SCALAPACK 2.2.2
LAPACK: Linear Algebra PACKage
Loading...
Searching...
No Matches
pslaschk.f
Go to the documentation of this file.
1 SUBROUTINE pslaschk( SYMM, DIAG, N, NRHS, X, IX, JX, DESCX,
2 $ IASEED, IA, JA, DESCA, IBSEED, ANORM, RESID,
3 $ WORK )
4*
5* -- ScaLAPACK auxiliary routine (version 1.7) --
6* University of Tennessee, Knoxville, Oak Ridge National Laboratory,
7* and University of California, Berkeley.
8* May 1, 1997
9*
10* .. Scalar Arguments ..
11 CHARACTER DIAG, SYMM
12 INTEGER IA, IASEED, IBSEED, IX, JA, JX, N, NRHS
13 REAL ANORM, RESID
14* ..
15* .. Array Arguments ..
16 INTEGER DESCA( * ), DESCX( * )
17 REAL WORK( * ), X( * )
18* ..
19*
20* Purpose
21* =======
22*
23* PSLASCHK computes the residual
24* || sub( A )*sub( X ) - B || / (|| sub( A ) ||*|| sub( X ) ||*eps*N)
25* to check the accuracy of the factorization and solve steps in the
26* LU and Cholesky decompositions, where sub( A ) denotes
27* A(IA:IA+N-1,JA,JA+N-1), sub( X ) denotes X(IX:IX+N-1, JX:JX+NRHS-1).
28*
29* Notes
30* =====
31*
32* Each global data object is described by an associated description
33* vector. This vector stores the information required to establish
34* the mapping between an object element and its corresponding process
35* and memory location.
36*
37* Let A be a generic term for any 2D block cyclicly distributed array.
38* Such a global array has an associated description vector DESCA.
39* In the following comments, the character _ should be read as
40* "of the global array".
41*
42* NOTATION STORED IN EXPLANATION
43* --------------- -------------- --------------------------------------
44* DTYPE_A(global) DESCA( DTYPE_ )The descriptor type. In this case,
45* DTYPE_A = 1.
46* CTXT_A (global) DESCA( CTXT_ ) The BLACS context handle, indicating
47* the BLACS process grid A is distribu-
48* ted over. The context itself is glo-
49* bal, but the handle (the integer
50* value) may vary.
51* M_A (global) DESCA( M_ ) The number of rows in the global
52* array A.
53* N_A (global) DESCA( N_ ) The number of columns in the global
54* array A.
55* MB_A (global) DESCA( MB_ ) The blocking factor used to distribute
56* the rows of the array.
57* NB_A (global) DESCA( NB_ ) The blocking factor used to distribute
58* the columns of the array.
59* RSRC_A (global) DESCA( RSRC_ ) The process row over which the first
60* row of the array A is distributed.
61* CSRC_A (global) DESCA( CSRC_ ) The process column over which the
62* first column of the array A is
63* distributed.
64* LLD_A (local) DESCA( LLD_ ) The leading dimension of the local
65* array. LLD_A >= MAX(1,LOCr(M_A)).
66*
67* Let K be the number of rows or columns of a distributed matrix,
68* and assume that its process grid has dimension p x q.
69* LOCr( K ) denotes the number of elements of K that a process
70* would receive if K were distributed over the p processes of its
71* process column.
72* Similarly, LOCc( K ) denotes the number of elements of K that a
73* process would receive if K were distributed over the q processes of
74* its process row.
75* The values of LOCr() and LOCc() may be determined via a call to the
76* ScaLAPACK tool function, NUMROC:
77* LOCr( M ) = NUMROC( M, MB_A, MYROW, RSRC_A, NPROW ),
78* LOCc( N ) = NUMROC( N, NB_A, MYCOL, CSRC_A, NPCOL ).
79* An upper bound for these quantities may be computed by:
80* LOCr( M ) <= ceil( ceil(M/MB_A)/NPROW )*MB_A
81* LOCc( N ) <= ceil( ceil(N/NB_A)/NPCOL )*NB_A
82*
83* Arguments
84* =========
85*
86* SYMM (global input) CHARACTER
87* if SYMM = 'S', sub( A ) is a symmetric distributed matrix,
88* otherwise sub( A ) is a general distributed matrix.
89*
90* DIAG (global input) CHARACTER
91* If DIAG = 'D', sub( A ) is diagonally dominant.
92*
93* N (global input) INTEGER
94* The number of columns to be operated on, i.e. the number of
95* columns of the distributed submatrix sub( A ). N >= 0.
96*
97* NRHS (global input) INTEGER
98* The number of right-hand-sides, i.e the number of columns
99* of the distributed matrix sub( X ). NRHS >= 0.
100*
101* X (local input) REAL pointer into the local memory
102* to an array of dimension (LLD_X,LOCc(JX+NRHS-1). This array
103* contains the local pieces of the answer vector(s) sub( X ) of
104* sub( A ) sub( X ) - B, split up over a column of processes.
105*
106* IX (global input) INTEGER
107* The row index in the global array X indicating the first
108* row of sub( X ).
109*
110* JX (global input) INTEGER
111* The column index in the global array X indicating the
112* first column of sub( X ).
113*
114* DESCX (global and local input) INTEGER array of dimension DLEN_.
115* The array descriptor for the distributed matrix X.
116*
117* IASEED (global input) INTEGER
118* The seed number to generate the original matrix Ao.
119*
120* IA (global input) INTEGER
121* The row index in the global array A indicating the first
122* row of sub( A ).
123*
124* JA (global input) INTEGER
125* The column index in the global array A indicating the
126* first column of sub( A ).
127*
128* DESCA (global and local input) INTEGER array of dimension DLEN_.
129* The array descriptor for the distributed matrix A.
130*
131* IBSEED (global input) INTEGER
132* The seed number to generate the original matrix B.
133*
134* ANORM (global input) REAL
135* The 1-norm or infinity norm of the distributed matrix
136* sub( A ).
137*
138* RESID (global output) REAL
139* The residual error:
140* ||sub( A )*sub( X )-B|| / (||sub( A )||*||sub( X )||*eps*N).
141*
142* WORK (local workspace) REAL array, dimension (LWORK)
143* LWORK >= MAX(1,Np)*NB_X + Nq*NB_X + MAX( MAX(NQ*MB_A,2*NB_X),
144* NB_X * NUMROC( NUMROC(N,MB_X,0,0,NPCOL), MB_X, 0, 0, LCMQ ) )
145*
146* =====================================================================
147*
148* .. Parameters ..
149 INTEGER BLOCK_CYCLIC_2D, CSRC_, CTXT_, DLEN_, DTYPE_,
150 $ LLD_, MB_, M_, NB_, N_, RSRC_
151 parameter( block_cyclic_2d = 1, dlen_ = 9, dtype_ = 1,
152 $ ctxt_ = 2, m_ = 3, n_ = 4, mb_ = 5, nb_ = 6,
153 $ rsrc_ = 7, csrc_ = 8, lld_ = 9 )
154 REAL ZERO, ONE
155 PARAMETER ( ONE = 1.0e+0, zero = 0.0e+0 )
156* ..
157* .. Local Scalars ..
158 INTEGER IACOL, IAROW, IB, ICOFF, ICTXT, ICURCOL, IDUMM,
159 $ II, IIA, IIX, IOFFX, IPA, IPB, IPW, IPX, IROFF,
160 $ ixcol, ixrow, j, jbrhs, jj, jja, jjx, ldx,
161 $ mycol, myrow, np, npcol, nprow, nq
162 REAL BETA, DIVISOR, EPS, RESID1
163* ..
164* .. External Subroutines ..
165 EXTERNAL blacs_gridinfo, pbstran, psmatgen,
166 $ sgamx2d, sgebr2d, sgebs2d, sgemm,
167 $ sgerv2d, sgesd2d, sgsum2d, slaset
168* ..
169* .. External Functions ..
170 INTEGER ISAMAX, NUMROC
171 REAL PSLAMCH
172 EXTERNAL isamax, numroc, pslamch
173* ..
174* .. Intrinsic Functions ..
175 INTRINSIC abs, max, min, mod, real
176* ..
177* .. Executable Statements ..
178*
179* Get needed initial parameters
180*
181 ictxt = desca( ctxt_ )
182 CALL blacs_gridinfo( ictxt, nprow, npcol, myrow, mycol )
183*
184 eps = pslamch( ictxt, 'eps' )
185 resid = 0.0e+0
186 divisor = anorm * eps * real( n )
187*
188 CALL infog2l( ia, ja, desca, nprow, npcol, myrow, mycol, iia, jja,
189 $ iarow, iacol )
190 CALL infog2l( ix, jx, descx, nprow, npcol, myrow, mycol, iix, jjx,
191 $ ixrow, ixcol )
192 iroff = mod( ia-1, desca( mb_ ) )
193 icoff = mod( ja-1, desca( nb_ ) )
194 np = numroc( n+iroff, desca( mb_ ), myrow, iarow, nprow )
195 nq = numroc( n+icoff, desca( nb_ ), mycol, iacol, npcol )
196*
197 ldx = max( 1, np )
198 ipb = 1
199 ipx = ipb + np * descx( nb_ )
200 ipa = ipx + nq * descx( nb_ )
201*
202 IF( myrow.EQ.iarow )
203 $ np = np - iroff
204 IF( mycol.EQ.iacol )
205 $ nq = nq - icoff
206*
207 icurcol = ixcol
208*
209* Loop over the rhs
210*
211 DO 40 j = 1, nrhs, descx( nb_ )
212 jbrhs = min( descx( nb_ ), nrhs-j+1 )
213*
214* Transpose x from ICURCOL to all rows
215*
216 ioffx = iix + ( jjx - 1 ) * descx( lld_ )
217 CALL pbstran( ictxt, 'Column', 'Transpose', n, jbrhs,
218 $ descx( mb_ ), x( ioffx ), descx( lld_ ), zero,
219 $ work( ipx ), jbrhs, ixrow, icurcol, -1, iacol,
220 $ work( ipa ) )
221*
222* Regenerate B in IXCOL
223*
224 IF( mycol.EQ.icurcol ) THEN
225 CALL psmatgen( ictxt, 'N', 'N', descx( m_ ), descx( n_ ),
226 $ descx( mb_ ), descx( nb_ ), work( ipb ), ldx,
227 $ ixrow, ixcol, ibseed, iix-1, np, jjx-1,
228 $ jbrhs, myrow, mycol, nprow, npcol )
229 beta = one
230 ELSE
231 beta = zero
232 END IF
233*
234 IF( nq.GT.0 ) THEN
235 DO 10 ii = iia, iia+np-1, desca( mb_ )
236 ib = min( desca( mb_ ), iia+np-ii )
237*
238* Regenerate ib rows of the matrix A(IA:IA+N-1,JA:JA+N-1).
239*
240 CALL psmatgen( ictxt, symm, diag, desca( m_ ),
241 $ desca( n_ ), desca( mb_ ), desca( nb_ ),
242 $ work( ipa ), ib, desca( rsrc_ ),
243 $ desca( csrc_ ), iaseed, ii-1, ib,
244 $ jja-1, nq, myrow, mycol, nprow, npcol )
245*
246* Compute B <= B - A * X.
247*
248 CALL sgemm( 'No transpose', 'Transpose', ib, jbrhs, nq,
249 $ -one, work( ipa ), ib, work( ipx ), jbrhs,
250 $ beta, work( ipb+ii-iia ), ldx )
251*
252 10 CONTINUE
253*
254 ELSE IF( mycol.NE.icurcol ) THEN
255*
256 CALL slaset( 'All', np, jbrhs, zero, zero, work( ipb ),
257 $ ldx )
258*
259 END IF
260*
261* Add B rowwise to ICURCOL
262*
263 CALL sgsum2d( ictxt, 'Row', ' ', np, jbrhs, work( ipb ), ldx,
264 $ myrow, icurcol )
265*
266 IF( mycol.EQ.icurcol ) THEN
267*
268* Figure || A * X - B || & || X ||
269*
270 ipw = ipa + jbrhs
271 DO 20 jj = 0, jbrhs - 1
272 IF( np.GT.0 ) THEN
273 ii = isamax( np, work( ipb+jj*ldx ), 1 )
274 work( ipa+jj ) = abs( work( ipb+ii-1+jj*ldx ) )
275 work( ipw+jj ) = abs( x( ioffx + isamax( np,
276 $ x( ioffx + jj*descx( lld_ ) ), 1 )-1+jj*
277 $ descx( lld_ ) ) )
278 ELSE
279 work( ipa+jj ) = zero
280 work( ipw+jj ) = zero
281 END IF
282 20 CONTINUE
283*
284* After SGAMX2D computation,
285* WORK(IPB) has the maximum of || Ax - b ||, and
286* WORK(IPX) has the maximum of || X ||.
287*
288 CALL sgamx2d( ictxt, 'Column', ' ', 1, 2*jbrhs,
289 $ work( ipa ), 1, idumm, idumm, -1, 0, icurcol )
290*
291* Calculate residual = ||Ax-b|| / (||x||*||A||*eps*N)
292*
293 IF( myrow.EQ.0 ) THEN
294 DO 30 jj = 0, jbrhs - 1
295 resid1 = work( ipa+jj ) / ( work( ipw+jj )*divisor )
296 IF( resid.LT.resid1 )
297 $ resid = resid1
298 30 CONTINUE
299 IF( mycol.NE.0 )
300 $ CALL sgesd2d( ictxt, 1, 1, resid, 1, 0, 0 )
301 END IF
302*
303 ELSE IF( myrow.EQ.0 .AND. mycol.EQ.0 ) THEN
304*
305 CALL sgerv2d( ictxt, 1, 1, resid1, 1, 0, icurcol )
306 IF( resid.LT.resid1 )
307 $ resid = resid1
308*
309 END IF
310*
311 IF( mycol.EQ.icurcol )
312 $ jjx = jjx + jbrhs
313 icurcol = mod( icurcol+1, npcol )
314*
315 40 CONTINUE
316*
317 IF( myrow.EQ.0 .AND. mycol.EQ.0 ) THEN
318 CALL sgebs2d( ictxt, 'All', ' ', 1, 1, resid, 1 )
319 ELSE
320 CALL sgebr2d( ictxt, 'All', ' ', 1, 1, resid, 1, 0, 0 )
321 END IF
322*
323 RETURN
324*
325* End of PSLASCHK
326*
327 END
subroutine psmatgen(ictxt, aform, diag, m, n, mb, nb, a, lda, iarow, iacol, iseed, iroff, irnum, icoff, icnum, myrow, mycol, nprow, npcol)
Definition psmatgen.f:4
subroutine infog2l(grindx, gcindx, desc, nprow, npcol, myrow, mycol, lrindx, lcindx, rsrc, csrc)
Definition infog2l.f:3
subroutine pbstran(icontxt, adist, trans, m, n, nb, a, lda, beta, c, ldc, iarow, iacol, icrow, iccol, work)
Definition pbstran.f:3
#define max(A, B)
Definition pcgemr.c:180
#define min(A, B)
Definition pcgemr.c:181
subroutine pslaschk(symm, diag, n, nrhs, x, ix, jx, descx, iaseed, ia, ja, desca, ibseed, anorm, resid, work)
Definition pslaschk.f:4