zchdc(l) — SunSoft Performance Library

NAME

zchdc - compute the Cholesky decomposition of a symmetric positive definite matrix A.

SYNOPSIS

CALL DCHDC (DA, LDA, N, DWORK, IPIVOT, JOB, INFO)

CALL SCHDC (SA, LDA, N, SWORK, IPIVOT, JOB, INFO)

CALL ZCHDC (ZA, LDA, N, ZWORK, IPIVOT, JOB, INFO)

CALL CCHDC (CA, LDA, N, CWORK, IPIVOT, JOB, INFO)

ARGUMENTS

xAOn entry, the upper triangle of A.
On exit, the upper triangle of the Cholesky factor, as permuted by pivoting if pivoting was selected. The strict lower triangle of A is not referenced.

LDALeading dimension of the array A as specified in a dimension or type statement. LDA >= max(1,N).

NOrder of the matrix A. N >= 0.

xWORKScratch array with a dimension of N.

IPIVOTIf JOB selected no pivoting then IPIVOT is not referenced.
On entry, if JOB selected pivoting, IPIVOT contains integers representing array indices that control the selection of pivot elements from the diagonal of A according to the system below:
   IPIVOT(k) > 0 A(k,k) is an initial element.
   IPIVOT(k) = 0 A(k,k) is a free element.
   IPIVOT(k) < 0 A(k,k) is a final element.
Before the decomposition is computed, symmetric row and column interchanges are used to move initial elements to the beginning of A and final elements to the end of A. During the computation, symmetric row and column interchanges are used to move the largest remaining free diagonal element into the pivot position. On exit, IPIVOT(k) contains the index of the diagonal element of A that was moved into the kth position.

JOBDetermines how decomposition is done:
0 without pivoting
not 0 with pivoting

INFOOn exit:
Returns the index of the last positive diagonal element of the Cholesky factor. The subroutine was successful if INFO = N.

SAMPLE PROGRAM

      PROGRAM TEST
      IMPLICIT NONE
C
      INTEGER           LDA, N, NOPIV
      PARAMETER        (N = 4)
      PARAMETER        (LDA = N)
      PARAMETER        (NOPIV = 0)
C
      DOUBLE PRECISION A(LDA,N), WORK(N)
      INTEGER           I, INFO, IPIVOT(N), J, JOB
C
      EXTERNAL          DCHDC
C
C         4 3 2 1
C     A = 3 4 3 2
C         2 3 4 3
C         1 2 3 4
C
      DATA A / 4.0D0, 3∗8D8, 3.0D0, 4.0D0, 2∗8D8, 2.0D0, 3.0D0, 4.0D0,
     $         8D8, 1.0D0, 2.0D0, 3.0D0, 4.0D0 /
C
      PRINT 1000
      DO 100, I = 1, N
        PRINT 1010, (A(J,I), J = 1, I), (A(I,J), J = I + 1, N)
100 CONTINUE
      PRINT 1020
      PRINT 1010, ((A(I,J), J = 1, N), I = 1, N)
      JOB = NOPIV
      CALL DCHDC (A, LDA, N, WORK, IPIVOT, JOB, INFO)
      IF (INFO .EQ. N) THEN
        PRINT 1030
        PRINT 1010, A(1,1), A(1,2), A(1,3), A(1,4)
        PRINT 1040,         A(2,2), A(2,3), A(2,4)
        PRINT 1050,                 A(3,3), A(3,4)
        PRINT 1060,                         A(4,4)
      ELSE
        PRINT 1070
      END IF
C
1000 FORMAT (1X, ’A in full form:’)
1010 FORMAT (4(3X, F7.3))
1020 FORMAT (/1X, ’A in symmetric form: (∗ in unused entries)’)
1030 FORMAT (/1X, ’Upper Cholesky factor:’)
1040 FORMAT (10X, 3(3X, F7.3))
1050 FORMAT (20X, 2(3X, F7.3))
1060 FORMAT (30X, 1(3X, F7.3))
1070 FORMAT (/1X, ’A is not positive definite.’)
C
      END

SAMPLE OUTPUT

A in full form:
     4.000     3.000     2.000     1.000
     3.000     4.000     3.000     2.000
     2.000     3.000     4.000     3.000
     1.000     2.000     3.000     4.000

A in symmetric form: (∗ in unused entries)
     4.000     3.000     2.000     1.000
   ∗∗∗∗∗∗∗     4.000     3.000     2.000
   ∗∗∗∗∗∗∗   ∗∗∗∗∗∗∗     4.000     3.000
   ∗∗∗∗∗∗∗   ∗∗∗∗∗∗∗   ∗∗∗∗∗∗∗     4.000

Upper Cholesky factor:
     2.000     1.500     1.000     0.500
               1.323     1.134     0.945
                         1.309     1.091
                                   1.291