dsyevd_2stage.f

Go to the documentation of this file.

*> \brief <b> DSYEVD_2STAGE computes the eigenvalues and, optionally, the left and/or right eigenvectors for SY matrices</b>
*
*  @precisions fortran d -> s
*
*  =========== DOCUMENTATION ===========
*
* Online html documentation available at
*            http://www.netlib.org/lapack/explore-html/
*
*> \htmlonly
*> Download DSYEVD_2STAGE + dependencies
*> <a href="http://www.netlib.org/cgi-bin/netlibfiles.tgz?format=tgz&filename=/lapack/lapack_routine/dsyevd_2stage.f">
*> [TGZ]</a>
*> <a href="http://www.netlib.org/cgi-bin/netlibfiles.zip?format=zip&filename=/lapack/lapack_routine/dsyevd_2stage.f">
*> [ZIP]</a>
*> <a href="http://www.netlib.org/cgi-bin/netlibfiles.txt?format=txt&filename=/lapack/lapack_routine/dsyevd_2stage.f">
*> [TXT]</a>
*> \endhtmlonly
*
*  Definition:
*  ===========
*
*       SUBROUTINE DSYEVD_2STAGE( JOBZ, UPLO, N, A, LDA, W, WORK, LWORK,
*                                IWORK, LIWORK, INFO )
*
*       IMPLICIT NONE
*
*       .. Scalar Arguments ..
*       CHARACTER          JOBZ, UPLO
*       INTEGER            INFO, LDA, LIWORK, LWORK, N
*       ..
*       .. Array Arguments ..
*       INTEGER            IWORK( * )
*       DOUBLE PRECISION   A( LDA, * ), W( * ), WORK( * )
*       ..
*
*
*> \par Purpose:
*  =============
*>
*> \verbatim
*>
*> DSYEVD_2STAGE computes all eigenvalues and, optionally, eigenvectors of a
*> real symmetric matrix A using the 2stage technique for
*> the reduction to tridiagonal. If eigenvectors are desired, it uses a
*> divide and conquer algorithm.
*>
*> The divide and conquer algorithm makes very mild assumptions about
*> floating point arithmetic. It will work on machines with a guard
*> digit in add/subtract, or on those binary machines without guard
*> digits which subtract like the Cray X-MP, Cray Y-MP, Cray C-90, or
*> Cray-2. It could conceivably fail on hexadecimal or decimal machines
*> without guard digits, but we know of none.
*> \endverbatim
*
*  Arguments:
*  ==========
*
*> \param[in] JOBZ
*> \verbatim
*>          JOBZ is CHARACTER*1
*>          = 'N':  Compute eigenvalues only;
*>          = 'V':  Compute eigenvalues and eigenvectors.
*>                  Not available in this release.
*> \endverbatim
*>
*> \param[in] UPLO
*> \verbatim
*>          UPLO is CHARACTER*1
*>          = 'U':  Upper triangle of A is stored;
*>          = 'L':  Lower triangle of A is stored.
*> \endverbatim
*>
*> \param[in] N
*> \verbatim
*>          N is INTEGER
*>          The order of the matrix A.  N >= 0.
*> \endverbatim
*>
*> \param[in,out] A
*> \verbatim
*>          A is DOUBLE PRECISION array, dimension (LDA, N)
*>          On entry, the symmetric matrix A.  If UPLO = 'U', the
*>          leading N-by-N upper triangular part of A contains the
*>          upper triangular part of the matrix A.  If UPLO = 'L',
*>          the leading N-by-N lower triangular part of A contains
*>          the lower triangular part of the matrix A.
*>          On exit, if JOBZ = 'V', then if INFO = 0, A contains the
*>          orthonormal eigenvectors of the matrix A.
*>          If JOBZ = 'N', then on exit the lower triangle (if UPLO='L')
*>          or the upper triangle (if UPLO='U') of A, including the
*>          diagonal, is destroyed.
*> \endverbatim
*>
*> \param[in] LDA
*> \verbatim
*>          LDA is INTEGER
*>          The leading dimension of the array A.  LDA >= max(1,N).
*> \endverbatim
*>
*> \param[out] W
*> \verbatim
*>          W is DOUBLE PRECISION array, dimension (N)
*>          If INFO = 0, the eigenvalues in ascending order.
*> \endverbatim
*>
*> \param[out] WORK
*> \verbatim
*>          WORK is DOUBLE PRECISION array,
*>                                         dimension (LWORK)
*>          On exit, if INFO = 0, WORK(1) returns the optimal LWORK.
*> \endverbatim
*>
*> \param[in] LWORK
*> \verbatim
*>          LWORK is INTEGER
*>          The dimension of the array WORK.
*>          If N <= 1,               LWORK must be at least 1.
*>          If JOBZ = 'N' and N > 1, LWORK must be queried.
*>                                   LWORK = MAX(1, dimension) where
*>                                   dimension = max(stage1,stage2) + (KD+1)*N + 2*N+1
*>                                             = N*KD + N*max(KD+1,FACTOPTNB) 
*>                                               + max(2*KD*KD, KD*NTHREADS) 
*>                                               + (KD+1)*N + 2*N+1
*>                                   where KD is the blocking size of the reduction,
*>                                   FACTOPTNB is the blocking used by the QR or LQ
*>                                   algorithm, usually FACTOPTNB=128 is a good choice
*>                                   NTHREADS is the number of threads used when
*>                                   openMP compilation is enabled, otherwise =1.
*>          If JOBZ = 'V' and N > 1, LWORK must be at least
*>                                                1 + 6*N + 2*N**2.
*>
*>          If LWORK = -1, then a workspace query is assumed; the routine
*>          only calculates the optimal sizes of the WORK and IWORK
*>          arrays, returns these values as the first entries of the WORK
*>          and IWORK arrays, and no error message related to LWORK or
*>          LIWORK is issued by XERBLA.
*> \endverbatim
*>
*> \param[out] IWORK
*> \verbatim
*>          IWORK is INTEGER array, dimension (MAX(1,LIWORK))
*>          On exit, if INFO = 0, IWORK(1) returns the optimal LIWORK.
*> \endverbatim
*>
*> \param[in] LIWORK
*> \verbatim
*>          LIWORK is INTEGER
*>          The dimension of the array IWORK.
*>          If N <= 1,                LIWORK must be at least 1.
*>          If JOBZ  = 'N' and N > 1, LIWORK must be at least 1.
*>          If JOBZ  = 'V' and N > 1, LIWORK must be at least 3 + 5*N.
*>
*>          If LIWORK = -1, then a workspace query is assumed; the
*>          routine only calculates the optimal sizes of the WORK and
*>          IWORK arrays, returns these values as the first entries of
*>          the WORK and IWORK arrays, and no error message related to
*>          LWORK or LIWORK is issued by XERBLA.
*> \endverbatim
*>
*> \param[out] INFO
*> \verbatim
*>          INFO is INTEGER
*>          = 0:  successful exit
*>          < 0:  if INFO = -i, the i-th argument had an illegal value
*>          > 0:  if INFO = i and JOBZ = 'N', then the algorithm failed
*>                to converge; i off-diagonal elements of an intermediate
*>                tridiagonal form did not converge to zero;
*>                if INFO = i and JOBZ = 'V', then the algorithm failed
*>                to compute an eigenvalue while working on the submatrix
*>                lying in rows and columns INFO/(N+1) through
*>                mod(INFO,N+1).
*> \endverbatim
*
*  Authors:
*  ========
*
*> \author Univ. of Tennessee
*> \author Univ. of California Berkeley
*> \author Univ. of Colorado Denver
*> \author NAG Ltd.
*
*> \ingroup doubleSYeigen
*
*> \par Contributors:
*  ==================
*>
*> Jeff Rutter, Computer Science Division, University of California
*> at Berkeley, USA \n
*>  Modified by Francoise Tisseur, University of Tennessee \n
*>  Modified description of INFO. Sven, 16 Feb 05. \n
*> \par Further Details:
*  =====================
*>
*> \verbatim
*>
*>  All details about the 2stage techniques are available in:
*>
*>  Azzam Haidar, Hatem Ltaief, and Jack Dongarra.
*>  Parallel reduction to condensed forms for symmetric eigenvalue problems
*>  using aggregated fine-grained and memory-aware kernels. In Proceedings
*>  of 2011 International Conference for High Performance Computing,
*>  Networking, Storage and Analysis (SC '11), New York, NY, USA,
*>  Article 8 , 11 pages.
*>  http://doi.acm.org/10.1145/2063384.2063394
*>
*>  A. Haidar, J. Kurzak, P. Luszczek, 2013.
*>  An improved parallel singular value algorithm and its implementation 
*>  for multicore hardware, In Proceedings of 2013 International Conference
*>  for High Performance Computing, Networking, Storage and Analysis (SC '13).
*>  Denver, Colorado, USA, 2013.
*>  Article 90, 12 pages.
*>  http://doi.acm.org/10.1145/2503210.2503292
*>
*>  A. Haidar, R. Solca, S. Tomov, T. Schulthess and J. Dongarra.
*>  A novel hybrid CPU-GPU generalized eigensolver for electronic structure 
*>  calculations based on fine-grained memory aware tasks.
*>  International Journal of High Performance Computing Applications.
*>  Volume 28 Issue 2, Pages 196-209, May 2014.
*>  http://hpc.sagepub.com/content/28/2/196 
*>
*> \endverbatim
*
*  =====================================================================
      SUBROUTINE dsyevd_2stage( JOBZ, UPLO, N, A, LDA, W, WORK, LWORK,
     $                          IWORK, LIWORK, INFO )
*
      IMPLICIT NONE
*
*  -- LAPACK driver routine --
*  -- LAPACK is a software package provided by Univ. of Tennessee,    --
*  -- Univ. of California Berkeley, Univ. of Colorado Denver and NAG Ltd..--
*
*     .. Scalar Arguments ..
      CHARACTER          JOBZ, UPLO
      INTEGER            INFO, LDA, LIWORK, LWORK, N
*     ..
*     .. Array Arguments ..
      INTEGER            IWORK( * )
      DOUBLE PRECISION   A( LDA, * ), W( * ), WORK( * )
*     ..
*
*  =====================================================================
*
*     .. Parameters ..
      DOUBLE PRECISION   ZERO, ONE
      parameter( zero = 0.0d+0, one = 1.0d+0 )
*     ..
*     .. Local Scalars ..
*
      LOGICAL            LOWER, LQUERY, WANTZ
      INTEGER            IINFO, INDE, INDTAU, INDWK2, INDWRK, ISCALE,
     $                   liwmin, llwork, llwrk2, lwmin,
     $                   lhtrd, lwtrd, kd, ib, indhous
      DOUBLE PRECISION   ANRM, BIGNUM, EPS, RMAX, RMIN, SAFMIN, SIGMA,
     $                   smlnum
*     ..
*     .. External Functions ..
      LOGICAL            LSAME
      INTEGER            ILAENV2STAGE
      DOUBLE PRECISION   DLAMCH, DLANSY
      EXTERNAL           lsame, dlamch, dlansy, ilaenv2stage
*     ..
*     .. External Subroutines ..
      EXTERNAL           dlacpy, dlascl, dormtr, dscal, dstedc, dsterf,
     $                   dsytrd_2stage, xerbla
*     ..
*     .. Intrinsic Functions ..
      INTRINSIC          max, sqrt
*     ..
*     .. Executable Statements ..
*
*     Test the input parameters.
*
      wantz = lsame( jobz, 'V' )
      lower = lsame( uplo, 'L' )
      lquery = ( lwork.EQ.-1 .OR. liwork.EQ.-1 )
*
      info = 0
      IF( .NOT.( lsame( jobz, 'N' ) ) ) THEN
         info = -1
      ELSE IF( .NOT.( lower .OR. lsame( uplo, 'U' ) ) ) THEN
         info = -2
      ELSE IF( n.LT.0 ) THEN
         info = -3
      ELSE IF( lda.LT.max( 1, n ) ) THEN
         info = -5
      END IF
*
      IF( info.EQ.0 ) THEN
         IF( n.LE.1 ) THEN
            liwmin = 1
            lwmin = 1
         ELSE
            kd    = ilaenv2stage( 1, 'DSYTRD_2STAGE', jobz,
     $                            n, -1, -1, -1 )
            ib    = ilaenv2stage( 2, 'DSYTRD_2STAGE', jobz,
     $                            n, kd, -1, -1 )
            lhtrd = ilaenv2stage( 3, 'DSYTRD_2STAGE', jobz,
     $                            n, kd, ib, -1 )
            lwtrd = ilaenv2stage( 4, 'DSYTRD_2STAGE', jobz,
     $                            n, kd, ib, -1 )
            IF( wantz ) THEN
               liwmin = 3 + 5*n
               lwmin = 1 + 6*n + 2*n**2
            ELSE
               liwmin = 1
               lwmin = 2*n + 1 + lhtrd + lwtrd
            END IF
         END IF
         work( 1 )  = lwmin
         iwork( 1 ) = liwmin
*
         IF( lwork.LT.lwmin .AND. .NOT.lquery ) THEN
            info = -8
         ELSE IF( liwork.LT.liwmin .AND. .NOT.lquery ) THEN
            info = -10
         END IF
      END IF
*
      IF( info.NE.0 ) THEN
         CALL xerbla( 'DSYEVD_2STAGE', -info )
         RETURN
      ELSE IF( lquery ) THEN
         RETURN
      END IF
*
*     Quick return if possible
*
      IF( n.EQ.0 )
     $   RETURN
*
      IF( n.EQ.1 ) THEN
         w( 1 ) = a( 1, 1 )
         IF( wantz )
     $      a( 1, 1 ) = one
         RETURN
      END IF
*
*     Get machine constants.
*
      safmin = dlamch( 'Safe minimum' )
      eps    = dlamch( 'Precision' )
      smlnum = safmin / eps
      bignum = one / smlnum
      rmin   = sqrt( smlnum )
      rmax   = sqrt( bignum )
*
*     Scale matrix to allowable range, if necessary.
*
      anrm = dlansy( 'M', uplo, n, a, lda, work )
      iscale = 0
      IF( anrm.GT.zero .AND. anrm.LT.rmin ) THEN
         iscale = 1
         sigma = rmin / anrm
      ELSE IF( anrm.GT.rmax ) THEN
         iscale = 1
         sigma = rmax / anrm
      END IF
      IF( iscale.EQ.1 )
     $   CALL dlascl( uplo, 0, 0, one, sigma, n, n, a, lda, info )
*
*     Call DSYTRD_2STAGE to reduce symmetric matrix to tridiagonal form.
*
      inde    = 1
      indtau  = inde + n
      indhous = indtau + n
      indwrk  = indhous + lhtrd
      llwork  = lwork - indwrk + 1
      indwk2  = indwrk + n*n
      llwrk2  = lwork - indwk2 + 1
*
      CALL dsytrd_2stage( jobz, uplo, n, a, lda, w, work( inde ),
     $                    work( indtau ), work( indhous ), lhtrd, 
     $                    work( indwrk ), llwork, iinfo )
*
*     For eigenvalues only, call DSTERF.  For eigenvectors, first call
*     DSTEDC to generate the eigenvector matrix, WORK(INDWRK), of the
*     tridiagonal matrix, then call DORMTR to multiply it by the
*     Householder transformations stored in A.
*
      IF( .NOT.wantz ) THEN
         CALL dsterf( n, w, work( inde ), info )
      ELSE
*        Not available in this release, and argument checking should not
*        let it getting here
         RETURN
         CALL dstedc( 'I', n, w, work( inde ), work( indwrk ), n,
     $                work( indwk2 ), llwrk2, iwork, liwork, info )
         CALL dormtr( 'L', uplo, 'N', n, n, a, lda, work( indtau ),
     $                work( indwrk ), n, work( indwk2 ), llwrk2, iinfo )
         CALL dlacpy( 'A', n, n, work( indwrk ), n, a, lda )
      END IF
*
*     If matrix was scaled, then rescale eigenvalues appropriately.
*
      IF( iscale.EQ.1 )
     $   CALL dscal( n, one / sigma, w, 1 )
*
      work( 1 )  = lwmin
      iwork( 1 ) = liwmin
*
      RETURN
*
*     End of DSYEVD_2STAGE
*
      END