File:  [local] / rpl / lapack / lapack / dlarfb_gett.f
Revision 1.1: download - view: text, annotated - select for diffs - revision graph
Mon Aug 7 08:55:30 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 DLARFB_GETT
    2: *
    3: *  =========== DOCUMENTATION ===========
    4: *
    5: * Online html documentation available at
    6: *            http://www.netlib.org/lapack/explore-html/
    7: *
    8: *> \htmlonly
    9: *> Download DLARFB_GETT + dependencies
   10: *> <a href="http://www.netlib.org/cgi-bin/netlibfiles.tgz?format=tgz&filename=/lapack/lapack_routine/dlarfb_gett.f">
   11: *> [TGZ]</a>
   12: *> <a href="http://www.netlib.org/cgi-bin/netlibfiles.zip?format=zip&filename=/lapack/lapack_routine/dlarfb_gett.f">
   13: *> [ZIP]</a>
   14: *> <a href="http://www.netlib.org/cgi-bin/netlibfiles.txt?format=txt&filename=/lapack/lapack_routine/dlarfb_gett.f">
   15: *> [TXT]</a>
   16: *> \endhtmlonly
   17: *
   18: *  Definition:
   19: *  ===========
   20: *
   21: *       SUBROUTINE DLARFB_GETT( IDENT, M, N, K, T, LDT, A, LDA, B, LDB,
   22: *      $                        WORK, LDWORK )
   23: *       IMPLICIT NONE
   24: *
   25: *       .. Scalar Arguments ..
   26: *       CHARACTER          IDENT
   27: *       INTEGER            K, LDA, LDB, LDT, LDWORK, M, N
   28: *       ..
   29: *       .. Array Arguments ..
   30: *       DOUBLE PRECISION   A( LDA, * ), B( LDB, * ), T( LDT, * ),
   31: *      $                   WORK( LDWORK, * )
   32: *       ..
   33: *
   34: *> \par Purpose:
   35: *  =============
   36: *>
   37: *> \verbatim
   38: *>
   39: *> DLARFB_GETT applies a real Householder block reflector H from the
   40: *> left to a real (K+M)-by-N  "triangular-pentagonal" matrix
   41: *> composed of two block matrices: an upper trapezoidal K-by-N matrix A
   42: *> stored in the array A, and a rectangular M-by-(N-K) matrix B, stored
   43: *> in the array B. The block reflector H is stored in a compact
   44: *> WY-representation, where the elementary reflectors are in the
   45: *> arrays A, B and T. See Further Details section.
   46: *> \endverbatim
   47: *
   48: *  Arguments:
   49: *  ==========
   50: *
   51: *> \param[in] IDENT
   52: *> \verbatim
   53: *>          IDENT is CHARACTER*1
   54: *>          If IDENT = not 'I', or not 'i', then V1 is unit
   55: *>             lower-triangular and stored in the left K-by-K block of
   56: *>             the input matrix A,
   57: *>          If IDENT = 'I' or 'i', then  V1 is an identity matrix and
   58: *>             not stored.
   59: *>          See Further Details section.
   60: *> \endverbatim
   61: *>
   62: *> \param[in] M
   63: *> \verbatim
   64: *>          M is INTEGER
   65: *>          The number of rows of the matrix B.
   66: *>          M >= 0.
   67: *> \endverbatim
   68: *>
   69: *> \param[in] N
   70: *> \verbatim
   71: *>          N is INTEGER
   72: *>          The number of columns of the matrices A and B.
   73: *>          N >= 0.
   74: *> \endverbatim
   75: *>
   76: *> \param[in] K
   77: *> \verbatim
   78: *>          K is INTEGER
   79: *>          The number or rows of the matrix A.
   80: *>          K is also order of the matrix T, i.e. the number of
   81: *>          elementary reflectors whose product defines the block
   82: *>          reflector. 0 <= K <= N.
   83: *> \endverbatim
   84: *>
   85: *> \param[in] T
   86: *> \verbatim
   87: *>          T is DOUBLE PRECISION array, dimension (LDT,K)
   88: *>          The upper-triangular K-by-K matrix T in the representation
   89: *>          of the block reflector.
   90: *> \endverbatim
   91: *>
   92: *> \param[in] LDT
   93: *> \verbatim
   94: *>          LDT is INTEGER
   95: *>          The leading dimension of the array T. LDT >= K.
   96: *> \endverbatim
   97: *>
   98: *> \param[in,out] A
   99: *> \verbatim
  100: *>          A is DOUBLE PRECISION array, dimension (LDA,N)
  101: *>
  102: *>          On entry:
  103: *>           a) In the K-by-N upper-trapezoidal part A: input matrix A.
  104: *>           b) In the columns below the diagonal: columns of V1
  105: *>              (ones are not stored on the diagonal).
  106: *>
  107: *>          On exit:
  108: *>            A is overwritten by rectangular K-by-N product H*A.
  109: *>
  110: *>          See Further Details section.
  111: *> \endverbatim
  112: *>
  113: *> \param[in] LDA
  114: *> \verbatim
  115: *>          LDB is INTEGER
  116: *>          The leading dimension of the array A. LDA >= max(1,K).
  117: *> \endverbatim
  118: *>
  119: *> \param[in,out] B
  120: *> \verbatim
  121: *>          B is DOUBLE PRECISION array, dimension (LDB,N)
  122: *>
  123: *>          On entry:
  124: *>            a) In the M-by-(N-K) right block: input matrix B.
  125: *>            b) In the M-by-N left block: columns of V2.
  126: *>
  127: *>          On exit:
  128: *>            B is overwritten by rectangular M-by-N product H*B.
  129: *>
  130: *>          See Further Details section.
  131: *> \endverbatim
  132: *>
  133: *> \param[in] LDB
  134: *> \verbatim
  135: *>          LDB is INTEGER
  136: *>          The leading dimension of the array B. LDB >= max(1,M).
  137: *> \endverbatim
  138: *>
  139: *> \param[out] WORK
  140: *> \verbatim
  141: *>          WORK is DOUBLE PRECISION array,
  142: *>          dimension (LDWORK,max(K,N-K))
  143: *> \endverbatim
  144: *>
  145: *> \param[in] LDWORK
  146: *> \verbatim
  147: *>          LDWORK is INTEGER
  148: *>          The leading dimension of the array WORK. LDWORK>=max(1,K).
  149: *>
  150: *> \endverbatim
  151: *
  152: *  Authors:
  153: *  ========
  154: *
  155: *> \author Univ. of Tennessee
  156: *> \author Univ. of California Berkeley
  157: *> \author Univ. of Colorado Denver
  158: *> \author NAG Ltd.
  159: *
  160: *> \ingroup doubleOTHERauxiliary
  161: *
  162: *> \par Contributors:
  163: *  ==================
  164: *>
  165: *> \verbatim
  166: *>
  167: *> November 2020, Igor Kozachenko,
  168: *>                Computer Science Division,
  169: *>                University of California, Berkeley
  170: *>
  171: *> \endverbatim
  172: *
  173: *> \par Further Details:
  174: *  =====================
  175: *>
  176: *> \verbatim
  177: *>
  178: *>    (1) Description of the Algebraic Operation.
  179: *>
  180: *>    The matrix A is a K-by-N matrix composed of two column block
  181: *>    matrices, A1, which is K-by-K, and A2, which is K-by-(N-K):
  182: *>    A = ( A1, A2 ).
  183: *>    The matrix B is an M-by-N matrix composed of two column block
  184: *>    matrices, B1, which is M-by-K, and B2, which is M-by-(N-K):
  185: *>    B = ( B1, B2 ).
  186: *>
  187: *>    Perform the operation:
  188: *>
  189: *>       ( A_out ) := H * ( A_in ) = ( I - V * T * V**T ) * ( A_in ) =
  190: *>       ( B_out )        ( B_in )                          ( B_in )
  191: *>                  = ( I - ( V1 ) * T * ( V1**T, V2**T ) ) * ( A_in )
  192: *>                          ( V2 )                            ( B_in )
  193: *>     On input:
  194: *>
  195: *>    a) ( A_in )  consists of two block columns:
  196: *>       ( B_in )
  197: *>
  198: *>       ( A_in ) = (( A1_in ) ( A2_in )) = (( A1_in ) ( A2_in ))
  199: *>       ( B_in )   (( B1_in ) ( B2_in ))   ((     0 ) ( B2_in )),
  200: *>
  201: *>       where the column blocks are:
  202: *>
  203: *>       (  A1_in )  is a K-by-K upper-triangular matrix stored in the
  204: *>                   upper triangular part of the array A(1:K,1:K).
  205: *>       (  B1_in )  is an M-by-K rectangular ZERO matrix and not stored.
  206: *>
  207: *>       ( A2_in )  is a K-by-(N-K) rectangular matrix stored
  208: *>                  in the array A(1:K,K+1:N).
  209: *>       ( B2_in )  is an M-by-(N-K) rectangular matrix stored
  210: *>                  in the array B(1:M,K+1:N).
  211: *>
  212: *>    b) V = ( V1 )
  213: *>           ( V2 )
  214: *>
  215: *>       where:
  216: *>       1) if IDENT == 'I',V1 is a K-by-K identity matrix, not stored;
  217: *>       2) if IDENT != 'I',V1 is a K-by-K unit lower-triangular matrix,
  218: *>          stored in the lower-triangular part of the array
  219: *>          A(1:K,1:K) (ones are not stored),
  220: *>       and V2 is an M-by-K rectangular stored the array B(1:M,1:K),
  221: *>                 (because on input B1_in is a rectangular zero
  222: *>                  matrix that is not stored and the space is
  223: *>                  used to store V2).
  224: *>
  225: *>    c) T is a K-by-K upper-triangular matrix stored
  226: *>       in the array T(1:K,1:K).
  227: *>
  228: *>    On output:
  229: *>
  230: *>    a) ( A_out ) consists of two  block columns:
  231: *>       ( B_out )
  232: *>
  233: *>       ( A_out ) = (( A1_out ) ( A2_out ))
  234: *>       ( B_out )   (( B1_out ) ( B2_out )),
  235: *>
  236: *>       where the column blocks are:
  237: *>
  238: *>       ( A1_out )  is a K-by-K square matrix, or a K-by-K
  239: *>                   upper-triangular matrix, if V1 is an
  240: *>                   identity matrix. AiOut is stored in
  241: *>                   the array A(1:K,1:K).
  242: *>       ( B1_out )  is an M-by-K rectangular matrix stored
  243: *>                   in the array B(1:M,K:N).
  244: *>
  245: *>       ( A2_out )  is a K-by-(N-K) rectangular matrix stored
  246: *>                   in the array A(1:K,K+1:N).
  247: *>       ( B2_out )  is an M-by-(N-K) rectangular matrix stored
  248: *>                   in the array B(1:M,K+1:N).
  249: *>
  250: *>
  251: *>    The operation above can be represented as the same operation
  252: *>    on each block column:
  253: *>
  254: *>       ( A1_out ) := H * ( A1_in ) = ( I - V * T * V**T ) * ( A1_in )
  255: *>       ( B1_out )        (     0 )                          (     0 )
  256: *>
  257: *>       ( A2_out ) := H * ( A2_in ) = ( I - V * T * V**T ) * ( A2_in )
  258: *>       ( B2_out )        ( B2_in )                          ( B2_in )
  259: *>
  260: *>    If IDENT != 'I':
  261: *>
  262: *>       The computation for column block 1:
  263: *>
  264: *>       A1_out: = A1_in - V1*T*(V1**T)*A1_in
  265: *>
  266: *>       B1_out: = - V2*T*(V1**T)*A1_in
  267: *>
  268: *>       The computation for column block 2, which exists if N > K:
  269: *>
  270: *>       A2_out: = A2_in - V1*T*( (V1**T)*A2_in + (V2**T)*B2_in )
  271: *>
  272: *>       B2_out: = B2_in - V2*T*( (V1**T)*A2_in + (V2**T)*B2_in )
  273: *>
  274: *>    If IDENT == 'I':
  275: *>
  276: *>       The operation for column block 1:
  277: *>
  278: *>       A1_out: = A1_in - V1*T**A1_in
  279: *>
  280: *>       B1_out: = - V2*T**A1_in
  281: *>
  282: *>       The computation for column block 2, which exists if N > K:
  283: *>
  284: *>       A2_out: = A2_in - T*( A2_in + (V2**T)*B2_in )
  285: *>
  286: *>       B2_out: = B2_in - V2*T*( A2_in + (V2**T)*B2_in )
  287: *>
  288: *>    (2) Description of the Algorithmic Computation.
  289: *>
  290: *>    In the first step, we compute column block 2, i.e. A2 and B2.
  291: *>    Here, we need to use the K-by-(N-K) rectangular workspace
  292: *>    matrix W2 that is of the same size as the matrix A2.
  293: *>    W2 is stored in the array WORK(1:K,1:(N-K)).
  294: *>
  295: *>    In the second step, we compute column block 1, i.e. A1 and B1.
  296: *>    Here, we need to use the K-by-K square workspace matrix W1
  297: *>    that is of the same size as the as the matrix A1.
  298: *>    W1 is stored in the array WORK(1:K,1:K).
  299: *>
  300: *>    NOTE: Hence, in this routine, we need the workspace array WORK
  301: *>    only of size WORK(1:K,1:max(K,N-K)) so it can hold both W2 from
  302: *>    the first step and W1 from the second step.
  303: *>
  304: *>    Case (A), when V1 is unit lower-triangular, i.e. IDENT != 'I',
  305: *>    more computations than in the Case (B).
  306: *>
  307: *>    if( IDENT != 'I' ) then
  308: *>     if ( N > K ) then
  309: *>       (First Step - column block 2)
  310: *>       col2_(1) W2: = A2
  311: *>       col2_(2) W2: = (V1**T) * W2 = (unit_lower_tr_of_(A1)**T) * W2
  312: *>       col2_(3) W2: = W2 + (V2**T) * B2 = W2 + (B1**T) * B2
  313: *>       col2_(4) W2: = T * W2
  314: *>       col2_(5) B2: = B2 - V2 * W2 = B2 - B1 * W2
  315: *>       col2_(6) W2: = V1 * W2 = unit_lower_tr_of_(A1) * W2
  316: *>       col2_(7) A2: = A2 - W2
  317: *>     else
  318: *>       (Second Step - column block 1)
  319: *>       col1_(1) W1: = A1
  320: *>       col1_(2) W1: = (V1**T) * W1 = (unit_lower_tr_of_(A1)**T) * W1
  321: *>       col1_(3) W1: = T * W1
  322: *>       col1_(4) B1: = - V2 * W1 = - B1 * W1
  323: *>       col1_(5) square W1: = V1 * W1 = unit_lower_tr_of_(A1) * W1
  324: *>       col1_(6) square A1: = A1 - W1
  325: *>     end if
  326: *>    end if
  327: *>
  328: *>    Case (B), when V1 is an identity matrix, i.e. IDENT == 'I',
  329: *>    less computations than in the Case (A)
  330: *>
  331: *>    if( IDENT == 'I' ) then
  332: *>     if ( N > K ) then
  333: *>       (First Step - column block 2)
  334: *>       col2_(1) W2: = A2
  335: *>       col2_(3) W2: = W2 + (V2**T) * B2 = W2 + (B1**T) * B2
  336: *>       col2_(4) W2: = T * W2
  337: *>       col2_(5) B2: = B2 - V2 * W2 = B2 - B1 * W2
  338: *>       col2_(7) A2: = A2 - W2
  339: *>     else
  340: *>       (Second Step - column block 1)
  341: *>       col1_(1) W1: = A1
  342: *>       col1_(3) W1: = T * W1
  343: *>       col1_(4) B1: = - V2 * W1 = - B1 * W1
  344: *>       col1_(6) upper-triangular_of_(A1): = A1 - W1
  345: *>     end if
  346: *>    end if
  347: *>
  348: *>    Combine these cases (A) and (B) together, this is the resulting
  349: *>    algorithm:
  350: *>
  351: *>    if ( N > K ) then
  352: *>
  353: *>      (First Step - column block 2)
  354: *>
  355: *>      col2_(1)  W2: = A2
  356: *>      if( IDENT != 'I' ) then
  357: *>        col2_(2)  W2: = (V1**T) * W2
  358: *>                      = (unit_lower_tr_of_(A1)**T) * W2
  359: *>      end if
  360: *>      col2_(3)  W2: = W2 + (V2**T) * B2 = W2 + (B1**T) * B2]
  361: *>      col2_(4)  W2: = T * W2
  362: *>      col2_(5)  B2: = B2 - V2 * W2 = B2 - B1 * W2
  363: *>      if( IDENT != 'I' ) then
  364: *>        col2_(6)    W2: = V1 * W2 = unit_lower_tr_of_(A1) * W2
  365: *>      end if
  366: *>      col2_(7) A2: = A2 - W2
  367: *>
  368: *>    else
  369: *>
  370: *>    (Second Step - column block 1)
  371: *>
  372: *>      col1_(1) W1: = A1
  373: *>      if( IDENT != 'I' ) then
  374: *>        col1_(2) W1: = (V1**T) * W1
  375: *>                    = (unit_lower_tr_of_(A1)**T) * W1
  376: *>      end if
  377: *>      col1_(3) W1: = T * W1
  378: *>      col1_(4) B1: = - V2 * W1 = - B1 * W1
  379: *>      if( IDENT != 'I' ) then
  380: *>        col1_(5) square W1: = V1 * W1 = unit_lower_tr_of_(A1) * W1
  381: *>        col1_(6_a) below_diag_of_(A1): =  - below_diag_of_(W1)
  382: *>      end if
  383: *>      col1_(6_b) up_tr_of_(A1): = up_tr_of_(A1) - up_tr_of_(W1)
  384: *>
  385: *>    end if
  386: *>
  387: *> \endverbatim
  388: *>
  389: *  =====================================================================
  390:       SUBROUTINE DLARFB_GETT( IDENT, M, N, K, T, LDT, A, LDA, B, LDB,
  391:      $                        WORK, LDWORK )
  392:       IMPLICIT NONE
  393: *
  394: *  -- LAPACK auxiliary routine --
  395: *  -- LAPACK is a software package provided by Univ. of Tennessee,    --
  396: *  -- Univ. of California Berkeley, Univ. of Colorado Denver and NAG Ltd..--
  397: *
  398: *     .. Scalar Arguments ..
  399:       CHARACTER          IDENT
  400:       INTEGER            K, LDA, LDB, LDT, LDWORK, M, N
  401: *     ..
  402: *     .. Array Arguments ..
  403:       DOUBLE PRECISION   A( LDA, * ), B( LDB, * ), T( LDT, * ),
  404:      $                   WORK( LDWORK, * )
  405: *     ..
  406: *
  407: *  =====================================================================
  408: *
  409: *     .. Parameters ..
  410:       DOUBLE PRECISION   ONE, ZERO
  411:       PARAMETER          ( ONE = 1.0D+0, ZERO = 0.0D+0 )
  412: *     ..
  413: *     .. Local Scalars ..
  414:       LOGICAL            LNOTIDENT
  415:       INTEGER            I, J
  416: *     ..
  417: *     .. EXTERNAL FUNCTIONS ..
  418:       LOGICAL            LSAME
  419:       EXTERNAL           LSAME
  420: *     ..
  421: *     .. External Subroutines ..
  422:       EXTERNAL           DCOPY, DGEMM, DTRMM
  423: *     ..
  424: *     .. Executable Statements ..
  425: *
  426: *     Quick return if possible
  427: *
  428:       IF( M.LT.0 .OR. N.LE.0 .OR. K.EQ.0 .OR. K.GT.N )
  429:      $   RETURN
  430: *
  431:       LNOTIDENT = .NOT.LSAME( IDENT, 'I' )
  432: *
  433: *     ------------------------------------------------------------------
  434: *
  435: *     First Step. Computation of the Column Block 2:
  436: *
  437: *        ( A2 ) := H * ( A2 )
  438: *        ( B2 )        ( B2 )
  439: *
  440: *     ------------------------------------------------------------------
  441: *
  442:       IF( N.GT.K ) THEN
  443: *
  444: *        col2_(1) Compute W2: = A2. Therefore, copy A2 = A(1:K, K+1:N)
  445: *        into W2=WORK(1:K, 1:N-K) column-by-column.
  446: *
  447:          DO J = 1, N-K
  448:             CALL DCOPY( K, A( 1, K+J ), 1, WORK( 1, J ), 1 )
  449:          END DO
  450: 
  451:          IF( LNOTIDENT ) THEN
  452: *
  453: *           col2_(2) Compute W2: = (V1**T) * W2 = (A1**T) * W2,
  454: *           V1 is not an identy matrix, but unit lower-triangular
  455: *           V1 stored in A1 (diagonal ones are not stored).
  456: *
  457: *
  458:             CALL DTRMM( 'L', 'L', 'T', 'U', K, N-K, ONE, A, LDA,
  459:      $                  WORK, LDWORK )
  460:          END IF
  461: *
  462: *        col2_(3) Compute W2: = W2 + (V2**T) * B2 = W2 + (B1**T) * B2
  463: *        V2 stored in B1.
  464: *
  465:          IF( M.GT.0 ) THEN
  466:             CALL DGEMM( 'T', 'N', K, N-K, M, ONE, B, LDB,
  467:      $                  B( 1, K+1 ), LDB, ONE, WORK, LDWORK )
  468:          END IF
  469: *
  470: *        col2_(4) Compute W2: = T * W2,
  471: *        T is upper-triangular.
  472: *
  473:          CALL DTRMM( 'L', 'U', 'N', 'N', K, N-K, ONE, T, LDT,
  474:      $               WORK, LDWORK )
  475: *
  476: *        col2_(5) Compute B2: = B2 - V2 * W2 = B2 - B1 * W2,
  477: *        V2 stored in B1.
  478: *
  479:          IF( M.GT.0 ) THEN
  480:             CALL DGEMM( 'N', 'N', M, N-K, K, -ONE, B, LDB,
  481:      $                   WORK, LDWORK, ONE, B( 1, K+1 ), LDB )
  482:          END IF
  483: *
  484:          IF( LNOTIDENT ) THEN
  485: *
  486: *           col2_(6) Compute W2: = V1 * W2 = A1 * W2,
  487: *           V1 is not an identity matrix, but unit lower-triangular,
  488: *           V1 stored in A1 (diagonal ones are not stored).
  489: *
  490:             CALL DTRMM( 'L', 'L', 'N', 'U', K, N-K, ONE, A, LDA,
  491:      $                  WORK, LDWORK )
  492:          END IF
  493: *
  494: *        col2_(7) Compute A2: = A2 - W2 =
  495: *                             = A(1:K, K+1:N-K) - WORK(1:K, 1:N-K),
  496: *        column-by-column.
  497: *
  498:          DO J = 1, N-K
  499:             DO I = 1, K
  500:                A( I, K+J ) = A( I, K+J ) - WORK( I, J )
  501:             END DO
  502:          END DO
  503: *
  504:       END IF
  505: *
  506: *     ------------------------------------------------------------------
  507: *
  508: *     Second Step. Computation of the Column Block 1:
  509: *
  510: *        ( A1 ) := H * ( A1 )
  511: *        ( B1 )        (  0 )
  512: *
  513: *     ------------------------------------------------------------------
  514: *
  515: *     col1_(1) Compute W1: = A1. Copy the upper-triangular
  516: *     A1 = A(1:K, 1:K) into the upper-triangular
  517: *     W1 = WORK(1:K, 1:K) column-by-column.
  518: *
  519:       DO J = 1, K
  520:          CALL DCOPY( J, A( 1, J ), 1, WORK( 1, J ), 1 )
  521:       END DO
  522: *
  523: *     Set the subdiagonal elements of W1 to zero column-by-column.
  524: *
  525:       DO J = 1, K - 1
  526:          DO I = J + 1, K
  527:             WORK( I, J ) = ZERO
  528:          END DO
  529:       END DO
  530: *
  531:       IF( LNOTIDENT ) THEN
  532: *
  533: *        col1_(2) Compute W1: = (V1**T) * W1 = (A1**T) * W1,
  534: *        V1 is not an identity matrix, but unit lower-triangular
  535: *        V1 stored in A1 (diagonal ones are not stored),
  536: *        W1 is upper-triangular with zeroes below the diagonal.
  537: *
  538:          CALL DTRMM( 'L', 'L', 'T', 'U', K, K, ONE, A, LDA,
  539:      $               WORK, LDWORK )
  540:       END IF
  541: *
  542: *     col1_(3) Compute W1: = T * W1,
  543: *     T is upper-triangular,
  544: *     W1 is upper-triangular with zeroes below the diagonal.
  545: *
  546:       CALL DTRMM( 'L', 'U', 'N', 'N', K, K, ONE, T, LDT,
  547:      $            WORK, LDWORK )
  548: *
  549: *     col1_(4) Compute B1: = - V2 * W1 = - B1 * W1,
  550: *     V2 = B1, W1 is upper-triangular with zeroes below the diagonal.
  551: *
  552:       IF( M.GT.0 ) THEN
  553:          CALL DTRMM( 'R', 'U', 'N', 'N', M, K, -ONE, WORK, LDWORK,
  554:      $               B, LDB )
  555:       END IF
  556: *
  557:       IF( LNOTIDENT ) THEN
  558: *
  559: *        col1_(5) Compute W1: = V1 * W1 = A1 * W1,
  560: *        V1 is not an identity matrix, but unit lower-triangular
  561: *        V1 stored in A1 (diagonal ones are not stored),
  562: *        W1 is upper-triangular on input with zeroes below the diagonal,
  563: *        and square on output.
  564: *
  565:          CALL DTRMM( 'L', 'L', 'N', 'U', K, K, ONE, A, LDA,
  566:      $               WORK, LDWORK )
  567: *
  568: *        col1_(6) Compute A1: = A1 - W1 = A(1:K, 1:K) - WORK(1:K, 1:K)
  569: *        column-by-column. A1 is upper-triangular on input.
  570: *        If IDENT, A1 is square on output, and W1 is square,
  571: *        if NOT IDENT, A1 is upper-triangular on output,
  572: *        W1 is upper-triangular.
  573: *
  574: *        col1_(6)_a Compute elements of A1 below the diagonal.
  575: *
  576:          DO J = 1, K - 1
  577:             DO I = J + 1, K
  578:                A( I, J ) = - WORK( I, J )
  579:             END DO
  580:          END DO
  581: *
  582:       END IF
  583: *
  584: *     col1_(6)_b Compute elements of A1 on and above the diagonal.
  585: *
  586:       DO J = 1, K
  587:          DO I = 1, J
  588:             A( I, J ) = A( I, J ) - WORK( I, J )
  589:          END DO
  590:       END DO
  591: *
  592:       RETURN
  593: *
  594: *     End of DLARFB_GETT
  595: *
  596:       END

CVSweb interface <joel.bertrand@systella.fr>