1CLAIC1(1) LAPACK auxiliary routine (version 3.2) CLAIC1(1)
2
3
4
6 CLAIC1 - applies one step of incremental condition estimation in its
7 simplest version
8
10 SUBROUTINE CLAIC1( JOB, J, X, SEST, W, GAMMA, SESTPR, S, C )
11
12 INTEGER J, JOB
13
14 REAL SEST, SESTPR
15
16 COMPLEX C, GAMMA, S
17
18 COMPLEX W( J ), X( J )
19
21 CLAIC1 applies one step of incremental condition estimation in its sim‐
22 plest version: Let x, twonorm(x) = 1, be an approximate singular vector
23 of an j-by-j lower triangular matrix L, such that
24 twonorm(L*x) = sest
25 Then CLAIC1 computes sestpr, s, c such that
26 the vector
27 [ s*x ]
28 xhat = [ c ]
29 is an approximate singular vector of
30 [ L 0 ]
31 Lhat = [ w' gamma ]
32 in the sense that
33 twonorm(Lhat*xhat) = sestpr.
34 Depending on JOB, an estimate for the largest or smallest singular
35 value is computed.
36 Note that [s c]' and sestpr**2 is an eigenpair of the system
37 diag(sest*sest, 0) + [alpha gamma] * [ conjg(alpha) ]
38 [ conjg(gamma) ] where alpha
39 = conjg(x)'*w.
40
42 JOB (input) INTEGER
43 = 1: an estimate for the largest singular value is computed.
44 = 2: an estimate for the smallest singular value is computed.
45
46 J (input) INTEGER
47 Length of X and W
48
49 X (input) COMPLEX array, dimension (J)
50 The j-vector x.
51
52 SEST (input) REAL
53 Estimated singular value of j by j matrix L
54
55 W (input) COMPLEX array, dimension (J)
56 The j-vector w.
57
58 GAMMA (input) COMPLEX
59 The diagonal element gamma.
60
61 SESTPR (output) REAL
62 Estimated singular value of (j+1) by (j+1) matrix Lhat.
63
64 S (output) COMPLEX
65 Sine needed in forming xhat.
66
67 C (output) COMPLEX
68 Cosine needed in forming xhat.
69
70
71
72 LAPACK auxiliary routine (versionNo3v.e2m)ber 2008 CLAIC1(1)