File:  [local] / rpl / lapack / lapack / zheevd.f
Revision 1.11: download - view: text, annotated - select for diffs - revision graph
Fri Dec 14 14:22:47 2012 UTC (11 years, 5 months ago) by bertrand
Branches: MAIN
CVS tags: rpl-4_1_16, rpl-4_1_15, rpl-4_1_14, rpl-4_1_13, rpl-4_1_12, rpl-4_1_11, HEAD
Mise à jour de lapack.

    1: *> \brief <b> ZHEEVD computes the eigenvalues and, optionally, the left and/or right eigenvectors for HE matrices</b>
    2: *
    3: *  =========== DOCUMENTATION ===========
    4: *
    5: * Online html documentation available at 
    6: *            http://www.netlib.org/lapack/explore-html/ 
    7: *
    8: *> \htmlonly
    9: *> Download ZHEEVD + dependencies 
   10: *> <a href="http://www.netlib.org/cgi-bin/netlibfiles.tgz?format=tgz&filename=/lapack/lapack_routine/zheevd.f"> 
   11: *> [TGZ]</a> 
   12: *> <a href="http://www.netlib.org/cgi-bin/netlibfiles.zip?format=zip&filename=/lapack/lapack_routine/zheevd.f"> 
   13: *> [ZIP]</a> 
   14: *> <a href="http://www.netlib.org/cgi-bin/netlibfiles.txt?format=txt&filename=/lapack/lapack_routine/zheevd.f"> 
   15: *> [TXT]</a>
   16: *> \endhtmlonly 
   17: *
   18: *  Definition:
   19: *  ===========
   20: *
   21: *       SUBROUTINE ZHEEVD( JOBZ, UPLO, N, A, LDA, W, WORK, LWORK, RWORK,
   22: *                          LRWORK, IWORK, LIWORK, INFO )
   23:    24: *       .. Scalar Arguments ..
   25: *       CHARACTER          JOBZ, UPLO
   26: *       INTEGER            INFO, LDA, LIWORK, LRWORK, LWORK, N
   27: *       ..
   28: *       .. Array Arguments ..
   29: *       INTEGER            IWORK( * )
   30: *       DOUBLE PRECISION   RWORK( * ), W( * )
   31: *       COMPLEX*16         A( LDA, * ), WORK( * )
   32: *       ..
   33: *  
   34: *
   35: *> \par Purpose:
   36: *  =============
   37: *>
   38: *> \verbatim
   39: *>
   40: *> ZHEEVD computes all eigenvalues and, optionally, eigenvectors of a
   41: *> complex Hermitian matrix A.  If eigenvectors are desired, it uses a
   42: *> divide and conquer algorithm.
   43: *>
   44: *> The divide and conquer algorithm makes very mild assumptions about
   45: *> floating point arithmetic. It will work on machines with a guard
   46: *> digit in add/subtract, or on those binary machines without guard
   47: *> digits which subtract like the Cray X-MP, Cray Y-MP, Cray C-90, or
   48: *> Cray-2. It could conceivably fail on hexadecimal or decimal machines
   49: *> without guard digits, but we know of none.
   50: *> \endverbatim
   51: *
   52: *  Arguments:
   53: *  ==========
   54: *
   55: *> \param[in] JOBZ
   56: *> \verbatim
   57: *>          JOBZ is CHARACTER*1
   58: *>          = 'N':  Compute eigenvalues only;
   59: *>          = 'V':  Compute eigenvalues and eigenvectors.
   60: *> \endverbatim
   61: *>
   62: *> \param[in] UPLO
   63: *> \verbatim
   64: *>          UPLO is CHARACTER*1
   65: *>          = 'U':  Upper triangle of A is stored;
   66: *>          = 'L':  Lower triangle of A is stored.
   67: *> \endverbatim
   68: *>
   69: *> \param[in] N
   70: *> \verbatim
   71: *>          N is INTEGER
   72: *>          The order of the matrix A.  N >= 0.
   73: *> \endverbatim
   74: *>
   75: *> \param[in,out] A
   76: *> \verbatim
   77: *>          A is COMPLEX*16 array, dimension (LDA, N)
   78: *>          On entry, the Hermitian matrix A.  If UPLO = 'U', the
   79: *>          leading N-by-N upper triangular part of A contains the
   80: *>          upper triangular part of the matrix A.  If UPLO = 'L',
   81: *>          the leading N-by-N lower triangular part of A contains
   82: *>          the lower triangular part of the matrix A.
   83: *>          On exit, if JOBZ = 'V', then if INFO = 0, A contains the
   84: *>          orthonormal eigenvectors of the matrix A.
   85: *>          If JOBZ = 'N', then on exit the lower triangle (if UPLO='L')
   86: *>          or the upper triangle (if UPLO='U') of A, including the
   87: *>          diagonal, is destroyed.
   88: *> \endverbatim
   89: *>
   90: *> \param[in] LDA
   91: *> \verbatim
   92: *>          LDA is INTEGER
   93: *>          The leading dimension of the array A.  LDA >= max(1,N).
   94: *> \endverbatim
   95: *>
   96: *> \param[out] W
   97: *> \verbatim
   98: *>          W is DOUBLE PRECISION array, dimension (N)
   99: *>          If INFO = 0, the eigenvalues in ascending order.
  100: *> \endverbatim
  101: *>
  102: *> \param[out] WORK
  103: *> \verbatim
  104: *>          WORK is COMPLEX*16 array, dimension (MAX(1,LWORK))
  105: *>          On exit, if INFO = 0, WORK(1) returns the optimal LWORK.
  106: *> \endverbatim
  107: *>
  108: *> \param[in] LWORK
  109: *> \verbatim
  110: *>          LWORK is INTEGER
  111: *>          The length of the array WORK.
  112: *>          If N <= 1,                LWORK must be at least 1.
  113: *>          If JOBZ  = 'N' and N > 1, LWORK must be at least N + 1.
  114: *>          If JOBZ  = 'V' and N > 1, LWORK must be at least 2*N + N**2.
  115: *>
  116: *>          If LWORK = -1, then a workspace query is assumed; the routine
  117: *>          only calculates the optimal sizes of the WORK, RWORK and
  118: *>          IWORK arrays, returns these values as the first entries of
  119: *>          the WORK, RWORK and IWORK arrays, and no error message
  120: *>          related to LWORK or LRWORK or LIWORK is issued by XERBLA.
  121: *> \endverbatim
  122: *>
  123: *> \param[out] RWORK
  124: *> \verbatim
  125: *>          RWORK is DOUBLE PRECISION array,
  126: *>                                         dimension (LRWORK)
  127: *>          On exit, if INFO = 0, RWORK(1) returns the optimal LRWORK.
  128: *> \endverbatim
  129: *>
  130: *> \param[in] LRWORK
  131: *> \verbatim
  132: *>          LRWORK is INTEGER
  133: *>          The dimension of the array RWORK.
  134: *>          If N <= 1,                LRWORK must be at least 1.
  135: *>          If JOBZ  = 'N' and N > 1, LRWORK must be at least N.
  136: *>          If JOBZ  = 'V' and N > 1, LRWORK must be at least
  137: *>                         1 + 5*N + 2*N**2.
  138: *>
  139: *>          If LRWORK = -1, then a workspace query is assumed; the
  140: *>          routine only calculates the optimal sizes of the WORK, RWORK
  141: *>          and IWORK arrays, returns these values as the first entries
  142: *>          of the WORK, RWORK and IWORK arrays, and no error message
  143: *>          related to LWORK or LRWORK or LIWORK is issued by XERBLA.
  144: *> \endverbatim
  145: *>
  146: *> \param[out] IWORK
  147: *> \verbatim
  148: *>          IWORK is INTEGER array, dimension (MAX(1,LIWORK))
  149: *>          On exit, if INFO = 0, IWORK(1) returns the optimal LIWORK.
  150: *> \endverbatim
  151: *>
  152: *> \param[in] LIWORK
  153: *> \verbatim
  154: *>          LIWORK is INTEGER
  155: *>          The dimension of the array IWORK.
  156: *>          If N <= 1,                LIWORK must be at least 1.
  157: *>          If JOBZ  = 'N' and N > 1, LIWORK must be at least 1.
  158: *>          If JOBZ  = 'V' and N > 1, LIWORK must be at least 3 + 5*N.
  159: *>
  160: *>          If LIWORK = -1, then a workspace query is assumed; the
  161: *>          routine only calculates the optimal sizes of the WORK, RWORK
  162: *>          and IWORK arrays, returns these values as the first entries
  163: *>          of the WORK, RWORK and IWORK arrays, and no error message
  164: *>          related to LWORK or LRWORK or LIWORK is issued by XERBLA.
  165: *> \endverbatim
  166: *>
  167: *> \param[out] INFO
  168: *> \verbatim
  169: *>          INFO is INTEGER
  170: *>          = 0:  successful exit
  171: *>          < 0:  if INFO = -i, the i-th argument had an illegal value
  172: *>          > 0:  if INFO = i and JOBZ = 'N', then the algorithm failed
  173: *>                to converge; i off-diagonal elements of an intermediate
  174: *>                tridiagonal form did not converge to zero;
  175: *>                if INFO = i and JOBZ = 'V', then the algorithm failed
  176: *>                to compute an eigenvalue while working on the submatrix
  177: *>                lying in rows and columns INFO/(N+1) through
  178: *>                mod(INFO,N+1).
  179: *> \endverbatim
  180: *
  181: *  Authors:
  182: *  ========
  183: *
  184: *> \author Univ. of Tennessee 
  185: *> \author Univ. of California Berkeley 
  186: *> \author Univ. of Colorado Denver 
  187: *> \author NAG Ltd. 
  188: *
  189: *> \date November 2011
  190: *
  191: *> \ingroup complex16HEeigen
  192: *
  193: *> \par Further Details:
  194: *  =====================
  195: *>
  196: *>  Modified description of INFO. Sven, 16 Feb 05.
  197: *
  198: *> \par Contributors:
  199: *  ==================
  200: *>
  201: *> Jeff Rutter, Computer Science Division, University of California
  202: *> at Berkeley, USA
  203: *>
  204: *  =====================================================================
  205:       SUBROUTINE ZHEEVD( JOBZ, UPLO, N, A, LDA, W, WORK, LWORK, RWORK,
  206:      $                   LRWORK, IWORK, LIWORK, INFO )
  207: *
  208: *  -- LAPACK driver routine (version 3.4.0) --
  209: *  -- LAPACK is a software package provided by Univ. of Tennessee,    --
  210: *  -- Univ. of California Berkeley, Univ. of Colorado Denver and NAG Ltd..--
  211: *     November 2011
  212: *
  213: *     .. Scalar Arguments ..
  214:       CHARACTER          JOBZ, UPLO
  215:       INTEGER            INFO, LDA, LIWORK, LRWORK, LWORK, N
  216: *     ..
  217: *     .. Array Arguments ..
  218:       INTEGER            IWORK( * )
  219:       DOUBLE PRECISION   RWORK( * ), W( * )
  220:       COMPLEX*16         A( LDA, * ), WORK( * )
  221: *     ..
  222: *
  223: *  =====================================================================
  224: *
  225: *     .. Parameters ..
  226:       DOUBLE PRECISION   ZERO, ONE
  227:       PARAMETER          ( ZERO = 0.0D0, ONE = 1.0D0 )
  228:       COMPLEX*16         CONE
  229:       PARAMETER          ( CONE = ( 1.0D0, 0.0D0 ) )
  230: *     ..
  231: *     .. Local Scalars ..
  232:       LOGICAL            LOWER, LQUERY, WANTZ
  233:       INTEGER            IINFO, IMAX, INDE, INDRWK, INDTAU, INDWK2,
  234:      $                   INDWRK, ISCALE, LIOPT, LIWMIN, LLRWK, LLWORK,
  235:      $                   LLWRK2, LOPT, LROPT, LRWMIN, LWMIN
  236:       DOUBLE PRECISION   ANRM, BIGNUM, EPS, RMAX, RMIN, SAFMIN, SIGMA,
  237:      $                   SMLNUM
  238: *     ..
  239: *     .. External Functions ..
  240:       LOGICAL            LSAME
  241:       INTEGER            ILAENV
  242:       DOUBLE PRECISION   DLAMCH, ZLANHE
  243:       EXTERNAL           LSAME, ILAENV, DLAMCH, ZLANHE
  244: *     ..
  245: *     .. External Subroutines ..
  246:       EXTERNAL           DSCAL, DSTERF, XERBLA, ZHETRD, ZLACPY, ZLASCL,
  247:      $                   ZSTEDC, ZUNMTR
  248: *     ..
  249: *     .. Intrinsic Functions ..
  250:       INTRINSIC          MAX, SQRT
  251: *     ..
  252: *     .. Executable Statements ..
  253: *
  254: *     Test the input parameters.
  255: *
  256:       WANTZ = LSAME( JOBZ, 'V' )
  257:       LOWER = LSAME( UPLO, 'L' )
  258:       LQUERY = ( LWORK.EQ.-1 .OR. LRWORK.EQ.-1 .OR. LIWORK.EQ.-1 )
  259: *
  260:       INFO = 0
  261:       IF( .NOT.( WANTZ .OR. LSAME( JOBZ, 'N' ) ) ) THEN
  262:          INFO = -1
  263:       ELSE IF( .NOT.( LOWER .OR. LSAME( UPLO, 'U' ) ) ) THEN
  264:          INFO = -2
  265:       ELSE IF( N.LT.0 ) THEN
  266:          INFO = -3
  267:       ELSE IF( LDA.LT.MAX( 1, N ) ) THEN
  268:          INFO = -5
  269:       END IF
  270: *
  271:       IF( INFO.EQ.0 ) THEN
  272:          IF( N.LE.1 ) THEN
  273:             LWMIN = 1
  274:             LRWMIN = 1
  275:             LIWMIN = 1
  276:             LOPT = LWMIN
  277:             LROPT = LRWMIN
  278:             LIOPT = LIWMIN
  279:          ELSE
  280:             IF( WANTZ ) THEN
  281:                LWMIN = 2*N + N*N
  282:                LRWMIN = 1 + 5*N + 2*N**2
  283:                LIWMIN = 3 + 5*N
  284:             ELSE
  285:                LWMIN = N + 1
  286:                LRWMIN = N
  287:                LIWMIN = 1
  288:             END IF
  289:             LOPT = MAX( LWMIN, N +
  290:      $                  ILAENV( 1, 'ZHETRD', UPLO, N, -1, -1, -1 ) )
  291:             LROPT = LRWMIN
  292:             LIOPT = LIWMIN
  293:          END IF
  294:          WORK( 1 ) = LOPT
  295:          RWORK( 1 ) = LROPT
  296:          IWORK( 1 ) = LIOPT
  297: *
  298:          IF( LWORK.LT.LWMIN .AND. .NOT.LQUERY ) THEN
  299:             INFO = -8
  300:          ELSE IF( LRWORK.LT.LRWMIN .AND. .NOT.LQUERY ) THEN
  301:             INFO = -10
  302:          ELSE IF( LIWORK.LT.LIWMIN .AND. .NOT.LQUERY ) THEN
  303:             INFO = -12
  304:          END IF
  305:       END IF
  306: *
  307:       IF( INFO.NE.0 ) THEN
  308:          CALL XERBLA( 'ZHEEVD', -INFO )
  309:          RETURN
  310:       ELSE IF( LQUERY ) THEN
  311:          RETURN
  312:       END IF
  313: *
  314: *     Quick return if possible
  315: *
  316:       IF( N.EQ.0 )
  317:      $   RETURN
  318: *
  319:       IF( N.EQ.1 ) THEN
  320:          W( 1 ) = A( 1, 1 )
  321:          IF( WANTZ )
  322:      $      A( 1, 1 ) = CONE
  323:          RETURN
  324:       END IF
  325: *
  326: *     Get machine constants.
  327: *
  328:       SAFMIN = DLAMCH( 'Safe minimum' )
  329:       EPS = DLAMCH( 'Precision' )
  330:       SMLNUM = SAFMIN / EPS
  331:       BIGNUM = ONE / SMLNUM
  332:       RMIN = SQRT( SMLNUM )
  333:       RMAX = SQRT( BIGNUM )
  334: *
  335: *     Scale matrix to allowable range, if necessary.
  336: *
  337:       ANRM = ZLANHE( 'M', UPLO, N, A, LDA, RWORK )
  338:       ISCALE = 0
  339:       IF( ANRM.GT.ZERO .AND. ANRM.LT.RMIN ) THEN
  340:          ISCALE = 1
  341:          SIGMA = RMIN / ANRM
  342:       ELSE IF( ANRM.GT.RMAX ) THEN
  343:          ISCALE = 1
  344:          SIGMA = RMAX / ANRM
  345:       END IF
  346:       IF( ISCALE.EQ.1 )
  347:      $   CALL ZLASCL( UPLO, 0, 0, ONE, SIGMA, N, N, A, LDA, INFO )
  348: *
  349: *     Call ZHETRD to reduce Hermitian matrix to tridiagonal form.
  350: *
  351:       INDE = 1
  352:       INDTAU = 1
  353:       INDWRK = INDTAU + N
  354:       INDRWK = INDE + N
  355:       INDWK2 = INDWRK + N*N
  356:       LLWORK = LWORK - INDWRK + 1
  357:       LLWRK2 = LWORK - INDWK2 + 1
  358:       LLRWK = LRWORK - INDRWK + 1
  359:       CALL ZHETRD( UPLO, N, A, LDA, W, RWORK( INDE ), WORK( INDTAU ),
  360:      $             WORK( INDWRK ), LLWORK, IINFO )
  361: *
  362: *     For eigenvalues only, call DSTERF.  For eigenvectors, first call
  363: *     ZSTEDC to generate the eigenvector matrix, WORK(INDWRK), of the
  364: *     tridiagonal matrix, then call ZUNMTR to multiply it to the
  365: *     Householder transformations represented as Householder vectors in
  366: *     A.
  367: *
  368:       IF( .NOT.WANTZ ) THEN
  369:          CALL DSTERF( N, W, RWORK( INDE ), INFO )
  370:       ELSE
  371:          CALL ZSTEDC( 'I', N, W, RWORK( INDE ), WORK( INDWRK ), N,
  372:      $                WORK( INDWK2 ), LLWRK2, RWORK( INDRWK ), LLRWK,
  373:      $                IWORK, LIWORK, INFO )
  374:          CALL ZUNMTR( 'L', UPLO, 'N', N, N, A, LDA, WORK( INDTAU ),
  375:      $                WORK( INDWRK ), N, WORK( INDWK2 ), LLWRK2, IINFO )
  376:          CALL ZLACPY( 'A', N, N, WORK( INDWRK ), N, A, LDA )
  377:       END IF
  378: *
  379: *     If matrix was scaled, then rescale eigenvalues appropriately.
  380: *
  381:       IF( ISCALE.EQ.1 ) THEN
  382:          IF( INFO.EQ.0 ) THEN
  383:             IMAX = N
  384:          ELSE
  385:             IMAX = INFO - 1
  386:          END IF
  387:          CALL DSCAL( IMAX, ONE / SIGMA, W, 1 )
  388:       END IF
  389: *
  390:       WORK( 1 ) = LOPT
  391:       RWORK( 1 ) = LROPT
  392:       IWORK( 1 ) = LIOPT
  393: *
  394:       RETURN
  395: *
  396: *     End of ZHEEVD
  397: *
  398:       END

CVSweb interface <joel.bertrand@systella.fr>