1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
|
<HTML>
<HEAD><TITLE>MB04WD - SLICOT Library Routine Documentation</TITLE>
</HEAD>
<BODY>
<H2><A Name="MB04WD">MB04WD</A></H2>
<H3>
Generating an orthogonal basis spanning an isotropic subspace
</H3>
<A HREF ="#Specification"><B>[Specification]</B></A>
<A HREF ="#Arguments"><B>[Arguments]</B></A>
<A HREF ="#Method"><B>[Method]</B></A>
<A HREF ="#References"><B>[References]</B></A>
<A HREF ="#Comments"><B>[Comments]</B></A>
<A HREF ="#Example"><B>[Example]</B></A>
<P>
<B><FONT SIZE="+1">Purpose</FONT></B>
<PRE>
To generate a matrix Q with orthogonal columns (spanning an
isotropic subspace), which is defined as the first n columns
of a product of symplectic reflectors and Givens rotations,
Q = diag( H(1),H(1) ) G(1) diag( F(1),F(1) )
diag( H(2),H(2) ) G(2) diag( F(2),F(2) )
....
diag( H(k),H(k) ) G(k) diag( F(k),F(k) ).
The matrix Q is returned in terms of its first 2*M rows
[ op( Q1 ) op( Q2 ) ]
Q = [ ].
[ -op( Q2 ) op( Q1 ) ]
Blocked version of the SLICOT Library routine MB04WU.
</PRE>
<A name="Specification"><B><FONT SIZE="+1">Specification</FONT></B></A>
<PRE>
SUBROUTINE MB04WD( TRANQ1, TRANQ2, M, N, K, Q1, LDQ1, Q2, LDQ2,
$ CS, TAU, DWORK, LDWORK, INFO )
C .. Scalar Arguments ..
CHARACTER TRANQ1, TRANQ2
INTEGER INFO, K, LDQ1, LDQ2, LDWORK, M, N
C .. Array Arguments ..
DOUBLE PRECISION CS(*), DWORK(*), Q1(LDQ1,*), Q2(LDQ2,*), TAU(*)
</PRE>
<A name="Arguments"><B><FONT SIZE="+1">Arguments</FONT></B></A>
<P>
<B>Mode Parameters</B>
<PRE>
TRANQ1 CHARACTER*1
Specifies the form of op( Q1 ) as follows:
= 'N': op( Q1 ) = Q1;
= 'T': op( Q1 ) = Q1';
= 'C': op( Q1 ) = Q1'.
TRANQ2 CHARACTER*1
Specifies the form of op( Q2 ) as follows:
= 'N': op( Q2 ) = Q2;
= 'T': op( Q2 ) = Q2';
= 'C': op( Q2 ) = Q2'.
</PRE>
<B>Input/Output Parameters</B>
<PRE>
M (input) INTEGER
The number of rows of the matrices Q1 and Q2. M >= 0.
N (input) INTEGER
The number of columns of the matrices Q1 and Q2.
M >= N >= 0.
K (input) INTEGER
The number of symplectic Givens rotations whose product
partly defines the matrix Q. N >= K >= 0.
Q1 (input/output) DOUBLE PRECISION array, dimension
(LDQ1,N) if TRANQ1 = 'N',
(LDQ1,M) if TRANQ1 = 'T' or TRANQ1 = 'C'
On entry with TRANQ1 = 'N', the leading M-by-K part of
this array must contain in its i-th column the vector
which defines the elementary reflector F(i).
On entry with TRANQ1 = 'T' or TRANQ1 = 'C', the leading
K-by-M part of this array must contain in its i-th row
the vector which defines the elementary reflector F(i).
On exit with TRANQ1 = 'N', the leading M-by-N part of this
array contains the matrix Q1.
On exit with TRANQ1 = 'T' or TRANQ1 = 'C', the leading
N-by-M part of this array contains the matrix Q1'.
LDQ1 INTEGER
The leading dimension of the array Q1.
LDQ1 >= MAX(1,M), if TRANQ1 = 'N';
LDQ1 >= MAX(1,N), if TRANQ1 = 'T' or TRANQ1 = 'C'.
Q2 (input/output) DOUBLE PRECISION array, dimension
(LDQ2,N) if TRANQ2 = 'N',
(LDQ2,M) if TRANQ2 = 'T' or TRANQ2 = 'C'
On entry with TRANQ2 = 'N', the leading M-by-K part of
this array must contain in its i-th column the vector
which defines the elementary reflector H(i) and, on the
diagonal, the scalar factor of H(i).
On entry with TRANQ2 = 'T' or TRANQ2 = 'C', the leading
K-by-M part of this array must contain in its i-th row the
vector which defines the elementary reflector H(i) and, on
the diagonal, the scalar factor of H(i).
On exit with TRANQ2 = 'N', the leading M-by-N part of this
array contains the matrix Q2.
On exit with TRANQ2 = 'T' or TRANQ2 = 'C', the leading
N-by-M part of this array contains the matrix Q2'.
LDQ2 INTEGER
The leading dimension of the array Q2.
LDQ2 >= MAX(1,M), if TRANQ2 = 'N';
LDQ2 >= MAX(1,N), if TRANQ2 = 'T' or TRANQ2 = 'C'.
CS (input) DOUBLE PRECISION array, dimension (2*K)
On entry, the first 2*K elements of this array must
contain the cosines and sines of the symplectic Givens
rotations G(i).
TAU (input) DOUBLE PRECISION array, dimension (K)
On entry, the first K elements of this array must
contain the scalar factors of the elementary reflectors
F(i).
</PRE>
<B>Workspace</B>
<PRE>
DWORK DOUBLE PRECISION array, dimension (LDWORK)
On exit, if INFO = 0, DWORK(1) returns the optimal
value of LDWORK, MAX(M+N,8*N*NB + 15*NB*NB), where NB is
the optimal block size determined by the function UE01MD.
On exit, if INFO = -13, DWORK(1) returns the minimum
value of LDWORK.
LDWORK INTEGER
The length of the array DWORK. LDWORK >= MAX(1,M+N).
If LDWORK = -1, then a workspace query is assumed;
the routine only calculates the optimal size of the
DWORK array, returns this value as the first entry of
the DWORK array, and no error message related to LDWORK
is issued by XERBLA.
</PRE>
<B>Error Indicator</B>
<PRE>
INFO INTEGER
= 0: successful exit;
< 0: if INFO = -i, the i-th argument had an illegal
value.
</PRE>
<A name="References"><B><FONT SIZE="+1">References</FONT></B></A>
<PRE>
[1] Kressner, D.
Block algorithms for orthogonal symplectic factorizations.
BIT, 43 (4), pp. 775-790, 2003.
</PRE>
<A name="Comments"><B><FONT SIZE="+1">Further Comments</FONT></B></A>
<PRE>
None
</PRE>
<A name="Example"><B><FONT SIZE="+1">Example</FONT></B></A>
<P>
<B>Program Text</B>
<PRE>
None
</PRE>
<B>Program Data</B>
<PRE>
None
</PRE>
<B>Program Results</B>
<PRE>
None
</PRE>
<HR>
<A HREF=support.html><B>Return to Supporting Routines index</B></A></BODY>
</HTML>
|
