LAPACK  3.4.2
LAPACK: Linear Algebra PACKage
 All Files Functions Groups
cdrvpo.f
Go to the documentation of this file.
1 *> \brief \b CDRVPO
2 *
3 * =========== DOCUMENTATION ===========
4 *
5 * Online html documentation available at
6 * http://www.netlib.org/lapack/explore-html/
7 *
8 * Definition:
9 * ===========
10 *
11 * SUBROUTINE CDRVPO( DOTYPE, NN, NVAL, NRHS, THRESH, TSTERR, NMAX,
12 * A, AFAC, ASAV, B, BSAV, X, XACT, S, WORK,
13 * RWORK, NOUT )
14 *
15 * .. Scalar Arguments ..
16 * LOGICAL TSTERR
17 * INTEGER NMAX, NN, NOUT, NRHS
18 * REAL THRESH
19 * ..
20 * .. Array Arguments ..
21 * LOGICAL DOTYPE( * )
22 * INTEGER NVAL( * )
23 * REAL RWORK( * ), S( * )
24 * COMPLEX A( * ), AFAC( * ), ASAV( * ), B( * ),
25 * $ BSAV( * ), WORK( * ), X( * ), XACT( * )
26 * ..
27 *
28 *
29 *> \par Purpose:
30 * =============
31 *>
32 *> \verbatim
33 *>
34 *> CDRVPO tests the driver routines CPOSV and -SVX.
35 *> \endverbatim
36 *
37 * Arguments:
38 * ==========
39 *
40 *> \param[in] DOTYPE
41 *> \verbatim
42 *> DOTYPE is LOGICAL array, dimension (NTYPES)
43 *> The matrix types to be used for testing. Matrices of type j
44 *> (for 1 <= j <= NTYPES) are used for testing if DOTYPE(j) =
45 *> .TRUE.; if DOTYPE(j) = .FALSE., then type j is not used.
46 *> \endverbatim
47 *>
48 *> \param[in] NN
49 *> \verbatim
50 *> NN is INTEGER
51 *> The number of values of N contained in the vector NVAL.
52 *> \endverbatim
53 *>
54 *> \param[in] NVAL
55 *> \verbatim
56 *> NVAL is INTEGER array, dimension (NN)
57 *> The values of the matrix dimension N.
58 *> \endverbatim
59 *>
60 *> \param[in] NRHS
61 *> \verbatim
62 *> NRHS is INTEGER
63 *> The number of right hand side vectors to be generated for
64 *> each linear system.
65 *> \endverbatim
66 *>
67 *> \param[in] THRESH
68 *> \verbatim
69 *> THRESH is REAL
70 *> The threshold value for the test ratios. A result is
71 *> included in the output file if RESULT >= THRESH. To have
72 *> every test ratio printed, use THRESH = 0.
73 *> \endverbatim
74 *>
75 *> \param[in] TSTERR
76 *> \verbatim
77 *> TSTERR is LOGICAL
78 *> Flag that indicates whether error exits are to be tested.
79 *> \endverbatim
80 *>
81 *> \param[in] NMAX
82 *> \verbatim
83 *> NMAX is INTEGER
84 *> The maximum value permitted for N, used in dimensioning the
85 *> work arrays.
86 *> \endverbatim
87 *>
88 *> \param[out] A
89 *> \verbatim
90 *> A is COMPLEX array, dimension (NMAX*NMAX)
91 *> \endverbatim
92 *>
93 *> \param[out] AFAC
94 *> \verbatim
95 *> AFAC is COMPLEX array, dimension (NMAX*NMAX)
96 *> \endverbatim
97 *>
98 *> \param[out] ASAV
99 *> \verbatim
100 *> ASAV is COMPLEX array, dimension (NMAX*NMAX)
101 *> \endverbatim
102 *>
103 *> \param[out] B
104 *> \verbatim
105 *> B is COMPLEX array, dimension (NMAX*NRHS)
106 *> \endverbatim
107 *>
108 *> \param[out] BSAV
109 *> \verbatim
110 *> BSAV is COMPLEX array, dimension (NMAX*NRHS)
111 *> \endverbatim
112 *>
113 *> \param[out] X
114 *> \verbatim
115 *> X is COMPLEX array, dimension (NMAX*NRHS)
116 *> \endverbatim
117 *>
118 *> \param[out] XACT
119 *> \verbatim
120 *> XACT is COMPLEX array, dimension (NMAX*NRHS)
121 *> \endverbatim
122 *>
123 *> \param[out] S
124 *> \verbatim
125 *> S is REAL array, dimension (NMAX)
126 *> \endverbatim
127 *>
128 *> \param[out] WORK
129 *> \verbatim
130 *> WORK is COMPLEX array, dimension
131 *> (NMAX*max(3,NRHS))
132 *> \endverbatim
133 *>
134 *> \param[out] RWORK
135 *> \verbatim
136 *> RWORK is REAL array, dimension (NMAX+2*NRHS)
137 *> \endverbatim
138 *>
139 *> \param[in] NOUT
140 *> \verbatim
141 *> NOUT is INTEGER
142 *> The unit number for output.
143 *> \endverbatim
144 *
145 * Authors:
146 * ========
147 *
148 *> \author Univ. of Tennessee
149 *> \author Univ. of California Berkeley
150 *> \author Univ. of Colorado Denver
151 *> \author NAG Ltd.
152 *
153 *> \date November 2011
154 *
155 *> \ingroup complex_lin
156 *
157 * =====================================================================
158  SUBROUTINE cdrvpo( DOTYPE, NN, NVAL, NRHS, THRESH, TSTERR, NMAX,
159  $ a, afac, asav, b, bsav, x, xact, s, work,
160  $ rwork, nout )
161 *
162 * -- LAPACK test routine (version 3.4.0) --
163 * -- LAPACK is a software package provided by Univ. of Tennessee, --
164 * -- Univ. of California Berkeley, Univ. of Colorado Denver and NAG Ltd..--
165 * November 2011
166 *
167 * .. Scalar Arguments ..
168  LOGICAL tsterr
169  INTEGER nmax, nn, nout, nrhs
170  REAL thresh
171 * ..
172 * .. Array Arguments ..
173  LOGICAL dotype( * )
174  INTEGER nval( * )
175  REAL rwork( * ), s( * )
176  COMPLEX a( * ), afac( * ), asav( * ), b( * ),
177  $ bsav( * ), work( * ), x( * ), xact( * )
178 * ..
179 *
180 * =====================================================================
181 *
182 * .. Parameters ..
183  REAL one, zero
184  parameter( one = 1.0e+0, zero = 0.0e+0 )
185  INTEGER ntypes
186  parameter( ntypes = 9 )
187  INTEGER ntests
188  parameter( ntests = 6 )
189 * ..
190 * .. Local Scalars ..
191  LOGICAL equil, nofact, prefac, zerot
192  CHARACTER dist, equed, fact, type, uplo, xtype
193  CHARACTER*3 path
194  INTEGER i, iequed, ifact, imat, in, info, ioff, iuplo,
195  $ izero, k, k1, kl, ku, lda, mode, n, nb, nbmin,
196  $ nerrs, nfact, nfail, nimat, nrun, nt
197  REAL ainvnm, amax, anorm, cndnum, rcond, rcondc,
198  $ roldc, scond
199 * ..
200 * .. Local Arrays ..
201  CHARACTER equeds( 2 ), facts( 3 ), uplos( 2 )
202  INTEGER iseed( 4 ), iseedy( 4 )
203  REAL result( ntests )
204 * ..
205 * .. External Functions ..
206  LOGICAL lsame
207  REAL clanhe, sget06
208  EXTERNAL lsame, clanhe, sget06
209 * ..
210 * .. External Subroutines ..
211  EXTERNAL aladhd, alaerh, alasvm, cerrvx, cget04, clacpy,
212  $ claipd, claqhe, clarhs, claset, clatb4, clatms,
213  $ cpoequ, cposv, cposvx, cpot01, cpot02, cpot05,
214  $ cpotrf, cpotri, xlaenv
215 * ..
216 * .. Scalars in Common ..
217  LOGICAL lerr, ok
218  CHARACTER*32 srnamt
219  INTEGER infot, nunit
220 * ..
221 * .. Common blocks ..
222  common / infoc / infot, nunit, ok, lerr
223  common / srnamc / srnamt
224 * ..
225 * .. Intrinsic Functions ..
226  INTRINSIC cmplx, max
227 * ..
228 * .. Data statements ..
229  DATA iseedy / 1988, 1989, 1990, 1991 /
230  DATA uplos / 'U', 'L' /
231  DATA facts / 'F', 'N', 'E' /
232  DATA equeds / 'N', 'Y' /
233 * ..
234 * .. Executable Statements ..
235 *
236 * Initialize constants and the random number seed.
237 *
238  path( 1: 1 ) = 'Complex precision'
239  path( 2: 3 ) = 'PO'
240  nrun = 0
241  nfail = 0
242  nerrs = 0
243  DO 10 i = 1, 4
244  iseed( i ) = iseedy( i )
245  10 continue
246 *
247 * Test the error exits
248 *
249  IF( tsterr )
250  $ CALL cerrvx( path, nout )
251  infot = 0
252 *
253 * Set the block size and minimum block size for testing.
254 *
255  nb = 1
256  nbmin = 2
257  CALL xlaenv( 1, nb )
258  CALL xlaenv( 2, nbmin )
259 *
260 * Do for each value of N in NVAL
261 *
262  DO 130 in = 1, nn
263  n = nval( in )
264  lda = max( n, 1 )
265  xtype = 'N'
266  nimat = ntypes
267  IF( n.LE.0 )
268  $ nimat = 1
269 *
270  DO 120 imat = 1, nimat
271 *
272 * Do the tests only if DOTYPE( IMAT ) is true.
273 *
274  IF( .NOT.dotype( imat ) )
275  $ go to 120
276 *
277 * Skip types 3, 4, or 5 if the matrix size is too small.
278 *
279  zerot = imat.GE.3 .AND. imat.LE.5
280  IF( zerot .AND. n.LT.imat-2 )
281  $ go to 120
282 *
283 * Do first for UPLO = 'U', then for UPLO = 'L'
284 *
285  DO 110 iuplo = 1, 2
286  uplo = uplos( iuplo )
287 *
288 * Set up parameters with CLATB4 and generate a test matrix
289 * with CLATMS.
290 *
291  CALL clatb4( path, imat, n, n, type, kl, ku, anorm, mode,
292  $ cndnum, dist )
293 *
294  srnamt = 'CLATMS'
295  CALL clatms( n, n, dist, iseed, type, rwork, mode,
296  $ cndnum, anorm, kl, ku, uplo, a, lda, work,
297  $ info )
298 *
299 * Check error code from CLATMS.
300 *
301  IF( info.NE.0 ) THEN
302  CALL alaerh( path, 'CLATMS', info, 0, uplo, n, n, -1,
303  $ -1, -1, imat, nfail, nerrs, nout )
304  go to 110
305  END IF
306 *
307 * For types 3-5, zero one row and column of the matrix to
308 * test that INFO is returned correctly.
309 *
310  IF( zerot ) THEN
311  IF( imat.EQ.3 ) THEN
312  izero = 1
313  ELSE IF( imat.EQ.4 ) THEN
314  izero = n
315  ELSE
316  izero = n / 2 + 1
317  END IF
318  ioff = ( izero-1 )*lda
319 *
320 * Set row and column IZERO of A to 0.
321 *
322  IF( iuplo.EQ.1 ) THEN
323  DO 20 i = 1, izero - 1
324  a( ioff+i ) = zero
325  20 continue
326  ioff = ioff + izero
327  DO 30 i = izero, n
328  a( ioff ) = zero
329  ioff = ioff + lda
330  30 continue
331  ELSE
332  ioff = izero
333  DO 40 i = 1, izero - 1
334  a( ioff ) = zero
335  ioff = ioff + lda
336  40 continue
337  ioff = ioff - izero
338  DO 50 i = izero, n
339  a( ioff+i ) = zero
340  50 continue
341  END IF
342  ELSE
343  izero = 0
344  END IF
345 *
346 * Set the imaginary part of the diagonals.
347 *
348  CALL claipd( n, a, lda+1, 0 )
349 *
350 * Save a copy of the matrix A in ASAV.
351 *
352  CALL clacpy( uplo, n, n, a, lda, asav, lda )
353 *
354  DO 100 iequed = 1, 2
355  equed = equeds( iequed )
356  IF( iequed.EQ.1 ) THEN
357  nfact = 3
358  ELSE
359  nfact = 1
360  END IF
361 *
362  DO 90 ifact = 1, nfact
363  fact = facts( ifact )
364  prefac = lsame( fact, 'F' )
365  nofact = lsame( fact, 'N' )
366  equil = lsame( fact, 'E' )
367 *
368  IF( zerot ) THEN
369  IF( prefac )
370  $ go to 90
371  rcondc = zero
372 *
373  ELSE IF( .NOT.lsame( fact, 'N' ) ) THEN
374 *
375 * Compute the condition number for comparison with
376 * the value returned by CPOSVX (FACT = 'N' reuses
377 * the condition number from the previous iteration
378 * with FACT = 'F').
379 *
380  CALL clacpy( uplo, n, n, asav, lda, afac, lda )
381  IF( equil .OR. iequed.GT.1 ) THEN
382 *
383 * Compute row and column scale factors to
384 * equilibrate the matrix A.
385 *
386  CALL cpoequ( n, afac, lda, s, scond, amax,
387  $ info )
388  IF( info.EQ.0 .AND. n.GT.0 ) THEN
389  IF( iequed.GT.1 )
390  $ scond = zero
391 *
392 * Equilibrate the matrix.
393 *
394  CALL claqhe( uplo, n, afac, lda, s, scond,
395  $ amax, equed )
396  END IF
397  END IF
398 *
399 * Save the condition number of the
400 * non-equilibrated system for use in CGET04.
401 *
402  IF( equil )
403  $ roldc = rcondc
404 *
405 * Compute the 1-norm of A.
406 *
407  anorm = clanhe( '1', uplo, n, afac, lda, rwork )
408 *
409 * Factor the matrix A.
410 *
411  CALL cpotrf( uplo, n, afac, lda, info )
412 *
413 * Form the inverse of A.
414 *
415  CALL clacpy( uplo, n, n, afac, lda, a, lda )
416  CALL cpotri( uplo, n, a, lda, info )
417 *
418 * Compute the 1-norm condition number of A.
419 *
420  ainvnm = clanhe( '1', uplo, n, a, lda, rwork )
421  IF( anorm.LE.zero .OR. ainvnm.LE.zero ) THEN
422  rcondc = one
423  ELSE
424  rcondc = ( one / anorm ) / ainvnm
425  END IF
426  END IF
427 *
428 * Restore the matrix A.
429 *
430  CALL clacpy( uplo, n, n, asav, lda, a, lda )
431 *
432 * Form an exact solution and set the right hand side.
433 *
434  srnamt = 'CLARHS'
435  CALL clarhs( path, xtype, uplo, ' ', n, n, kl, ku,
436  $ nrhs, a, lda, xact, lda, b, lda,
437  $ iseed, info )
438  xtype = 'C'
439  CALL clacpy( 'Full', n, nrhs, b, lda, bsav, lda )
440 *
441  IF( nofact ) THEN
442 *
443 * --- Test CPOSV ---
444 *
445 * Compute the L*L' or U'*U factorization of the
446 * matrix and solve the system.
447 *
448  CALL clacpy( uplo, n, n, a, lda, afac, lda )
449  CALL clacpy( 'Full', n, nrhs, b, lda, x, lda )
450 *
451  srnamt = 'CPOSV '
452  CALL cposv( uplo, n, nrhs, afac, lda, x, lda,
453  $ info )
454 *
455 * Check error code from CPOSV .
456 *
457  IF( info.NE.izero ) THEN
458  CALL alaerh( path, 'CPOSV ', info, izero,
459  $ uplo, n, n, -1, -1, nrhs, imat,
460  $ nfail, nerrs, nout )
461  go to 70
462  ELSE IF( info.NE.0 ) THEN
463  go to 70
464  END IF
465 *
466 * Reconstruct matrix from factors and compute
467 * residual.
468 *
469  CALL cpot01( uplo, n, a, lda, afac, lda, rwork,
470  $ result( 1 ) )
471 *
472 * Compute residual of the computed solution.
473 *
474  CALL clacpy( 'Full', n, nrhs, b, lda, work,
475  $ lda )
476  CALL cpot02( uplo, n, nrhs, a, lda, x, lda,
477  $ work, lda, rwork, result( 2 ) )
478 *
479 * Check solution from generated exact solution.
480 *
481  CALL cget04( n, nrhs, x, lda, xact, lda, rcondc,
482  $ result( 3 ) )
483  nt = 3
484 *
485 * Print information about the tests that did not
486 * pass the threshold.
487 *
488  DO 60 k = 1, nt
489  IF( result( k ).GE.thresh ) THEN
490  IF( nfail.EQ.0 .AND. nerrs.EQ.0 )
491  $ CALL aladhd( nout, path )
492  WRITE( nout, fmt = 9999 )'CPOSV ', uplo,
493  $ n, imat, k, result( k )
494  nfail = nfail + 1
495  END IF
496  60 continue
497  nrun = nrun + nt
498  70 continue
499  END IF
500 *
501 * --- Test CPOSVX ---
502 *
503  IF( .NOT.prefac )
504  $ CALL claset( uplo, n, n, cmplx( zero ),
505  $ cmplx( zero ), afac, lda )
506  CALL claset( 'Full', n, nrhs, cmplx( zero ),
507  $ cmplx( zero ), x, lda )
508  IF( iequed.GT.1 .AND. n.GT.0 ) THEN
509 *
510 * Equilibrate the matrix if FACT='F' and
511 * EQUED='Y'.
512 *
513  CALL claqhe( uplo, n, a, lda, s, scond, amax,
514  $ equed )
515  END IF
516 *
517 * Solve the system and compute the condition number
518 * and error bounds using CPOSVX.
519 *
520  srnamt = 'CPOSVX'
521  CALL cposvx( fact, uplo, n, nrhs, a, lda, afac,
522  $ lda, equed, s, b, lda, x, lda, rcond,
523  $ rwork, rwork( nrhs+1 ), work,
524  $ rwork( 2*nrhs+1 ), info )
525 *
526 * Check the error code from CPOSVX.
527 *
528  IF( info.NE.izero ) THEN
529  CALL alaerh( path, 'CPOSVX', info, izero,
530  $ fact // uplo, n, n, -1, -1, nrhs,
531  $ imat, nfail, nerrs, nout )
532  go to 90
533  END IF
534 *
535  IF( info.EQ.0 ) THEN
536  IF( .NOT.prefac ) THEN
537 *
538 * Reconstruct matrix from factors and compute
539 * residual.
540 *
541  CALL cpot01( uplo, n, a, lda, afac, lda,
542  $ rwork( 2*nrhs+1 ), result( 1 ) )
543  k1 = 1
544  ELSE
545  k1 = 2
546  END IF
547 *
548 * Compute residual of the computed solution.
549 *
550  CALL clacpy( 'Full', n, nrhs, bsav, lda, work,
551  $ lda )
552  CALL cpot02( uplo, n, nrhs, asav, lda, x, lda,
553  $ work, lda, rwork( 2*nrhs+1 ),
554  $ result( 2 ) )
555 *
556 * Check solution from generated exact solution.
557 *
558  IF( nofact .OR. ( prefac .AND. lsame( equed,
559  $ 'N' ) ) ) THEN
560  CALL cget04( n, nrhs, x, lda, xact, lda,
561  $ rcondc, result( 3 ) )
562  ELSE
563  CALL cget04( n, nrhs, x, lda, xact, lda,
564  $ roldc, result( 3 ) )
565  END IF
566 *
567 * Check the error bounds from iterative
568 * refinement.
569 *
570  CALL cpot05( uplo, n, nrhs, asav, lda, b, lda,
571  $ x, lda, xact, lda, rwork,
572  $ rwork( nrhs+1 ), result( 4 ) )
573  ELSE
574  k1 = 6
575  END IF
576 *
577 * Compare RCOND from CPOSVX with the computed value
578 * in RCONDC.
579 *
580  result( 6 ) = sget06( rcond, rcondc )
581 *
582 * Print information about the tests that did not pass
583 * the threshold.
584 *
585  DO 80 k = k1, 6
586  IF( result( k ).GE.thresh ) THEN
587  IF( nfail.EQ.0 .AND. nerrs.EQ.0 )
588  $ CALL aladhd( nout, path )
589  IF( prefac ) THEN
590  WRITE( nout, fmt = 9997 )'CPOSVX', fact,
591  $ uplo, n, equed, imat, k, result( k )
592  ELSE
593  WRITE( nout, fmt = 9998 )'CPOSVX', fact,
594  $ uplo, n, imat, k, result( k )
595  END IF
596  nfail = nfail + 1
597  END IF
598  80 continue
599  nrun = nrun + 7 - k1
600  90 continue
601  100 continue
602  110 continue
603  120 continue
604  130 continue
605 *
606 * Print a summary of the results.
607 *
608  CALL alasvm( path, nout, nfail, nrun, nerrs )
609 *
610  9999 format( 1x, a, ', UPLO=''', a1, ''', N =', i5, ', type ', i1,
611  $ ', test(', i1, ')=', g12.5 )
612  9998 format( 1x, a, ', FACT=''', a1, ''', UPLO=''', a1, ''', N=', i5,
613  $ ', type ', i1, ', test(', i1, ')=', g12.5 )
614  9997 format( 1x, a, ', FACT=''', a1, ''', UPLO=''', a1, ''', N=', i5,
615  $ ', EQUED=''', a1, ''', type ', i1, ', test(', i1, ') =',
616  $ g12.5 )
617  return
618 *
619 * End of CDRVPO
620 *
621  END