1: *> \brief \b DLAQR5 performs a single small-bulge multi-shift QR sweep.
2: *
3: * =========== DOCUMENTATION ===========
4: *
5: * Online html documentation available at
6: * http://www.netlib.org/lapack/explore-html/
7: *
8: *> \htmlonly
9: *> Download DLAQR5 + dependencies
10: *> <a href="http://www.netlib.org/cgi-bin/netlibfiles.tgz?format=tgz&filename=/lapack/lapack_routine/dlaqr5.f">
11: *> [TGZ]</a>
12: *> <a href="http://www.netlib.org/cgi-bin/netlibfiles.zip?format=zip&filename=/lapack/lapack_routine/dlaqr5.f">
13: *> [ZIP]</a>
14: *> <a href="http://www.netlib.org/cgi-bin/netlibfiles.txt?format=txt&filename=/lapack/lapack_routine/dlaqr5.f">
15: *> [TXT]</a>
16: *> \endhtmlonly
17: *
18: * Definition:
19: * ===========
20: *
21: * SUBROUTINE DLAQR5( WANTT, WANTZ, KACC22, N, KTOP, KBOT, NSHFTS,
22: * SR, SI, H, LDH, ILOZ, IHIZ, Z, LDZ, V, LDV, U,
23: * LDU, NV, WV, LDWV, NH, WH, LDWH )
24: *
25: * .. Scalar Arguments ..
26: * INTEGER IHIZ, ILOZ, KACC22, KBOT, KTOP, LDH, LDU, LDV,
27: * $ LDWH, LDWV, LDZ, N, NH, NSHFTS, NV
28: * LOGICAL WANTT, WANTZ
29: * ..
30: * .. Array Arguments ..
31: * DOUBLE PRECISION H( LDH, * ), SI( * ), SR( * ), U( LDU, * ),
32: * $ V( LDV, * ), WH( LDWH, * ), WV( LDWV, * ),
33: * $ Z( LDZ, * )
34: * ..
35: *
36: *
37: *> \par Purpose:
38: * =============
39: *>
40: *> \verbatim
41: *>
42: *> DLAQR5, called by DLAQR0, performs a
43: *> single small-bulge multi-shift QR sweep.
44: *> \endverbatim
45: *
46: * Arguments:
47: * ==========
48: *
49: *> \param[in] WANTT
50: *> \verbatim
51: *> WANTT is LOGICAL
52: *> WANTT = .true. if the quasi-triangular Schur factor
53: *> is being computed. WANTT is set to .false. otherwise.
54: *> \endverbatim
55: *>
56: *> \param[in] WANTZ
57: *> \verbatim
58: *> WANTZ is LOGICAL
59: *> WANTZ = .true. if the orthogonal Schur factor is being
60: *> computed. WANTZ is set to .false. otherwise.
61: *> \endverbatim
62: *>
63: *> \param[in] KACC22
64: *> \verbatim
65: *> KACC22 is INTEGER with value 0, 1, or 2.
66: *> Specifies the computation mode of far-from-diagonal
67: *> orthogonal updates.
68: *> = 0: DLAQR5 does not accumulate reflections and does not
69: *> use matrix-matrix multiply to update far-from-diagonal
70: *> matrix entries.
71: *> = 1: DLAQR5 accumulates reflections and uses matrix-matrix
72: *> multiply to update the far-from-diagonal matrix entries.
73: *> = 2: Same as KACC22 = 1. This option used to enable exploiting
74: *> the 2-by-2 structure during matrix multiplications, but
75: *> this is no longer supported.
76: *> \endverbatim
77: *>
78: *> \param[in] N
79: *> \verbatim
80: *> N is INTEGER
81: *> N is the order of the Hessenberg matrix H upon which this
82: *> subroutine operates.
83: *> \endverbatim
84: *>
85: *> \param[in] KTOP
86: *> \verbatim
87: *> KTOP is INTEGER
88: *> \endverbatim
89: *>
90: *> \param[in] KBOT
91: *> \verbatim
92: *> KBOT is INTEGER
93: *> These are the first and last rows and columns of an
94: *> isolated diagonal block upon which the QR sweep is to be
95: *> applied. It is assumed without a check that
96: *> either KTOP = 1 or H(KTOP,KTOP-1) = 0
97: *> and
98: *> either KBOT = N or H(KBOT+1,KBOT) = 0.
99: *> \endverbatim
100: *>
101: *> \param[in] NSHFTS
102: *> \verbatim
103: *> NSHFTS is INTEGER
104: *> NSHFTS gives the number of simultaneous shifts. NSHFTS
105: *> must be positive and even.
106: *> \endverbatim
107: *>
108: *> \param[in,out] SR
109: *> \verbatim
110: *> SR is DOUBLE PRECISION array, dimension (NSHFTS)
111: *> \endverbatim
112: *>
113: *> \param[in,out] SI
114: *> \verbatim
115: *> SI is DOUBLE PRECISION array, dimension (NSHFTS)
116: *> SR contains the real parts and SI contains the imaginary
117: *> parts of the NSHFTS shifts of origin that define the
118: *> multi-shift QR sweep. On output SR and SI may be
119: *> reordered.
120: *> \endverbatim
121: *>
122: *> \param[in,out] H
123: *> \verbatim
124: *> H is DOUBLE PRECISION array, dimension (LDH,N)
125: *> On input H contains a Hessenberg matrix. On output a
126: *> multi-shift QR sweep with shifts SR(J)+i*SI(J) is applied
127: *> to the isolated diagonal block in rows and columns KTOP
128: *> through KBOT.
129: *> \endverbatim
130: *>
131: *> \param[in] LDH
132: *> \verbatim
133: *> LDH is INTEGER
134: *> LDH is the leading dimension of H just as declared in the
135: *> calling procedure. LDH >= MAX(1,N).
136: *> \endverbatim
137: *>
138: *> \param[in] ILOZ
139: *> \verbatim
140: *> ILOZ is INTEGER
141: *> \endverbatim
142: *>
143: *> \param[in] IHIZ
144: *> \verbatim
145: *> IHIZ is INTEGER
146: *> Specify the rows of Z to which transformations must be
147: *> applied if WANTZ is .TRUE.. 1 <= ILOZ <= IHIZ <= N
148: *> \endverbatim
149: *>
150: *> \param[in,out] Z
151: *> \verbatim
152: *> Z is DOUBLE PRECISION array, dimension (LDZ,IHIZ)
153: *> If WANTZ = .TRUE., then the QR Sweep orthogonal
154: *> similarity transformation is accumulated into
155: *> Z(ILOZ:IHIZ,ILOZ:IHIZ) from the right.
156: *> If WANTZ = .FALSE., then Z is unreferenced.
157: *> \endverbatim
158: *>
159: *> \param[in] LDZ
160: *> \verbatim
161: *> LDZ is INTEGER
162: *> LDA is the leading dimension of Z just as declared in
163: *> the calling procedure. LDZ >= N.
164: *> \endverbatim
165: *>
166: *> \param[out] V
167: *> \verbatim
168: *> V is DOUBLE PRECISION array, dimension (LDV,NSHFTS/2)
169: *> \endverbatim
170: *>
171: *> \param[in] LDV
172: *> \verbatim
173: *> LDV is INTEGER
174: *> LDV is the leading dimension of V as declared in the
175: *> calling procedure. LDV >= 3.
176: *> \endverbatim
177: *>
178: *> \param[out] U
179: *> \verbatim
180: *> U is DOUBLE PRECISION array, dimension (LDU,2*NSHFTS)
181: *> \endverbatim
182: *>
183: *> \param[in] LDU
184: *> \verbatim
185: *> LDU is INTEGER
186: *> LDU is the leading dimension of U just as declared in the
187: *> in the calling subroutine. LDU >= 2*NSHFTS.
188: *> \endverbatim
189: *>
190: *> \param[in] NV
191: *> \verbatim
192: *> NV is INTEGER
193: *> NV is the number of rows in WV agailable for workspace.
194: *> NV >= 1.
195: *> \endverbatim
196: *>
197: *> \param[out] WV
198: *> \verbatim
199: *> WV is DOUBLE PRECISION array, dimension (LDWV,2*NSHFTS)
200: *> \endverbatim
201: *>
202: *> \param[in] LDWV
203: *> \verbatim
204: *> LDWV is INTEGER
205: *> LDWV is the leading dimension of WV as declared in the
206: *> in the calling subroutine. LDWV >= NV.
207: *> \endverbatim
208: *
209: *> \param[in] NH
210: *> \verbatim
211: *> NH is INTEGER
212: *> NH is the number of columns in array WH available for
213: *> workspace. NH >= 1.
214: *> \endverbatim
215: *>
216: *> \param[out] WH
217: *> \verbatim
218: *> WH is DOUBLE PRECISION array, dimension (LDWH,NH)
219: *> \endverbatim
220: *>
221: *> \param[in] LDWH
222: *> \verbatim
223: *> LDWH is INTEGER
224: *> Leading dimension of WH just as declared in the
225: *> calling procedure. LDWH >= 2*NSHFTS.
226: *> \endverbatim
227: *>
228: * Authors:
229: * ========
230: *
231: *> \author Univ. of Tennessee
232: *> \author Univ. of California Berkeley
233: *> \author Univ. of Colorado Denver
234: *> \author NAG Ltd.
235: *
236: *> \ingroup doubleOTHERauxiliary
237: *
238: *> \par Contributors:
239: * ==================
240: *>
241: *> Karen Braman and Ralph Byers, Department of Mathematics,
242: *> University of Kansas, USA
243: *>
244: *> Lars Karlsson, Daniel Kressner, and Bruno Lang
245: *>
246: *> Thijs Steel, Department of Computer science,
247: *> KU Leuven, Belgium
248: *
249: *> \par References:
250: * ================
251: *>
252: *> K. Braman, R. Byers and R. Mathias, The Multi-Shift QR
253: *> Algorithm Part I: Maintaining Well Focused Shifts, and Level 3
254: *> Performance, SIAM Journal of Matrix Analysis, volume 23, pages
255: *> 929--947, 2002.
256: *>
257: *> Lars Karlsson, Daniel Kressner, and Bruno Lang, Optimally packed
258: *> chains of bulges in multishift QR algorithms.
259: *> ACM Trans. Math. Softw. 40, 2, Article 12 (February 2014).
260: *>
261: * =====================================================================
262: SUBROUTINE DLAQR5( WANTT, WANTZ, KACC22, N, KTOP, KBOT, NSHFTS,
263: $ SR, SI, H, LDH, ILOZ, IHIZ, Z, LDZ, V, LDV, U,
264: $ LDU, NV, WV, LDWV, NH, WH, LDWH )
265: IMPLICIT NONE
266: *
267: * -- LAPACK auxiliary routine --
268: * -- LAPACK is a software package provided by Univ. of Tennessee, --
269: * -- Univ. of California Berkeley, Univ. of Colorado Denver and NAG Ltd..--
270: *
271: * .. Scalar Arguments ..
272: INTEGER IHIZ, ILOZ, KACC22, KBOT, KTOP, LDH, LDU, LDV,
273: $ LDWH, LDWV, LDZ, N, NH, NSHFTS, NV
274: LOGICAL WANTT, WANTZ
275: * ..
276: * .. Array Arguments ..
277: DOUBLE PRECISION H( LDH, * ), SI( * ), SR( * ), U( LDU, * ),
278: $ V( LDV, * ), WH( LDWH, * ), WV( LDWV, * ),
279: $ Z( LDZ, * )
280: * ..
281: *
282: * ================================================================
283: * .. Parameters ..
284: DOUBLE PRECISION ZERO, ONE
285: PARAMETER ( ZERO = 0.0d0, ONE = 1.0d0 )
286: * ..
287: * .. Local Scalars ..
288: DOUBLE PRECISION ALPHA, BETA, H11, H12, H21, H22, REFSUM,
289: $ SAFMAX, SAFMIN, SCL, SMLNUM, SWAP, T1, T2,
290: $ T3, TST1, TST2, ULP
291: INTEGER I, I2, I4, INCOL, J, JBOT, JCOL, JLEN,
292: $ JROW, JTOP, K, K1, KDU, KMS, KRCOL,
293: $ M, M22, MBOT, MTOP, NBMPS, NDCOL,
294: $ NS, NU
295: LOGICAL ACCUM, BMP22
296: * ..
297: * .. External Functions ..
298: DOUBLE PRECISION DLAMCH
299: EXTERNAL DLAMCH
300: * ..
301: * .. Intrinsic Functions ..
302: *
303: INTRINSIC ABS, DBLE, MAX, MIN, MOD
304: * ..
305: * .. Local Arrays ..
306: DOUBLE PRECISION VT( 3 )
307: * ..
308: * .. External Subroutines ..
309: EXTERNAL DGEMM, DLABAD, DLACPY, DLAQR1, DLARFG, DLASET,
310: $ DTRMM
311: * ..
312: * .. Executable Statements ..
313: *
314: * ==== If there are no shifts, then there is nothing to do. ====
315: *
316: IF( NSHFTS.LT.2 )
317: $ RETURN
318: *
319: * ==== If the active block is empty or 1-by-1, then there
320: * . is nothing to do. ====
321: *
322: IF( KTOP.GE.KBOT )
323: $ RETURN
324: *
325: * ==== Shuffle shifts into pairs of real shifts and pairs
326: * . of complex conjugate shifts assuming complex
327: * . conjugate shifts are already adjacent to one
328: * . another. ====
329: *
330: DO 10 I = 1, NSHFTS - 2, 2
331: IF( SI( I ).NE.-SI( I+1 ) ) THEN
332: *
333: SWAP = SR( I )
334: SR( I ) = SR( I+1 )
335: SR( I+1 ) = SR( I+2 )
336: SR( I+2 ) = SWAP
337: *
338: SWAP = SI( I )
339: SI( I ) = SI( I+1 )
340: SI( I+1 ) = SI( I+2 )
341: SI( I+2 ) = SWAP
342: END IF
343: 10 CONTINUE
344: *
345: * ==== NSHFTS is supposed to be even, but if it is odd,
346: * . then simply reduce it by one. The shuffle above
347: * . ensures that the dropped shift is real and that
348: * . the remaining shifts are paired. ====
349: *
350: NS = NSHFTS - MOD( NSHFTS, 2 )
351: *
352: * ==== Machine constants for deflation ====
353: *
354: SAFMIN = DLAMCH( 'SAFE MINIMUM' )
355: SAFMAX = ONE / SAFMIN
356: CALL DLABAD( SAFMIN, SAFMAX )
357: ULP = DLAMCH( 'PRECISION' )
358: SMLNUM = SAFMIN*( DBLE( N ) / ULP )
359: *
360: * ==== Use accumulated reflections to update far-from-diagonal
361: * . entries ? ====
362: *
363: ACCUM = ( KACC22.EQ.1 ) .OR. ( KACC22.EQ.2 )
364: *
365: * ==== clear trash ====
366: *
367: IF( KTOP+2.LE.KBOT )
368: $ H( KTOP+2, KTOP ) = ZERO
369: *
370: * ==== NBMPS = number of 2-shift bulges in the chain ====
371: *
372: NBMPS = NS / 2
373: *
374: * ==== KDU = width of slab ====
375: *
376: KDU = 4*NBMPS
377: *
378: * ==== Create and chase chains of NBMPS bulges ====
379: *
380: DO 180 INCOL = KTOP - 2*NBMPS + 1, KBOT - 2, 2*NBMPS
381: *
382: * JTOP = Index from which updates from the right start.
383: *
384: IF( ACCUM ) THEN
385: JTOP = MAX( KTOP, INCOL )
386: ELSE IF( WANTT ) THEN
387: JTOP = 1
388: ELSE
389: JTOP = KTOP
390: END IF
391: *
392: NDCOL = INCOL + KDU
393: IF( ACCUM )
394: $ CALL DLASET( 'ALL', KDU, KDU, ZERO, ONE, U, LDU )
395: *
396: * ==== Near-the-diagonal bulge chase. The following loop
397: * . performs the near-the-diagonal part of a small bulge
398: * . multi-shift QR sweep. Each 4*NBMPS column diagonal
399: * . chunk extends from column INCOL to column NDCOL
400: * . (including both column INCOL and column NDCOL). The
401: * . following loop chases a 2*NBMPS+1 column long chain of
402: * . NBMPS bulges 2*NBMPS columns to the right. (INCOL
403: * . may be less than KTOP and and NDCOL may be greater than
404: * . KBOT indicating phantom columns from which to chase
405: * . bulges before they are actually introduced or to which
406: * . to chase bulges beyond column KBOT.) ====
407: *
408: DO 145 KRCOL = INCOL, MIN( INCOL+2*NBMPS-1, KBOT-2 )
409: *
410: * ==== Bulges number MTOP to MBOT are active double implicit
411: * . shift bulges. There may or may not also be small
412: * . 2-by-2 bulge, if there is room. The inactive bulges
413: * . (if any) must wait until the active bulges have moved
414: * . down the diagonal to make room. The phantom matrix
415: * . paradigm described above helps keep track. ====
416: *
417: MTOP = MAX( 1, ( KTOP-KRCOL ) / 2+1 )
418: MBOT = MIN( NBMPS, ( KBOT-KRCOL-1 ) / 2 )
419: M22 = MBOT + 1
420: BMP22 = ( MBOT.LT.NBMPS ) .AND. ( KRCOL+2*( M22-1 ) ).EQ.
421: $ ( KBOT-2 )
422: *
423: * ==== Generate reflections to chase the chain right
424: * . one column. (The minimum value of K is KTOP-1.) ====
425: *
426: IF ( BMP22 ) THEN
427: *
428: * ==== Special case: 2-by-2 reflection at bottom treated
429: * . separately ====
430: *
431: K = KRCOL + 2*( M22-1 )
432: IF( K.EQ.KTOP-1 ) THEN
433: CALL DLAQR1( 2, H( K+1, K+1 ), LDH, SR( 2*M22-1 ),
434: $ SI( 2*M22-1 ), SR( 2*M22 ), SI( 2*M22 ),
435: $ V( 1, M22 ) )
436: BETA = V( 1, M22 )
437: CALL DLARFG( 2, BETA, V( 2, M22 ), 1, V( 1, M22 ) )
438: ELSE
439: BETA = H( K+1, K )
440: V( 2, M22 ) = H( K+2, K )
441: CALL DLARFG( 2, BETA, V( 2, M22 ), 1, V( 1, M22 ) )
442: H( K+1, K ) = BETA
443: H( K+2, K ) = ZERO
444: END IF
445:
446: *
447: * ==== Perform update from right within
448: * . computational window. ====
449: *
450: T1 = V( 1, M22 )
451: T2 = T1*V( 2, M22 )
452: DO 30 J = JTOP, MIN( KBOT, K+3 )
453: REFSUM = H( J, K+1 ) + V( 2, M22 )*H( J, K+2 )
454: H( J, K+1 ) = H( J, K+1 ) - REFSUM*T1
455: H( J, K+2 ) = H( J, K+2 ) - REFSUM*T2
456: 30 CONTINUE
457: *
458: * ==== Perform update from left within
459: * . computational window. ====
460: *
461: IF( ACCUM ) THEN
462: JBOT = MIN( NDCOL, KBOT )
463: ELSE IF( WANTT ) THEN
464: JBOT = N
465: ELSE
466: JBOT = KBOT
467: END IF
468: T1 = V( 1, M22 )
469: T2 = T1*V( 2, M22 )
470: DO 40 J = K+1, JBOT
471: REFSUM = H( K+1, J ) + V( 2, M22 )*H( K+2, J )
472: H( K+1, J ) = H( K+1, J ) - REFSUM*T1
473: H( K+2, J ) = H( K+2, J ) - REFSUM*T2
474: 40 CONTINUE
475: *
476: * ==== The following convergence test requires that
477: * . the tradition small-compared-to-nearby-diagonals
478: * . criterion and the Ahues & Tisseur (LAWN 122, 1997)
479: * . criteria both be satisfied. The latter improves
480: * . accuracy in some examples. Falling back on an
481: * . alternate convergence criterion when TST1 or TST2
482: * . is zero (as done here) is traditional but probably
483: * . unnecessary. ====
484: *
485: IF( K.GE.KTOP ) THEN
486: IF( H( K+1, K ).NE.ZERO ) THEN
487: TST1 = ABS( H( K, K ) ) + ABS( H( K+1, K+1 ) )
488: IF( TST1.EQ.ZERO ) THEN
489: IF( K.GE.KTOP+1 )
490: $ TST1 = TST1 + ABS( H( K, K-1 ) )
491: IF( K.GE.KTOP+2 )
492: $ TST1 = TST1 + ABS( H( K, K-2 ) )
493: IF( K.GE.KTOP+3 )
494: $ TST1 = TST1 + ABS( H( K, K-3 ) )
495: IF( K.LE.KBOT-2 )
496: $ TST1 = TST1 + ABS( H( K+2, K+1 ) )
497: IF( K.LE.KBOT-3 )
498: $ TST1 = TST1 + ABS( H( K+3, K+1 ) )
499: IF( K.LE.KBOT-4 )
500: $ TST1 = TST1 + ABS( H( K+4, K+1 ) )
501: END IF
502: IF( ABS( H( K+1, K ) )
503: $ .LE.MAX( SMLNUM, ULP*TST1 ) ) THEN
504: H12 = MAX( ABS( H( K+1, K ) ),
505: $ ABS( H( K, K+1 ) ) )
506: H21 = MIN( ABS( H( K+1, K ) ),
507: $ ABS( H( K, K+1 ) ) )
508: H11 = MAX( ABS( H( K+1, K+1 ) ),
509: $ ABS( H( K, K )-H( K+1, K+1 ) ) )
510: H22 = MIN( ABS( H( K+1, K+1 ) ),
511: $ ABS( H( K, K )-H( K+1, K+1 ) ) )
512: SCL = H11 + H12
513: TST2 = H22*( H11 / SCL )
514: *
515: IF( TST2.EQ.ZERO .OR. H21*( H12 / SCL ).LE.
516: $ MAX( SMLNUM, ULP*TST2 ) ) THEN
517: H( K+1, K ) = ZERO
518: END IF
519: END IF
520: END IF
521: END IF
522: *
523: * ==== Accumulate orthogonal transformations. ====
524: *
525: IF( ACCUM ) THEN
526: KMS = K - INCOL
527: T1 = V( 1, M22 )
528: T2 = T1*V( 2, M22 )
529: DO 50 J = MAX( 1, KTOP-INCOL ), KDU
530: REFSUM = U( J, KMS+1 ) + V( 2, M22 )*U( J, KMS+2 )
531: U( J, KMS+1 ) = U( J, KMS+1 ) - REFSUM*T1
532: U( J, KMS+2 ) = U( J, KMS+2 ) - REFSUM*T2
533: 50 CONTINUE
534: ELSE IF( WANTZ ) THEN
535: T1 = V( 1, M22 )
536: T2 = T1*V( 2, M22 )
537: DO 60 J = ILOZ, IHIZ
538: REFSUM = Z( J, K+1 )+V( 2, M22 )*Z( J, K+2 )
539: Z( J, K+1 ) = Z( J, K+1 ) - REFSUM*T1
540: Z( J, K+2 ) = Z( J, K+2 ) - REFSUM*T2
541: 60 CONTINUE
542: END IF
543: END IF
544: *
545: * ==== Normal case: Chain of 3-by-3 reflections ====
546: *
547: DO 80 M = MBOT, MTOP, -1
548: K = KRCOL + 2*( M-1 )
549: IF( K.EQ.KTOP-1 ) THEN
550: CALL DLAQR1( 3, H( KTOP, KTOP ), LDH, SR( 2*M-1 ),
551: $ SI( 2*M-1 ), SR( 2*M ), SI( 2*M ),
552: $ V( 1, M ) )
553: ALPHA = V( 1, M )
554: CALL DLARFG( 3, ALPHA, V( 2, M ), 1, V( 1, M ) )
555: ELSE
556: *
557: * ==== Perform delayed transformation of row below
558: * . Mth bulge. Exploit fact that first two elements
559: * . of row are actually zero. ====
560: *
561: REFSUM = V( 1, M )*V( 3, M )*H( K+3, K+2 )
562: H( K+3, K ) = -REFSUM
563: H( K+3, K+1 ) = -REFSUM*V( 2, M )
564: H( K+3, K+2 ) = H( K+3, K+2 ) - REFSUM*V( 3, M )
565: *
566: * ==== Calculate reflection to move
567: * . Mth bulge one step. ====
568: *
569: BETA = H( K+1, K )
570: V( 2, M ) = H( K+2, K )
571: V( 3, M ) = H( K+3, K )
572: CALL DLARFG( 3, BETA, V( 2, M ), 1, V( 1, M ) )
573: *
574: * ==== A Bulge may collapse because of vigilant
575: * . deflation or destructive underflow. In the
576: * . underflow case, try the two-small-subdiagonals
577: * . trick to try to reinflate the bulge. ====
578: *
579: IF( H( K+3, K ).NE.ZERO .OR. H( K+3, K+1 ).NE.
580: $ ZERO .OR. H( K+3, K+2 ).EQ.ZERO ) THEN
581: *
582: * ==== Typical case: not collapsed (yet). ====
583: *
584: H( K+1, K ) = BETA
585: H( K+2, K ) = ZERO
586: H( K+3, K ) = ZERO
587: ELSE
588: *
589: * ==== Atypical case: collapsed. Attempt to
590: * . reintroduce ignoring H(K+1,K) and H(K+2,K).
591: * . If the fill resulting from the new
592: * . reflector is too large, then abandon it.
593: * . Otherwise, use the new one. ====
594: *
595: CALL DLAQR1( 3, H( K+1, K+1 ), LDH, SR( 2*M-1 ),
596: $ SI( 2*M-1 ), SR( 2*M ), SI( 2*M ),
597: $ VT )
598: ALPHA = VT( 1 )
599: CALL DLARFG( 3, ALPHA, VT( 2 ), 1, VT( 1 ) )
600: REFSUM = VT( 1 )*( H( K+1, K )+VT( 2 )*
601: $ H( K+2, K ) )
602: *
603: IF( ABS( H( K+2, K )-REFSUM*VT( 2 ) )+
604: $ ABS( REFSUM*VT( 3 ) ).GT.ULP*
605: $ ( ABS( H( K, K ) )+ABS( H( K+1,
606: $ K+1 ) )+ABS( H( K+2, K+2 ) ) ) ) THEN
607: *
608: * ==== Starting a new bulge here would
609: * . create non-negligible fill. Use
610: * . the old one with trepidation. ====
611: *
612: H( K+1, K ) = BETA
613: H( K+2, K ) = ZERO
614: H( K+3, K ) = ZERO
615: ELSE
616: *
617: * ==== Starting a new bulge here would
618: * . create only negligible fill.
619: * . Replace the old reflector with
620: * . the new one. ====
621: *
622: H( K+1, K ) = H( K+1, K ) - REFSUM
623: H( K+2, K ) = ZERO
624: H( K+3, K ) = ZERO
625: V( 1, M ) = VT( 1 )
626: V( 2, M ) = VT( 2 )
627: V( 3, M ) = VT( 3 )
628: END IF
629: END IF
630: END IF
631: *
632: * ==== Apply reflection from the right and
633: * . the first column of update from the left.
634: * . These updates are required for the vigilant
635: * . deflation check. We still delay most of the
636: * . updates from the left for efficiency. ====
637: *
638: T1 = V( 1, M )
639: T2 = T1*V( 2, M )
640: T3 = T1*V( 3, M )
641: DO 70 J = JTOP, MIN( KBOT, K+3 )
642: REFSUM = H( J, K+1 ) + V( 2, M )*H( J, K+2 )
643: $ + V( 3, M )*H( J, K+3 )
644: H( J, K+1 ) = H( J, K+1 ) - REFSUM*T1
645: H( J, K+2 ) = H( J, K+2 ) - REFSUM*T2
646: H( J, K+3 ) = H( J, K+3 ) - REFSUM*T3
647: 70 CONTINUE
648: *
649: * ==== Perform update from left for subsequent
650: * . column. ====
651: *
652: REFSUM = H( K+1, K+1 ) + V( 2, M )*H( K+2, K+1 )
653: $ + V( 3, M )*H( K+3, K+1 )
654: H( K+1, K+1 ) = H( K+1, K+1 ) - REFSUM*T1
655: H( K+2, K+1 ) = H( K+2, K+1 ) - REFSUM*T2
656: H( K+3, K+1 ) = H( K+3, K+1 ) - REFSUM*T3
657: *
658: * ==== The following convergence test requires that
659: * . the tradition small-compared-to-nearby-diagonals
660: * . criterion and the Ahues & Tisseur (LAWN 122, 1997)
661: * . criteria both be satisfied. The latter improves
662: * . accuracy in some examples. Falling back on an
663: * . alternate convergence criterion when TST1 or TST2
664: * . is zero (as done here) is traditional but probably
665: * . unnecessary. ====
666: *
667: IF( K.LT.KTOP)
668: $ CYCLE
669: IF( H( K+1, K ).NE.ZERO ) THEN
670: TST1 = ABS( H( K, K ) ) + ABS( H( K+1, K+1 ) )
671: IF( TST1.EQ.ZERO ) THEN
672: IF( K.GE.KTOP+1 )
673: $ TST1 = TST1 + ABS( H( K, K-1 ) )
674: IF( K.GE.KTOP+2 )
675: $ TST1 = TST1 + ABS( H( K, K-2 ) )
676: IF( K.GE.KTOP+3 )
677: $ TST1 = TST1 + ABS( H( K, K-3 ) )
678: IF( K.LE.KBOT-2 )
679: $ TST1 = TST1 + ABS( H( K+2, K+1 ) )
680: IF( K.LE.KBOT-3 )
681: $ TST1 = TST1 + ABS( H( K+3, K+1 ) )
682: IF( K.LE.KBOT-4 )
683: $ TST1 = TST1 + ABS( H( K+4, K+1 ) )
684: END IF
685: IF( ABS( H( K+1, K ) ).LE.MAX( SMLNUM, ULP*TST1 ) )
686: $ THEN
687: H12 = MAX( ABS( H( K+1, K ) ), ABS( H( K, K+1 ) ) )
688: H21 = MIN( ABS( H( K+1, K ) ), ABS( H( K, K+1 ) ) )
689: H11 = MAX( ABS( H( K+1, K+1 ) ),
690: $ ABS( H( K, K )-H( K+1, K+1 ) ) )
691: H22 = MIN( ABS( H( K+1, K+1 ) ),
692: $ ABS( H( K, K )-H( K+1, K+1 ) ) )
693: SCL = H11 + H12
694: TST2 = H22*( H11 / SCL )
695: *
696: IF( TST2.EQ.ZERO .OR. H21*( H12 / SCL ).LE.
697: $ MAX( SMLNUM, ULP*TST2 ) ) THEN
698: H( K+1, K ) = ZERO
699: END IF
700: END IF
701: END IF
702: 80 CONTINUE
703: *
704: * ==== Multiply H by reflections from the left ====
705: *
706: IF( ACCUM ) THEN
707: JBOT = MIN( NDCOL, KBOT )
708: ELSE IF( WANTT ) THEN
709: JBOT = N
710: ELSE
711: JBOT = KBOT
712: END IF
713: *
714: DO 100 M = MBOT, MTOP, -1
715: K = KRCOL + 2*( M-1 )
716: T1 = V( 1, M )
717: T2 = T1*V( 2, M )
718: T3 = T1*V( 3, M )
719: DO 90 J = MAX( KTOP, KRCOL + 2*M ), JBOT
720: REFSUM = H( K+1, J ) + V( 2, M )*H( K+2, J )
721: $ + V( 3, M )*H( K+3, J )
722: H( K+1, J ) = H( K+1, J ) - REFSUM*T1
723: H( K+2, J ) = H( K+2, J ) - REFSUM*T2
724: H( K+3, J ) = H( K+3, J ) - REFSUM*T3
725: 90 CONTINUE
726: 100 CONTINUE
727: *
728: * ==== Accumulate orthogonal transformations. ====
729: *
730: IF( ACCUM ) THEN
731: *
732: * ==== Accumulate U. (If needed, update Z later
733: * . with an efficient matrix-matrix
734: * . multiply.) ====
735: *
736: DO 120 M = MBOT, MTOP, -1
737: K = KRCOL + 2*( M-1 )
738: KMS = K - INCOL
739: I2 = MAX( 1, KTOP-INCOL )
740: I2 = MAX( I2, KMS-(KRCOL-INCOL)+1 )
741: I4 = MIN( KDU, KRCOL + 2*( MBOT-1 ) - INCOL + 5 )
742: T1 = V( 1, M )
743: T2 = T1*V( 2, M )
744: T3 = T1*V( 3, M )
745: DO 110 J = I2, I4
746: REFSUM = U( J, KMS+1 ) + V( 2, M )*U( J, KMS+2 )
747: $ + V( 3, M )*U( J, KMS+3 )
748: U( J, KMS+1 ) = U( J, KMS+1 ) - REFSUM*T1
749: U( J, KMS+2 ) = U( J, KMS+2 ) - REFSUM*T2
750: U( J, KMS+3 ) = U( J, KMS+3 ) - REFSUM*T3
751: 110 CONTINUE
752: 120 CONTINUE
753: ELSE IF( WANTZ ) THEN
754: *
755: * ==== U is not accumulated, so update Z
756: * . now by multiplying by reflections
757: * . from the right. ====
758: *
759: DO 140 M = MBOT, MTOP, -1
760: K = KRCOL + 2*( M-1 )
761: T1 = V( 1, M )
762: T2 = T1*V( 2, M )
763: T3 = T1*V( 3, M )
764: DO 130 J = ILOZ, IHIZ
765: REFSUM = Z( J, K+1 ) + V( 2, M )*Z( J, K+2 )
766: $ + V( 3, M )*Z( J, K+3 )
767: Z( J, K+1 ) = Z( J, K+1 ) - REFSUM*T1
768: Z( J, K+2 ) = Z( J, K+2 ) - REFSUM*T2
769: Z( J, K+3 ) = Z( J, K+3 ) - REFSUM*T3
770: 130 CONTINUE
771: 140 CONTINUE
772: END IF
773: *
774: * ==== End of near-the-diagonal bulge chase. ====
775: *
776: 145 CONTINUE
777: *
778: * ==== Use U (if accumulated) to update far-from-diagonal
779: * . entries in H. If required, use U to update Z as
780: * . well. ====
781: *
782: IF( ACCUM ) THEN
783: IF( WANTT ) THEN
784: JTOP = 1
785: JBOT = N
786: ELSE
787: JTOP = KTOP
788: JBOT = KBOT
789: END IF
790: K1 = MAX( 1, KTOP-INCOL )
791: NU = ( KDU-MAX( 0, NDCOL-KBOT ) ) - K1 + 1
792: *
793: * ==== Horizontal Multiply ====
794: *
795: DO 150 JCOL = MIN( NDCOL, KBOT ) + 1, JBOT, NH
796: JLEN = MIN( NH, JBOT-JCOL+1 )
797: CALL DGEMM( 'C', 'N', NU, JLEN, NU, ONE, U( K1, K1 ),
798: $ LDU, H( INCOL+K1, JCOL ), LDH, ZERO, WH,
799: $ LDWH )
800: CALL DLACPY( 'ALL', NU, JLEN, WH, LDWH,
801: $ H( INCOL+K1, JCOL ), LDH )
802: 150 CONTINUE
803: *
804: * ==== Vertical multiply ====
805: *
806: DO 160 JROW = JTOP, MAX( KTOP, INCOL ) - 1, NV
807: JLEN = MIN( NV, MAX( KTOP, INCOL )-JROW )
808: CALL DGEMM( 'N', 'N', JLEN, NU, NU, ONE,
809: $ H( JROW, INCOL+K1 ), LDH, U( K1, K1 ),
810: $ LDU, ZERO, WV, LDWV )
811: CALL DLACPY( 'ALL', JLEN, NU, WV, LDWV,
812: $ H( JROW, INCOL+K1 ), LDH )
813: 160 CONTINUE
814: *
815: * ==== Z multiply (also vertical) ====
816: *
817: IF( WANTZ ) THEN
818: DO 170 JROW = ILOZ, IHIZ, NV
819: JLEN = MIN( NV, IHIZ-JROW+1 )
820: CALL DGEMM( 'N', 'N', JLEN, NU, NU, ONE,
821: $ Z( JROW, INCOL+K1 ), LDZ, U( K1, K1 ),
822: $ LDU, ZERO, WV, LDWV )
823: CALL DLACPY( 'ALL', JLEN, NU, WV, LDWV,
824: $ Z( JROW, INCOL+K1 ), LDZ )
825: 170 CONTINUE
826: END IF
827: END IF
828: 180 CONTINUE
829: *
830: * ==== End of DLAQR5 ====
831: *
832: END
CVSweb interface <joel.bertrand@systella.fr>