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
      SUBROUTINE DPPEQU( UPLO, N, AP, S, SCOND, AMAX, INFO )
*
*  -- LAPACK routine (version 3.2) --
*  -- LAPACK is a software package provided by Univ. of Tennessee,    --
*  -- Univ. of California Berkeley, Univ. of Colorado Denver and NAG Ltd..--
*     November 2006
*
*     .. Scalar Arguments ..
      CHARACTER          UPLO
      INTEGER            INFO, N
      DOUBLE PRECISION   AMAX, SCOND
*     ..
*     .. Array Arguments ..
      DOUBLE PRECISION   AP( * ), S( * )
*     ..
*
*  Purpose
*  =======
*
*  DPPEQU computes row and column scalings intended to equilibrate a
*  symmetric positive definite matrix A in packed storage and reduce
*  its condition number (with respect to the two-norm).  S contains the
*  scale factors, S(i)=1/sqrt(A(i,i)), chosen so that the scaled matrix
*  B with elements B(i,j)=S(i)*A(i,j)*S(j) has ones on the diagonal.
*  This choice of S puts the condition number of B within a factor N of
*  the smallest possible condition number over all possible diagonal
*  scalings.
*
*  Arguments
*  =========
*
*  UPLO    (input) CHARACTER*1
*          = 'U':  Upper triangle of A is stored;
*          = 'L':  Lower triangle of A is stored.
*
*  N       (input) INTEGER
*          The order of the matrix A.  N >= 0.
*
*  AP      (input) DOUBLE PRECISION array, dimension (N*(N+1)/2)
*          The upper or lower triangle of the symmetric matrix A, packed
*          columnwise in a linear array.  The j-th column of A is stored
*          in the array AP as follows:
*          if UPLO = 'U', AP(i + (j-1)*j/2) = A(i,j) for 1<=i<=j;
*          if UPLO = 'L', AP(i + (j-1)*(2n-j)/2) = A(i,j) for j<=i<=n.
*
*  S       (output) DOUBLE PRECISION array, dimension (N)
*          If INFO = 0, S contains the scale factors for A.
*
*  SCOND   (output) DOUBLE PRECISION
*          If INFO = 0, S contains the ratio of the smallest S(i) to
*          the largest S(i).  If SCOND >= 0.1 and AMAX is neither too
*          large nor too small, it is not worth scaling by S.
*
*  AMAX    (output) DOUBLE PRECISION
*          Absolute value of largest matrix element.  If AMAX is very
*          close to overflow or very close to underflow, the matrix
*          should be scaled.
*
*  INFO    (output) INTEGER
*          = 0:  successful exit
*          < 0:  if INFO = -i, the i-th argument had an illegal value
*          > 0:  if INFO = i, the i-th diagonal element is nonpositive.
*
*  =====================================================================
*
*     .. Parameters ..
      DOUBLE PRECISION   ONE, ZERO
      PARAMETER          ( ONE = 1.0D+0, ZERO = 0.0D+0 )
*     ..
*     .. Local Scalars ..
      LOGICAL            UPPER
      INTEGER            I, JJ
      DOUBLE PRECISION   SMIN
*     ..
*     .. External Functions ..
      LOGICAL            LSAME
      EXTERNAL           LSAME
*     ..
*     .. External Subroutines ..
      EXTERNAL           XERBLA
*     ..
*     .. Intrinsic Functions ..
      INTRINSIC          MAXMINSQRT
*     ..
*     .. Executable Statements ..
*
*     Test the input parameters.
*
      INFO = 0
      UPPER = LSAME( UPLO, 'U' )
      IF.NOT.UPPER .AND. .NOT.LSAME( UPLO, 'L' ) ) THEN
         INFO = -1
      ELSE IF( N.LT.0 ) THEN
         INFO = -2
      END IF
      IF( INFO.NE.0 ) THEN
         CALL XERBLA( 'DPPEQU'-INFO )
         RETURN
      END IF
*
*     Quick return if possible
*
      IF( N.EQ.0 ) THEN
         SCOND = ONE
         AMAX = ZERO
         RETURN
      END IF
*
*     Initialize SMIN and AMAX.
*
      S( 1 ) = AP( 1 )
      SMIN = S( 1 )
      AMAX = S( 1 )
*
      IF( UPPER ) THEN
*
*        UPLO = 'U':  Upper triangle of A is stored.
*        Find the minimum and maximum diagonal elements.
*
         JJ = 1
         DO 10 I = 2, N
            JJ = JJ + I
            S( I ) = AP( JJ )
            SMIN = MIN( SMIN, S( I ) )
            AMAX = MAX( AMAX, S( I ) )
   10    CONTINUE
*
      ELSE
*
*        UPLO = 'L':  Lower triangle of A is stored.
*        Find the minimum and maximum diagonal elements.
*
         JJ = 1
         DO 20 I = 2, N
            JJ = JJ + N - I + 2
            S( I ) = AP( JJ )
            SMIN = MIN( SMIN, S( I ) )
            AMAX = MAX( AMAX, S( I ) )
   20    CONTINUE
      END IF
*
      IF( SMIN.LE.ZERO ) THEN
*
*        Find the first non-positive diagonal element and return.
*
         DO 30 I = 1, N
            IF( S( I ).LE.ZERO ) THEN
               INFO = I
               RETURN
            END IF
   30    CONTINUE
      ELSE
*
*        Set the scale factors to the reciprocals
*        of the diagonal elements.
*
         DO 40 I = 1, N
            S( I ) = ONE / SQRT( S( I ) )
   40    CONTINUE
*
*        Compute SCOND = min(S(I)) / max(S(I))
*
         SCOND = SQRT( SMIN ) / SQRT( AMAX )
      END IF
      RETURN
*
*     End of DPPEQU
*
      END