1DLASD5(1) LAPACK auxiliary routine (version 3.2) DLASD5(1)
2
3
4
6 DLASD5 - subroutine compute the square root of the I-th eigenvalue of a
7 positive symmetric rank-one modification of a 2-by-2 diagonal matrix
8 diag( D ) * diag( D ) + RHO The diagonal entries in the array D are
9 assumed to satisfy 0 <= D(i) < D(j) for i < j
10
12 SUBROUTINE DLASD5( I, D, Z, DELTA, RHO, DSIGMA, WORK )
13
14 INTEGER I
15
16 DOUBLE PRECISION DSIGMA, RHO
17
18 DOUBLE PRECISION D( 2 ), DELTA( 2 ), WORK( 2 ), Z( 2 )
19
21 This subroutine computes the square root of the I-th eigenvalue of a
22 positive symmetric rank-one modification of a 2-by-2 diagonal matrix We
23 also assume RHO > 0 and that the Euclidean norm of the vector Z is one.
24
26 I (input) INTEGER
27 The index of the eigenvalue to be computed. I = 1 or I = 2.
28
29 D (input) DOUBLE PRECISION array, dimension ( 2 )
30 The original eigenvalues. We assume 0 <= D(1) < D(2).
31
32 Z (input) DOUBLE PRECISION array, dimension ( 2 )
33 The components of the updating vector.
34
35 DELTA (output) DOUBLE PRECISION array, dimension ( 2 )
36 Contains (D(j) - sigma_I) in its j-th component. The vector
37 DELTA contains the information necessary to construct the eigen‐
38 vectors.
39
40 RHO (input) DOUBLE PRECISION
41 The scalar in the symmetric updating formula. DSIGMA (output)
42 DOUBLE PRECISION The computed sigma_I, the I-th updated eigen‐
43 value.
44
45 WORK (workspace) DOUBLE PRECISION array, dimension ( 2 )
46 WORK contains (D(j) + sigma_I) in its j-th component.
47
49 Based on contributions by
50 Ren-Cang Li, Computer Science Division, University of California
51 at Berkeley, USA
52
53
54
55 LAPACK auxiliary routine (versionNo3v.e2m)ber 2008 DLASD5(1)