1: *> \brief \b DLASQ2 computes all the eigenvalues of the symmetric positive definite tridiagonal matrix associated with the qd Array Z to high relative accuracy. Used by sbdsqr and sstegr.
2: *
3: * =========== DOCUMENTATION ===========
4: *
5: * Online html documentation available at
6: * http://www.netlib.org/lapack/explore-html/
7: *
8: *> \htmlonly
9: *> Download DLASQ2 + dependencies
10: *> <a href="http://www.netlib.org/cgi-bin/netlibfiles.tgz?format=tgz&filename=/lapack/lapack_routine/dlasq2.f">
11: *> [TGZ]</a>
12: *> <a href="http://www.netlib.org/cgi-bin/netlibfiles.zip?format=zip&filename=/lapack/lapack_routine/dlasq2.f">
13: *> [ZIP]</a>
14: *> <a href="http://www.netlib.org/cgi-bin/netlibfiles.txt?format=txt&filename=/lapack/lapack_routine/dlasq2.f">
15: *> [TXT]</a>
16: *> \endhtmlonly
17: *
18: * Definition:
19: * ===========
20: *
21: * SUBROUTINE DLASQ2( N, Z, INFO )
22: *
23: * .. Scalar Arguments ..
24: * INTEGER INFO, N
25: * ..
26: * .. Array Arguments ..
27: * DOUBLE PRECISION Z( * )
28: * ..
29: *
30: *
31: *> \par Purpose:
32: * =============
33: *>
34: *> \verbatim
35: *>
36: *> DLASQ2 computes all the eigenvalues of the symmetric positive
37: *> definite tridiagonal matrix associated with the qd array Z to high
38: *> relative accuracy are computed to high relative accuracy, in the
39: *> absence of denormalization, underflow and overflow.
40: *>
41: *> To see the relation of Z to the tridiagonal matrix, let L be a
42: *> unit lower bidiagonal matrix with subdiagonals Z(2,4,6,,..) and
43: *> let U be an upper bidiagonal matrix with 1's above and diagonal
44: *> Z(1,3,5,,..). The tridiagonal is L*U or, if you prefer, the
45: *> symmetric tridiagonal to which it is similar.
46: *>
47: *> Note : DLASQ2 defines a logical variable, IEEE, which is true
48: *> on machines which follow ieee-754 floating-point standard in their
49: *> handling of infinities and NaNs, and false otherwise. This variable
50: *> is passed to DLASQ3.
51: *> \endverbatim
52: *
53: * Arguments:
54: * ==========
55: *
56: *> \param[in] N
57: *> \verbatim
58: *> N is INTEGER
59: *> The number of rows and columns in the matrix. N >= 0.
60: *> \endverbatim
61: *>
62: *> \param[in,out] Z
63: *> \verbatim
64: *> Z is DOUBLE PRECISION array, dimension ( 4*N )
65: *> On entry Z holds the qd array. On exit, entries 1 to N hold
66: *> the eigenvalues in decreasing order, Z( 2*N+1 ) holds the
67: *> trace, and Z( 2*N+2 ) holds the sum of the eigenvalues. If
68: *> N > 2, then Z( 2*N+3 ) holds the iteration count, Z( 2*N+4 )
69: *> holds NDIVS/NIN^2, and Z( 2*N+5 ) holds the percentage of
70: *> shifts that failed.
71: *> \endverbatim
72: *>
73: *> \param[out] INFO
74: *> \verbatim
75: *> INFO is INTEGER
76: *> = 0: successful exit
77: *> < 0: if the i-th argument is a scalar and had an illegal
78: *> value, then INFO = -i, if the i-th argument is an
79: *> array and the j-entry had an illegal value, then
80: *> INFO = -(i*100+j)
81: *> > 0: the algorithm failed
82: *> = 1, a split was marked by a positive value in E
83: *> = 2, current block of Z not diagonalized after 100*N
84: *> iterations (in inner while loop). On exit Z holds
85: *> a qd array with the same eigenvalues as the given Z.
86: *> = 3, termination criterion of outer while loop not met
87: *> (program created more than N unreduced blocks)
88: *> \endverbatim
89: *
90: * Authors:
91: * ========
92: *
93: *> \author Univ. of Tennessee
94: *> \author Univ. of California Berkeley
95: *> \author Univ. of Colorado Denver
96: *> \author NAG Ltd.
97: *
98: *> \date September 2012
99: *
100: *> \ingroup auxOTHERcomputational
101: *
102: *> \par Further Details:
103: * =====================
104: *>
105: *> \verbatim
106: *>
107: *> Local Variables: I0:N0 defines a current unreduced segment of Z.
108: *> The shifts are accumulated in SIGMA. Iteration count is in ITER.
109: *> Ping-pong is controlled by PP (alternates between 0 and 1).
110: *> \endverbatim
111: *>
112: * =====================================================================
113: SUBROUTINE DLASQ2( N, Z, INFO )
114: *
115: * -- LAPACK computational routine (version 3.4.2) --
116: * -- LAPACK is a software package provided by Univ. of Tennessee, --
117: * -- Univ. of California Berkeley, Univ. of Colorado Denver and NAG Ltd..--
118: * September 2012
119: *
120: * .. Scalar Arguments ..
121: INTEGER INFO, N
122: * ..
123: * .. Array Arguments ..
124: DOUBLE PRECISION Z( * )
125: * ..
126: *
127: * =====================================================================
128: *
129: * .. Parameters ..
130: DOUBLE PRECISION CBIAS
131: PARAMETER ( CBIAS = 1.50D0 )
132: DOUBLE PRECISION ZERO, HALF, ONE, TWO, FOUR, HUNDRD
133: PARAMETER ( ZERO = 0.0D0, HALF = 0.5D0, ONE = 1.0D0,
134: $ TWO = 2.0D0, FOUR = 4.0D0, HUNDRD = 100.0D0 )
135: * ..
136: * .. Local Scalars ..
137: LOGICAL IEEE
138: INTEGER I0, I1, I4, IINFO, IPN4, ITER, IWHILA, IWHILB,
139: $ K, KMIN, N0, N1, NBIG, NDIV, NFAIL, PP, SPLT,
140: $ TTYPE
141: DOUBLE PRECISION D, DEE, DEEMIN, DESIG, DMIN, DMIN1, DMIN2, DN,
142: $ DN1, DN2, E, EMAX, EMIN, EPS, G, OLDEMN, QMAX,
143: $ QMIN, S, SAFMIN, SIGMA, T, TAU, TEMP, TOL,
144: $ TOL2, TRACE, ZMAX, TEMPE, TEMPQ
145: * ..
146: * .. External Subroutines ..
147: EXTERNAL DLASQ3, DLASRT, XERBLA
148: * ..
149: * .. External Functions ..
150: INTEGER ILAENV
151: DOUBLE PRECISION DLAMCH
152: EXTERNAL DLAMCH, ILAENV
153: * ..
154: * .. Intrinsic Functions ..
155: INTRINSIC ABS, DBLE, MAX, MIN, SQRT
156: * ..
157: * .. Executable Statements ..
158: *
159: * Test the input arguments.
160: * (in case DLASQ2 is not called by DLASQ1)
161: *
162: INFO = 0
163: EPS = DLAMCH( 'Precision' )
164: SAFMIN = DLAMCH( 'Safe minimum' )
165: TOL = EPS*HUNDRD
166: TOL2 = TOL**2
167: *
168: IF( N.LT.0 ) THEN
169: INFO = -1
170: CALL XERBLA( 'DLASQ2', 1 )
171: RETURN
172: ELSE IF( N.EQ.0 ) THEN
173: RETURN
174: ELSE IF( N.EQ.1 ) THEN
175: *
176: * 1-by-1 case.
177: *
178: IF( Z( 1 ).LT.ZERO ) THEN
179: INFO = -201
180: CALL XERBLA( 'DLASQ2', 2 )
181: END IF
182: RETURN
183: ELSE IF( N.EQ.2 ) THEN
184: *
185: * 2-by-2 case.
186: *
187: IF( Z( 2 ).LT.ZERO .OR. Z( 3 ).LT.ZERO ) THEN
188: INFO = -2
189: CALL XERBLA( 'DLASQ2', 2 )
190: RETURN
191: ELSE IF( Z( 3 ).GT.Z( 1 ) ) THEN
192: D = Z( 3 )
193: Z( 3 ) = Z( 1 )
194: Z( 1 ) = D
195: END IF
196: Z( 5 ) = Z( 1 ) + Z( 2 ) + Z( 3 )
197: IF( Z( 2 ).GT.Z( 3 )*TOL2 ) THEN
198: T = HALF*( ( Z( 1 )-Z( 3 ) )+Z( 2 ) )
199: S = Z( 3 )*( Z( 2 ) / T )
200: IF( S.LE.T ) THEN
201: S = Z( 3 )*( Z( 2 ) / ( T*( ONE+SQRT( ONE+S / T ) ) ) )
202: ELSE
203: S = Z( 3 )*( Z( 2 ) / ( T+SQRT( T )*SQRT( T+S ) ) )
204: END IF
205: T = Z( 1 ) + ( S+Z( 2 ) )
206: Z( 3 ) = Z( 3 )*( Z( 1 ) / T )
207: Z( 1 ) = T
208: END IF
209: Z( 2 ) = Z( 3 )
210: Z( 6 ) = Z( 2 ) + Z( 1 )
211: RETURN
212: END IF
213: *
214: * Check for negative data and compute sums of q's and e's.
215: *
216: Z( 2*N ) = ZERO
217: EMIN = Z( 2 )
218: QMAX = ZERO
219: ZMAX = ZERO
220: D = ZERO
221: E = ZERO
222: *
223: DO 10 K = 1, 2*( N-1 ), 2
224: IF( Z( K ).LT.ZERO ) THEN
225: INFO = -( 200+K )
226: CALL XERBLA( 'DLASQ2', 2 )
227: RETURN
228: ELSE IF( Z( K+1 ).LT.ZERO ) THEN
229: INFO = -( 200+K+1 )
230: CALL XERBLA( 'DLASQ2', 2 )
231: RETURN
232: END IF
233: D = D + Z( K )
234: E = E + Z( K+1 )
235: QMAX = MAX( QMAX, Z( K ) )
236: EMIN = MIN( EMIN, Z( K+1 ) )
237: ZMAX = MAX( QMAX, ZMAX, Z( K+1 ) )
238: 10 CONTINUE
239: IF( Z( 2*N-1 ).LT.ZERO ) THEN
240: INFO = -( 200+2*N-1 )
241: CALL XERBLA( 'DLASQ2', 2 )
242: RETURN
243: END IF
244: D = D + Z( 2*N-1 )
245: QMAX = MAX( QMAX, Z( 2*N-1 ) )
246: ZMAX = MAX( QMAX, ZMAX )
247: *
248: * Check for diagonality.
249: *
250: IF( E.EQ.ZERO ) THEN
251: DO 20 K = 2, N
252: Z( K ) = Z( 2*K-1 )
253: 20 CONTINUE
254: CALL DLASRT( 'D', N, Z, IINFO )
255: Z( 2*N-1 ) = D
256: RETURN
257: END IF
258: *
259: TRACE = D + E
260: *
261: * Check for zero data.
262: *
263: IF( TRACE.EQ.ZERO ) THEN
264: Z( 2*N-1 ) = ZERO
265: RETURN
266: END IF
267: *
268: * Check whether the machine is IEEE conformable.
269: *
270: IEEE = ILAENV( 10, 'DLASQ2', 'N', 1, 2, 3, 4 ).EQ.1 .AND.
271: $ ILAENV( 11, 'DLASQ2', 'N', 1, 2, 3, 4 ).EQ.1
272: *
273: * Rearrange data for locality: Z=(q1,qq1,e1,ee1,q2,qq2,e2,ee2,...).
274: *
275: DO 30 K = 2*N, 2, -2
276: Z( 2*K ) = ZERO
277: Z( 2*K-1 ) = Z( K )
278: Z( 2*K-2 ) = ZERO
279: Z( 2*K-3 ) = Z( K-1 )
280: 30 CONTINUE
281: *
282: I0 = 1
283: N0 = N
284: *
285: * Reverse the qd-array, if warranted.
286: *
287: IF( CBIAS*Z( 4*I0-3 ).LT.Z( 4*N0-3 ) ) THEN
288: IPN4 = 4*( I0+N0 )
289: DO 40 I4 = 4*I0, 2*( I0+N0-1 ), 4
290: TEMP = Z( I4-3 )
291: Z( I4-3 ) = Z( IPN4-I4-3 )
292: Z( IPN4-I4-3 ) = TEMP
293: TEMP = Z( I4-1 )
294: Z( I4-1 ) = Z( IPN4-I4-5 )
295: Z( IPN4-I4-5 ) = TEMP
296: 40 CONTINUE
297: END IF
298: *
299: * Initial split checking via dqd and Li's test.
300: *
301: PP = 0
302: *
303: DO 80 K = 1, 2
304: *
305: D = Z( 4*N0+PP-3 )
306: DO 50 I4 = 4*( N0-1 ) + PP, 4*I0 + PP, -4
307: IF( Z( I4-1 ).LE.TOL2*D ) THEN
308: Z( I4-1 ) = -ZERO
309: D = Z( I4-3 )
310: ELSE
311: D = Z( I4-3 )*( D / ( D+Z( I4-1 ) ) )
312: END IF
313: 50 CONTINUE
314: *
315: * dqd maps Z to ZZ plus Li's test.
316: *
317: EMIN = Z( 4*I0+PP+1 )
318: D = Z( 4*I0+PP-3 )
319: DO 60 I4 = 4*I0 + PP, 4*( N0-1 ) + PP, 4
320: Z( I4-2*PP-2 ) = D + Z( I4-1 )
321: IF( Z( I4-1 ).LE.TOL2*D ) THEN
322: Z( I4-1 ) = -ZERO
323: Z( I4-2*PP-2 ) = D
324: Z( I4-2*PP ) = ZERO
325: D = Z( I4+1 )
326: ELSE IF( SAFMIN*Z( I4+1 ).LT.Z( I4-2*PP-2 ) .AND.
327: $ SAFMIN*Z( I4-2*PP-2 ).LT.Z( I4+1 ) ) THEN
328: TEMP = Z( I4+1 ) / Z( I4-2*PP-2 )
329: Z( I4-2*PP ) = Z( I4-1 )*TEMP
330: D = D*TEMP
331: ELSE
332: Z( I4-2*PP ) = Z( I4+1 )*( Z( I4-1 ) / Z( I4-2*PP-2 ) )
333: D = Z( I4+1 )*( D / Z( I4-2*PP-2 ) )
334: END IF
335: EMIN = MIN( EMIN, Z( I4-2*PP ) )
336: 60 CONTINUE
337: Z( 4*N0-PP-2 ) = D
338: *
339: * Now find qmax.
340: *
341: QMAX = Z( 4*I0-PP-2 )
342: DO 70 I4 = 4*I0 - PP + 2, 4*N0 - PP - 2, 4
343: QMAX = MAX( QMAX, Z( I4 ) )
344: 70 CONTINUE
345: *
346: * Prepare for the next iteration on K.
347: *
348: PP = 1 - PP
349: 80 CONTINUE
350: *
351: * Initialise variables to pass to DLASQ3.
352: *
353: TTYPE = 0
354: DMIN1 = ZERO
355: DMIN2 = ZERO
356: DN = ZERO
357: DN1 = ZERO
358: DN2 = ZERO
359: G = ZERO
360: TAU = ZERO
361: *
362: ITER = 2
363: NFAIL = 0
364: NDIV = 2*( N0-I0 )
365: *
366: DO 160 IWHILA = 1, N + 1
367: IF( N0.LT.1 )
368: $ GO TO 170
369: *
370: * While array unfinished do
371: *
372: * E(N0) holds the value of SIGMA when submatrix in I0:N0
373: * splits from the rest of the array, but is negated.
374: *
375: DESIG = ZERO
376: IF( N0.EQ.N ) THEN
377: SIGMA = ZERO
378: ELSE
379: SIGMA = -Z( 4*N0-1 )
380: END IF
381: IF( SIGMA.LT.ZERO ) THEN
382: INFO = 1
383: RETURN
384: END IF
385: *
386: * Find last unreduced submatrix's top index I0, find QMAX and
387: * EMIN. Find Gershgorin-type bound if Q's much greater than E's.
388: *
389: EMAX = ZERO
390: IF( N0.GT.I0 ) THEN
391: EMIN = ABS( Z( 4*N0-5 ) )
392: ELSE
393: EMIN = ZERO
394: END IF
395: QMIN = Z( 4*N0-3 )
396: QMAX = QMIN
397: DO 90 I4 = 4*N0, 8, -4
398: IF( Z( I4-5 ).LE.ZERO )
399: $ GO TO 100
400: IF( QMIN.GE.FOUR*EMAX ) THEN
401: QMIN = MIN( QMIN, Z( I4-3 ) )
402: EMAX = MAX( EMAX, Z( I4-5 ) )
403: END IF
404: QMAX = MAX( QMAX, Z( I4-7 )+Z( I4-5 ) )
405: EMIN = MIN( EMIN, Z( I4-5 ) )
406: 90 CONTINUE
407: I4 = 4
408: *
409: 100 CONTINUE
410: I0 = I4 / 4
411: PP = 0
412: *
413: IF( N0-I0.GT.1 ) THEN
414: DEE = Z( 4*I0-3 )
415: DEEMIN = DEE
416: KMIN = I0
417: DO 110 I4 = 4*I0+1, 4*N0-3, 4
418: DEE = Z( I4 )*( DEE /( DEE+Z( I4-2 ) ) )
419: IF( DEE.LE.DEEMIN ) THEN
420: DEEMIN = DEE
421: KMIN = ( I4+3 )/4
422: END IF
423: 110 CONTINUE
424: IF( (KMIN-I0)*2.LT.N0-KMIN .AND.
425: $ DEEMIN.LE.HALF*Z(4*N0-3) ) THEN
426: IPN4 = 4*( I0+N0 )
427: PP = 2
428: DO 120 I4 = 4*I0, 2*( I0+N0-1 ), 4
429: TEMP = Z( I4-3 )
430: Z( I4-3 ) = Z( IPN4-I4-3 )
431: Z( IPN4-I4-3 ) = TEMP
432: TEMP = Z( I4-2 )
433: Z( I4-2 ) = Z( IPN4-I4-2 )
434: Z( IPN4-I4-2 ) = TEMP
435: TEMP = Z( I4-1 )
436: Z( I4-1 ) = Z( IPN4-I4-5 )
437: Z( IPN4-I4-5 ) = TEMP
438: TEMP = Z( I4 )
439: Z( I4 ) = Z( IPN4-I4-4 )
440: Z( IPN4-I4-4 ) = TEMP
441: 120 CONTINUE
442: END IF
443: END IF
444: *
445: * Put -(initial shift) into DMIN.
446: *
447: DMIN = -MAX( ZERO, QMIN-TWO*SQRT( QMIN )*SQRT( EMAX ) )
448: *
449: * Now I0:N0 is unreduced.
450: * PP = 0 for ping, PP = 1 for pong.
451: * PP = 2 indicates that flipping was applied to the Z array and
452: * and that the tests for deflation upon entry in DLASQ3
453: * should not be performed.
454: *
455: NBIG = 100*( N0-I0+1 )
456: DO 140 IWHILB = 1, NBIG
457: IF( I0.GT.N0 )
458: $ GO TO 150
459: *
460: * While submatrix unfinished take a good dqds step.
461: *
462: CALL DLASQ3( I0, N0, Z, PP, DMIN, SIGMA, DESIG, QMAX, NFAIL,
463: $ ITER, NDIV, IEEE, TTYPE, DMIN1, DMIN2, DN, DN1,
464: $ DN2, G, TAU )
465: *
466: PP = 1 - PP
467: *
468: * When EMIN is very small check for splits.
469: *
470: IF( PP.EQ.0 .AND. N0-I0.GE.3 ) THEN
471: IF( Z( 4*N0 ).LE.TOL2*QMAX .OR.
472: $ Z( 4*N0-1 ).LE.TOL2*SIGMA ) THEN
473: SPLT = I0 - 1
474: QMAX = Z( 4*I0-3 )
475: EMIN = Z( 4*I0-1 )
476: OLDEMN = Z( 4*I0 )
477: DO 130 I4 = 4*I0, 4*( N0-3 ), 4
478: IF( Z( I4 ).LE.TOL2*Z( I4-3 ) .OR.
479: $ Z( I4-1 ).LE.TOL2*SIGMA ) THEN
480: Z( I4-1 ) = -SIGMA
481: SPLT = I4 / 4
482: QMAX = ZERO
483: EMIN = Z( I4+3 )
484: OLDEMN = Z( I4+4 )
485: ELSE
486: QMAX = MAX( QMAX, Z( I4+1 ) )
487: EMIN = MIN( EMIN, Z( I4-1 ) )
488: OLDEMN = MIN( OLDEMN, Z( I4 ) )
489: END IF
490: 130 CONTINUE
491: Z( 4*N0-1 ) = EMIN
492: Z( 4*N0 ) = OLDEMN
493: I0 = SPLT + 1
494: END IF
495: END IF
496: *
497: 140 CONTINUE
498: *
499: INFO = 2
500: *
501: * Maximum number of iterations exceeded, restore the shift
502: * SIGMA and place the new d's and e's in a qd array.
503: * This might need to be done for several blocks
504: *
505: I1 = I0
506: N1 = N0
507: 145 CONTINUE
508: TEMPQ = Z( 4*I0-3 )
509: Z( 4*I0-3 ) = Z( 4*I0-3 ) + SIGMA
510: DO K = I0+1, N0
511: TEMPE = Z( 4*K-5 )
512: Z( 4*K-5 ) = Z( 4*K-5 ) * (TEMPQ / Z( 4*K-7 ))
513: TEMPQ = Z( 4*K-3 )
514: Z( 4*K-3 ) = Z( 4*K-3 ) + SIGMA + TEMPE - Z( 4*K-5 )
515: END DO
516: *
517: * Prepare to do this on the previous block if there is one
518: *
519: IF( I1.GT.1 ) THEN
520: N1 = I1-1
521: DO WHILE( ( I1.GE.2 ) .AND. ( Z(4*I1-5).GE.ZERO ) )
522: I1 = I1 - 1
523: END DO
524: SIGMA = -Z(4*N1-1)
525: GO TO 145
526: END IF
527:
528: DO K = 1, N
529: Z( 2*K-1 ) = Z( 4*K-3 )
530: *
531: * Only the block 1..N0 is unfinished. The rest of the e's
532: * must be essentially zero, although sometimes other data
533: * has been stored in them.
534: *
535: IF( K.LT.N0 ) THEN
536: Z( 2*K ) = Z( 4*K-1 )
537: ELSE
538: Z( 2*K ) = 0
539: END IF
540: END DO
541: RETURN
542: *
543: * end IWHILB
544: *
545: 150 CONTINUE
546: *
547: 160 CONTINUE
548: *
549: INFO = 3
550: RETURN
551: *
552: * end IWHILA
553: *
554: 170 CONTINUE
555: *
556: * Move q's to the front.
557: *
558: DO 180 K = 2, N
559: Z( K ) = Z( 4*K-3 )
560: 180 CONTINUE
561: *
562: * Sort and compute sum of eigenvalues.
563: *
564: CALL DLASRT( 'D', N, Z, IINFO )
565: *
566: E = ZERO
567: DO 190 K = N, 1, -1
568: E = E + Z( K )
569: 190 CONTINUE
570: *
571: * Store trace, sum(eigenvalues) and information on performance.
572: *
573: Z( 2*N+1 ) = TRACE
574: Z( 2*N+2 ) = E
575: Z( 2*N+3 ) = DBLE( ITER )
576: Z( 2*N+4 ) = DBLE( NDIV ) / DBLE( N**2 )
577: Z( 2*N+5 ) = HUNDRD*NFAIL / DBLE( ITER )
578: RETURN
579: *
580: * End of DLASQ2
581: *
582: END
CVSweb interface <joel.bertrand@systella.fr>