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
|
/* ---------------------------------------------------------------------
*
* -- PBLAS auxiliary 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 PB_CInV2( PBTYP_T * TYPE, char * CONJUG, char * ROWCOL, int M,
int N, int * DESCA, int K, char * X, int IX, int JX,
int * DESCX, char * XROC, char * XAPTR, int IJXA,
int * DXA )
#else
void PB_CInV2( TYPE, CONJUG, ROWCOL, M, N, DESCA, K, X, IX, JX, DESCX,
XROC, XAPTR, IJXA, DXA )
/*
* .. Scalar Arguments ..
*/
char * CONJUG, * ROWCOL, * XROC;
int IJXA, IX, JX, K, M, N;
PBTYP_T * TYPE;
/*
* .. Array Arguments ..
*/
int * DESCA, * DESCX, * DXA;
char * X, * XAPTR;
#endif
{
/*
* Purpose
* =======
*
* PB_CInV2 adds data to an array that contains a one-dimensional input
* only subvector which is replicated over the rows or columns of a sub-
* matrix described by DESCA. A subvector is specified on input to this
* routine that is added to the replicated buffer. This routine is spe-
* cifically designed for LCM hybrid variants.
*
* 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
* =========
*
* TYPE (local input) pointer to a PBTYP_T structure
* On entry, TYPE is a pointer to a structure of type PBTYP_T,
* that contains type information (See pblas.h).
*
* CONJUG (global input) pointer to CHAR
* On entry, CONJUG specifies if this routine should conjugate
* the subvector as follows:
* = 'N' or 'n': The initial subvector is copied,
* = 'Z' or 'z': The conjugate subvector is copied.
*
* ROWCOL (global input) pointer to CHAR
* On entry, ROWCOL specifies if the existing buffer pointed to
* XAPTR is a row or column subvector replicated over the under-
* lying submatrix as follows:
* = 'R' or 'r': XAPTR is a row subvector,
* = 'C' or 'c': XAPTR is a column subvector.
*
* M (global input) INTEGER
* On entry, M specifies the number of rows of the underlying
* submatrix described by DESCA. M must be at least zero.
*
* N (global input) INTEGER
* On entry, N specifies the number of columns of the underlying
* submatrix described by DESCA. N must be at least zero.
*
* DESCA (global and local input) INTEGER array
* On entry, DESCA is an integer array of dimension DLEN_. This
* is the array descriptor for the matrix A.
*
* K (global input) INTEGER
* On entry, K specifies the length of the non-distributed di-
* mension of the subvector sub( X ). K must be at least zero.
*
* X (local input) pointer to CHAR
* On entry, X is an array of dimension (LLD_X, Kx), where LLD_X
* is at least MAX( 1, Lr( K, IX ) ) when XROC is 'R' or 'r'
* and MAX( 1, Lr( 1, IX+Lx-1 ) ) otherwise, and, Kx is at least
* Lc( 1, JX+Lx-1 ) when INCX = M_X and Lc( K, JX ) otherwise.
* Lx is N when ROWCOL = 'R' or 'r' and M otherwise. Before en-
* try, 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.
*
* XROC (global input) pointer to CHAR
* On entry, XROC specifies the orientation of the subvector
* sub( X ). When XROC is 'R' or 'r', sub( X ) is a row vector,
* and a column vector otherwise.
*
* XAPTR (local input/local output) pointer to CHAR
* On entry, XAPTR is an array containing some initial data. On
* exit, the subvector sub( X ) is copied into this array which
* is replicated over the rows or columns of the underlying ma-
* trix as specified by ROWCOL and DESCA.
*
* IJXA (global input) INTEGER
* On entry, IJXA specifies XA global row or column index depen-
* ding on ROWCOL in the array pointed to by XAPTR, where the
* subvector sub( X ) should copied.
*
* DXA (global and local input) INTEGER array
* On entry, DXA is a descriptor array of dimension DLEN_ des-
* cribing the data layout of the data pointed to by XAPTR.
*
* -- Written on April 1, 1998 by
* Antoine Petitet, University of Tennessee, Knoxville 37996, USA.
*
* ---------------------------------------------------------------------
*/
/*
* .. Local Scalars ..
*/
char * Xptr = NULL, * top;
int AColSpan, ARowSpan, Acol, Aimb, Ainb, AisD, Amb, Amp, Anb,
Anq, Arow, XAld, Xcol, Xii, Ximb1, Xinb1, XisD, XisR, XisRow,
Xjj, Xld=1, Xmb, Xnb, Xrow, ctxt, mycol, myrow, npcol, nprow,
size;
/* ..
* .. Executable Statements ..
*
*/
/*
* Quick return if possible
*/
if( ( M <= 0 ) || ( N <= 0 ) || ( K <= 0 ) ) return;
/*
* Retrieve process grid information
*/
Cblacs_gridinfo( ( ctxt = DESCX[CTXT_] ), &nprow, &npcol, &myrow, &mycol );
/*
* Retrieve sub( X )'s local information: Xii, Xjj, Xrow, Ycol
*/
Minfog2l( IX, JX, DESCX, nprow, npcol, myrow, mycol, Xii, Xjj, Xrow, Xcol );
/*
* Is sub( X ) distributed or not, replicated or not ?
*/
if( ( XisRow = ( Mupcase( XROC[0] ) == CROW ) ) != 0 )
{
XisD = ( ( Xcol >= 0 ) && ( npcol > 1 ) );
XisR = ( ( Xrow == -1 ) || ( nprow == 1 ) );
}
else
{
XisD = ( ( Xrow >= 0 ) && ( nprow > 1 ) );
XisR = ( ( Xcol == -1 ) || ( npcol == 1 ) );
}
Arow = DESCA[ RSRC_ ]; Acol = DESCA[ CSRC_ ];
if( Mupcase( ROWCOL[0] ) == CROW )
{
/*
* Want a row vector. It is possible to reuse sub( X ) iff sub( X ) is already
* a row vector and the data does not need to be conjugated.
*/
if( XisRow && ( Mupcase( CONJUG[0] ) == CNOCONJG ) )
{
AisD = ( ( Acol >= 0 ) && ( npcol > 1 ) );
Ainb = DESCA[INB_]; Anb = DESCA[NB_]; Xnb = DESCX[NB_];
Mfirstnb( Xinb1, N, JX, DESCX[INB_], Xnb );
/*
* sub( X ) is aligned with A (reuse condition) iff both operands are not
* distributed, or both of them are distributed and start in the same process
* column and either N is smaller than the first blocksize of sub( X ) and A,
* or their column blocking factors match.
*/
if( ( !AisD && !XisD ) ||
( ( AisD && XisD ) &&
( ( Acol == Xcol ) &&
( ( ( Ainb >= N ) && ( Xinb1 >= N ) ) ||
( ( Ainb == Xinb1 ) && ( Anb == Xnb ) ) ) ) ) )
{
/*
* sub( X ) is aligned with A. Does A spans multiples process rows ? It does
* if Arow < 0.
*/
ARowSpan = ( Arow < 0 ) ||
Mspan( M, 0, DESCA[IMB_], DESCA[MB_], Arow, nprow );
Mnumroc( Anq, N, 0, Ainb, Anb, mycol, Acol, npcol );
if( XisR || ( !ARowSpan && ( Arow == Xrow ) ) )
{
/*
* If sub( X ) is replicated, or, A spans only one process row and either
* sub( X ) is replicated or resides in the same process row than A, then
* sub( X ) is already at the correct place.
*/
if( ( Anq > 0 ) && ( ARowSpan || ( myrow == Arow ) ) )
{
size = TYPE->size; Xld = DESCX[ LLD_ ]; XAld = DXA[LLD_];
TYPE->Fmmadd( &K, &Anq, TYPE->one, Mptr( X, Xii, Xjj, Xld,
size ), &Xld, TYPE->zero, Mptr( XAPTR, IJXA,
0, XAld, size ), &XAld );
}
}
else if( ARowSpan )
{
/*
* Otherwise, we know that sub( X ) cannot be replicated, let suppose in
* addition that A spans all process rows. sub( X ) need simply to be broadcast
* over A.
*/
if( myrow == Xrow )
{
if( Anq > 0 )
{
top = PB_Ctop( &ctxt, BCAST, COLUMN, TOP_GET );
size = TYPE->size; Xld = DESCX[LLD_]; XAld = DXA[LLD_];
Xptr = Mptr( XAPTR, IJXA, 0, XAld, size );
TYPE->Fmmadd( &K, &Anq, TYPE->one, Mptr( X, Xii, Xjj, Xld,
size ), &Xld, TYPE->zero, Xptr, &XAld );
TYPE->Cgebs2d( ctxt, COLUMN, top, K, Anq, Xptr, XAld );
}
}
else
{
if( Anq > 0 )
{
top = PB_Ctop( &ctxt, BCAST, COLUMN, TOP_GET );
XAld = DXA[LLD_];
TYPE->Cgebr2d( ctxt, COLUMN, top, K, Anq, Mptr( XAPTR,
IJXA, 0, XAld, TYPE->size ), XAld, Xrow,
mycol );
}
}
}
else
{
/*
* Finally, sub( X ) is not replicated and A spans only one process row. There
* is no need to broadcast, a send/recv is sufficient.
*/
if( myrow == Xrow )
{
if( Anq > 0 )
{
Xld = DESCX[LLD_];
TYPE->Cgesd2d( ctxt, K, Anq, Mptr( X, Xii, Xjj, Xld,
TYPE->size ), Xld, Arow, mycol );
}
}
else if( myrow == Arow )
{
if( Anq > 0 )
{
XAld = DXA[LLD_];
TYPE->Cgerv2d( ctxt, K, Anq, Mptr( XAPTR, IJXA, 0, XAld,
TYPE->size ), XAld, Xrow, mycol );
}
}
}
return;
}
}
/*
* sub( X ) cannot be reused, too bad ... redistribute
*/
if( XisRow )
{
PB_Cpaxpby( TYPE, CONJUG, K, N, TYPE->one, X, IX, JX, DESCX, XROC,
TYPE->zero, XAPTR, IJXA, 0, DXA, ROW );
}
else
{
PB_Cpaxpby( TYPE, CONJUG, N, K, TYPE->one, X, IX, JX, DESCX, XROC,
TYPE->zero, XAPTR, IJXA, 0, DXA, ROW );
}
}
else
{
/*
* Want a column vector. It is possible to reuse sub( X ) iff sub( X ) is
* already a column vector and the data does not need to be conjugated
*/
if( !( XisRow ) && ( Mupcase( CONJUG[0] ) == CNOCONJG ) )
{
AisD = ( ( Arow >= 0 ) && ( nprow > 1 ) );
Aimb = DESCA[IMB_]; Amb = DESCA[MB_]; Xmb = DESCX[MB_];
Mfirstnb( Ximb1, M, IX, DESCX[IMB_], Xmb );
/*
* sub( X ) is aligned with A (reuse condition) iff both operands are not
* distributed, or both of them are distributed and start in the same process
* row and either M is smaller than the first blocksize of sub( X ) and A, or
* their row blocking factors match.
*/
if( ( !AisD && !XisD ) ||
( ( AisD && XisD ) &&
( ( Arow == Xrow ) &&
( ( ( Aimb >= M ) && ( Ximb1 >= M ) ) ||
( ( Aimb == Ximb1 ) && ( Amb == Xmb ) ) ) ) ) )
{
/*
* sub( X ) is aligned with A. Does A spans multiples process columns ? It
* does if Acol < 0.
*/
AColSpan = ( Acol < 0 ) ||
Mspan( N, 0, DESCA[INB_], DESCA[NB_], Acol, npcol );
Mnumroc( Amp, M, 0, Aimb, Amb, myrow, Arow, nprow );
if( XisR || ( !AColSpan && ( Acol == Xcol ) ) )
{
/*
* If sub( X ) is replicated, or, A spans only one process column and either
* sub( X ) is replicated or resides in the same process columns than A, then
* sub( X ) is already at the correct place.
*/
if( ( Amp > 0 ) && ( AColSpan || ( mycol == Acol ) ) )
{
size = TYPE->size; Xld = DESCX[ LLD_ ]; XAld = DXA[LLD_];
TYPE->Fmmadd( &Amp, &K, TYPE->one, Mptr( X, Xii, Xjj, Xld,
size ), &Xld, TYPE->zero, Mptr( XAPTR, 0, IJXA,
XAld, size ), &XAld );
}
}
else if( AColSpan )
{
/*
* Otherwise, we know that sub( X ) is not be replicated, let suppose in
* addition that A spans all process columns. sub( X ) need simply to be
* broadcast over A.
*/
if( mycol == Xcol )
{
if( Amp > 0 )
{
top = PB_Ctop( &ctxt, BCAST, ROW, TOP_GET );
size = TYPE->size; Xld = DESCX[LLD_]; XAld = DXA[LLD_];
Xptr = Mptr( XAPTR, 0, IJXA, XAld, size );
TYPE->Fmmadd( &Amp, &K, TYPE->one, Mptr( X, Xii, Xjj, Xld,
size ), &Xld, TYPE->zero, Xptr, &XAld );
TYPE->Cgebs2d( ctxt, ROW, top, Amp, K, Xptr, XAld );
}
}
else
{
if( Amp > 0 )
{
top = PB_Ctop( &ctxt, BCAST, ROW, TOP_GET );
XAld = DXA[LLD_];
TYPE->Cgebr2d( ctxt, ROW, top, Amp, K, Mptr( XAPTR, 0,
IJXA, XAld, TYPE->size ), XAld, myrow,
Xcol );
}
}
}
else
{
/*
* Finally, sub( X ) is not replicated and A spans only one process column.
* There is no need to broadcast, a send/recv is sufficient.
*/
if( mycol == Xcol )
{
if( Amp > 0 )
{
Xld = DESCX[LLD_];
TYPE->Cgesd2d( ctxt, Amp, K, Mptr( X, Xii, Xjj, Xld,
TYPE->size ), Xld, myrow, Acol );
}
}
else if( mycol == Acol )
{
if( Amp > 0 )
{
XAld = DXA[LLD_];
TYPE->Cgerv2d( ctxt, Amp, K, Mptr( XAPTR, 0, IJXA, XAld,
TYPE->size ), XAld, myrow, Xcol );
}
}
}
return;
}
}
/*
* sub( X ) cannot be reused, too bad ... redistribute
*/
if( XisRow )
{
PB_Cpaxpby( TYPE, CONJUG, K, M, TYPE->one, X, IX, JX, DESCX, XROC,
TYPE->zero, XAPTR, 0, IJXA, DXA, COLUMN );
}
else
{
PB_Cpaxpby( TYPE, CONJUG, M, K, TYPE->one, X, IX, JX, DESCX, XROC,
TYPE->zero, XAPTR, 0, IJXA, DXA, COLUMN );
}
}
/*
* End of PB_CInV2
*/
}
|
