Annotation of rpl/lapack/lapack/dlasda.f, revision 1.10

1.10    ! bertrand    1: *> \brief \b DLASDA
        !             2: *
        !             3: *  =========== DOCUMENTATION ===========
        !             4: *
        !             5: * Online html documentation available at 
        !             6: *            http://www.netlib.org/lapack/explore-html/ 
        !             7: *
        !             8: *> \htmlonly
        !             9: *> Download DLASDA + dependencies 
        !            10: *> <a href="http://www.netlib.org/cgi-bin/netlibfiles.tgz?format=tgz&filename=/lapack/lapack_routine/dlasda.f"> 
        !            11: *> [TGZ]</a> 
        !            12: *> <a href="http://www.netlib.org/cgi-bin/netlibfiles.zip?format=zip&filename=/lapack/lapack_routine/dlasda.f"> 
        !            13: *> [ZIP]</a> 
        !            14: *> <a href="http://www.netlib.org/cgi-bin/netlibfiles.txt?format=txt&filename=/lapack/lapack_routine/dlasda.f"> 
        !            15: *> [TXT]</a>
        !            16: *> \endhtmlonly 
        !            17: *
        !            18: *  Definition:
        !            19: *  ===========
        !            20: *
        !            21: *       SUBROUTINE DLASDA( ICOMPQ, SMLSIZ, N, SQRE, D, E, U, LDU, VT, K,
        !            22: *                          DIFL, DIFR, Z, POLES, GIVPTR, GIVCOL, LDGCOL,
        !            23: *                          PERM, GIVNUM, C, S, WORK, IWORK, INFO )
        !            24: * 
        !            25: *       .. Scalar Arguments ..
        !            26: *       INTEGER            ICOMPQ, INFO, LDGCOL, LDU, N, SMLSIZ, SQRE
        !            27: *       ..
        !            28: *       .. Array Arguments ..
        !            29: *       INTEGER            GIVCOL( LDGCOL, * ), GIVPTR( * ), IWORK( * ),
        !            30: *      $                   K( * ), PERM( LDGCOL, * )
        !            31: *       DOUBLE PRECISION   C( * ), D( * ), DIFL( LDU, * ), DIFR( LDU, * ),
        !            32: *      $                   E( * ), GIVNUM( LDU, * ), POLES( LDU, * ),
        !            33: *      $                   S( * ), U( LDU, * ), VT( LDU, * ), WORK( * ),
        !            34: *      $                   Z( LDU, * )
        !            35: *       ..
        !            36: *  
        !            37: *
        !            38: *> \par Purpose:
        !            39: *  =============
        !            40: *>
        !            41: *> \verbatim
        !            42: *>
        !            43: *> Using a divide and conquer approach, DLASDA computes the singular
        !            44: *> value decomposition (SVD) of a real upper bidiagonal N-by-M matrix
        !            45: *> B with diagonal D and offdiagonal E, where M = N + SQRE. The
        !            46: *> algorithm computes the singular values in the SVD B = U * S * VT.
        !            47: *> The orthogonal matrices U and VT are optionally computed in
        !            48: *> compact form.
        !            49: *>
        !            50: *> A related subroutine, DLASD0, computes the singular values and
        !            51: *> the singular vectors in explicit form.
        !            52: *> \endverbatim
        !            53: *
        !            54: *  Arguments:
        !            55: *  ==========
        !            56: *
        !            57: *> \param[in] ICOMPQ
        !            58: *> \verbatim
        !            59: *>          ICOMPQ is INTEGER
        !            60: *>         Specifies whether singular vectors are to be computed
        !            61: *>         in compact form, as follows
        !            62: *>         = 0: Compute singular values only.
        !            63: *>         = 1: Compute singular vectors of upper bidiagonal
        !            64: *>              matrix in compact form.
        !            65: *> \endverbatim
        !            66: *>
        !            67: *> \param[in] SMLSIZ
        !            68: *> \verbatim
        !            69: *>          SMLSIZ is INTEGER
        !            70: *>         The maximum size of the subproblems at the bottom of the
        !            71: *>         computation tree.
        !            72: *> \endverbatim
        !            73: *>
        !            74: *> \param[in] N
        !            75: *> \verbatim
        !            76: *>          N is INTEGER
        !            77: *>         The row dimension of the upper bidiagonal matrix. This is
        !            78: *>         also the dimension of the main diagonal array D.
        !            79: *> \endverbatim
        !            80: *>
        !            81: *> \param[in] SQRE
        !            82: *> \verbatim
        !            83: *>          SQRE is INTEGER
        !            84: *>         Specifies the column dimension of the bidiagonal matrix.
        !            85: *>         = 0: The bidiagonal matrix has column dimension M = N;
        !            86: *>         = 1: The bidiagonal matrix has column dimension M = N + 1.
        !            87: *> \endverbatim
        !            88: *>
        !            89: *> \param[in,out] D
        !            90: *> \verbatim
        !            91: *>          D is DOUBLE PRECISION array, dimension ( N )
        !            92: *>         On entry D contains the main diagonal of the bidiagonal
        !            93: *>         matrix. On exit D, if INFO = 0, contains its singular values.
        !            94: *> \endverbatim
        !            95: *>
        !            96: *> \param[in] E
        !            97: *> \verbatim
        !            98: *>          E is DOUBLE PRECISION array, dimension ( M-1 )
        !            99: *>         Contains the subdiagonal entries of the bidiagonal matrix.
        !           100: *>         On exit, E has been destroyed.
        !           101: *> \endverbatim
        !           102: *>
        !           103: *> \param[out] U
        !           104: *> \verbatim
        !           105: *>          U is DOUBLE PRECISION array,
        !           106: *>         dimension ( LDU, SMLSIZ ) if ICOMPQ = 1, and not referenced
        !           107: *>         if ICOMPQ = 0. If ICOMPQ = 1, on exit, U contains the left
        !           108: *>         singular vector matrices of all subproblems at the bottom
        !           109: *>         level.
        !           110: *> \endverbatim
        !           111: *>
        !           112: *> \param[in] LDU
        !           113: *> \verbatim
        !           114: *>          LDU is INTEGER, LDU = > N.
        !           115: *>         The leading dimension of arrays U, VT, DIFL, DIFR, POLES,
        !           116: *>         GIVNUM, and Z.
        !           117: *> \endverbatim
        !           118: *>
        !           119: *> \param[out] VT
        !           120: *> \verbatim
        !           121: *>          VT is DOUBLE PRECISION array,
        !           122: *>         dimension ( LDU, SMLSIZ+1 ) if ICOMPQ = 1, and not referenced
        !           123: *>         if ICOMPQ = 0. If ICOMPQ = 1, on exit, VT**T contains the right
        !           124: *>         singular vector matrices of all subproblems at the bottom
        !           125: *>         level.
        !           126: *> \endverbatim
        !           127: *>
        !           128: *> \param[out] K
        !           129: *> \verbatim
        !           130: *>          K is INTEGER array,
        !           131: *>         dimension ( N ) if ICOMPQ = 1 and dimension 1 if ICOMPQ = 0.
        !           132: *>         If ICOMPQ = 1, on exit, K(I) is the dimension of the I-th
        !           133: *>         secular equation on the computation tree.
        !           134: *> \endverbatim
        !           135: *>
        !           136: *> \param[out] DIFL
        !           137: *> \verbatim
        !           138: *>          DIFL is DOUBLE PRECISION array, dimension ( LDU, NLVL ),
        !           139: *>         where NLVL = floor(log_2 (N/SMLSIZ))).
        !           140: *> \endverbatim
        !           141: *>
        !           142: *> \param[out] DIFR
        !           143: *> \verbatim
        !           144: *>          DIFR is DOUBLE PRECISION array,
        !           145: *>                  dimension ( LDU, 2 * NLVL ) if ICOMPQ = 1 and
        !           146: *>                  dimension ( N ) if ICOMPQ = 0.
        !           147: *>         If ICOMPQ = 1, on exit, DIFL(1:N, I) and DIFR(1:N, 2 * I - 1)
        !           148: *>         record distances between singular values on the I-th
        !           149: *>         level and singular values on the (I -1)-th level, and
        !           150: *>         DIFR(1:N, 2 * I ) contains the normalizing factors for
        !           151: *>         the right singular vector matrix. See DLASD8 for details.
        !           152: *> \endverbatim
        !           153: *>
        !           154: *> \param[out] Z
        !           155: *> \verbatim
        !           156: *>          Z is DOUBLE PRECISION array,
        !           157: *>                  dimension ( LDU, NLVL ) if ICOMPQ = 1 and
        !           158: *>                  dimension ( N ) if ICOMPQ = 0.
        !           159: *>         The first K elements of Z(1, I) contain the components of
        !           160: *>         the deflation-adjusted updating row vector for subproblems
        !           161: *>         on the I-th level.
        !           162: *> \endverbatim
        !           163: *>
        !           164: *> \param[out] POLES
        !           165: *> \verbatim
        !           166: *>          POLES is DOUBLE PRECISION array,
        !           167: *>         dimension ( LDU, 2 * NLVL ) if ICOMPQ = 1, and not referenced
        !           168: *>         if ICOMPQ = 0. If ICOMPQ = 1, on exit, POLES(1, 2*I - 1) and
        !           169: *>         POLES(1, 2*I) contain  the new and old singular values
        !           170: *>         involved in the secular equations on the I-th level.
        !           171: *> \endverbatim
        !           172: *>
        !           173: *> \param[out] GIVPTR
        !           174: *> \verbatim
        !           175: *>          GIVPTR is INTEGER array,
        !           176: *>         dimension ( N ) if ICOMPQ = 1, and not referenced if
        !           177: *>         ICOMPQ = 0. If ICOMPQ = 1, on exit, GIVPTR( I ) records
        !           178: *>         the number of Givens rotations performed on the I-th
        !           179: *>         problem on the computation tree.
        !           180: *> \endverbatim
        !           181: *>
        !           182: *> \param[out] GIVCOL
        !           183: *> \verbatim
        !           184: *>          GIVCOL is INTEGER array,
        !           185: *>         dimension ( LDGCOL, 2 * NLVL ) if ICOMPQ = 1, and not
        !           186: *>         referenced if ICOMPQ = 0. If ICOMPQ = 1, on exit, for each I,
        !           187: *>         GIVCOL(1, 2 *I - 1) and GIVCOL(1, 2 *I) record the locations
        !           188: *>         of Givens rotations performed on the I-th level on the
        !           189: *>         computation tree.
        !           190: *> \endverbatim
        !           191: *>
        !           192: *> \param[in] LDGCOL
        !           193: *> \verbatim
        !           194: *>          LDGCOL is INTEGER, LDGCOL = > N.
        !           195: *>         The leading dimension of arrays GIVCOL and PERM.
        !           196: *> \endverbatim
        !           197: *>
        !           198: *> \param[out] PERM
        !           199: *> \verbatim
        !           200: *>          PERM is INTEGER array,
        !           201: *>         dimension ( LDGCOL, NLVL ) if ICOMPQ = 1, and not referenced
        !           202: *>         if ICOMPQ = 0. If ICOMPQ = 1, on exit, PERM(1, I) records
        !           203: *>         permutations done on the I-th level of the computation tree.
        !           204: *> \endverbatim
        !           205: *>
        !           206: *> \param[out] GIVNUM
        !           207: *> \verbatim
        !           208: *>          GIVNUM is DOUBLE PRECISION array,
        !           209: *>         dimension ( LDU,  2 * NLVL ) if ICOMPQ = 1, and not
        !           210: *>         referenced if ICOMPQ = 0. If ICOMPQ = 1, on exit, for each I,
        !           211: *>         GIVNUM(1, 2 *I - 1) and GIVNUM(1, 2 *I) record the C- and S-
        !           212: *>         values of Givens rotations performed on the I-th level on
        !           213: *>         the computation tree.
        !           214: *> \endverbatim
        !           215: *>
        !           216: *> \param[out] C
        !           217: *> \verbatim
        !           218: *>          C is DOUBLE PRECISION array,
        !           219: *>         dimension ( N ) if ICOMPQ = 1, and dimension 1 if ICOMPQ = 0.
        !           220: *>         If ICOMPQ = 1 and the I-th subproblem is not square, on exit,
        !           221: *>         C( I ) contains the C-value of a Givens rotation related to
        !           222: *>         the right null space of the I-th subproblem.
        !           223: *> \endverbatim
        !           224: *>
        !           225: *> \param[out] S
        !           226: *> \verbatim
        !           227: *>          S is DOUBLE PRECISION array, dimension ( N ) if
        !           228: *>         ICOMPQ = 1, and dimension 1 if ICOMPQ = 0. If ICOMPQ = 1
        !           229: *>         and the I-th subproblem is not square, on exit, S( I )
        !           230: *>         contains the S-value of a Givens rotation related to
        !           231: *>         the right null space of the I-th subproblem.
        !           232: *> \endverbatim
        !           233: *>
        !           234: *> \param[out] WORK
        !           235: *> \verbatim
        !           236: *>          WORK is DOUBLE PRECISION array, dimension
        !           237: *>         (6 * N + (SMLSIZ + 1)*(SMLSIZ + 1)).
        !           238: *> \endverbatim
        !           239: *>
        !           240: *> \param[out] IWORK
        !           241: *> \verbatim
        !           242: *>          IWORK is INTEGER array.
        !           243: *>         Dimension must be at least (7 * N).
        !           244: *> \endverbatim
        !           245: *>
        !           246: *> \param[out] INFO
        !           247: *> \verbatim
        !           248: *>          INFO is INTEGER
        !           249: *>          = 0:  successful exit.
        !           250: *>          < 0:  if INFO = -i, the i-th argument had an illegal value.
        !           251: *>          > 0:  if INFO = 1, a singular value did not converge
        !           252: *> \endverbatim
        !           253: *
        !           254: *  Authors:
        !           255: *  ========
        !           256: *
        !           257: *> \author Univ. of Tennessee 
        !           258: *> \author Univ. of California Berkeley 
        !           259: *> \author Univ. of Colorado Denver 
        !           260: *> \author NAG Ltd. 
        !           261: *
        !           262: *> \date November 2011
        !           263: *
        !           264: *> \ingroup auxOTHERauxiliary
        !           265: *
        !           266: *> \par Contributors:
        !           267: *  ==================
        !           268: *>
        !           269: *>     Ming Gu and Huan Ren, Computer Science Division, University of
        !           270: *>     California at Berkeley, USA
        !           271: *>
        !           272: *  =====================================================================
1.1       bertrand  273:       SUBROUTINE DLASDA( ICOMPQ, SMLSIZ, N, SQRE, D, E, U, LDU, VT, K,
                    274:      $                   DIFL, DIFR, Z, POLES, GIVPTR, GIVCOL, LDGCOL,
                    275:      $                   PERM, GIVNUM, C, S, WORK, IWORK, INFO )
                    276: *
1.10    ! bertrand  277: *  -- LAPACK auxiliary routine (version 3.4.0) --
1.1       bertrand  278: *  -- LAPACK is a software package provided by Univ. of Tennessee,    --
                    279: *  -- Univ. of California Berkeley, Univ. of Colorado Denver and NAG Ltd..--
1.10    ! bertrand  280: *     November 2011
1.1       bertrand  281: *
                    282: *     .. Scalar Arguments ..
                    283:       INTEGER            ICOMPQ, INFO, LDGCOL, LDU, N, SMLSIZ, SQRE
                    284: *     ..
                    285: *     .. Array Arguments ..
                    286:       INTEGER            GIVCOL( LDGCOL, * ), GIVPTR( * ), IWORK( * ),
                    287:      $                   K( * ), PERM( LDGCOL, * )
                    288:       DOUBLE PRECISION   C( * ), D( * ), DIFL( LDU, * ), DIFR( LDU, * ),
                    289:      $                   E( * ), GIVNUM( LDU, * ), POLES( LDU, * ),
                    290:      $                   S( * ), U( LDU, * ), VT( LDU, * ), WORK( * ),
                    291:      $                   Z( LDU, * )
                    292: *     ..
                    293: *
                    294: *  =====================================================================
                    295: *
                    296: *     .. Parameters ..
                    297:       DOUBLE PRECISION   ZERO, ONE
                    298:       PARAMETER          ( ZERO = 0.0D+0, ONE = 1.0D+0 )
                    299: *     ..
                    300: *     .. Local Scalars ..
                    301:       INTEGER            I, I1, IC, IDXQ, IDXQI, IM1, INODE, ITEMP, IWK,
                    302:      $                   J, LF, LL, LVL, LVL2, M, NCC, ND, NDB1, NDIML,
                    303:      $                   NDIMR, NL, NLF, NLP1, NLVL, NR, NRF, NRP1, NRU,
                    304:      $                   NWORK1, NWORK2, SMLSZP, SQREI, VF, VFI, VL, VLI
                    305:       DOUBLE PRECISION   ALPHA, BETA
                    306: *     ..
                    307: *     .. External Subroutines ..
                    308:       EXTERNAL           DCOPY, DLASD6, DLASDQ, DLASDT, DLASET, XERBLA
                    309: *     ..
                    310: *     .. Executable Statements ..
                    311: *
                    312: *     Test the input parameters.
                    313: *
                    314:       INFO = 0
                    315: *
                    316:       IF( ( ICOMPQ.LT.0 ) .OR. ( ICOMPQ.GT.1 ) ) THEN
                    317:          INFO = -1
                    318:       ELSE IF( SMLSIZ.LT.3 ) THEN
                    319:          INFO = -2
                    320:       ELSE IF( N.LT.0 ) THEN
                    321:          INFO = -3
                    322:       ELSE IF( ( SQRE.LT.0 ) .OR. ( SQRE.GT.1 ) ) THEN
                    323:          INFO = -4
                    324:       ELSE IF( LDU.LT.( N+SQRE ) ) THEN
                    325:          INFO = -8
                    326:       ELSE IF( LDGCOL.LT.N ) THEN
                    327:          INFO = -17
                    328:       END IF
                    329:       IF( INFO.NE.0 ) THEN
                    330:          CALL XERBLA( 'DLASDA', -INFO )
                    331:          RETURN
                    332:       END IF
                    333: *
                    334:       M = N + SQRE
                    335: *
                    336: *     If the input matrix is too small, call DLASDQ to find the SVD.
                    337: *
                    338:       IF( N.LE.SMLSIZ ) THEN
                    339:          IF( ICOMPQ.EQ.0 ) THEN
                    340:             CALL DLASDQ( 'U', SQRE, N, 0, 0, 0, D, E, VT, LDU, U, LDU,
                    341:      $                   U, LDU, WORK, INFO )
                    342:          ELSE
                    343:             CALL DLASDQ( 'U', SQRE, N, M, N, 0, D, E, VT, LDU, U, LDU,
                    344:      $                   U, LDU, WORK, INFO )
                    345:          END IF
                    346:          RETURN
                    347:       END IF
                    348: *
                    349: *     Book-keeping and  set up the computation tree.
                    350: *
                    351:       INODE = 1
                    352:       NDIML = INODE + N
                    353:       NDIMR = NDIML + N
                    354:       IDXQ = NDIMR + N
                    355:       IWK = IDXQ + N
                    356: *
                    357:       NCC = 0
                    358:       NRU = 0
                    359: *
                    360:       SMLSZP = SMLSIZ + 1
                    361:       VF = 1
                    362:       VL = VF + M
                    363:       NWORK1 = VL + M
                    364:       NWORK2 = NWORK1 + SMLSZP*SMLSZP
                    365: *
                    366:       CALL DLASDT( N, NLVL, ND, IWORK( INODE ), IWORK( NDIML ),
                    367:      $             IWORK( NDIMR ), SMLSIZ )
                    368: *
                    369: *     for the nodes on bottom level of the tree, solve
                    370: *     their subproblems by DLASDQ.
                    371: *
                    372:       NDB1 = ( ND+1 ) / 2
                    373:       DO 30 I = NDB1, ND
                    374: *
                    375: *        IC : center row of each node
                    376: *        NL : number of rows of left  subproblem
                    377: *        NR : number of rows of right subproblem
                    378: *        NLF: starting row of the left   subproblem
                    379: *        NRF: starting row of the right  subproblem
                    380: *
                    381:          I1 = I - 1
                    382:          IC = IWORK( INODE+I1 )
                    383:          NL = IWORK( NDIML+I1 )
                    384:          NLP1 = NL + 1
                    385:          NR = IWORK( NDIMR+I1 )
                    386:          NLF = IC - NL
                    387:          NRF = IC + 1
                    388:          IDXQI = IDXQ + NLF - 2
                    389:          VFI = VF + NLF - 1
                    390:          VLI = VL + NLF - 1
                    391:          SQREI = 1
                    392:          IF( ICOMPQ.EQ.0 ) THEN
                    393:             CALL DLASET( 'A', NLP1, NLP1, ZERO, ONE, WORK( NWORK1 ),
                    394:      $                   SMLSZP )
                    395:             CALL DLASDQ( 'U', SQREI, NL, NLP1, NRU, NCC, D( NLF ),
                    396:      $                   E( NLF ), WORK( NWORK1 ), SMLSZP,
                    397:      $                   WORK( NWORK2 ), NL, WORK( NWORK2 ), NL,
                    398:      $                   WORK( NWORK2 ), INFO )
                    399:             ITEMP = NWORK1 + NL*SMLSZP
                    400:             CALL DCOPY( NLP1, WORK( NWORK1 ), 1, WORK( VFI ), 1 )
                    401:             CALL DCOPY( NLP1, WORK( ITEMP ), 1, WORK( VLI ), 1 )
                    402:          ELSE
                    403:             CALL DLASET( 'A', NL, NL, ZERO, ONE, U( NLF, 1 ), LDU )
                    404:             CALL DLASET( 'A', NLP1, NLP1, ZERO, ONE, VT( NLF, 1 ), LDU )
                    405:             CALL DLASDQ( 'U', SQREI, NL, NLP1, NL, NCC, D( NLF ),
                    406:      $                   E( NLF ), VT( NLF, 1 ), LDU, U( NLF, 1 ), LDU,
                    407:      $                   U( NLF, 1 ), LDU, WORK( NWORK1 ), INFO )
                    408:             CALL DCOPY( NLP1, VT( NLF, 1 ), 1, WORK( VFI ), 1 )
                    409:             CALL DCOPY( NLP1, VT( NLF, NLP1 ), 1, WORK( VLI ), 1 )
                    410:          END IF
                    411:          IF( INFO.NE.0 ) THEN
                    412:             RETURN
                    413:          END IF
                    414:          DO 10 J = 1, NL
                    415:             IWORK( IDXQI+J ) = J
                    416:    10    CONTINUE
                    417:          IF( ( I.EQ.ND ) .AND. ( SQRE.EQ.0 ) ) THEN
                    418:             SQREI = 0
                    419:          ELSE
                    420:             SQREI = 1
                    421:          END IF
                    422:          IDXQI = IDXQI + NLP1
                    423:          VFI = VFI + NLP1
                    424:          VLI = VLI + NLP1
                    425:          NRP1 = NR + SQREI
                    426:          IF( ICOMPQ.EQ.0 ) THEN
                    427:             CALL DLASET( 'A', NRP1, NRP1, ZERO, ONE, WORK( NWORK1 ),
                    428:      $                   SMLSZP )
                    429:             CALL DLASDQ( 'U', SQREI, NR, NRP1, NRU, NCC, D( NRF ),
                    430:      $                   E( NRF ), WORK( NWORK1 ), SMLSZP,
                    431:      $                   WORK( NWORK2 ), NR, WORK( NWORK2 ), NR,
                    432:      $                   WORK( NWORK2 ), INFO )
                    433:             ITEMP = NWORK1 + ( NRP1-1 )*SMLSZP
                    434:             CALL DCOPY( NRP1, WORK( NWORK1 ), 1, WORK( VFI ), 1 )
                    435:             CALL DCOPY( NRP1, WORK( ITEMP ), 1, WORK( VLI ), 1 )
                    436:          ELSE
                    437:             CALL DLASET( 'A', NR, NR, ZERO, ONE, U( NRF, 1 ), LDU )
                    438:             CALL DLASET( 'A', NRP1, NRP1, ZERO, ONE, VT( NRF, 1 ), LDU )
                    439:             CALL DLASDQ( 'U', SQREI, NR, NRP1, NR, NCC, D( NRF ),
                    440:      $                   E( NRF ), VT( NRF, 1 ), LDU, U( NRF, 1 ), LDU,
                    441:      $                   U( NRF, 1 ), LDU, WORK( NWORK1 ), INFO )
                    442:             CALL DCOPY( NRP1, VT( NRF, 1 ), 1, WORK( VFI ), 1 )
                    443:             CALL DCOPY( NRP1, VT( NRF, NRP1 ), 1, WORK( VLI ), 1 )
                    444:          END IF
                    445:          IF( INFO.NE.0 ) THEN
                    446:             RETURN
                    447:          END IF
                    448:          DO 20 J = 1, NR
                    449:             IWORK( IDXQI+J ) = J
                    450:    20    CONTINUE
                    451:    30 CONTINUE
                    452: *
                    453: *     Now conquer each subproblem bottom-up.
                    454: *
                    455:       J = 2**NLVL
                    456:       DO 50 LVL = NLVL, 1, -1
                    457:          LVL2 = LVL*2 - 1
                    458: *
                    459: *        Find the first node LF and last node LL on
                    460: *        the current level LVL.
                    461: *
                    462:          IF( LVL.EQ.1 ) THEN
                    463:             LF = 1
                    464:             LL = 1
                    465:          ELSE
                    466:             LF = 2**( LVL-1 )
                    467:             LL = 2*LF - 1
                    468:          END IF
                    469:          DO 40 I = LF, LL
                    470:             IM1 = I - 1
                    471:             IC = IWORK( INODE+IM1 )
                    472:             NL = IWORK( NDIML+IM1 )
                    473:             NR = IWORK( NDIMR+IM1 )
                    474:             NLF = IC - NL
                    475:             NRF = IC + 1
                    476:             IF( I.EQ.LL ) THEN
                    477:                SQREI = SQRE
                    478:             ELSE
                    479:                SQREI = 1
                    480:             END IF
                    481:             VFI = VF + NLF - 1
                    482:             VLI = VL + NLF - 1
                    483:             IDXQI = IDXQ + NLF - 1
                    484:             ALPHA = D( IC )
                    485:             BETA = E( IC )
                    486:             IF( ICOMPQ.EQ.0 ) THEN
                    487:                CALL DLASD6( ICOMPQ, NL, NR, SQREI, D( NLF ),
                    488:      $                      WORK( VFI ), WORK( VLI ), ALPHA, BETA,
                    489:      $                      IWORK( IDXQI ), PERM, GIVPTR( 1 ), GIVCOL,
                    490:      $                      LDGCOL, GIVNUM, LDU, POLES, DIFL, DIFR, Z,
                    491:      $                      K( 1 ), C( 1 ), S( 1 ), WORK( NWORK1 ),
                    492:      $                      IWORK( IWK ), INFO )
                    493:             ELSE
                    494:                J = J - 1
                    495:                CALL DLASD6( ICOMPQ, NL, NR, SQREI, D( NLF ),
                    496:      $                      WORK( VFI ), WORK( VLI ), ALPHA, BETA,
                    497:      $                      IWORK( IDXQI ), PERM( NLF, LVL ),
                    498:      $                      GIVPTR( J ), GIVCOL( NLF, LVL2 ), LDGCOL,
                    499:      $                      GIVNUM( NLF, LVL2 ), LDU,
                    500:      $                      POLES( NLF, LVL2 ), DIFL( NLF, LVL ),
                    501:      $                      DIFR( NLF, LVL2 ), Z( NLF, LVL ), K( J ),
                    502:      $                      C( J ), S( J ), WORK( NWORK1 ),
                    503:      $                      IWORK( IWK ), INFO )
                    504:             END IF
                    505:             IF( INFO.NE.0 ) THEN
                    506:                RETURN
                    507:             END IF
                    508:    40    CONTINUE
                    509:    50 CONTINUE
                    510: *
                    511:       RETURN
                    512: *
                    513: *     End of DLASDA
                    514: *
                    515:       END

CVSweb interface <joel.bertrand@systella.fr>