1SLAIC1(1) LAPACK auxiliary routine (version 3.2) SLAIC1(1)
2
3
4
6 SLAIC1 - applies one step of incremental condition estimation in its
7 simplest version
8
10 SUBROUTINE SLAIC1( JOB, J, X, SEST, W, GAMMA, SESTPR, S, C )
11
12 INTEGER J, JOB
13
14 REAL C, GAMMA, S, SEST, SESTPR
15
16 REAL W( J ), X( J )
17
19 SLAIC1 applies one step of incremental condition estimation in its sim‐
20 plest version: Let x, twonorm(x) = 1, be an approximate singular vector
21 of an j-by-j lower triangular matrix L, such that
22 twonorm(L*x) = sest
23 Then SLAIC1 computes sestpr, s, c such that
24 the vector
25 [ s*x ]
26 xhat = [ c ]
27 is an approximate singular vector of
28 [ L 0 ]
29 Lhat = [ w' gamma ]
30 in the sense that
31 twonorm(Lhat*xhat) = sestpr.
32 Depending on JOB, an estimate for the largest or smallest singular
33 value is computed.
34 Note that [s c]' and sestpr**2 is an eigenpair of the system
35 diag(sest*sest, 0) + [alpha gamma] * [ alpha ]
36 [ gamma ]
37 where alpha = x'*w.
38
40 JOB (input) INTEGER
41 = 1: an estimate for the largest singular value is computed.
42 = 2: an estimate for the smallest singular value is computed.
43
44 J (input) INTEGER
45 Length of X and W
46
47 X (input) REAL array, dimension (J)
48 The j-vector x.
49
50 SEST (input) REAL
51 Estimated singular value of j by j matrix L
52
53 W (input) REAL array, dimension (J)
54 The j-vector w.
55
56 GAMMA (input) REAL
57 The diagonal element gamma.
58
59 SESTPR (output) REAL
60 Estimated singular value of (j+1) by (j+1) matrix Lhat.
61
62 S (output) REAL
63 Sine needed in forming xhat.
64
65 C (output) REAL
66 Cosine needed in forming xhat.
67
68
69
70 LAPACK auxiliary routine (versionNo3v.e2m)ber 2008 SLAIC1(1)