File:  [local] / rpl / lapack / lapack / dgetsqrhrt.f
Revision 1.1: download - view: text, annotated - select for diffs - revision graph
Mon Aug 7 08:55:29 2023 UTC (8 months, 3 weeks ago) by bertrand
Branches: MAIN
CVS tags: rpl-4_1_35, rpl-4_1_34, HEAD
Ajout de fichiers de lapack 3.11

    1: *> \brief \b DGETSQRHRT
    2: *
    3: *  =========== DOCUMENTATION ===========
    4: *
    5: * Online html documentation available at
    6: *            http://www.netlib.org/lapack/explore-html/
    7: *
    8: *> \htmlonly
    9: *> Download DGETSQRHRT + dependencies
   10: *> <a href="http://www.netlib.org/cgi-bin/netlibfiles.tgz?format=tgz&filename=/lapack/lapack_routine/dgetsqrhrt.f">
   11: *> [TGZ]</a>
   12: *> <a href="http://www.netlib.org/cgi-bin/netlibfiles.zip?format=zip&filename=/lapack/lapack_routine/dgetsqrhrt.f">
   13: *> [ZIP]</a>
   14: *> <a href="http://www.netlib.org/cgi-bin/netlibfiles.txt?format=txt&filename=/lapack/lapack_routine/dgetsqrhrt.f">
   15: *> [TXT]</a>
   16: *> \endhtmlonly
   17: *
   18: *  Definition:
   19: *  ===========
   20: *
   21: *       SUBROUTINE DGETSQRHRT( M, N, MB1, NB1, NB2, A, LDA, T, LDT, WORK,
   22: *      $                       LWORK, INFO )
   23: *       IMPLICIT NONE
   24: *
   25: *       .. Scalar Arguments ..
   26: *       INTEGER           INFO, LDA, LDT, LWORK, M, N, NB1, NB2, MB1
   27: *       ..
   28: *       .. Array Arguments ..
   29: *       DOUBLE PRECISION  A( LDA, * ), T( LDT, * ), WORK( * )
   30: *       ..
   31: *
   32: *
   33: *> \par Purpose:
   34: *  =============
   35: *>
   36: *> \verbatim
   37: *>
   38: *> DGETSQRHRT computes a NB2-sized column blocked QR-factorization
   39: *> of a real M-by-N matrix A with M >= N,
   40: *>
   41: *>    A = Q * R.
   42: *>
   43: *> The routine uses internally a NB1-sized column blocked and MB1-sized
   44: *> row blocked TSQR-factorization and perfors the reconstruction
   45: *> of the Householder vectors from the TSQR output. The routine also
   46: *> converts the R_tsqr factor from the TSQR-factorization output into
   47: *> the R factor that corresponds to the Householder QR-factorization,
   48: *>
   49: *>    A = Q_tsqr * R_tsqr = Q * R.
   50: *>
   51: *> The output Q and R factors are stored in the same format as in DGEQRT
   52: *> (Q is in blocked compact WY-representation). See the documentation
   53: *> of DGEQRT for more details on the format.
   54: *> \endverbatim
   55: *
   56: *  Arguments:
   57: *  ==========
   58: *
   59: *> \param[in] M
   60: *> \verbatim
   61: *>          M is INTEGER
   62: *>          The number of rows of the matrix A.  M >= 0.
   63: *> \endverbatim
   64: *>
   65: *> \param[in] N
   66: *> \verbatim
   67: *>          N is INTEGER
   68: *>          The number of columns of the matrix A. M >= N >= 0.
   69: *> \endverbatim
   70: *>
   71: *> \param[in] MB1
   72: *> \verbatim
   73: *>          MB1 is INTEGER
   74: *>          The row block size to be used in the blocked TSQR.
   75: *>          MB1 > N.
   76: *> \endverbatim
   77: *>
   78: *> \param[in] NB1
   79: *> \verbatim
   80: *>          NB1 is INTEGER
   81: *>          The column block size to be used in the blocked TSQR.
   82: *>          N >= NB1 >= 1.
   83: *> \endverbatim
   84: *>
   85: *> \param[in] NB2
   86: *> \verbatim
   87: *>          NB2 is INTEGER
   88: *>          The block size to be used in the blocked QR that is
   89: *>          output. NB2 >= 1.
   90: *> \endverbatim
   91: *>
   92: *> \param[in,out] A
   93: *> \verbatim
   94: *>          A is DOUBLE PRECISION array, dimension (LDA,N)
   95: *>
   96: *>          On entry: an M-by-N matrix A.
   97: *>
   98: *>          On exit:
   99: *>           a) the elements on and above the diagonal
  100: *>              of the array contain the N-by-N upper-triangular
  101: *>              matrix R corresponding to the Householder QR;
  102: *>           b) the elements below the diagonal represent Q by
  103: *>              the columns of blocked V (compact WY-representation).
  104: *> \endverbatim
  105: *>
  106: *> \param[in] LDA
  107: *> \verbatim
  108: *>          LDA is INTEGER
  109: *>          The leading dimension of the array A.  LDA >= max(1,M).
  110: *> \endverbatim
  111: *>
  112: *> \param[out] T
  113: *> \verbatim
  114: *>          T is DOUBLE PRECISION array, dimension (LDT,N))
  115: *>          The upper triangular block reflectors stored in compact form
  116: *>          as a sequence of upper triangular blocks.
  117: *> \endverbatim
  118: *>
  119: *> \param[in] LDT
  120: *> \verbatim
  121: *>          LDT is INTEGER
  122: *>          The leading dimension of the array T.  LDT >= NB2.
  123: *> \endverbatim
  124: *>
  125: *> \param[out] WORK
  126: *> \verbatim
  127: *>          (workspace) DOUBLE PRECISION array, dimension (MAX(1,LWORK))
  128: *>          On exit, if INFO = 0, WORK(1) returns the optimal LWORK.
  129: *> \endverbatim
  130: *>
  131: *> \param[in] LWORK
  132: *> \verbatim
  133: *>          The dimension of the array WORK.
  134: *>          LWORK >= MAX( LWT + LW1, MAX( LWT+N*N+LW2, LWT+N*N+N ) ),
  135: *>          where
  136: *>             NUM_ALL_ROW_BLOCKS = CEIL((M-N)/(MB1-N)),
  137: *>             NB1LOCAL = MIN(NB1,N).
  138: *>             LWT = NUM_ALL_ROW_BLOCKS * N * NB1LOCAL,
  139: *>             LW1 = NB1LOCAL * N,
  140: *>             LW2 = NB1LOCAL * MAX( NB1LOCAL, ( N - NB1LOCAL ) ),
  141: *>          If LWORK = -1, then a workspace query is assumed.
  142: *>          The routine only calculates the optimal size of the WORK
  143: *>          array, returns this value as the first entry of the WORK
  144: *>          array, and no error message related to LWORK is issued
  145: *>          by XERBLA.
  146: *> \endverbatim
  147: *>
  148: *> \param[out] INFO
  149: *> \verbatim
  150: *>          INFO is INTEGER
  151: *>          = 0:  successful exit
  152: *>          < 0:  if INFO = -i, the i-th argument had an illegal value
  153: *> \endverbatim
  154: *
  155: *  Authors:
  156: *  ========
  157: *
  158: *> \author Univ. of Tennessee
  159: *> \author Univ. of California Berkeley
  160: *> \author Univ. of Colorado Denver
  161: *> \author NAG Ltd.
  162: *
  163: *> \ingroup doubleOTHERcomputational
  164: *
  165: *> \par Contributors:
  166: *  ==================
  167: *>
  168: *> \verbatim
  169: *>
  170: *> November 2020, Igor Kozachenko,
  171: *>                Computer Science Division,
  172: *>                University of California, Berkeley
  173: *>
  174: *> \endverbatim
  175: *>
  176: *  =====================================================================
  177:       SUBROUTINE DGETSQRHRT( M, N, MB1, NB1, NB2, A, LDA, T, LDT, WORK,
  178:      $                       LWORK, INFO )
  179:       IMPLICIT NONE
  180: *
  181: *  -- LAPACK computational routine --
  182: *  -- LAPACK is a software package provided by Univ. of Tennessee,    --
  183: *  -- Univ. of California Berkeley, Univ. of Colorado Denver and NAG Ltd..--
  184: *
  185: *     .. Scalar Arguments ..
  186:       INTEGER           INFO, LDA, LDT, LWORK, M, N, NB1, NB2, MB1
  187: *     ..
  188: *     .. Array Arguments ..
  189:       DOUBLE PRECISION  A( LDA, * ), T( LDT, * ), WORK( * )
  190: *     ..
  191: *
  192: *  =====================================================================
  193: *
  194: *     .. Parameters ..
  195:       DOUBLE PRECISION   ONE
  196:       PARAMETER          ( ONE = 1.0D+0 )
  197: *     ..
  198: *     .. Local Scalars ..
  199:       LOGICAL            LQUERY
  200:       INTEGER            I, IINFO, J, LW1, LW2, LWT, LDWT, LWORKOPT,
  201:      $                   NB1LOCAL, NB2LOCAL, NUM_ALL_ROW_BLOCKS
  202: *     ..
  203: *     .. External Subroutines ..
  204:       EXTERNAL           DCOPY, DLATSQR, DORGTSQR_ROW, DORHR_COL,
  205:      $                   XERBLA
  206: *     ..
  207: *     .. Intrinsic Functions ..
  208:       INTRINSIC          CEILING, DBLE, MAX, MIN
  209: *     ..
  210: *     .. Executable Statements ..
  211: *
  212: *     Test the input arguments
  213: *
  214:       INFO = 0
  215:       LQUERY  = LWORK.EQ.-1
  216:       IF( M.LT.0 ) THEN
  217:          INFO = -1
  218:       ELSE IF( N.LT.0 .OR. M.LT.N ) THEN
  219:          INFO = -2
  220:       ELSE IF( MB1.LE.N ) THEN
  221:          INFO = -3
  222:       ELSE IF( NB1.LT.1 ) THEN
  223:          INFO = -4
  224:       ELSE IF( NB2.LT.1 ) THEN
  225:          INFO = -5
  226:       ELSE IF( LDA.LT.MAX( 1, M ) ) THEN
  227:          INFO = -7
  228:       ELSE IF( LDT.LT.MAX( 1,  MIN( NB2, N ) ) ) THEN
  229:          INFO = -9
  230:       ELSE
  231: *
  232: *        Test the input LWORK for the dimension of the array WORK.
  233: *        This workspace is used to store array:
  234: *        a) Matrix T and WORK for DLATSQR;
  235: *        b) N-by-N upper-triangular factor R_tsqr;
  236: *        c) Matrix T and array WORK for DORGTSQR_ROW;
  237: *        d) Diagonal D for DORHR_COL.
  238: *
  239:          IF( LWORK.LT.N*N+1 .AND. .NOT.LQUERY ) THEN
  240:             INFO = -11
  241:          ELSE
  242: *
  243: *           Set block size for column blocks
  244: *
  245:             NB1LOCAL = MIN( NB1, N )
  246: *
  247:             NUM_ALL_ROW_BLOCKS = MAX( 1,
  248:      $                   CEILING( DBLE( M - N ) / DBLE( MB1 - N ) ) )
  249: *
  250: *           Length and leading dimension of WORK array to place
  251: *           T array in TSQR.
  252: *
  253:             LWT = NUM_ALL_ROW_BLOCKS * N * NB1LOCAL
  254: 
  255:             LDWT = NB1LOCAL
  256: *
  257: *           Length of TSQR work array
  258: *
  259:             LW1 = NB1LOCAL * N
  260: *
  261: *           Length of DORGTSQR_ROW work array.
  262: *
  263:             LW2 = NB1LOCAL * MAX( NB1LOCAL, ( N - NB1LOCAL ) )
  264: *
  265:             LWORKOPT = MAX( LWT + LW1, MAX( LWT+N*N+LW2, LWT+N*N+N ) )
  266: *
  267:             IF( ( LWORK.LT.MAX( 1, LWORKOPT ) ).AND.(.NOT.LQUERY) ) THEN
  268:                INFO = -11
  269:             END IF
  270: *
  271:          END IF
  272:       END IF
  273: *
  274: *     Handle error in the input parameters and return workspace query.
  275: *
  276:       IF( INFO.NE.0 ) THEN
  277:          CALL XERBLA( 'DGETSQRHRT', -INFO )
  278:          RETURN
  279:       ELSE IF ( LQUERY ) THEN
  280:          WORK( 1 ) = DBLE( LWORKOPT )
  281:          RETURN
  282:       END IF
  283: *
  284: *     Quick return if possible
  285: *
  286:       IF( MIN( M, N ).EQ.0 ) THEN
  287:          WORK( 1 ) = DBLE( LWORKOPT )
  288:          RETURN
  289:       END IF
  290: *
  291:       NB2LOCAL = MIN( NB2, N )
  292: *
  293: *
  294: *     (1) Perform TSQR-factorization of the M-by-N matrix A.
  295: *
  296:       CALL DLATSQR( M, N, MB1, NB1LOCAL, A, LDA, WORK, LDWT,
  297:      $              WORK(LWT+1), LW1, IINFO )
  298: *
  299: *     (2) Copy the factor R_tsqr stored in the upper-triangular part
  300: *         of A into the square matrix in the work array
  301: *         WORK(LWT+1:LWT+N*N) column-by-column.
  302: *
  303:       DO J = 1, N
  304:          CALL DCOPY( J, A( 1, J ), 1, WORK( LWT + N*(J-1)+1 ), 1 )
  305:       END DO
  306: *
  307: *     (3) Generate a M-by-N matrix Q with orthonormal columns from
  308: *     the result stored below the diagonal in the array A in place.
  309: *
  310: 
  311:       CALL DORGTSQR_ROW( M, N, MB1, NB1LOCAL, A, LDA, WORK, LDWT,
  312:      $                   WORK( LWT+N*N+1 ), LW2, IINFO )
  313: *
  314: *     (4) Perform the reconstruction of Householder vectors from
  315: *     the matrix Q (stored in A) in place.
  316: *
  317:       CALL DORHR_COL( M, N, NB2LOCAL, A, LDA, T, LDT,
  318:      $                WORK( LWT+N*N+1 ), IINFO )
  319: *
  320: *     (5) Copy the factor R_tsqr stored in the square matrix in the
  321: *     work array WORK(LWT+1:LWT+N*N) into the upper-triangular
  322: *     part of A.
  323: *
  324: *     (6) Compute from R_tsqr the factor R_hr corresponding to
  325: *     the reconstructed Householder vectors, i.e. R_hr = S * R_tsqr.
  326: *     This multiplication by the sign matrix S on the left means
  327: *     changing the sign of I-th row of the matrix R_tsqr according
  328: *     to sign of the I-th diagonal element DIAG(I) of the matrix S.
  329: *     DIAG is stored in WORK( LWT+N*N+1 ) from the DORHR_COL output.
  330: *
  331: *     (5) and (6) can be combined in a single loop, so the rows in A
  332: *     are accessed only once.
  333: *
  334:       DO I = 1, N
  335:          IF( WORK( LWT+N*N+I ).EQ.-ONE ) THEN
  336:             DO J = I, N
  337:                A( I, J ) = -ONE * WORK( LWT+N*(J-1)+I )
  338:             END DO
  339:          ELSE
  340:             CALL DCOPY( N-I+1, WORK(LWT+N*(I-1)+I), N, A( I, I ), LDA )
  341:          END IF
  342:       END DO
  343: *
  344:       WORK( 1 ) = DBLE( LWORKOPT )
  345:       RETURN
  346: *
  347: *     End of DGETSQRHRT
  348: *
  349:       END

CVSweb interface <joel.bertrand@systella.fr>