1
       2
       3
       4
       5
       6
       7
       8
       9
      10
      11
      12
      13
      14
      15
      16
      17
      18
      19
      20
      21
      22
      23
      24
      25
      26
      27
      28
      29
      30
      31
      32
      33
      34
      35
      36
      37
      38
      39
      40
      41
      42
      43
      44
      45
      46
      47
      48
      49
      50
      51
      52
      53
      54
      55
      56
      57
      58
      59
      60
      61
      62
      63
      64
      65
      66
      67
      68
      69
      70
      71
      72
      73
      74
      75
      76
      77
      78
      79
      80
      81
      82
      83
      84
      85
      86
      87
      88
      89
      90
      91
      92
      93
      94
      95
      96
      97
      98
      99
     100
     101
     102
     103
     104
     105
     106
     107
     108
     109
     110
     111
     112
     113
     114
     115
     116
     117
     118
     119
     120
     121
     122
     123
     124
     125
     126
     127
     128
     129
     130
     131
     132
     133
     134
     135
     136
     137
     138
     139
     140
     141
     142
     143
     144
     145
     146
     147
     148
     149
     150
     151
     152
     153
     154
     155
     156
     157
     158
     159
     160
     161
     162
     163
     164
     165
     166
     167
     168
     169
     170
     171
     172
     173
     174
     175
     176
     177
     178
     179
     180
     181
     182
     183
     184
     185
     186
     187
     188
     189
     190
     191
     192
     193
     194
     195
     196
     197
     198
     199
     200
     201
     202
     203
     204
     205
     206
     207
     208
     209
     210
     211
     212
     213
     214
     215
     216
     217
     218
     219
     220
     221
     222
     223
     224
     225
     226
     227
     228
     229
     230
     231
     232
     233
     234
     235
     236
     237
     238
     239
     240
     241
     242
     243
     244
     245
     246
     247
     248
     249
     250
     251
     252
     253
     254
     255
     256
     257
     258
     259
     260
     261
     262
     263
     264
     265
     266
     267
     268
     269
     270
     271
     272
     273
     274
     275
     276
     277
     278
     279
     280
     281
     282
     283
     284
     285
     286
     287
     288
     289
     290
     291
     292
     293
     294
     295
     296
     297
     298
     299
     300
     301
     302
     303
     304
     305
     306
     307
     308
     309
     310
     311
     312
     313
      SUBROUTINE DLASD6( ICOMPQ, NL, NR, SQRE, D, VF, VL, ALPHA, BETA,
     $                   IDXQ, PERM, GIVPTR, GIVCOL, LDGCOL, GIVNUM,
     $                   LDGNUM, POLES, DIFL, DIFR, Z, K, C, S, WORK,
     $                   IWORK, INFO )
*
*  -- LAPACK auxiliary routine (version 3.3.0) --
*  -- LAPACK is a software package provided by Univ. of Tennessee,    --
*  -- Univ. of California Berkeley, Univ. of Colorado Denver and NAG Ltd..--
*     November 2010
*
*     .. Scalar Arguments ..
      INTEGER            GIVPTR, ICOMPQ, INFO, K, LDGCOL, LDGNUM, NL,
     $                   NR, SQRE
      DOUBLE PRECISION   ALPHA, BETA, C, S
*     ..
*     .. Array Arguments ..
      INTEGER            GIVCOL( LDGCOL, * ), IDXQ( * ), IWORK( * ),
     $                   PERM( * )
      DOUBLE PRECISION   D( * ), DIFL( * ), DIFR( * ),
     $                   GIVNUM( LDGNUM, * ), POLES( LDGNUM, * ),
     $                   VF( * ), VL( * ), WORK( * ), Z( * )
*     ..
*
*  Purpose
*  =======
*
*  DLASD6 computes the SVD of an updated upper bidiagonal matrix B
*  obtained by merging two smaller ones by appending a row. This
*  routine is used only for the problem which requires all singular
*  values and optionally singular vector matrices in factored form.
*  B is an N-by-M matrix with N = NL + NR + 1 and M = N + SQRE.
*  A related subroutine, DLASD1, handles the case in which all singular
*  values and singular vectors of the bidiagonal matrix are desired.
*
*  DLASD6 computes the SVD as follows:
*
*                ( D1(in)    0    0       0 )
*    B = U(in) * (   Z1**T   a   Z2**T    b ) * VT(in)
*                (   0       0   D2(in)   0 )
*
*      = U(out) * ( D(out) 0) * VT(out)
*
*  where Z**T = (Z1**T a Z2**T b) = u**T VT**T, and u is a vector of dimension M
*  with ALPHA and BETA in the NL+1 and NL+2 th entries and zeros
*  elsewhere; and the entry b is empty if SQRE = 0.
*
*  The singular values of B can be computed using D1, D2, the first
*  components of all the right singular vectors of the lower block, and
*  the last components of all the right singular vectors of the upper
*  block. These components are stored and updated in VF and VL,
*  respectively, in DLASD6. Hence U and VT are not explicitly
*  referenced.
*
*  The singular values are stored in D. The algorithm consists of two
*  stages:
*
*        The first stage consists of deflating the size of the problem
*        when there are multiple singular values or if there is a zero
*        in the Z vector. For each such occurence the dimension of the
*        secular equation problem is reduced by one. This stage is
*        performed by the routine DLASD7.
*
*        The second stage consists of calculating the updated
*        singular values. This is done by finding the roots of the
*        secular equation via the routine DLASD4 (as called by DLASD8).
*        This routine also updates VF and VL and computes the distances
*        between the updated singular values and the old singular
*        values.
*
*  DLASD6 is called from DLASDA.
*
*  Arguments
*  =========
*
*  ICOMPQ (input) INTEGER
*         Specifies whether singular vectors are to be computed in
*         factored form:
*         = 0: Compute singular values only.
*         = 1: Compute singular vectors in factored form as well.
*
*  NL     (input) INTEGER
*         The row dimension of the upper block.  NL >= 1.
*
*  NR     (input) INTEGER
*         The row dimension of the lower block.  NR >= 1.
*
*  SQRE   (input) INTEGER
*         = 0: the lower block is an NR-by-NR square matrix.
*         = 1: the lower block is an NR-by-(NR+1) rectangular matrix.
*
*         The bidiagonal matrix has row dimension N = NL + NR + 1,
*         and column dimension M = N + SQRE.
*
*  D      (input/output) DOUBLE PRECISION array, dimension ( NL+NR+1 ).
*         On entry D(1:NL,1:NL) contains the singular values of the
*         upper block, and D(NL+2:N) contains the singular values
*         of the lower block. On exit D(1:N) contains the singular
*         values of the modified matrix.
*
*  VF     (input/output) DOUBLE PRECISION array, dimension ( M )
*         On entry, VF(1:NL+1) contains the first components of all
*         right singular vectors of the upper block; and VF(NL+2:M)
*         contains the first components of all right singular vectors
*         of the lower block. On exit, VF contains the first components
*         of all right singular vectors of the bidiagonal matrix.
*
*  VL     (input/output) DOUBLE PRECISION array, dimension ( M )
*         On entry, VL(1:NL+1) contains the  last components of all
*         right singular vectors of the upper block; and VL(NL+2:M)
*         contains the last components of all right singular vectors of
*         the lower block. On exit, VL contains the last components of
*         all right singular vectors of the bidiagonal matrix.
*
*  ALPHA  (input/output) DOUBLE PRECISION
*         Contains the diagonal element associated with the added row.
*
*  BETA   (input/output) DOUBLE PRECISION
*         Contains the off-diagonal element associated with the added
*         row.
*
*  IDXQ   (output) INTEGER array, dimension ( N )
*         This contains the permutation which will reintegrate the
*         subproblem just solved back into sorted order, i.e.
*         D( IDXQ( I = 1, N ) ) will be in ascending order.
*
*  PERM   (output) INTEGER array, dimension ( N )
*         The permutations (from deflation and sorting) to be applied
*         to each block. Not referenced if ICOMPQ = 0.
*
*  GIVPTR (output) INTEGER
*         The number of Givens rotations which took place in this
*         subproblem. Not referenced if ICOMPQ = 0.
*
*  GIVCOL (output) INTEGER array, dimension ( LDGCOL, 2 )
*         Each pair of numbers indicates a pair of columns to take place
*         in a Givens rotation. Not referenced if ICOMPQ = 0.
*
*  LDGCOL (input) INTEGER
*         leading dimension of GIVCOL, must be at least N.
*
*  GIVNUM (output) DOUBLE PRECISION array, dimension ( LDGNUM, 2 )
*         Each number indicates the C or S value to be used in the
*         corresponding Givens rotation. Not referenced if ICOMPQ = 0.
*
*  LDGNUM (input) INTEGER
*         The leading dimension of GIVNUM and POLES, must be at least N.
*
*  POLES  (output) DOUBLE PRECISION array, dimension ( LDGNUM, 2 )
*         On exit, POLES(1,*) is an array containing the new singular
*         values obtained from solving the secular equation, and
*         POLES(2,*) is an array containing the poles in the secular
*         equation. Not referenced if ICOMPQ = 0.
*
*  DIFL   (output) DOUBLE PRECISION array, dimension ( N )
*         On exit, DIFL(I) is the distance between I-th updated
*         (undeflated) singular value and the I-th (undeflated) old
*         singular value.
*
*  DIFR   (output) DOUBLE PRECISION array,
*                  dimension ( LDGNUM, 2 ) if ICOMPQ = 1 and
*                  dimension ( N ) if ICOMPQ = 0.
*         On exit, DIFR(I, 1) is the distance between I-th updated
*         (undeflated) singular value and the I+1-th (undeflated) old
*         singular value.
*
*         If ICOMPQ = 1, DIFR(1:K,2) is an array containing the
*         normalizing factors for the right singular vector matrix.
*
*         See DLASD8 for details on DIFL and DIFR.
*
*  Z      (output) DOUBLE PRECISION array, dimension ( M )
*         The first elements of this array contain the components
*         of the deflation-adjusted updating row vector.
*
*  K      (output) INTEGER
*         Contains the dimension of the non-deflated matrix,
*         This is the order of the related secular equation. 1 <= K <=N.
*
*  C      (output) DOUBLE PRECISION
*         C contains garbage if SQRE =0 and the C-value of a Givens
*         rotation related to the right null space if SQRE = 1.
*
*  S      (output) DOUBLE PRECISION
*         S contains garbage if SQRE =0 and the S-value of a Givens
*         rotation related to the right null space if SQRE = 1.
*
*  WORK   (workspace) DOUBLE PRECISION array, dimension ( 4 * M )
*
*  IWORK  (workspace) INTEGER array, dimension ( 3 * N )
*
*  INFO   (output) INTEGER
*          = 0:  successful exit.
*          < 0:  if INFO = -i, the i-th argument had an illegal value.
*          > 0:  if INFO = 1, a singular value did not converge
*
*  Further Details
*  ===============
*
*  Based on contributions by
*     Ming Gu and Huan Ren, Computer Science Division, University of
*     California at Berkeley, USA
*
*  =====================================================================
*
*     .. Parameters ..
      DOUBLE PRECISION   ONE, ZERO
      PARAMETER          ( ONE = 1.0D+0, ZERO = 0.0D+0 )
*     ..
*     .. Local Scalars ..
      INTEGER            I, IDX, IDXC, IDXP, ISIGMA, IVFW, IVLW, IW, M,
     $                   N, N1, N2
      DOUBLE PRECISION   ORGNRM
*     ..
*     .. External Subroutines ..
      EXTERNAL           DCOPY, DLAMRG, DLASCL, DLASD7, DLASD8, XERBLA
*     ..
*     .. Intrinsic Functions ..
      INTRINSIC          ABSMAX
*     ..
*     .. Executable Statements ..
*
*     Test the input parameters.
*
      INFO = 0
      N = NL + NR + 1
      M = N + SQRE
*
      IF( ( ICOMPQ.LT.0 ) .OR. ( ICOMPQ.GT.1 ) ) THEN
         INFO = -1
      ELSE IF( NL.LT.1 ) THEN
         INFO = -2
      ELSE IF( NR.LT.1 ) THEN
         INFO = -3
      ELSE IF( ( SQRE.LT.0 ) .OR. ( SQRE.GT.1 ) ) THEN
         INFO = -4
      ELSE IF( LDGCOL.LT.N ) THEN
         INFO = -14
      ELSE IF( LDGNUM.LT.N ) THEN
         INFO = -16
      END IF
      IF( INFO.NE.0 ) THEN
         CALL XERBLA( 'DLASD6'-INFO )
         RETURN
      END IF
*
*     The following values are for bookkeeping purposes only.  They are
*     integer pointers which indicate the portion of the workspace
*     used by a particular array in DLASD7 and DLASD8.
*
      ISIGMA = 1
      IW = ISIGMA + N
      IVFW = IW + M
      IVLW = IVFW + M
*
      IDX = 1
      IDXC = IDX + N
      IDXP = IDXC + N
*
*     Scale.
*
      ORGNRM = MAXABS( ALPHA ), ABS( BETA ) )
      D( NL+1 ) = ZERO
      DO 10 I = 1, N
         IFABS( D( I ) ).GT.ORGNRM ) THEN
            ORGNRM = ABS( D( I ) )
         END IF
   10 CONTINUE
      CALL DLASCL( 'G'00, ORGNRM, ONE, N, 1, D, N, INFO )
      ALPHA = ALPHA / ORGNRM
      BETA = BETA / ORGNRM
*
*     Sort and Deflate singular values.
*
      CALL DLASD7( ICOMPQ, NL, NR, SQRE, K, D, Z, WORK( IW ), VF,
     $             WORK( IVFW ), VL, WORK( IVLW ), ALPHA, BETA,
     $             WORK( ISIGMA ), IWORK( IDX ), IWORK( IDXP ), IDXQ,
     $             PERM, GIVPTR, GIVCOL, LDGCOL, GIVNUM, LDGNUM, C, S,
     $             INFO )
*
*     Solve Secular Equation, compute DIFL, DIFR, and update VF, VL.
*
      CALL DLASD8( ICOMPQ, K, D, Z, VF, VL, DIFL, DIFR, LDGNUM,
     $             WORK( ISIGMA ), WORK( IW ), INFO )
*
*     Handle error returned
*
      IF( INFO.NE.0 ) THEN
         CALL XERBLA( 'DLASD8'-INFO )
         RETURN
      END IF
*
*     Save the poles if ICOMPQ = 1.
*
      IF( ICOMPQ.EQ.1 ) THEN
         CALL DCOPY( K, D, 1, POLES( 11 ), 1 )
         CALL DCOPY( K, WORK( ISIGMA ), 1, POLES( 12 ), 1 )
      END IF
*
*     Unscale.
*
      CALL DLASCL( 'G'00, ONE, ORGNRM, N, 1, D, N, INFO )
*
*     Prepare the IDXQ sorting permutation.
*
      N1 = K
      N2 = N - K
      CALL DLAMRG( N1, N2, D, 1-1, IDXQ )
*
      RETURN
*
*     End of DLASD6
*
      END