1CPPEQU(1)                LAPACK routine (version 3.2)                CPPEQU(1)
2
3
4

NAME

6       CPPEQU  -  computes  row  and column scalings intended to equilibrate a
7       Hermitian positive definite matrix A in packed storage and  reduce  its
8       condition number (with respect to the two-norm)
9

SYNOPSIS

11       SUBROUTINE CPPEQU( UPLO, N, AP, S, SCOND, AMAX, INFO )
12
13           CHARACTER      UPLO
14
15           INTEGER        INFO, N
16
17           REAL           AMAX, SCOND
18
19           REAL           S( * )
20
21           COMPLEX        AP( * )
22

PURPOSE

24       CPPEQU  computes row and column scalings intended to equilibrate a Her‐
25       mitian positive definite matrix A in packed storage and reduce its con‐
26       dition  number  (with  respect  to the two-norm).  S contains the scale
27       factors, S(i)=1/sqrt(A(i,i)), chosen so that the scaled matrix  B  with
28       elements B(i,j)=S(i)*A(i,j)*S(j) has ones on the diagonal.  This choice
29       of S puts the condition number of B within a factor N of  the  smallest
30       possible condition number over all possible diagonal scalings.
31

ARGUMENTS

33       UPLO    (input) CHARACTER*1
34               = 'U':  Upper triangle of A is stored;
35               = 'L':  Lower triangle of A is stored.
36
37       N       (input) INTEGER
38               The order of the matrix A.  N >= 0.
39
40       AP      (input) COMPLEX array, dimension (N*(N+1)/2)
41               The  upper  or lower triangle of the Hermitian matrix A, packed
42               columnwise in a linear array.  The j-th column of A  is  stored
43               in  the array AP as follows: if UPLO = 'U', AP(i + (j-1)*j/2) =
44               A(i,j) for 1<=i<=j; if UPLO = 'L',  AP(i  +  (j-1)*(2n-j)/2)  =
45               A(i,j) for j<=i<=n.
46
47       S       (output) REAL array, dimension (N)
48               If INFO = 0, S contains the scale factors for A.
49
50       SCOND   (output) REAL
51               If  INFO  = 0, S contains the ratio of the smallest S(i) to the
52               largest S(i).  If SCOND >= 0.1 and AMAX is  neither  too  large
53               nor too small, it is not worth scaling by S.
54
55       AMAX    (output) REAL
56               Absolute  value  of  largest  matrix  element.  If AMAX is very
57               close to overflow or very close to underflow, the matrix should
58               be scaled.
59
60       INFO    (output) INTEGER
61               = 0:  successful exit
62               < 0:  if INFO = -i, the i-th argument had an illegal value
63               > 0:  if INFO = i, the i-th diagonal element is nonpositive.
64
65
66
67 LAPACK routine (version 3.2)    November 2008                       CPPEQU(1)
Impressum