1ZLAESY(1) LAPACK auxiliary routine (version 3.1) ZLAESY(1)
2
3
4
6 ZLAESY - the eigendecomposition of a 2-by-2 symmetric matrix ( ( A, B
7 );( B, C ) ) provided the norm of the matrix of eigenvectors is larger
8 than some threshold value
9
11 SUBROUTINE ZLAESY( A, B, C, RT1, RT2, EVSCAL, CS1, SN1 )
12
13 COMPLEX*16 A, B, C, CS1, EVSCAL, RT1, RT2, SN1
14
16 ZLAESY computes the eigendecomposition of a 2-by-2 symmetric matrix
17 ( ( A, B );( B, C ) ) provided the norm of the matrix of eigenvec‐
18 tors is larger than some threshold value.
19
20 RT1 is the eigenvalue of larger absolute value, and RT2 of smaller
21 absolute value. If the eigenvectors are computed, then on return (
22 CS1, SN1 ) is the unit eigenvector for RT1, hence
23
24 [ CS1 SN1 ] . [ A B ] . [ CS1 -SN1 ] = [ RT1 0 ] [ -SN1
25 CS1 ] [ B C ] [ SN1 CS1 ] [ 0 RT2 ]
26
27
29 A (input) COMPLEX*16
30 The ( 1, 1 ) element of input matrix.
31
32 B (input) COMPLEX*16
33 The ( 1, 2 ) element of input matrix. The ( 2, 1 ) element is
34 also given by B, since the 2-by-2 matrix is symmetric.
35
36 C (input) COMPLEX*16
37 The ( 2, 2 ) element of input matrix.
38
39 RT1 (output) COMPLEX*16
40 The eigenvalue of larger modulus.
41
42 RT2 (output) COMPLEX*16
43 The eigenvalue of smaller modulus.
44
45 EVSCAL (output) COMPLEX*16
46 The complex value by which the eigenvector matrix was scaled to
47 make it orthonormal. If EVSCAL is zero, the eigenvectors were
48 not computed. This means one of two things: the 2-by-2 matrix
49 could not be diagonalized, or the norm of the matrix of eigen‐
50 vectors before scaling was larger than the threshold value
51 THRESH (set below).
52
53 CS1 (output) COMPLEX*16
54 SN1 (output) COMPLEX*16 If EVSCAL .NE. 0, ( CS1, SN1 ) is
55 the unit right eigenvector for RT1.
56
57
58
59 LAPACK auxiliary routine (versionNo3v.e1m)ber 2006 ZLAESY(1)