LAPACK  3.7.1
LAPACK: Linear Algebra PACKage
dtrexc.f
Go to the documentation of this file.
1 *> \brief \b DTREXC
2 *
3 * =========== DOCUMENTATION ===========
4 *
5 * Online html documentation available at
6 * http://www.netlib.org/lapack/explore-html/
7 *
8 *> \htmlonly
9 *> Download DTREXC + dependencies
10 *> <a href="http://www.netlib.org/cgi-bin/netlibfiles.tgz?format=tgz&filename=/lapack/lapack_routine/dtrexc.f">
11 *> [TGZ]</a>
12 *> <a href="http://www.netlib.org/cgi-bin/netlibfiles.zip?format=zip&filename=/lapack/lapack_routine/dtrexc.f">
13 *> [ZIP]</a>
14 *> <a href="http://www.netlib.org/cgi-bin/netlibfiles.txt?format=txt&filename=/lapack/lapack_routine/dtrexc.f">
15 *> [TXT]</a>
16 *> \endhtmlonly
17 *
18 * Definition:
19 * ===========
20 *
21 * SUBROUTINE DTREXC( COMPQ, N, T, LDT, Q, LDQ, IFST, ILST, WORK,
22 * INFO )
23 *
24 * .. Scalar Arguments ..
25 * CHARACTER COMPQ
26 * INTEGER IFST, ILST, INFO, LDQ, LDT, N
27 * ..
28 * .. Array Arguments ..
29 * DOUBLE PRECISION Q( LDQ, * ), T( LDT, * ), WORK( * )
30 * ..
31 *
32 *
33 *> \par Purpose:
34 * =============
35 *>
36 *> \verbatim
37 *>
38 *> DTREXC reorders the real Schur factorization of a real matrix
39 *> A = Q*T*Q**T, so that the diagonal block of T with row index IFST is
40 *> moved to row ILST.
41 *>
42 *> The real Schur form T is reordered by an orthogonal similarity
43 *> transformation Z**T*T*Z, and optionally the matrix Q of Schur vectors
44 *> is updated by postmultiplying it with Z.
45 *>
46 *> T must be in Schur canonical form (as returned by DHSEQR), that is,
47 *> block upper triangular with 1-by-1 and 2-by-2 diagonal blocks; each
48 *> 2-by-2 diagonal block has its diagonal elements equal and its
49 *> off-diagonal elements of opposite sign.
50 *> \endverbatim
51 *
52 * Arguments:
53 * ==========
54 *
55 *> \param[in] COMPQ
56 *> \verbatim
57 *> COMPQ is CHARACTER*1
58 *> = 'V': update the matrix Q of Schur vectors;
59 *> = 'N': do not update Q.
60 *> \endverbatim
61 *>
62 *> \param[in] N
63 *> \verbatim
64 *> N is INTEGER
65 *> The order of the matrix T. N >= 0.
66 *> If N == 0 arguments ILST and IFST may be any value.
67 *> \endverbatim
68 *>
69 *> \param[in,out] T
70 *> \verbatim
71 *> T is DOUBLE PRECISION array, dimension (LDT,N)
72 *> On entry, the upper quasi-triangular matrix T, in Schur
73 *> Schur canonical form.
74 *> On exit, the reordered upper quasi-triangular matrix, again
75 *> in Schur canonical form.
76 *> \endverbatim
77 *>
78 *> \param[in] LDT
79 *> \verbatim
80 *> LDT is INTEGER
81 *> The leading dimension of the array T. LDT >= max(1,N).
82 *> \endverbatim
83 *>
84 *> \param[in,out] Q
85 *> \verbatim
86 *> Q is DOUBLE PRECISION array, dimension (LDQ,N)
87 *> On entry, if COMPQ = 'V', the matrix Q of Schur vectors.
88 *> On exit, if COMPQ = 'V', Q has been postmultiplied by the
89 *> orthogonal transformation matrix Z which reorders T.
90 *> If COMPQ = 'N', Q is not referenced.
91 *> \endverbatim
92 *>
93 *> \param[in] LDQ
94 *> \verbatim
95 *> LDQ is INTEGER
96 *> The leading dimension of the array Q. LDQ >= 1, and if
97 *> COMPQ = 'V', LDQ >= max(1,N).
98 *> \endverbatim
99 *>
100 *> \param[in,out] IFST
101 *> \verbatim
102 *> IFST is INTEGER
103 *> \endverbatim
104 *>
105 *> \param[in,out] ILST
106 *> \verbatim
107 *> ILST is INTEGER
108 *>
109 *> Specify the reordering of the diagonal blocks of T.
110 *> The block with row index IFST is moved to row ILST, by a
111 *> sequence of transpositions between adjacent blocks.
112 *> On exit, if IFST pointed on entry to the second row of a
113 *> 2-by-2 block, it is changed to point to the first row; ILST
114 *> always points to the first row of the block in its final
115 *> position (which may differ from its input value by +1 or -1).
116 *> 1 <= IFST <= N; 1 <= ILST <= N.
117 *> \endverbatim
118 *>
119 *> \param[out] WORK
120 *> \verbatim
121 *> WORK is DOUBLE PRECISION array, dimension (N)
122 *> \endverbatim
123 *>
124 *> \param[out] INFO
125 *> \verbatim
126 *> INFO is INTEGER
127 *> = 0: successful exit
128 *> < 0: if INFO = -i, the i-th argument had an illegal value
129 *> = 1: two adjacent blocks were too close to swap (the problem
130 *> is very ill-conditioned); T may have been partially
131 *> reordered, and ILST points to the first row of the
132 *> current position of the block being moved.
133 *> \endverbatim
134 *
135 * Authors:
136 * ========
137 *
138 *> \author Univ. of Tennessee
139 *> \author Univ. of California Berkeley
140 *> \author Univ. of Colorado Denver
141 *> \author NAG Ltd.
142 *
143 *> \date December 2016
144 *
145 *> \ingroup doubleOTHERcomputational
146 *
147 * =====================================================================
148  SUBROUTINE dtrexc( COMPQ, N, T, LDT, Q, LDQ, IFST, ILST, WORK,
149  $ INFO )
150 *
151 * -- LAPACK computational routine (version 3.7.0) --
152 * -- LAPACK is a software package provided by Univ. of Tennessee, --
153 * -- Univ. of California Berkeley, Univ. of Colorado Denver and NAG Ltd..--
154 * December 2016
155 *
156 * .. Scalar Arguments ..
157  CHARACTER COMPQ
158  INTEGER IFST, ILST, INFO, LDQ, LDT, N
159 * ..
160 * .. Array Arguments ..
161  DOUBLE PRECISION Q( ldq, * ), T( ldt, * ), WORK( * )
162 * ..
163 *
164 * =====================================================================
165 *
166 * .. Parameters ..
167  DOUBLE PRECISION ZERO
168  parameter( zero = 0.0d+0 )
169 * ..
170 * .. Local Scalars ..
171  LOGICAL WANTQ
172  INTEGER HERE, NBF, NBL, NBNEXT
173 * ..
174 * .. External Functions ..
175  LOGICAL LSAME
176  EXTERNAL lsame
177 * ..
178 * .. External Subroutines ..
179  EXTERNAL dlaexc, xerbla
180 * ..
181 * .. Intrinsic Functions ..
182  INTRINSIC max
183 * ..
184 * .. Executable Statements ..
185 *
186 * Decode and test the input arguments.
187 *
188  info = 0
189  wantq = lsame( compq, 'V' )
190  IF( .NOT.wantq .AND. .NOT.lsame( compq, 'N' ) ) THEN
191  info = -1
192  ELSE IF( n.LT.0 ) THEN
193  info = -2
194  ELSE IF( ldt.LT.max( 1, n ) ) THEN
195  info = -4
196  ELSE IF( ldq.LT.1 .OR. ( wantq .AND. ldq.LT.max( 1, n ) ) ) THEN
197  info = -6
198  ELSE IF(( ifst.LT.1 .OR. ifst.GT.n ).AND.( n.GT.0 )) THEN
199  info = -7
200  ELSE IF(( ilst.LT.1 .OR. ilst.GT.n ).AND.( n.GT.0 )) THEN
201  info = -8
202  END IF
203  IF( info.NE.0 ) THEN
204  CALL xerbla( 'DTREXC', -info )
205  RETURN
206  END IF
207 *
208 * Quick return if possible
209 *
210  IF( n.LE.1 )
211  $ RETURN
212 *
213 * Determine the first row of specified block
214 * and find out it is 1 by 1 or 2 by 2.
215 *
216  IF( ifst.GT.1 ) THEN
217  IF( t( ifst, ifst-1 ).NE.zero )
218  $ ifst = ifst - 1
219  END IF
220  nbf = 1
221  IF( ifst.LT.n ) THEN
222  IF( t( ifst+1, ifst ).NE.zero )
223  $ nbf = 2
224  END IF
225 *
226 * Determine the first row of the final block
227 * and find out it is 1 by 1 or 2 by 2.
228 *
229  IF( ilst.GT.1 ) THEN
230  IF( t( ilst, ilst-1 ).NE.zero )
231  $ ilst = ilst - 1
232  END IF
233  nbl = 1
234  IF( ilst.LT.n ) THEN
235  IF( t( ilst+1, ilst ).NE.zero )
236  $ nbl = 2
237  END IF
238 *
239  IF( ifst.EQ.ilst )
240  $ RETURN
241 *
242  IF( ifst.LT.ilst ) THEN
243 *
244 * Update ILST
245 *
246  IF( nbf.EQ.2 .AND. nbl.EQ.1 )
247  $ ilst = ilst - 1
248  IF( nbf.EQ.1 .AND. nbl.EQ.2 )
249  $ ilst = ilst + 1
250 *
251  here = ifst
252 *
253  10 CONTINUE
254 *
255 * Swap block with next one below
256 *
257  IF( nbf.EQ.1 .OR. nbf.EQ.2 ) THEN
258 *
259 * Current block either 1 by 1 or 2 by 2
260 *
261  nbnext = 1
262  IF( here+nbf+1.LE.n ) THEN
263  IF( t( here+nbf+1, here+nbf ).NE.zero )
264  $ nbnext = 2
265  END IF
266  CALL dlaexc( wantq, n, t, ldt, q, ldq, here, nbf, nbnext,
267  $ work, info )
268  IF( info.NE.0 ) THEN
269  ilst = here
270  RETURN
271  END IF
272  here = here + nbnext
273 *
274 * Test if 2 by 2 block breaks into two 1 by 1 blocks
275 *
276  IF( nbf.EQ.2 ) THEN
277  IF( t( here+1, here ).EQ.zero )
278  $ nbf = 3
279  END IF
280 *
281  ELSE
282 *
283 * Current block consists of two 1 by 1 blocks each of which
284 * must be swapped individually
285 *
286  nbnext = 1
287  IF( here+3.LE.n ) THEN
288  IF( t( here+3, here+2 ).NE.zero )
289  $ nbnext = 2
290  END IF
291  CALL dlaexc( wantq, n, t, ldt, q, ldq, here+1, 1, nbnext,
292  $ work, info )
293  IF( info.NE.0 ) THEN
294  ilst = here
295  RETURN
296  END IF
297  IF( nbnext.EQ.1 ) THEN
298 *
299 * Swap two 1 by 1 blocks, no problems possible
300 *
301  CALL dlaexc( wantq, n, t, ldt, q, ldq, here, 1, nbnext,
302  $ work, info )
303  here = here + 1
304  ELSE
305 *
306 * Recompute NBNEXT in case 2 by 2 split
307 *
308  IF( t( here+2, here+1 ).EQ.zero )
309  $ nbnext = 1
310  IF( nbnext.EQ.2 ) THEN
311 *
312 * 2 by 2 Block did not split
313 *
314  CALL dlaexc( wantq, n, t, ldt, q, ldq, here, 1,
315  $ nbnext, work, info )
316  IF( info.NE.0 ) THEN
317  ilst = here
318  RETURN
319  END IF
320  here = here + 2
321  ELSE
322 *
323 * 2 by 2 Block did split
324 *
325  CALL dlaexc( wantq, n, t, ldt, q, ldq, here, 1, 1,
326  $ work, info )
327  CALL dlaexc( wantq, n, t, ldt, q, ldq, here+1, 1, 1,
328  $ work, info )
329  here = here + 2
330  END IF
331  END IF
332  END IF
333  IF( here.LT.ilst )
334  $ GO TO 10
335 *
336  ELSE
337 *
338  here = ifst
339  20 CONTINUE
340 *
341 * Swap block with next one above
342 *
343  IF( nbf.EQ.1 .OR. nbf.EQ.2 ) THEN
344 *
345 * Current block either 1 by 1 or 2 by 2
346 *
347  nbnext = 1
348  IF( here.GE.3 ) THEN
349  IF( t( here-1, here-2 ).NE.zero )
350  $ nbnext = 2
351  END IF
352  CALL dlaexc( wantq, n, t, ldt, q, ldq, here-nbnext, nbnext,
353  $ nbf, work, info )
354  IF( info.NE.0 ) THEN
355  ilst = here
356  RETURN
357  END IF
358  here = here - nbnext
359 *
360 * Test if 2 by 2 block breaks into two 1 by 1 blocks
361 *
362  IF( nbf.EQ.2 ) THEN
363  IF( t( here+1, here ).EQ.zero )
364  $ nbf = 3
365  END IF
366 *
367  ELSE
368 *
369 * Current block consists of two 1 by 1 blocks each of which
370 * must be swapped individually
371 *
372  nbnext = 1
373  IF( here.GE.3 ) THEN
374  IF( t( here-1, here-2 ).NE.zero )
375  $ nbnext = 2
376  END IF
377  CALL dlaexc( wantq, n, t, ldt, q, ldq, here-nbnext, nbnext,
378  $ 1, work, info )
379  IF( info.NE.0 ) THEN
380  ilst = here
381  RETURN
382  END IF
383  IF( nbnext.EQ.1 ) THEN
384 *
385 * Swap two 1 by 1 blocks, no problems possible
386 *
387  CALL dlaexc( wantq, n, t, ldt, q, ldq, here, nbnext, 1,
388  $ work, info )
389  here = here - 1
390  ELSE
391 *
392 * Recompute NBNEXT in case 2 by 2 split
393 *
394  IF( t( here, here-1 ).EQ.zero )
395  $ nbnext = 1
396  IF( nbnext.EQ.2 ) THEN
397 *
398 * 2 by 2 Block did not split
399 *
400  CALL dlaexc( wantq, n, t, ldt, q, ldq, here-1, 2, 1,
401  $ work, info )
402  IF( info.NE.0 ) THEN
403  ilst = here
404  RETURN
405  END IF
406  here = here - 2
407  ELSE
408 *
409 * 2 by 2 Block did split
410 *
411  CALL dlaexc( wantq, n, t, ldt, q, ldq, here, 1, 1,
412  $ work, info )
413  CALL dlaexc( wantq, n, t, ldt, q, ldq, here-1, 1, 1,
414  $ work, info )
415  here = here - 2
416  END IF
417  END IF
418  END IF
419  IF( here.GT.ilst )
420  $ GO TO 20
421  END IF
422  ilst = here
423 *
424  RETURN
425 *
426 * End of DTREXC
427 *
428  END
dtrexc
subroutine dtrexc(COMPQ, N, T, LDT, Q, LDQ, IFST, ILST, WORK, INFO)
DTREXC
Definition: dtrexc.f:150
xerbla
subroutine xerbla(SRNAME, INFO)
XERBLA
Definition: xerbla.f:62
dlaexc
subroutine dlaexc(WANTQ, N, T, LDT, Q, LDQ, J1, N1, N2, WORK, INFO)
DLAEXC swaps adjacent diagonal blocks of a real upper quasi-triangular matrix in Schur canonical form...
Definition: dlaexc.f:140