File:  [local] / rpl / lapack / lapack / zsteqr.f
Revision 1.17: download - view: text, annotated - select for diffs - revision graph
Mon Aug 7 08:39:37 2023 UTC (9 months ago) by bertrand
Branches: MAIN
CVS tags: rpl-4_1_35, rpl-4_1_34, HEAD
Première mise à jour de lapack et blas.

    1: *> \brief \b ZSTEQR
    2: *
    3: *  =========== DOCUMENTATION ===========
    4: *
    5: * Online html documentation available at
    6: *            http://www.netlib.org/lapack/explore-html/
    7: *
    8: *> \htmlonly
    9: *> Download ZSTEQR + dependencies
   10: *> <a href="http://www.netlib.org/cgi-bin/netlibfiles.tgz?format=tgz&filename=/lapack/lapack_routine/zsteqr.f">
   11: *> [TGZ]</a>
   12: *> <a href="http://www.netlib.org/cgi-bin/netlibfiles.zip?format=zip&filename=/lapack/lapack_routine/zsteqr.f">
   13: *> [ZIP]</a>
   14: *> <a href="http://www.netlib.org/cgi-bin/netlibfiles.txt?format=txt&filename=/lapack/lapack_routine/zsteqr.f">
   15: *> [TXT]</a>
   16: *> \endhtmlonly
   17: *
   18: *  Definition:
   19: *  ===========
   20: *
   21: *       SUBROUTINE ZSTEQR( COMPZ, N, D, E, Z, LDZ, WORK, INFO )
   22: *
   23: *       .. Scalar Arguments ..
   24: *       CHARACTER          COMPZ
   25: *       INTEGER            INFO, LDZ, N
   26: *       ..
   27: *       .. Array Arguments ..
   28: *       DOUBLE PRECISION   D( * ), E( * ), WORK( * )
   29: *       COMPLEX*16         Z( LDZ, * )
   30: *       ..
   31: *
   32: *
   33: *> \par Purpose:
   34: *  =============
   35: *>
   36: *> \verbatim
   37: *>
   38: *> ZSTEQR computes all eigenvalues and, optionally, eigenvectors of a
   39: *> symmetric tridiagonal matrix using the implicit QL or QR method.
   40: *> The eigenvectors of a full or band complex Hermitian matrix can also
   41: *> be found if ZHETRD or ZHPTRD or ZHBTRD has been used to reduce this
   42: *> matrix to tridiagonal form.
   43: *> \endverbatim
   44: *
   45: *  Arguments:
   46: *  ==========
   47: *
   48: *> \param[in] COMPZ
   49: *> \verbatim
   50: *>          COMPZ is CHARACTER*1
   51: *>          = 'N':  Compute eigenvalues only.
   52: *>          = 'V':  Compute eigenvalues and eigenvectors of the original
   53: *>                  Hermitian matrix.  On entry, Z must contain the
   54: *>                  unitary matrix used to reduce the original matrix
   55: *>                  to tridiagonal form.
   56: *>          = 'I':  Compute eigenvalues and eigenvectors of the
   57: *>                  tridiagonal matrix.  Z is initialized to the identity
   58: *>                  matrix.
   59: *> \endverbatim
   60: *>
   61: *> \param[in] N
   62: *> \verbatim
   63: *>          N is INTEGER
   64: *>          The order of the matrix.  N >= 0.
   65: *> \endverbatim
   66: *>
   67: *> \param[in,out] D
   68: *> \verbatim
   69: *>          D is DOUBLE PRECISION array, dimension (N)
   70: *>          On entry, the diagonal elements of the tridiagonal matrix.
   71: *>          On exit, if INFO = 0, the eigenvalues in ascending order.
   72: *> \endverbatim
   73: *>
   74: *> \param[in,out] E
   75: *> \verbatim
   76: *>          E is DOUBLE PRECISION array, dimension (N-1)
   77: *>          On entry, the (n-1) subdiagonal elements of the tridiagonal
   78: *>          matrix.
   79: *>          On exit, E has been destroyed.
   80: *> \endverbatim
   81: *>
   82: *> \param[in,out] Z
   83: *> \verbatim
   84: *>          Z is COMPLEX*16 array, dimension (LDZ, N)
   85: *>          On entry, if  COMPZ = 'V', then Z contains the unitary
   86: *>          matrix used in the reduction to tridiagonal form.
   87: *>          On exit, if INFO = 0, then if COMPZ = 'V', Z contains the
   88: *>          orthonormal eigenvectors of the original Hermitian matrix,
   89: *>          and if COMPZ = 'I', Z contains the orthonormal eigenvectors
   90: *>          of the symmetric tridiagonal matrix.
   91: *>          If COMPZ = 'N', then Z is not referenced.
   92: *> \endverbatim
   93: *>
   94: *> \param[in] LDZ
   95: *> \verbatim
   96: *>          LDZ is INTEGER
   97: *>          The leading dimension of the array Z.  LDZ >= 1, and if
   98: *>          eigenvectors are desired, then  LDZ >= max(1,N).
   99: *> \endverbatim
  100: *>
  101: *> \param[out] WORK
  102: *> \verbatim
  103: *>          WORK is DOUBLE PRECISION array, dimension (max(1,2*N-2))
  104: *>          If COMPZ = 'N', then WORK is not referenced.
  105: *> \endverbatim
  106: *>
  107: *> \param[out] INFO
  108: *> \verbatim
  109: *>          INFO is INTEGER
  110: *>          = 0:  successful exit
  111: *>          < 0:  if INFO = -i, the i-th argument had an illegal value
  112: *>          > 0:  the algorithm has failed to find all the eigenvalues in
  113: *>                a total of 30*N iterations; if INFO = i, then i
  114: *>                elements of E have not converged to zero; on exit, D
  115: *>                and E contain the elements of a symmetric tridiagonal
  116: *>                matrix which is unitarily similar to the original
  117: *>                matrix.
  118: *> \endverbatim
  119: *
  120: *  Authors:
  121: *  ========
  122: *
  123: *> \author Univ. of Tennessee
  124: *> \author Univ. of California Berkeley
  125: *> \author Univ. of Colorado Denver
  126: *> \author NAG Ltd.
  127: *
  128: *> \ingroup complex16OTHERcomputational
  129: *
  130: *  =====================================================================
  131:       SUBROUTINE ZSTEQR( COMPZ, N, D, E, Z, LDZ, WORK, INFO )
  132: *
  133: *  -- LAPACK computational routine --
  134: *  -- LAPACK is a software package provided by Univ. of Tennessee,    --
  135: *  -- Univ. of California Berkeley, Univ. of Colorado Denver and NAG Ltd..--
  136: *
  137: *     .. Scalar Arguments ..
  138:       CHARACTER          COMPZ
  139:       INTEGER            INFO, LDZ, N
  140: *     ..
  141: *     .. Array Arguments ..
  142:       DOUBLE PRECISION   D( * ), E( * ), WORK( * )
  143:       COMPLEX*16         Z( LDZ, * )
  144: *     ..
  145: *
  146: *  =====================================================================
  147: *
  148: *     .. Parameters ..
  149:       DOUBLE PRECISION   ZERO, ONE, TWO, THREE
  150:       PARAMETER          ( ZERO = 0.0D0, ONE = 1.0D0, TWO = 2.0D0,
  151:      $                   THREE = 3.0D0 )
  152:       COMPLEX*16         CZERO, CONE
  153:       PARAMETER          ( CZERO = ( 0.0D0, 0.0D0 ),
  154:      $                   CONE = ( 1.0D0, 0.0D0 ) )
  155:       INTEGER            MAXIT
  156:       PARAMETER          ( MAXIT = 30 )
  157: *     ..
  158: *     .. Local Scalars ..
  159:       INTEGER            I, ICOMPZ, II, ISCALE, J, JTOT, K, L, L1, LEND,
  160:      $                   LENDM1, LENDP1, LENDSV, LM1, LSV, M, MM, MM1,
  161:      $                   NM1, NMAXIT
  162:       DOUBLE PRECISION   ANORM, B, C, EPS, EPS2, F, G, P, R, RT1, RT2,
  163:      $                   S, SAFMAX, SAFMIN, SSFMAX, SSFMIN, TST
  164: *     ..
  165: *     .. External Functions ..
  166:       LOGICAL            LSAME
  167:       DOUBLE PRECISION   DLAMCH, DLANST, DLAPY2
  168:       EXTERNAL           LSAME, DLAMCH, DLANST, DLAPY2
  169: *     ..
  170: *     .. External Subroutines ..
  171:       EXTERNAL           DLAE2, DLAEV2, DLARTG, DLASCL, DLASRT, XERBLA,
  172:      $                   ZLASET, ZLASR, ZSWAP
  173: *     ..
  174: *     .. Intrinsic Functions ..
  175:       INTRINSIC          ABS, MAX, SIGN, SQRT
  176: *     ..
  177: *     .. Executable Statements ..
  178: *
  179: *     Test the input parameters.
  180: *
  181:       INFO = 0
  182: *
  183:       IF( LSAME( COMPZ, 'N' ) ) THEN
  184:          ICOMPZ = 0
  185:       ELSE IF( LSAME( COMPZ, 'V' ) ) THEN
  186:          ICOMPZ = 1
  187:       ELSE IF( LSAME( COMPZ, 'I' ) ) THEN
  188:          ICOMPZ = 2
  189:       ELSE
  190:          ICOMPZ = -1
  191:       END IF
  192:       IF( ICOMPZ.LT.0 ) THEN
  193:          INFO = -1
  194:       ELSE IF( N.LT.0 ) THEN
  195:          INFO = -2
  196:       ELSE IF( ( LDZ.LT.1 ) .OR. ( ICOMPZ.GT.0 .AND. LDZ.LT.MAX( 1,
  197:      $         N ) ) ) THEN
  198:          INFO = -6
  199:       END IF
  200:       IF( INFO.NE.0 ) THEN
  201:          CALL XERBLA( 'ZSTEQR', -INFO )
  202:          RETURN
  203:       END IF
  204: *
  205: *     Quick return if possible
  206: *
  207:       IF( N.EQ.0 )
  208:      $   RETURN
  209: *
  210:       IF( N.EQ.1 ) THEN
  211:          IF( ICOMPZ.EQ.2 )
  212:      $      Z( 1, 1 ) = CONE
  213:          RETURN
  214:       END IF
  215: *
  216: *     Determine the unit roundoff and over/underflow thresholds.
  217: *
  218:       EPS = DLAMCH( 'E' )
  219:       EPS2 = EPS**2
  220:       SAFMIN = DLAMCH( 'S' )
  221:       SAFMAX = ONE / SAFMIN
  222:       SSFMAX = SQRT( SAFMAX ) / THREE
  223:       SSFMIN = SQRT( SAFMIN ) / EPS2
  224: *
  225: *     Compute the eigenvalues and eigenvectors of the tridiagonal
  226: *     matrix.
  227: *
  228:       IF( ICOMPZ.EQ.2 )
  229:      $   CALL ZLASET( 'Full', N, N, CZERO, CONE, Z, LDZ )
  230: *
  231:       NMAXIT = N*MAXIT
  232:       JTOT = 0
  233: *
  234: *     Determine where the matrix splits and choose QL or QR iteration
  235: *     for each block, according to whether top or bottom diagonal
  236: *     element is smaller.
  237: *
  238:       L1 = 1
  239:       NM1 = N - 1
  240: *
  241:    10 CONTINUE
  242:       IF( L1.GT.N )
  243:      $   GO TO 160
  244:       IF( L1.GT.1 )
  245:      $   E( L1-1 ) = ZERO
  246:       IF( L1.LE.NM1 ) THEN
  247:          DO 20 M = L1, NM1
  248:             TST = ABS( E( M ) )
  249:             IF( TST.EQ.ZERO )
  250:      $         GO TO 30
  251:             IF( TST.LE.( SQRT( ABS( D( M ) ) )*SQRT( ABS( D( M+
  252:      $          1 ) ) ) )*EPS ) THEN
  253:                E( M ) = ZERO
  254:                GO TO 30
  255:             END IF
  256:    20    CONTINUE
  257:       END IF
  258:       M = N
  259: *
  260:    30 CONTINUE
  261:       L = L1
  262:       LSV = L
  263:       LEND = M
  264:       LENDSV = LEND
  265:       L1 = M + 1
  266:       IF( LEND.EQ.L )
  267:      $   GO TO 10
  268: *
  269: *     Scale submatrix in rows and columns L to LEND
  270: *
  271:       ANORM = DLANST( 'I', LEND-L+1, D( L ), E( L ) )
  272:       ISCALE = 0
  273:       IF( ANORM.EQ.ZERO )
  274:      $   GO TO 10
  275:       IF( ANORM.GT.SSFMAX ) THEN
  276:          ISCALE = 1
  277:          CALL DLASCL( 'G', 0, 0, ANORM, SSFMAX, LEND-L+1, 1, D( L ), N,
  278:      $                INFO )
  279:          CALL DLASCL( 'G', 0, 0, ANORM, SSFMAX, LEND-L, 1, E( L ), N,
  280:      $                INFO )
  281:       ELSE IF( ANORM.LT.SSFMIN ) THEN
  282:          ISCALE = 2
  283:          CALL DLASCL( 'G', 0, 0, ANORM, SSFMIN, LEND-L+1, 1, D( L ), N,
  284:      $                INFO )
  285:          CALL DLASCL( 'G', 0, 0, ANORM, SSFMIN, LEND-L, 1, E( L ), N,
  286:      $                INFO )
  287:       END IF
  288: *
  289: *     Choose between QL and QR iteration
  290: *
  291:       IF( ABS( D( LEND ) ).LT.ABS( D( L ) ) ) THEN
  292:          LEND = LSV
  293:          L = LENDSV
  294:       END IF
  295: *
  296:       IF( LEND.GT.L ) THEN
  297: *
  298: *        QL Iteration
  299: *
  300: *        Look for small subdiagonal element.
  301: *
  302:    40    CONTINUE
  303:          IF( L.NE.LEND ) THEN
  304:             LENDM1 = LEND - 1
  305:             DO 50 M = L, LENDM1
  306:                TST = ABS( E( M ) )**2
  307:                IF( TST.LE.( EPS2*ABS( D( M ) ) )*ABS( D( M+1 ) )+
  308:      $             SAFMIN )GO TO 60
  309:    50       CONTINUE
  310:          END IF
  311: *
  312:          M = LEND
  313: *
  314:    60    CONTINUE
  315:          IF( M.LT.LEND )
  316:      $      E( M ) = ZERO
  317:          P = D( L )
  318:          IF( M.EQ.L )
  319:      $      GO TO 80
  320: *
  321: *        If remaining matrix is 2-by-2, use DLAE2 or SLAEV2
  322: *        to compute its eigensystem.
  323: *
  324:          IF( M.EQ.L+1 ) THEN
  325:             IF( ICOMPZ.GT.0 ) THEN
  326:                CALL DLAEV2( D( L ), E( L ), D( L+1 ), RT1, RT2, C, S )
  327:                WORK( L ) = C
  328:                WORK( N-1+L ) = S
  329:                CALL ZLASR( 'R', 'V', 'B', N, 2, WORK( L ),
  330:      $                     WORK( N-1+L ), Z( 1, L ), LDZ )
  331:             ELSE
  332:                CALL DLAE2( D( L ), E( L ), D( L+1 ), RT1, RT2 )
  333:             END IF
  334:             D( L ) = RT1
  335:             D( L+1 ) = RT2
  336:             E( L ) = ZERO
  337:             L = L + 2
  338:             IF( L.LE.LEND )
  339:      $         GO TO 40
  340:             GO TO 140
  341:          END IF
  342: *
  343:          IF( JTOT.EQ.NMAXIT )
  344:      $      GO TO 140
  345:          JTOT = JTOT + 1
  346: *
  347: *        Form shift.
  348: *
  349:          G = ( D( L+1 )-P ) / ( TWO*E( L ) )
  350:          R = DLAPY2( G, ONE )
  351:          G = D( M ) - P + ( E( L ) / ( G+SIGN( R, G ) ) )
  352: *
  353:          S = ONE
  354:          C = ONE
  355:          P = ZERO
  356: *
  357: *        Inner loop
  358: *
  359:          MM1 = M - 1
  360:          DO 70 I = MM1, L, -1
  361:             F = S*E( I )
  362:             B = C*E( I )
  363:             CALL DLARTG( G, F, C, S, R )
  364:             IF( I.NE.M-1 )
  365:      $         E( I+1 ) = R
  366:             G = D( I+1 ) - P
  367:             R = ( D( I )-G )*S + TWO*C*B
  368:             P = S*R
  369:             D( I+1 ) = G + P
  370:             G = C*R - B
  371: *
  372: *           If eigenvectors are desired, then save rotations.
  373: *
  374:             IF( ICOMPZ.GT.0 ) THEN
  375:                WORK( I ) = C
  376:                WORK( N-1+I ) = -S
  377:             END IF
  378: *
  379:    70    CONTINUE
  380: *
  381: *        If eigenvectors are desired, then apply saved rotations.
  382: *
  383:          IF( ICOMPZ.GT.0 ) THEN
  384:             MM = M - L + 1
  385:             CALL ZLASR( 'R', 'V', 'B', N, MM, WORK( L ), WORK( N-1+L ),
  386:      $                  Z( 1, L ), LDZ )
  387:          END IF
  388: *
  389:          D( L ) = D( L ) - P
  390:          E( L ) = G
  391:          GO TO 40
  392: *
  393: *        Eigenvalue found.
  394: *
  395:    80    CONTINUE
  396:          D( L ) = P
  397: *
  398:          L = L + 1
  399:          IF( L.LE.LEND )
  400:      $      GO TO 40
  401:          GO TO 140
  402: *
  403:       ELSE
  404: *
  405: *        QR Iteration
  406: *
  407: *        Look for small superdiagonal element.
  408: *
  409:    90    CONTINUE
  410:          IF( L.NE.LEND ) THEN
  411:             LENDP1 = LEND + 1
  412:             DO 100 M = L, LENDP1, -1
  413:                TST = ABS( E( M-1 ) )**2
  414:                IF( TST.LE.( EPS2*ABS( D( M ) ) )*ABS( D( M-1 ) )+
  415:      $             SAFMIN )GO TO 110
  416:   100       CONTINUE
  417:          END IF
  418: *
  419:          M = LEND
  420: *
  421:   110    CONTINUE
  422:          IF( M.GT.LEND )
  423:      $      E( M-1 ) = ZERO
  424:          P = D( L )
  425:          IF( M.EQ.L )
  426:      $      GO TO 130
  427: *
  428: *        If remaining matrix is 2-by-2, use DLAE2 or SLAEV2
  429: *        to compute its eigensystem.
  430: *
  431:          IF( M.EQ.L-1 ) THEN
  432:             IF( ICOMPZ.GT.0 ) THEN
  433:                CALL DLAEV2( D( L-1 ), E( L-1 ), D( L ), RT1, RT2, C, S )
  434:                WORK( M ) = C
  435:                WORK( N-1+M ) = S
  436:                CALL ZLASR( 'R', 'V', 'F', N, 2, WORK( M ),
  437:      $                     WORK( N-1+M ), Z( 1, L-1 ), LDZ )
  438:             ELSE
  439:                CALL DLAE2( D( L-1 ), E( L-1 ), D( L ), RT1, RT2 )
  440:             END IF
  441:             D( L-1 ) = RT1
  442:             D( L ) = RT2
  443:             E( L-1 ) = ZERO
  444:             L = L - 2
  445:             IF( L.GE.LEND )
  446:      $         GO TO 90
  447:             GO TO 140
  448:          END IF
  449: *
  450:          IF( JTOT.EQ.NMAXIT )
  451:      $      GO TO 140
  452:          JTOT = JTOT + 1
  453: *
  454: *        Form shift.
  455: *
  456:          G = ( D( L-1 )-P ) / ( TWO*E( L-1 ) )
  457:          R = DLAPY2( G, ONE )
  458:          G = D( M ) - P + ( E( L-1 ) / ( G+SIGN( R, G ) ) )
  459: *
  460:          S = ONE
  461:          C = ONE
  462:          P = ZERO
  463: *
  464: *        Inner loop
  465: *
  466:          LM1 = L - 1
  467:          DO 120 I = M, LM1
  468:             F = S*E( I )
  469:             B = C*E( I )
  470:             CALL DLARTG( G, F, C, S, R )
  471:             IF( I.NE.M )
  472:      $         E( I-1 ) = R
  473:             G = D( I ) - P
  474:             R = ( D( I+1 )-G )*S + TWO*C*B
  475:             P = S*R
  476:             D( I ) = G + P
  477:             G = C*R - B
  478: *
  479: *           If eigenvectors are desired, then save rotations.
  480: *
  481:             IF( ICOMPZ.GT.0 ) THEN
  482:                WORK( I ) = C
  483:                WORK( N-1+I ) = S
  484:             END IF
  485: *
  486:   120    CONTINUE
  487: *
  488: *        If eigenvectors are desired, then apply saved rotations.
  489: *
  490:          IF( ICOMPZ.GT.0 ) THEN
  491:             MM = L - M + 1
  492:             CALL ZLASR( 'R', 'V', 'F', N, MM, WORK( M ), WORK( N-1+M ),
  493:      $                  Z( 1, M ), LDZ )
  494:          END IF
  495: *
  496:          D( L ) = D( L ) - P
  497:          E( LM1 ) = G
  498:          GO TO 90
  499: *
  500: *        Eigenvalue found.
  501: *
  502:   130    CONTINUE
  503:          D( L ) = P
  504: *
  505:          L = L - 1
  506:          IF( L.GE.LEND )
  507:      $      GO TO 90
  508:          GO TO 140
  509: *
  510:       END IF
  511: *
  512: *     Undo scaling if necessary
  513: *
  514:   140 CONTINUE
  515:       IF( ISCALE.EQ.1 ) THEN
  516:          CALL DLASCL( 'G', 0, 0, SSFMAX, ANORM, LENDSV-LSV+1, 1,
  517:      $                D( LSV ), N, INFO )
  518:          CALL DLASCL( 'G', 0, 0, SSFMAX, ANORM, LENDSV-LSV, 1, E( LSV ),
  519:      $                N, INFO )
  520:       ELSE IF( ISCALE.EQ.2 ) THEN
  521:          CALL DLASCL( 'G', 0, 0, SSFMIN, ANORM, LENDSV-LSV+1, 1,
  522:      $                D( LSV ), N, INFO )
  523:          CALL DLASCL( 'G', 0, 0, SSFMIN, ANORM, LENDSV-LSV, 1, E( LSV ),
  524:      $                N, INFO )
  525:       END IF
  526: *
  527: *     Check for no convergence to an eigenvalue after a total
  528: *     of N*MAXIT iterations.
  529: *
  530:       IF( JTOT.EQ.NMAXIT ) THEN
  531:          DO 150 I = 1, N - 1
  532:             IF( E( I ).NE.ZERO )
  533:      $         INFO = INFO + 1
  534:   150    CONTINUE
  535:          RETURN
  536:       END IF
  537:       GO TO 10
  538: *
  539: *     Order eigenvalues and eigenvectors.
  540: *
  541:   160 CONTINUE
  542:       IF( ICOMPZ.EQ.0 ) THEN
  543: *
  544: *        Use Quick Sort
  545: *
  546:          CALL DLASRT( 'I', N, D, INFO )
  547: *
  548:       ELSE
  549: *
  550: *        Use Selection Sort to minimize swaps of eigenvectors
  551: *
  552:          DO 180 II = 2, N
  553:             I = II - 1
  554:             K = I
  555:             P = D( I )
  556:             DO 170 J = II, N
  557:                IF( D( J ).LT.P ) THEN
  558:                   K = J
  559:                   P = D( J )
  560:                END IF
  561:   170       CONTINUE
  562:             IF( K.NE.I ) THEN
  563:                D( K ) = D( I )
  564:                D( I ) = P
  565:                CALL ZSWAP( N, Z( 1, I ), 1, Z( 1, K ), 1 )
  566:             END IF
  567:   180    CONTINUE
  568:       END IF
  569:       RETURN
  570: *
  571: *     End of ZSTEQR
  572: *
  573:       END

CVSweb interface <joel.bertrand@systella.fr>