Debian Codegraph @

src/util/math/zquatev/unblocked.cc

raw
1 //
2 // ZQUATEV: Diagonalization of quaternionic matrices
3 // File : unblocked.cc
4 // Copyright (c) 2013, Toru Shiozaki ([email protected])
5 // All rights reserved.
6 //
7 // Redistribution and use in source and binary forms, with or without
8 // modification, are permitted provided that the following conditions are met:
9 //
10 // 1. Redistributions of source code must retain the above copyright notice, this
11 // list of conditions and the following disclaimer.
12 // 2. Redistributions in binary form must reproduce the above copyright notice,
13 // this list of conditions and the following disclaimer in the documentation
14 // and/or other materials provided with the distribution.
15 //
16 // THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND
17 // ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
18 // WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
19 // DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR
20 // ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
21 // (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
22 // LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND
23 // ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
24 // (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
25 // SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
26 //
27 // The views and conclusions contained in the software and documentation are those
28 // of the authors and should not be interpreted as representing official policies,
29 // either expressed or implied, of the FreeBSD Project.
30 //
31
32 #include "zquatev.h"
33 #include "f77.h"
34 #include <cassert>
35 #include <algorithm>
36
37 using namespace std;
38
39 namespace ts {
40 namespace impl {
41
42 // implementation...
43
44 void unblocked_update(const int n, complex<double>* const D0, complex<double>* const D1, complex<double>* const Q0, complex<double>* const Q1, const int ld,
45 const int norig, complex<double>* const work) {
46
47 complex<double>* tmp = work;
48 complex<double>* vec = work + norig;
49 complex<double>* cvec = work + norig+n;
50
51 for (int k = 0; k != n-1; ++k) {
52 const int len = n-k-1;
53 if (len > 1) {
54 copy_n(D1+k*ld+k+2, len-1, vec+1);
55 complex<double> tau;
56 complex<double> alpha = D1[k*ld+k+1];
57 vec[0] = 1.0;
58 zlarfg_(len, alpha, vec+1, 1, tau);
59 tau = conj(tau);
60
61 for (int i = 0; i != len; ++i) cvec[i] = conj(vec[i]);
62
63 // 00
64 zgemv_("C", len, len+1, 1.0, D0+k+1+(k)*ld, ld, cvec, 1, 0.0, tmp, 1);
65 zaxpy_(len, -conj(tau)*0.5*zdotc_(len, tmp+1, 1, cvec, 1), cvec, 1, tmp+1, 1);
66 zgerc_(len, len+1, -conj(tau), cvec, 1, tmp, 1, D0+k+1+(k)*ld, ld);
67 zgeru_(len+1, len, -tau, tmp, 1, vec, 1, D0+(k+1)*ld+(k), ld);
68
69 // 10
70 zgemv_("N", len+1, len, 1.0, D1+k+(k+1)*ld, ld, cvec, 1, 0.0, tmp, 1);
71 zgeru_(len, len+1, tau, vec, 1, tmp, 1, D1+k+1+(k)*ld, ld);
72 zgeru_(len+1, len, -tau, tmp, 1, vec, 1, D1+(k+1)*ld+(k), ld);
73
74 // 00-2
75 zgemv_("N", norig, len, 1.0, Q0+(k+1)*ld, ld, cvec, 1, 0.0, tmp, 1);
76 zgeru_(norig, len, -tau, tmp, 1, vec, 1, Q0+(k+1)*ld, ld);
77
78 // 10-2
79 zgemv_("N", norig, len, 1.0, Q1+(k+1)*ld, ld, cvec, 1, 0.0, tmp, 1);
80 zgeru_(norig, len, -tau, tmp, 1, vec, 1, Q1+(k+1)*ld, ld);
81 }
82
83 // symplectic Givens rotation to clear out D(k+n, k)
84 double c;
85 complex<double> s, dum;
86 zlartg_(D0[k+1+k*ld], D1[k+1+k*ld], c, s, dum);
87
88 zrot_(len+1, D0+k+1+k*ld, ld, D1+k+1+k*ld, ld, c, s);
89
90 conj_n(D1+(k+1)*ld+k, len+1);
91 zrot_(len+1, D1+(k+1)*ld+k, 1, D0+(k+1)*ld+k, 1, c, s);
92 conj_n(D1+(k+1)*ld+k, len+1);
93
94 conj_n(Q1+(k+1)*ld, norig);
95 zrot_(norig, Q1+(k+1)*ld, 1, Q0+(k+1)*ld, 1, c, s);
96 conj_n(Q1+(k+1)*ld, norig);
97
98 // Householder to fix top half in column k
99 if (len > 1) {
100 copy_n(D0+k*ld+k+2, len-1, vec+1);
101 complex<double> tau;
102 complex<double> alpha = D0[k*ld+k+1];
103 vec[0] = 1.0;
104 zlarfg_(len, alpha, vec+1, 1, tau);
105 tau = conj(tau);
106
107 for (int i = 0; i != len; ++i) cvec[i] = conj(vec[i]);
108
109 // 00
110 zgemv_("C", len, len+1, 1.0, D0+k+1+(k)*ld, ld, vec, 1, 0.0, tmp, 1);
111 zaxpy_(len, -tau*0.5*zdotc_(len, vec, 1, tmp+1, 1), vec, 1, tmp+1, 1);
112 zgerc_(len, len+1, -tau, vec, 1, tmp, 1, D0+k+1+(k)*ld, ld);
113 zgerc_(len+1, len, -conj(tau), tmp, 1, vec, 1, D0+(k+1)*ld+(k), ld);
114
115 // 01-1
116 zgemv_("T", len, len+1, 1.0, D1+k+1+(k)*ld, ld, vec, 1, 0.0, tmp, 1);
117 zgeru_(len, len+1, -conj(tau), cvec, 1, tmp, 1, D1+k+1+(k)*ld, ld);
118 zgerc_(len+1, len, conj(tau), tmp, 1, vec, 1, D1+(k+1)*ld+(k), ld);
119
120 // 00-2
121 zgemv_("N", norig, len, 1.0, Q0+(k+1)*ld, ld, vec, 1, 0.0, tmp, 1);
122 zgerc_(norig, len, -conj(tau), tmp, 1, vec, 1, Q0+(k+1)*ld, ld);
123
124 // 01-2
125 zgemv_("N", norig, len, -1.0, Q1+(k+1)*ld, ld, vec, 1, 0.0, tmp, 1);
126 zgerc_(norig, len, conj(tau), tmp, 1, vec, 1, Q1+(k+1)*ld, ld);
127
128 }
129
130 }
131 }
132
133 }}