------------------------
*
* -- PBLAS routine (version 2.0) --
* University of Tennessee, Knoxville, Oak Ridge National Laboratory,
* and University of California, Berkeley.
* April 1, 1998
*
* ---------------------------------------------------------------------
*/
/*
* Include files
*/
#include "pblas.h"
#include "PBpblas.h"
#include "PBtools.h"
#include "PBblacs.h"
#include "PBblas.h"
#ifdef __STDC__
void pcdotu_( Int * N,
float * DOT,
float * X, Int * IX, Int * JX, Int * DESCX, Int * INCX,
float * Y, Int * IY, Int * JY, Int * DESCY, Int * INCY )
#else
void pcdotu_( N, DOT, X, IX, JX, DESCX, INCX, Y, IY, JY, DESCY, INCY )
/*
* .. Scalar Arguments ..
*/
Int * INCX, * INCY, * IX, * IY, * JX, * JY, * N;
float * DOT;
/*
* .. Array Arguments ..
*/
Int * DESCX, * DESCY;
float * X, * Y;
#endif
{
/*
* Purpose
* =======
*
* PCDOTU forms the dot product of two subvectors,
*
* DOT := sub( X )**T * sub( Y ),
*
* where
*
* sub( X ) denotes X(IX,JX:JX+N-1) if INCX = M_X,
* X(IX:IX+N-1,JX) if INCX = 1 and INCX <> M_X, and,
*
* sub( Y ) denotes Y(IY,JY:JY+N-1) if INCY = M_Y,
* Y(IY:IY+N-1,JY) if INCY = 1 and INCY <> M_Y.
*
* Notes
* =====
*
* A description vector is associated with each 2D block-cyclicly dis-
* tributed matrix. This vector stores the information required to
* establish the mapping between a matrix entry and its corresponding
* process and memory location.
*
* In the following comments, the character _ should be read as
* "of the distributed matrix". Let A be a generic term for any 2D
* block cyclicly distributed matrix. Its description vector is DESC_A:
*
* NOTATION STORED IN EXPLANATION
* ---------------- --------------- ------------------------------------
* DTYPE_A (global) DESCA[ DTYPE_ ] The descriptor type.
* CTXT_A (global) DESCA[ CTXT_ ] The BLACS context handle, indicating
* the NPROW x NPCOL BLACS process grid
* A is distributed over. The context
* itself is global, but the handle
* (the integer value) may vary.
* M_A (global) DESCA[ M_ ] The number of rows in the distribu-
* ted matrix A, M_A >= 0.
* N_A (global) DESCA[ N_ ] The number of columns in the distri-
* buted matrix A, N_A >= 0.
* IMB_A (global) DESCA[ IMB_ ] The number of rows of the upper left
* block of the matrix A, IMB_A > 0.
* INB_A (global) DESCA[ INB_ ] The number of columns of the upper
* left block of the matrix A,
* INB_A > 0.
* MB_A (global) DESCA[ MB_ ] The blocking factor used to distri-
* bute the last M_A-IMB_A rows of A,
* MB_A > 0.
* NB_A (global) DESCA[ NB_ ] The blocking factor used to distri-
* bute the last N_A-INB_A columns of
* A, NB_A > 0.
* RSRC_A (global) DESCA[ RSRC_ ] The process row over which the first
* row of the matrix A is distributed,
* NPROW > RSRC_A >= 0.
* CSRC_A (global) DESCA[ CSRC_ ] The process column over which the
* first column of A is distributed.
* NPCOL > CSRC_A >= 0.
* LLD_A (local) DESCA[ LLD_ ] The leading dimension of the local
* array storing the local blocks of
* the distributed matrix A,
* IF( Lc( 1, N_A ) > 0 )
* LLD_A >= MAX( 1, Lr( 1, M_A ) )
* ELSE
* LLD_A >= 1.
*
* Let K be the number of rows of a matrix A starting at the global in-
* dex IA,i.e, A( IA:IA+K-1, : ). Lr( IA, K ) denotes the number of rows
* that the process of row coordinate MYROW ( 0 <= MYROW < NPROW ) would
* receive if these K rows were distributed over NPROW processes. If K
* is the number of columns of a matrix A starting at the global index
* JA, i.e, A( :, JA:JA+K-1, : ), Lc( JA, K ) denotes the number of co-
* lumns that the process MYCOL ( 0 <= MYCOL < NPCOL ) would receive if
* these K columns were distributed over NPCOL processes.
*
* The values of Lr() and Lc() may be determined via a call to the func-
* tion PB_Cnumroc:
* Lr( IA, K ) = PB_Cnumroc( K, IA, IMB_A, MB_A, MYROW, RSRC_A, NPROW )
* Lc( JA, K ) = PB_Cnumroc( K, JA, INB_A, NB_A, MYCOL, CSRC_A, NPCOL )
*
* Arguments
* =========
*
* N (global input) INTEGER
* On entry, N specifies the length of the subvectors to be
* multiplied. N must be at least zero.
*
* DOT (local output) COMPLEX array
* On exit, DOT specifies the dot product of the two subvectors
* sub( X ) and sub( Y ) only in their scope (See below for fur-
* ther details).
*
* X (local input) COMPLEX array
* On entry, X is an array of dimension (LLD_X, Kx), where LLD_X
* is at least MAX( 1, Lr( 1, IX ) ) when INCX = M_X and
* MAX( 1, Lr( 1, IX+N-1 ) ) otherwise, and, Kx is at least
* Lc( 1, JX+N-1 ) when INCX = M_X and Lc( 1, JX ) otherwise.
* Before entry, this array contains the local entries of the
* matrix X.
*
* IX (global input) INTEGER
* On entry, IX specifies X's global row index, which points to
* the beginning of the submatrix sub( X ).
*
* JX (global input) INTEGER
* On entry, JX specifies X's global column index, which points
* to the beginning of the submatrix sub( X ).
*
* DESCX (global and local input) INTEGER array
* On entry, DESCX is an integer array of dimension DLEN_. This
* is the array descriptor for the matrix X.
*
* INCX (global input) INTEGER
* On entry, INCX specifies the global increment for the
* elements of X. Only two values of INCX are supported in
* this version, namely 1 and M_X. INCX must not be zero.
*
* Y (local input) COMPLEX array
* On entry, Y is an array of dimension (LLD_Y, Ky), where LLD_Y
* is at least MAX( 1, Lr( 1, IY ) ) when INCY = M_Y and
* MAX( 1, Lr( 1, IY+N-1 ) ) otherwise, and, Ky is at least
* Lc( 1, JY+N-1 ) when INCY = M_Y and Lc( 1, JY ) otherwise.
* Before entry, this array contains the local entries of the
* matrix Y.
*
* IY (global input) INTEGER
* On entry, IY specifies Y's global row index, which points to
* the beginning of the submatrix sub( Y ).
*
* JY (global input) INTEGER
* On entry, JY specifies Y's global column index, which points
* to the beginning of the submatrix sub( Y ).
*
* DESCY (global and local input) INTEGER array
* On entry, DESCY is an integer array of dimension DLEN_. This
* is the array descriptor for the matrix Y.
*
* INCY (global input) INTEGER
* On entry, INCY specifies the global increment for the
* elements of Y. Only two values of INCY are supported in
* this version, namely 1 and M_Y. INCY must not be zero.
*
* Further Details
* ===============
*
* When the result of a vector-oriented PBLAS call is a scalar, this
* scalar is set only within the process scope which owns the vector(s)
* being operated on. Let sub( X ) be a generic term for the input vec-
* tor(s). Then, the processes owning the correct the answer is determi-
* ned as follows: if an operation involves more than one vector, the
* processes receiving the result will be the union of the following set
* of processes for each vector:
*
* If N = 1, M_X = 1 and INCX = 1, then one cannot determine if a pro-
* cess row or process column owns the vector operand, therefore only
* the process owning sub( X ) receives the correct result;
*
* If INCX = M_X, then sub( X ) is a vector distributed over a process
* row. Each process in this row receives the result;
*
* If INCX = 1, then sub( X ) is a vector distributed over a process
* column. Each process in this column receives the result;
*
* -- Written on April 1, 1998 by
* Antoine Petitet, University of Tennessee, Knoxville 37996, USA.
*
* ---------------------------------------------------------------------
*/
/*
* .. Local Scalars ..
*/
char scope, * top;
Int OneBlock, OneDgrid, RRorCC, Square, Xcol, Xi, Xii, XinbD,
Xinb1D, XisD, XisR, XisRow, Xj, Xjj, Xld, Xlinc, XmyprocD,
XmyprocR, XnbD, XnpD, XnprocsD, XnprocsR, XprocD, XprocR,
Xrow, Ycol, Yi, Yii, YinbD, Yinb1D, YisD, YisR, YisRow, Yj,
Yjj, Yld, Ylinc, YmyprocD, YmyprocR, YnbD, YnpD, YnprocsD,
YnprocsR, YprocD, YprocR, Yrow, cdst, csrc, ctxt, dst, info,
ione=1, mycol, myrow, npcol, nprow, rdst, rsrc, size, src;
PBTYP_T * type;
VVDOT_T dot;
/*
* .. Local Arrays ..
*/
char * buf = NULL;
Int Xd[DLEN_], Yd[DLEN_], dbuf[ DLEN_ ];
/* ..
* .. Executable Statements ..
*
*/
PB_CargFtoC( *IX, *JX, DESCX, &Xi, &Xj, Xd );
PB_CargFtoC( *IY, *JY, DESCY, &Yi, &Yj, Yd );
#ifndef NO_ARGCHK
/*
* Test the input parameters
*/
Cblacs_gridinfo( ( ctxt = Xd[CTXT_] ), &nprow, &npcol, &myrow, &mycol );
if( !( info = ( ( nprow == -1 ) ? -( 601 + CTXT_ ) : 0 ) ) )
{
PB_Cchkvec( ctxt, "PCDOTU", "X", *N, 1, Xi, Xj, Xd, *INCX, 6, &info );
PB_Cchkvec( ctxt, "PCDOTU", "Y", *N, 1, Yi, Yj, Yd, *INCY, 11, &info );
}
if( info ) { PB_Cabort( ctxt, "PCDOTU", info ); return; }
#endif
DOT[REAL_PART] = ZERO;
DOT[IMAG_PART] = ZERO;
/*
* Quick return if possible
*/
if( *N == 0 ) return;
/*
* Handle degenerate case
*/
if( ( *N == 1 ) && ( ( Xd[ M_ ] == 1 ) || ( Yd[ M_ ] == 1 ) ) )
{
type = PB_Cctypeset();
PB_Cpdot11( type, *N, ((char *) DOT), ((char *) X), Xi, Xj, Xd, *INCX,
((char *) Y), Yi, Yj, Yd, *INCY, type->Fvvdotu );
return;
}
/*
* Start the operations
*/
#ifdef NO_ARGCHK
Cblacs_gridinfo( ( ctxt = Xd[ CTXT_ ] ), &nprow, &npcol, &myrow, &mycol );
#endif
/*
* Determine if sub( X ) is distributed or not
*/
if( ( XisRow = ( *INCX == Xd[M_] ) ) != 0 )
XisD = ( ( Xd[CSRC_] >= 0 ) && ( ( XnprocsD = npcol ) > 1 ) );
else
XisD = ( ( Xd[RSRC_] >= 0 ) && ( ( XnprocsD = nprow ) > 1 ) );
/*
* Determine if sub( Y ) is distributed or not
*/
if( ( YisRow = ( *INCY == Yd[M_] ) ) != 0 )
YisD = ( ( Yd[CSRC_] >= 0 ) && ( ( YnprocsD = npcol ) > 1 ) );
else
YisD = ( ( Yd[RSRC_] >= 0 ) && ( ( YnprocsD = nprow ) > 1 ) );
/*
* Are sub( X ) and sub( Y ) both row or column vectors ?
*/
RRorCC = ( ( XisRow && YisRow ) || ( !( XisRow ) && !( YisRow ) ) );
/*
* XisD && YisD <=> both vector operands are indeed distributed
*/
if( XisD && YisD )
{
/*
* Retrieve sub( X )'s local information: Xii, Xjj, Xrow, Xcol
*/
PB_Cinfog2l( Xi, Xj, Xd, nprow, npcol, myrow, mycol, &Xii, &Xjj,
&Xrow, &Xcol );
if( XisRow )
{
XinbD = Xd[INB_]; XnbD = Xd[NB_];
Xld = Xd[LLD_]; Xlinc = Xld;
XprocD = Xcol; XmyprocD = mycol;
XprocR = Xrow; XmyprocR = myrow; XnprocsR = nprow;
XisR = ( ( Xrow == -1 ) || ( XnprocsR == 1 ) );
Mfirstnb( Xinb1D, *N, Xj, XinbD, XnbD );
}
else
{
XinbD = Xd[IMB_]; XnbD = Xd[MB_];
Xld = Xd[LLD_]; Xlinc = 1;
XprocD = Xrow; XmyprocD = myrow;
XprocR = Xcol; XmyprocR = mycol; XnprocsR = npcol;
XisR = ( ( Xcol == -1 ) || ( XnprocsR == 1 ) );
Mfirstnb( Xinb1D, *N, Xi, XinbD, XnbD );
}
/*
* Retrieve sub( Y )'s local information: Yii, Yjj, Yrow, Ycol
*/
PB_Cinfog2l( Yi, Yj, Yd, nprow, npcol, myrow, mycol, &Yii, &Yjj,
&Yrow, &Ycol );
if( YisRow )
{
YinbD = Yd[INB_]; YnbD = Yd[NB_];
Yld = Yd[LLD_]; Ylinc = Yld;
YprocD = Ycol; YmyprocD = mycol;
YprocR = Yrow; YmyprocR = myrow; YnprocsR = nprow;
YisR = ( ( Yrow == -1 ) || ( YnprocsR == 1 ) );
Mfirstnb( Yinb1D, *N, Yj, YinbD, YnbD );
}
else
{
YinbD = Yd[IMB_]; YnbD = Yd[MB_];
Yld = Yd[LLD_]; Ylinc = 1;
YprocD = Yrow; YmyprocD = myrow;
YprocR = Ycol; YmyprocR = mycol; YnprocsR = npcol;
YisR = ( ( Ycol == -1 ) || ( YnprocsR == 1 ) );
Mfirstnb( Yinb1D, *N, Yi, YinbD, YnbD );
}
/*
* Do sub( X ) and sub( Y ) span more than one process ?
*/
OneDgrid = ( ( XnprocsD == 1 ) && ( YnprocsD == 1 ) );
OneBlock = ( ( Xinb1D >= *N ) && ( Yinb1D >= *N ) );
/*
* Are sub( X ) and sub( Y ) distributed in the same manner ?
*/
Square = ( ( Xinb1D == Yinb1D ) && ( XnbD == YnbD ) &&
( XnprocsD == YnprocsD ) );
if( !( XisR ) )
{
/*
* sub( X ) is not replicated
*/
if( YisR )
{
/*
* If sub( X ) is not replicated, but sub( Y ) is, a process row or column
* YprocR need to be selected. It will contain the non-replicated vector used
* to perform the dot product computation.
*/
if( RRorCC )
{
/*
* sub( X ) and sub( Y ) are both row or column vectors
*/
if( ( OneDgrid || OneBlock || Square ) && ( XprocD == YprocD ) )
{
/*
* sub( X ) and sub( Y ) start in the same process row or column XprocD=YprocD.
* Enforce a purely local operation by choosing YprocR to be equal to XprocR.
*/
YprocR = XprocR;
}
else
{
/*
* Otherwise, communication has to occur, so choose the next process row or
* column for YprocR to maximize the number of links, i.e reduce contention.
*/
YprocR = MModAdd1( XprocR, XnprocsR );
}
}
else
{
/*
* sub( X ) and sub( Y ) are distributed in orthogonal directions, what is
* chosen for YprocR does not really matter. Select the process origin.
*/
YprocR = XprocD;
}
}
else
{
/*
* Neither sub( X ) nor sub( Y ) are replicated. If I am not in process row or
* column XprocR and not in process row or column YprocR, then quick return.
*/
if( ( XmyprocR != XprocR ) && ( YmyprocR != YprocR ) )
return;
}
}
else
{
/*
* sub( X ) is distributed and replicated (so no quick return possible)
*/
if( YisR )
{
/*
* sub( Y ) is distributed and replicated as well
*/
if( RRorCC )
{
/*
* sub( X ) and sub( Y ) are both row or column vectors
*/
if( ( OneDgrid || OneBlock || Square ) && ( XprocD == YprocD ) )
{
/*
* sub( X ) and sub( Y ) start in the same process row or column XprocD=YprocD.
* Enforce a purely local operation by choosing XprocR and YprocR to be equal
* to zero.
*/
XprocR = YprocR = 0;
}
else
{
/*
* Otherwise, communication has to occur, so select YprocR to be zero and the
* next process row or column for XprocR in order to maximize the number of
* used links, i.e reduce contention.
*/
YprocR = 0;
XprocR = MModAdd1( YprocR, YnprocsR );
}
}
else
{
/*
* sub( X ) and sub( Y ) are distributed in orthogonal directions, select the
* origin processes.
*/
XprocR = YprocD;
YprocR = XprocD;
}
}
else
{
/*
* sub( Y ) is distributed, but not replicated
*/
if( RRorCC )
{
/*
* sub( X ) and sub( Y ) are both row or column vectors
*/
if( ( OneDgrid || OneBlock || Square ) && ( XprocD == YprocD ) )
{
/*
* sub( X ) and sub( Y ) start in the same process row or column XprocD=YprocD.
* Enforce a purely local operation by choosing XprocR to be equal to YprocR.
*/
XprocR = YprocR;
}
else
{
/*
* Otherwise, communication has to occur, so choose the next process row or
* column for XprocR to maximize the number of links, i.e reduce contention.
*/
XprocR = MModAdd1( YprocR, YnprocsR );
}
}
else
{
/*
* sub( X ) and sub( Y ) are distributed in orthogonal directions, what is
* chosen for XprocR does not really matter. Select the origin process.
*/
XprocR = YprocD;
}
}
}
/*
* Even if sub( X ) and/or sub( Y ) are replicated, only two process row or
* column are active, namely XprocR and YprocR. If any of those operands is
* replicated, broadcast will occur (unless there is an easy way out).
*/
type = PB_Cctypeset(); size = type->size; dot = type->Fvvdotu;
/*
* A purely operation occurs iff the operands start in the same process and if
* either the grid is mono-dimensional or there is a single local block to be
* operated with or if both operands are aligned.
*/
if( ( ( RRorCC && ( XprocD == YprocD ) && ( XprocR == YprocR ) ) ||
( !( RRorCC ) && ( XprocD == YprocR ) && ( XprocR == YprocD ) ) ) &&
( OneDgrid || OneBlock || ( RRorCC && Square ) ) )
{
if( ( !XisR && ( XmyprocR == XprocR ) &&
!YisR && ( YmyprocR == YprocR ) ) ||
( !XisR && YisR && ( YmyprocR == YprocR ) ) ||
( !YisR && XisR && ( XmyprocR == XprocR ) ) ||
( XisR && YisR ) )
{
XnpD = PB_Cnumroc( *N, 0, Xinb1D, XnbD, XmyprocD, XprocD,
XnprocsD );
YnpD = PB_Cnumroc( *N, 0, Yinb1D, YnbD, YmyprocD, YprocD,
YnprocsD );
if( ( XnpD > 0 ) && ( YnpD > 0 ) )
{
dot( &XnpD, ((char *) DOT), Mptr( ((char *) X), Xii, Xjj, Xld,
size ), &Xlinc, Mptr( ((char *) Y), Yii, Yjj, Yld, size ),
&Ylinc );
}
}
/*
* Combine the local results in sub( X )'s scope
*/
if( ( XisR && YisR ) || ( XmyprocR == XprocR ) )
{
scope = ( XisRow ? CROW : CCOLUMN );
top = PB_Ctop( &ctxt, COMBINE, &scope, TOP_GET );
Ccgsum2d( ctxt, &scope, top, 1, 1, ((char *) DOT), 1, -1, 0 );
}
if( RRorCC && XisR && YisR ) return;
}
else if( ( RRorCC && OneDgrid ) || OneBlock || Square )
{
/*
* Otherwise, it may be possible to compute the desired dot-product in a single
* message exchange iff the grid is mono-dimensional and the operands are
* distributed in the same direction, or there is just one block to be exchanged
* or if both operands are similarly distributed in their respective direction.
*/
if( YmyprocR == YprocR )
{
/*
* The processes owning a piece of sub( Y ) send it to the corresponding
* process owning s piece of sub ( X ).
*/
YnpD = PB_Cnumroc( *N, 0, Yinb1D, YnbD, YmyprocD, YprocD,
YnprocsD );
if( YnpD > 0 )
{
dst = XprocD + MModSub( YmyprocD, YprocD, YnprocsD );
dst = MPosMod( dst, XnprocsD );
if( XisRow ) { rdst = XprocR; cdst = dst; }
else { rdst = dst; cdst = XprocR; }
if( ( myrow == rdst ) && ( mycol == cdst ) )
{
dot( &YnpD, ((char *) DOT), Mptr( ((char *) X), Xii, Xjj, Xld,
size ), &Xlinc, Mptr( ((char *) Y), Yii, Yjj, Yld,
size ), &Ylinc );
}
else
{
if( YisRow )
Ccgesd2d( ctxt, 1, YnpD, Mptr( ((char *) Y), Yii, Yjj,
Yld, size ), Yld, rdst, cdst );
else
Ccgesd2d( ctxt, YnpD, 1, Mptr( ((char *) Y), Yii, Yjj,
Yld, size ), Yld, rdst, cdst );
}
}
}
if( XmyprocR == XprocR )
{
/*
* The processes owning a piece of sub( X ) receive the corresponding local
* piece of sub( Y ), compute the local dot product and combine the results
* within sub( X )'s scope.
*/
XnpD = PB_Cnumroc( *N, 0, Xinb1D, XnbD, XmyprocD, XprocD,
XnprocsD );
if( XnpD > 0 )
{
src = YprocD + MModSub( XmyprocD, XprocD, XnprocsD );
src = MPosMod( src, YnprocsD );
if( YisRow ) { rsrc = YprocR; csrc = src; }
else { rsrc = src; csrc = YprocR; }
if( ( myrow != rsrc ) || ( mycol != csrc ) )
{
buf = PB_Cmalloc( XnpD * size );
if( YisRow )
Ccgerv2d( ctxt, 1, XnpD, buf, 1, rsrc, csrc );
else
Ccgerv2d( ctxt, XnpD, 1, buf, XnpD, rsrc, csrc );
dot( &XnpD, ((char *) DOT), Mptr( ((char *) X), Xii, Xjj, Xld,
size ), &Xlinc, buf, &ione );
if( buf ) free( buf );
}
}
if( XisRow )
{
top = PB_Ctop( &ctxt, COMBINE, ROW, TOP_GET );
Ccgsum2d( ctxt, ROW, top, 1, 1, ((char*)DOT), 1, -1, 0 );
}
else
{
top = PB_Ctop( &ctxt, COMBINE, COLUMN, TOP_GET );
Ccgsum2d( ctxt, COLUMN, top, 1, 1, ((char*)DOT), 1, -1, 0 );
}
}
}
else
{
/*
* General case, copy sub( Y ) within sub( X )'s scope, compute the local
* results and combine them within sub( X )'s scope.
*/
XnpD = PB_Cnumroc( *N, 0, Xinb1D, XnbD, XmyprocD, XprocD, XnprocsD );
if( XisRow )
{
PB_Cdescset( dbuf, 1, *N, 1, Xinb1D, 1, XnbD, XprocR, XprocD, ctxt,
1 );
}
else
{
PB_Cdescset( dbuf, *N, 1, Xinb1D, 1, XnbD, 1, XprocD, XprocR, ctxt,
MAX( 1, XnpD ) );
}
if( ( XmyprocR == XprocR ) && ( XnpD > 0 ) )
buf = PB_Cmalloc( XnpD * size );
if( YisRow )
{
PB_Cpaxpby( type, NOCONJG, 1, *N, type->one, ((char *) Y), Yi, Yj,
Yd, ROW, type->zero, buf, 0, 0, dbuf, ( XisRow ? ROW :
COLUMN ) );
}
else
{
PB_Cpaxpby( type, NOCONJG, *N, 1, type->one, ((char *) Y), Yi, Yj,
Yd, COLUMN, type->zero, buf, 0, 0, dbuf, ( XisRow ?
ROW : COLUMN ) );
}
if( XmyprocR == XprocR )
{
if( XnpD > 0 )
{
dot( &XnpD, ((char *) DOT), Mptr( ((char *) X), Xii, Xjj, Xld,
size ), &Xlinc, buf, &ione );
if( buf ) free( buf );
}
if( XisRow )
{
top = PB_Ctop( &ctxt, COMBINE, ROW, TOP_GET );
Ccgsum2d( ctxt, ROW, top, 1, 1, ((char*)DOT), 1, -1, 0 );
}
else
{
top = PB_Ctop( &ctxt, COMBINE, COLUMN, TOP_GET );
Ccgsum2d( ctxt, COLUMN, top, 1, 1, ((char*)DOT), 1, -1, 0 );
}
}
}
/*
* Send the DOT product result within sub( Y )'s scope
*/
if( XisR || YisR )
{
/*
* Either sub( X ) or sub( Y ) are replicated, so that every process should have
* the result -> broadcast it orthogonally from sub( X )'s direction.
*/
if( XisRow )
{
top = PB_Ctop( &ctxt, BCAST, COLUMN, TOP_GET );
if( XmyprocR == XprocR )
Ccgebs2d( ctxt, COLUMN, top, 1, 1, ((char*)DOT), 1 );
else
Ccgebr2d( ctxt, COLUMN, top, 1, 1, ((char*)DOT), 1, XprocR,
XmyprocD );
}
else
{
top = PB_Ctop( &ctxt, BCAST, ROW, TOP_GET );
if( XmyprocR == XprocR )
Ccgebs2d( ctxt, ROW, top, 1, 1, ((char*)DOT), 1 );
else
Ccgebr2d( ctxt, ROW, top, 1, 1, ((char*)DOT), 1, XmyprocD,
XprocR );
}
}
else
{
/*
* Neither sub( X ) nor sub( Y ) are replicated
*/
if( RRorCC )
{
/*
* Both sub( X ) are distributed in the same direction -> the process row or
* column XprocR sends the result to the process row or column YprocR.
*/
if( XprocR != YprocR )
{
if( XmyprocR == XprocR )
{
if( XisRow )
Ccgesd2d( ctxt, 1, 1, ((char *) DOT), 1, YprocR,
YmyprocD );
else
Ccgesd2d( ctxt, 1, 1, ((char *) DOT), 1, YmyprocD,
YprocR );
}
else if( YmyprocR == YprocR )
{
if( XisRow )
Ccgerv2d( ctxt, 1, 1, ((char *) DOT), 1, XprocR,
XmyprocD );
else
Ccgerv2d( ctxt, 1, 1, ((char *) DOT), 1, XmyprocD,
XprocR );
}
}
}
else
{
/*
* Otherwise, the process at the intersection of sub( X )'s and sub( Y )'s
* scope, broadcast the result within sub( Y )'s scope.
*/
if( YmyprocR == YprocR )
{
if( YisRow )
{
top = PB_Ctop( &ctxt, BCAST, ROW, TOP_GET );
if( YmyprocD == XprocR )
Ccgebs2d( ctxt, ROW, top, 1, 1, ((char*)DOT), 1 );
else
Ccgebr2d( ctxt, ROW, top, 1, 1, ((char*)DOT), 1,
YprocR, XprocR );
}
else
{
top = PB_Ctop( &ctxt, BCAST, COLUMN, TOP_GET );
if( YmyprocD == XprocR )
Ccgebs2d( ctxt, COLUMN, top, 1, 1, ((char*)DOT), 1 );
else
Ccgebr2d( ctxt, COLUMN, top, 1, 1, ((char*)DOT), 1,
XprocR, YprocR );
}
}
}
}
}
else if( !( XisD ) && YisD )
{
/*
* sub( X ) is not distributed and sub( Y ) is distributed.
*/
type = PB_Cctypeset();
PB_CpdotND( type, *N, ((char *) DOT), ((char *) X), Xi, Xj, Xd, *INCX,
((char *) Y), Yi, Yj, Yd, *INCY, type->Fvvdotu );
}
else if( XisD && !( YisD ) )
{
/*
* sub( X ) is distributed and sub( Y ) is not distributed.
*/
type = PB_Cctypeset();
PB_CpdotND( type, *N, ((char *) DOT), ((char *) Y), Yi, Yj, Yd, *INCY,
((char *) X), Xi, Xj, Xd, *INCX, type->Fvvdotu );
}
else
{
/*
* Neither sub( X ) nor sub( Y ) are distributed
*/
type = PB_Cctypeset();
PB_CpdotNN( type, *N, ((char *) DOT), ((char *) X), Xi, Xj, Xd, *INCX,
((char *) Y), Yi, Yj, Yd, *INCY, type->Fvvdotu );
}
/*
* End of PCDOTU
*/
}
|
