File:  [local] / rpl / lapack / lapack / dggsvp.f
Revision 1.11: download - view: text, annotated - select for diffs - revision graph
Wed Aug 22 09:48:15 2012 UTC (11 years, 9 months ago) by bertrand
Branches: MAIN
CVS tags: rpl-4_1_9, rpl-4_1_10, HEAD
Cohérence

    1: *> \brief \b DGGSVP
    2: *
    3: *  =========== DOCUMENTATION ===========
    4: *
    5: * Online html documentation available at 
    6: *            http://www.netlib.org/lapack/explore-html/ 
    7: *
    8: *> \htmlonly
    9: *> Download DGGSVP + dependencies 
   10: *> <a href="http://www.netlib.org/cgi-bin/netlibfiles.tgz?format=tgz&filename=/lapack/lapack_routine/dggsvp.f"> 
   11: *> [TGZ]</a> 
   12: *> <a href="http://www.netlib.org/cgi-bin/netlibfiles.zip?format=zip&filename=/lapack/lapack_routine/dggsvp.f"> 
   13: *> [ZIP]</a> 
   14: *> <a href="http://www.netlib.org/cgi-bin/netlibfiles.txt?format=txt&filename=/lapack/lapack_routine/dggsvp.f"> 
   15: *> [TXT]</a>
   16: *> \endhtmlonly 
   17: *
   18: *  Definition:
   19: *  ===========
   20: *
   21: *       SUBROUTINE DGGSVP( JOBU, JOBV, JOBQ, M, P, N, A, LDA, B, LDB,
   22: *                          TOLA, TOLB, K, L, U, LDU, V, LDV, Q, LDQ,
   23: *                          IWORK, TAU, WORK, INFO )
   24:    25: *       .. Scalar Arguments ..
   26: *       CHARACTER          JOBQ, JOBU, JOBV
   27: *       INTEGER            INFO, K, L, LDA, LDB, LDQ, LDU, LDV, M, N, P
   28: *       DOUBLE PRECISION   TOLA, TOLB
   29: *       ..
   30: *       .. Array Arguments ..
   31: *       INTEGER            IWORK( * )
   32: *       DOUBLE PRECISION   A( LDA, * ), B( LDB, * ), Q( LDQ, * ),
   33: *      $                   TAU( * ), U( LDU, * ), V( LDV, * ), WORK( * )
   34: *       ..
   35: *  
   36: *
   37: *> \par Purpose:
   38: *  =============
   39: *>
   40: *> \verbatim
   41: *>
   42: *> DGGSVP computes orthogonal matrices U, V and Q such that
   43: *>
   44: *>                    N-K-L  K    L
   45: *>  U**T*A*Q =     K ( 0    A12  A13 )  if M-K-L >= 0;
   46: *>                 L ( 0     0   A23 )
   47: *>             M-K-L ( 0     0    0  )
   48: *>
   49: *>                  N-K-L  K    L
   50: *>         =     K ( 0    A12  A13 )  if M-K-L < 0;
   51: *>             M-K ( 0     0   A23 )
   52: *>
   53: *>                  N-K-L  K    L
   54: *>  V**T*B*Q =   L ( 0     0   B13 )
   55: *>             P-L ( 0     0    0  )
   56: *>
   57: *> where the K-by-K matrix A12 and L-by-L matrix B13 are nonsingular
   58: *> upper triangular; A23 is L-by-L upper triangular if M-K-L >= 0,
   59: *> otherwise A23 is (M-K)-by-L upper trapezoidal.  K+L = the effective
   60: *> numerical rank of the (M+P)-by-N matrix (A**T,B**T)**T. 
   61: *>
   62: *> This decomposition is the preprocessing step for computing the
   63: *> Generalized Singular Value Decomposition (GSVD), see subroutine
   64: *> DGGSVD.
   65: *> \endverbatim
   66: *
   67: *  Arguments:
   68: *  ==========
   69: *
   70: *> \param[in] JOBU
   71: *> \verbatim
   72: *>          JOBU is CHARACTER*1
   73: *>          = 'U':  Orthogonal matrix U is computed;
   74: *>          = 'N':  U is not computed.
   75: *> \endverbatim
   76: *>
   77: *> \param[in] JOBV
   78: *> \verbatim
   79: *>          JOBV is CHARACTER*1
   80: *>          = 'V':  Orthogonal matrix V is computed;
   81: *>          = 'N':  V is not computed.
   82: *> \endverbatim
   83: *>
   84: *> \param[in] JOBQ
   85: *> \verbatim
   86: *>          JOBQ is CHARACTER*1
   87: *>          = 'Q':  Orthogonal matrix Q is computed;
   88: *>          = 'N':  Q is not computed.
   89: *> \endverbatim
   90: *>
   91: *> \param[in] M
   92: *> \verbatim
   93: *>          M is INTEGER
   94: *>          The number of rows of the matrix A.  M >= 0.
   95: *> \endverbatim
   96: *>
   97: *> \param[in] P
   98: *> \verbatim
   99: *>          P is INTEGER
  100: *>          The number of rows of the matrix B.  P >= 0.
  101: *> \endverbatim
  102: *>
  103: *> \param[in] N
  104: *> \verbatim
  105: *>          N is INTEGER
  106: *>          The number of columns of the matrices A and B.  N >= 0.
  107: *> \endverbatim
  108: *>
  109: *> \param[in,out] A
  110: *> \verbatim
  111: *>          A is DOUBLE PRECISION array, dimension (LDA,N)
  112: *>          On entry, the M-by-N matrix A.
  113: *>          On exit, A contains the triangular (or trapezoidal) matrix
  114: *>          described in the Purpose section.
  115: *> \endverbatim
  116: *>
  117: *> \param[in] LDA
  118: *> \verbatim
  119: *>          LDA is INTEGER
  120: *>          The leading dimension of the array A. LDA >= max(1,M).
  121: *> \endverbatim
  122: *>
  123: *> \param[in,out] B
  124: *> \verbatim
  125: *>          B is DOUBLE PRECISION array, dimension (LDB,N)
  126: *>          On entry, the P-by-N matrix B.
  127: *>          On exit, B contains the triangular matrix described in
  128: *>          the Purpose section.
  129: *> \endverbatim
  130: *>
  131: *> \param[in] LDB
  132: *> \verbatim
  133: *>          LDB is INTEGER
  134: *>          The leading dimension of the array B. LDB >= max(1,P).
  135: *> \endverbatim
  136: *>
  137: *> \param[in] TOLA
  138: *> \verbatim
  139: *>          TOLA is DOUBLE PRECISION
  140: *> \endverbatim
  141: *>
  142: *> \param[in] TOLB
  143: *> \verbatim
  144: *>          TOLB is DOUBLE PRECISION
  145: *>
  146: *>          TOLA and TOLB are the thresholds to determine the effective
  147: *>          numerical rank of matrix B and a subblock of A. Generally,
  148: *>          they are set to
  149: *>             TOLA = MAX(M,N)*norm(A)*MACHEPS,
  150: *>             TOLB = MAX(P,N)*norm(B)*MACHEPS.
  151: *>          The size of TOLA and TOLB may affect the size of backward
  152: *>          errors of the decomposition.
  153: *> \endverbatim
  154: *>
  155: *> \param[out] K
  156: *> \verbatim
  157: *>          K is INTEGER
  158: *> \endverbatim
  159: *>
  160: *> \param[out] L
  161: *> \verbatim
  162: *>          L is INTEGER
  163: *>
  164: *>          On exit, K and L specify the dimension of the subblocks
  165: *>          described in Purpose section.
  166: *>          K + L = effective numerical rank of (A**T,B**T)**T.
  167: *> \endverbatim
  168: *>
  169: *> \param[out] U
  170: *> \verbatim
  171: *>          U is DOUBLE PRECISION array, dimension (LDU,M)
  172: *>          If JOBU = 'U', U contains the orthogonal matrix U.
  173: *>          If JOBU = 'N', U is not referenced.
  174: *> \endverbatim
  175: *>
  176: *> \param[in] LDU
  177: *> \verbatim
  178: *>          LDU is INTEGER
  179: *>          The leading dimension of the array U. LDU >= max(1,M) if
  180: *>          JOBU = 'U'; LDU >= 1 otherwise.
  181: *> \endverbatim
  182: *>
  183: *> \param[out] V
  184: *> \verbatim
  185: *>          V is DOUBLE PRECISION array, dimension (LDV,P)
  186: *>          If JOBV = 'V', V contains the orthogonal matrix V.
  187: *>          If JOBV = 'N', V is not referenced.
  188: *> \endverbatim
  189: *>
  190: *> \param[in] LDV
  191: *> \verbatim
  192: *>          LDV is INTEGER
  193: *>          The leading dimension of the array V. LDV >= max(1,P) if
  194: *>          JOBV = 'V'; LDV >= 1 otherwise.
  195: *> \endverbatim
  196: *>
  197: *> \param[out] Q
  198: *> \verbatim
  199: *>          Q is DOUBLE PRECISION array, dimension (LDQ,N)
  200: *>          If JOBQ = 'Q', Q contains the orthogonal matrix Q.
  201: *>          If JOBQ = 'N', Q is not referenced.
  202: *> \endverbatim
  203: *>
  204: *> \param[in] LDQ
  205: *> \verbatim
  206: *>          LDQ is INTEGER
  207: *>          The leading dimension of the array Q. LDQ >= max(1,N) if
  208: *>          JOBQ = 'Q'; LDQ >= 1 otherwise.
  209: *> \endverbatim
  210: *>
  211: *> \param[out] IWORK
  212: *> \verbatim
  213: *>          IWORK is INTEGER array, dimension (N)
  214: *> \endverbatim
  215: *>
  216: *> \param[out] TAU
  217: *> \verbatim
  218: *>          TAU is DOUBLE PRECISION array, dimension (N)
  219: *> \endverbatim
  220: *>
  221: *> \param[out] WORK
  222: *> \verbatim
  223: *>          WORK is DOUBLE PRECISION array, dimension (max(3*N,M,P))
  224: *> \endverbatim
  225: *>
  226: *> \param[out] INFO
  227: *> \verbatim
  228: *>          INFO is INTEGER
  229: *>          = 0:  successful exit
  230: *>          < 0:  if INFO = -i, the i-th argument had an illegal value.
  231: *> \endverbatim
  232: *
  233: *  Authors:
  234: *  ========
  235: *
  236: *> \author Univ. of Tennessee 
  237: *> \author Univ. of California Berkeley 
  238: *> \author Univ. of Colorado Denver 
  239: *> \author NAG Ltd. 
  240: *
  241: *> \date November 2011
  242: *
  243: *> \ingroup doubleOTHERcomputational
  244: *
  245: *> \par Further Details:
  246: *  =====================
  247: *>
  248: *>  The subroutine uses LAPACK subroutine DGEQPF for the QR factorization
  249: *>  with column pivoting to detect the effective numerical rank of the
  250: *>  a matrix. It may be replaced by a better rank determination strategy.
  251: *>
  252: *  =====================================================================
  253:       SUBROUTINE DGGSVP( JOBU, JOBV, JOBQ, M, P, N, A, LDA, B, LDB,
  254:      $                   TOLA, TOLB, K, L, U, LDU, V, LDV, Q, LDQ,
  255:      $                   IWORK, TAU, WORK, INFO )
  256: *
  257: *  -- LAPACK computational routine (version 3.4.0) --
  258: *  -- LAPACK is a software package provided by Univ. of Tennessee,    --
  259: *  -- Univ. of California Berkeley, Univ. of Colorado Denver and NAG Ltd..--
  260: *     November 2011
  261: *
  262: *     .. Scalar Arguments ..
  263:       CHARACTER          JOBQ, JOBU, JOBV
  264:       INTEGER            INFO, K, L, LDA, LDB, LDQ, LDU, LDV, M, N, P
  265:       DOUBLE PRECISION   TOLA, TOLB
  266: *     ..
  267: *     .. Array Arguments ..
  268:       INTEGER            IWORK( * )
  269:       DOUBLE PRECISION   A( LDA, * ), B( LDB, * ), Q( LDQ, * ),
  270:      $                   TAU( * ), U( LDU, * ), V( LDV, * ), WORK( * )
  271: *     ..
  272: *
  273: *  =====================================================================
  274: *
  275: *     .. Parameters ..
  276:       DOUBLE PRECISION   ZERO, ONE
  277:       PARAMETER          ( ZERO = 0.0D+0, ONE = 1.0D+0 )
  278: *     ..
  279: *     .. Local Scalars ..
  280:       LOGICAL            FORWRD, WANTQ, WANTU, WANTV
  281:       INTEGER            I, J
  282: *     ..
  283: *     .. External Functions ..
  284:       LOGICAL            LSAME
  285:       EXTERNAL           LSAME
  286: *     ..
  287: *     .. External Subroutines ..
  288:       EXTERNAL           DGEQPF, DGEQR2, DGERQ2, DLACPY, DLAPMT, DLASET,
  289:      $                   DORG2R, DORM2R, DORMR2, XERBLA
  290: *     ..
  291: *     .. Intrinsic Functions ..
  292:       INTRINSIC          ABS, MAX, MIN
  293: *     ..
  294: *     .. Executable Statements ..
  295: *
  296: *     Test the input parameters
  297: *
  298:       WANTU = LSAME( JOBU, 'U' )
  299:       WANTV = LSAME( JOBV, 'V' )
  300:       WANTQ = LSAME( JOBQ, 'Q' )
  301:       FORWRD = .TRUE.
  302: *
  303:       INFO = 0
  304:       IF( .NOT.( WANTU .OR. LSAME( JOBU, 'N' ) ) ) THEN
  305:          INFO = -1
  306:       ELSE IF( .NOT.( WANTV .OR. LSAME( JOBV, 'N' ) ) ) THEN
  307:          INFO = -2
  308:       ELSE IF( .NOT.( WANTQ .OR. LSAME( JOBQ, 'N' ) ) ) THEN
  309:          INFO = -3
  310:       ELSE IF( M.LT.0 ) THEN
  311:          INFO = -4
  312:       ELSE IF( P.LT.0 ) THEN
  313:          INFO = -5
  314:       ELSE IF( N.LT.0 ) THEN
  315:          INFO = -6
  316:       ELSE IF( LDA.LT.MAX( 1, M ) ) THEN
  317:          INFO = -8
  318:       ELSE IF( LDB.LT.MAX( 1, P ) ) THEN
  319:          INFO = -10
  320:       ELSE IF( LDU.LT.1 .OR. ( WANTU .AND. LDU.LT.M ) ) THEN
  321:          INFO = -16
  322:       ELSE IF( LDV.LT.1 .OR. ( WANTV .AND. LDV.LT.P ) ) THEN
  323:          INFO = -18
  324:       ELSE IF( LDQ.LT.1 .OR. ( WANTQ .AND. LDQ.LT.N ) ) THEN
  325:          INFO = -20
  326:       END IF
  327:       IF( INFO.NE.0 ) THEN
  328:          CALL XERBLA( 'DGGSVP', -INFO )
  329:          RETURN
  330:       END IF
  331: *
  332: *     QR with column pivoting of B: B*P = V*( S11 S12 )
  333: *                                           (  0   0  )
  334: *
  335:       DO 10 I = 1, N
  336:          IWORK( I ) = 0
  337:    10 CONTINUE
  338:       CALL DGEQPF( P, N, B, LDB, IWORK, TAU, WORK, INFO )
  339: *
  340: *     Update A := A*P
  341: *
  342:       CALL DLAPMT( FORWRD, M, N, A, LDA, IWORK )
  343: *
  344: *     Determine the effective rank of matrix B.
  345: *
  346:       L = 0
  347:       DO 20 I = 1, MIN( P, N )
  348:          IF( ABS( B( I, I ) ).GT.TOLB )
  349:      $      L = L + 1
  350:    20 CONTINUE
  351: *
  352:       IF( WANTV ) THEN
  353: *
  354: *        Copy the details of V, and form V.
  355: *
  356:          CALL DLASET( 'Full', P, P, ZERO, ZERO, V, LDV )
  357:          IF( P.GT.1 )
  358:      $      CALL DLACPY( 'Lower', P-1, N, B( 2, 1 ), LDB, V( 2, 1 ),
  359:      $                   LDV )
  360:          CALL DORG2R( P, P, MIN( P, N ), V, LDV, TAU, WORK, INFO )
  361:       END IF
  362: *
  363: *     Clean up B
  364: *
  365:       DO 40 J = 1, L - 1
  366:          DO 30 I = J + 1, L
  367:             B( I, J ) = ZERO
  368:    30    CONTINUE
  369:    40 CONTINUE
  370:       IF( P.GT.L )
  371:      $   CALL DLASET( 'Full', P-L, N, ZERO, ZERO, B( L+1, 1 ), LDB )
  372: *
  373:       IF( WANTQ ) THEN
  374: *
  375: *        Set Q = I and Update Q := Q*P
  376: *
  377:          CALL DLASET( 'Full', N, N, ZERO, ONE, Q, LDQ )
  378:          CALL DLAPMT( FORWRD, N, N, Q, LDQ, IWORK )
  379:       END IF
  380: *
  381:       IF( P.GE.L .AND. N.NE.L ) THEN
  382: *
  383: *        RQ factorization of (S11 S12): ( S11 S12 ) = ( 0 S12 )*Z
  384: *
  385:          CALL DGERQ2( L, N, B, LDB, TAU, WORK, INFO )
  386: *
  387: *        Update A := A*Z**T
  388: *
  389:          CALL DORMR2( 'Right', 'Transpose', M, N, L, B, LDB, TAU, A,
  390:      $                LDA, WORK, INFO )
  391: *
  392:          IF( WANTQ ) THEN
  393: *
  394: *           Update Q := Q*Z**T
  395: *
  396:             CALL DORMR2( 'Right', 'Transpose', N, N, L, B, LDB, TAU, Q,
  397:      $                   LDQ, WORK, INFO )
  398:          END IF
  399: *
  400: *        Clean up B
  401: *
  402:          CALL DLASET( 'Full', L, N-L, ZERO, ZERO, B, LDB )
  403:          DO 60 J = N - L + 1, N
  404:             DO 50 I = J - N + L + 1, L
  405:                B( I, J ) = ZERO
  406:    50       CONTINUE
  407:    60    CONTINUE
  408: *
  409:       END IF
  410: *
  411: *     Let              N-L     L
  412: *                A = ( A11    A12 ) M,
  413: *
  414: *     then the following does the complete QR decomposition of A11:
  415: *
  416: *              A11 = U*(  0  T12 )*P1**T
  417: *                      (  0   0  )
  418: *
  419:       DO 70 I = 1, N - L
  420:          IWORK( I ) = 0
  421:    70 CONTINUE
  422:       CALL DGEQPF( M, N-L, A, LDA, IWORK, TAU, WORK, INFO )
  423: *
  424: *     Determine the effective rank of A11
  425: *
  426:       K = 0
  427:       DO 80 I = 1, MIN( M, N-L )
  428:          IF( ABS( A( I, I ) ).GT.TOLA )
  429:      $      K = K + 1
  430:    80 CONTINUE
  431: *
  432: *     Update A12 := U**T*A12, where A12 = A( 1:M, N-L+1:N )
  433: *
  434:       CALL DORM2R( 'Left', 'Transpose', M, L, MIN( M, N-L ), A, LDA,
  435:      $             TAU, A( 1, N-L+1 ), LDA, WORK, INFO )
  436: *
  437:       IF( WANTU ) THEN
  438: *
  439: *        Copy the details of U, and form U
  440: *
  441:          CALL DLASET( 'Full', M, M, ZERO, ZERO, U, LDU )
  442:          IF( M.GT.1 )
  443:      $      CALL DLACPY( 'Lower', M-1, N-L, A( 2, 1 ), LDA, U( 2, 1 ),
  444:      $                   LDU )
  445:          CALL DORG2R( M, M, MIN( M, N-L ), U, LDU, TAU, WORK, INFO )
  446:       END IF
  447: *
  448:       IF( WANTQ ) THEN
  449: *
  450: *        Update Q( 1:N, 1:N-L )  = Q( 1:N, 1:N-L )*P1
  451: *
  452:          CALL DLAPMT( FORWRD, N, N-L, Q, LDQ, IWORK )
  453:       END IF
  454: *
  455: *     Clean up A: set the strictly lower triangular part of
  456: *     A(1:K, 1:K) = 0, and A( K+1:M, 1:N-L ) = 0.
  457: *
  458:       DO 100 J = 1, K - 1
  459:          DO 90 I = J + 1, K
  460:             A( I, J ) = ZERO
  461:    90    CONTINUE
  462:   100 CONTINUE
  463:       IF( M.GT.K )
  464:      $   CALL DLASET( 'Full', M-K, N-L, ZERO, ZERO, A( K+1, 1 ), LDA )
  465: *
  466:       IF( N-L.GT.K ) THEN
  467: *
  468: *        RQ factorization of ( T11 T12 ) = ( 0 T12 )*Z1
  469: *
  470:          CALL DGERQ2( K, N-L, A, LDA, TAU, WORK, INFO )
  471: *
  472:          IF( WANTQ ) THEN
  473: *
  474: *           Update Q( 1:N,1:N-L ) = Q( 1:N,1:N-L )*Z1**T
  475: *
  476:             CALL DORMR2( 'Right', 'Transpose', N, N-L, K, A, LDA, TAU,
  477:      $                   Q, LDQ, WORK, INFO )
  478:          END IF
  479: *
  480: *        Clean up A
  481: *
  482:          CALL DLASET( 'Full', K, N-L-K, ZERO, ZERO, A, LDA )
  483:          DO 120 J = N - L - K + 1, N - L
  484:             DO 110 I = J - N + L + K + 1, K
  485:                A( I, J ) = ZERO
  486:   110       CONTINUE
  487:   120    CONTINUE
  488: *
  489:       END IF
  490: *
  491:       IF( M.GT.K ) THEN
  492: *
  493: *        QR factorization of A( K+1:M,N-L+1:N )
  494: *
  495:          CALL DGEQR2( M-K, L, A( K+1, N-L+1 ), LDA, TAU, WORK, INFO )
  496: *
  497:          IF( WANTU ) THEN
  498: *
  499: *           Update U(:,K+1:M) := U(:,K+1:M)*U1
  500: *
  501:             CALL DORM2R( 'Right', 'No transpose', M, M-K, MIN( M-K, L ),
  502:      $                   A( K+1, N-L+1 ), LDA, TAU, U( 1, K+1 ), LDU,
  503:      $                   WORK, INFO )
  504:          END IF
  505: *
  506: *        Clean up
  507: *
  508:          DO 140 J = N - L + 1, N
  509:             DO 130 I = J - N + K + L + 1, M
  510:                A( I, J ) = ZERO
  511:   130       CONTINUE
  512:   140    CONTINUE
  513: *
  514:       END IF
  515: *
  516:       RETURN
  517: *
  518: *     End of DGGSVP
  519: *
  520:       END

CVSweb interface <joel.bertrand@systella.fr>