Debian Codegraph @

src/util/math/zquatev/transpose.cc

raw
1 //
2 // ZQUATEV: Diagonalization of quaternionic matrices
3 // File: transpose.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 <complex>
33 #include <bagel_config.h>
34
35 using namespace std;
36
37 #ifdef HAVE_MKL_H
38 extern "C" {
39 void mkl_zomatcopy_(const char*, const char*, const int*, const int*, const complex<double>*, const complex<double>*, const int*, complex<double>*, const int*);
40 }
41 #endif
42
43 namespace ts {
44 namespace impl {
45
46 void transpose(const int m, const int n, const complex<double>* h, const int ld, complex<double>* vec, const int ldt) {
47 #ifdef HAVE_MKL_H
48 complex<double> one = 1.0;
49 mkl_zomatcopy_("c", "t", &m, &n, &one, h, &ld, vec, &ldt);
50 #else
51 const int mresidual = m % 10;
52 const int nresidual = n % 10;
53 const int mlim = m - mresidual;
54 const int nlim = n - nresidual;
55 for (int i = 0; i < nlim; i += 10) {
56 for (int j = 0; j < mlim; j += 10) {
57 vec[i + ldt*(j )] = h[j + ld*(i )];
58 vec[i+1 + ldt*(j )] = h[j + ld*(i+1)];
59 vec[i+2 + ldt*(j )] = h[j + ld*(i+2)];
60 vec[i+3 + ldt*(j )] = h[j + ld*(i+3)];
61 vec[i+4 + ldt*(j )] = h[j + ld*(i+4)];
62 vec[i+5 + ldt*(j )] = h[j + ld*(i+5)];
63 vec[i+6 + ldt*(j )] = h[j + ld*(i+6)];
64 vec[i+7 + ldt*(j )] = h[j + ld*(i+7)];
65 vec[i+8 + ldt*(j )] = h[j + ld*(i+8)];
66 vec[i+9 + ldt*(j )] = h[j + ld*(i+9)];
67 vec[i + ldt*(j+1)] = h[j+1 + ld*(i )];
68 vec[i+1 + ldt*(j+1)] = h[j+1 + ld*(i+1)];
69 vec[i+2 + ldt*(j+1)] = h[j+1 + ld*(i+2)];
70 vec[i+3 + ldt*(j+1)] = h[j+1 + ld*(i+3)];
71 vec[i+4 + ldt*(j+1)] = h[j+1 + ld*(i+4)];
72 vec[i+5 + ldt*(j+1)] = h[j+1 + ld*(i+5)];
73 vec[i+6 + ldt*(j+1)] = h[j+1 + ld*(i+6)];
74 vec[i+7 + ldt*(j+1)] = h[j+1 + ld*(i+7)];
75 vec[i+8 + ldt*(j+1)] = h[j+1 + ld*(i+8)];
76 vec[i+9 + ldt*(j+1)] = h[j+1 + ld*(i+9)];
77 vec[i + ldt*(j+2)] = h[j+2 + ld*(i )];
78 vec[i+1 + ldt*(j+2)] = h[j+2 + ld*(i+1)];
79 vec[i+2 + ldt*(j+2)] = h[j+2 + ld*(i+2)];
80 vec[i+3 + ldt*(j+2)] = h[j+2 + ld*(i+3)];
81 vec[i+4 + ldt*(j+2)] = h[j+2 + ld*(i+4)];
82 vec[i+5 + ldt*(j+2)] = h[j+2 + ld*(i+5)];
83 vec[i+6 + ldt*(j+2)] = h[j+2 + ld*(i+6)];
84 vec[i+7 + ldt*(j+2)] = h[j+2 + ld*(i+7)];
85 vec[i+8 + ldt*(j+2)] = h[j+2 + ld*(i+8)];
86 vec[i+9 + ldt*(j+2)] = h[j+2 + ld*(i+9)];
87 vec[i + ldt*(j+3)] = h[j+3 + ld*(i )];
88 vec[i+1 + ldt*(j+3)] = h[j+3 + ld*(i+1)];
89 vec[i+2 + ldt*(j+3)] = h[j+3 + ld*(i+2)];
90 vec[i+3 + ldt*(j+3)] = h[j+3 + ld*(i+3)];
91 vec[i+4 + ldt*(j+3)] = h[j+3 + ld*(i+4)];
92 vec[i+5 + ldt*(j+3)] = h[j+3 + ld*(i+5)];
93 vec[i+6 + ldt*(j+3)] = h[j+3 + ld*(i+6)];
94 vec[i+7 + ldt*(j+3)] = h[j+3 + ld*(i+7)];
95 vec[i+8 + ldt*(j+3)] = h[j+3 + ld*(i+8)];
96 vec[i+9 + ldt*(j+3)] = h[j+3 + ld*(i+9)];
97 vec[i + ldt*(j+4)] = h[j+4 + ld*(i )];
98 vec[i+1 + ldt*(j+4)] = h[j+4 + ld*(i+1)];
99 vec[i+2 + ldt*(j+4)] = h[j+4 + ld*(i+2)];
100 vec[i+3 + ldt*(j+4)] = h[j+4 + ld*(i+3)];
101 vec[i+4 + ldt*(j+4)] = h[j+4 + ld*(i+4)];
102 vec[i+5 + ldt*(j+4)] = h[j+4 + ld*(i+5)];
103 vec[i+6 + ldt*(j+4)] = h[j+4 + ld*(i+6)];
104 vec[i+7 + ldt*(j+4)] = h[j+4 + ld*(i+7)];
105 vec[i+8 + ldt*(j+4)] = h[j+4 + ld*(i+8)];
106 vec[i+9 + ldt*(j+4)] = h[j+4 + ld*(i+9)];
107 vec[i + ldt*(j+5)] = h[j+5 + ld*(i )];
108 vec[i+1 + ldt*(j+5)] = h[j+5 + ld*(i+1)];
109 vec[i+2 + ldt*(j+5)] = h[j+5 + ld*(i+2)];
110 vec[i+3 + ldt*(j+5)] = h[j+5 + ld*(i+3)];
111 vec[i+4 + ldt*(j+5)] = h[j+5 + ld*(i+4)];
112 vec[i+5 + ldt*(j+5)] = h[j+5 + ld*(i+5)];
113 vec[i+6 + ldt*(j+5)] = h[j+5 + ld*(i+6)];
114 vec[i+7 + ldt*(j+5)] = h[j+5 + ld*(i+7)];
115 vec[i+8 + ldt*(j+5)] = h[j+5 + ld*(i+8)];
116 vec[i+9 + ldt*(j+5)] = h[j+5 + ld*(i+9)];
117 vec[i + ldt*(j+6)] = h[j+6 + ld*(i )];
118 vec[i+1 + ldt*(j+6)] = h[j+6 + ld*(i+1)];
119 vec[i+2 + ldt*(j+6)] = h[j+6 + ld*(i+2)];
120 vec[i+3 + ldt*(j+6)] = h[j+6 + ld*(i+3)];
121 vec[i+4 + ldt*(j+6)] = h[j+6 + ld*(i+4)];
122 vec[i+5 + ldt*(j+6)] = h[j+6 + ld*(i+5)];
123 vec[i+6 + ldt*(j+6)] = h[j+6 + ld*(i+6)];
124 vec[i+7 + ldt*(j+6)] = h[j+6 + ld*(i+7)];
125 vec[i+8 + ldt*(j+6)] = h[j+6 + ld*(i+8)];
126 vec[i+9 + ldt*(j+6)] = h[j+6 + ld*(i+9)];
127 vec[i + ldt*(j+7)] = h[j+7 + ld*(i )];
128 vec[i+1 + ldt*(j+7)] = h[j+7 + ld*(i+1)];
129 vec[i+2 + ldt*(j+7)] = h[j+7 + ld*(i+2)];
130 vec[i+3 + ldt*(j+7)] = h[j+7 + ld*(i+3)];
131 vec[i+4 + ldt*(j+7)] = h[j+7 + ld*(i+4)];
132 vec[i+5 + ldt*(j+7)] = h[j+7 + ld*(i+5)];
133 vec[i+6 + ldt*(j+7)] = h[j+7 + ld*(i+6)];
134 vec[i+7 + ldt*(j+7)] = h[j+7 + ld*(i+7)];
135 vec[i+8 + ldt*(j+7)] = h[j+7 + ld*(i+8)];
136 vec[i+9 + ldt*(j+7)] = h[j+7 + ld*(i+9)];
137 vec[i + ldt*(j+8)] = h[j+8 + ld*(i )];
138 vec[i+1 + ldt*(j+8)] = h[j+8 + ld*(i+1)];
139 vec[i+2 + ldt*(j+8)] = h[j+8 + ld*(i+2)];
140 vec[i+3 + ldt*(j+8)] = h[j+8 + ld*(i+3)];
141 vec[i+4 + ldt*(j+8)] = h[j+8 + ld*(i+4)];
142 vec[i+5 + ldt*(j+8)] = h[j+8 + ld*(i+5)];
143 vec[i+6 + ldt*(j+8)] = h[j+8 + ld*(i+6)];
144 vec[i+7 + ldt*(j+8)] = h[j+8 + ld*(i+7)];
145 vec[i+8 + ldt*(j+8)] = h[j+8 + ld*(i+8)];
146 vec[i+9 + ldt*(j+8)] = h[j+8 + ld*(i+9)];
147 vec[i + ldt*(j+9)] = h[j+9 + ld*(i )];
148 vec[i+1 + ldt*(j+9)] = h[j+9 + ld*(i+1)];
149 vec[i+2 + ldt*(j+9)] = h[j+9 + ld*(i+2)];
150 vec[i+3 + ldt*(j+9)] = h[j+9 + ld*(i+3)];
151 vec[i+4 + ldt*(j+9)] = h[j+9 + ld*(i+4)];
152 vec[i+5 + ldt*(j+9)] = h[j+9 + ld*(i+5)];
153 vec[i+6 + ldt*(j+9)] = h[j+9 + ld*(i+6)];
154 vec[i+7 + ldt*(j+9)] = h[j+9 + ld*(i+7)];
155 vec[i+8 + ldt*(j+9)] = h[j+9 + ld*(i+8)];
156 vec[i+9 + ldt*(j+9)] = h[j+9 + ld*(i+9)];
157 }
158 for (int j = mlim; j != m; ++j) {
159 vec[i + ldt*j] = h[j + ld*(i )];
160 vec[i+1 + ldt*j] = h[j + ld*(i+1)];
161 vec[i+2 + ldt*j] = h[j + ld*(i+2)];
162 vec[i+3 + ldt*j] = h[j + ld*(i+3)];
163 vec[i+4 + ldt*j] = h[j + ld*(i+4)];
164 vec[i+5 + ldt*j] = h[j + ld*(i+5)];
165 vec[i+6 + ldt*j] = h[j + ld*(i+6)];
166 vec[i+7 + ldt*j] = h[j + ld*(i+7)];
167 vec[i+8 + ldt*j] = h[j + ld*(i+8)];
168 vec[i+9 + ldt*j] = h[j + ld*(i+9)];
169 }
170 }
171 for (int i = nlim; i != n; ++i) {
172 for (int j = 0; j < mlim; j += 10) {
173 vec[i + ldt*(j )] = h[j + ld*(i )];
174 vec[i + ldt*(j+1)] = h[j+1 + ld*(i )];
175 vec[i + ldt*(j+2)] = h[j+2 + ld*(i )];
176 vec[i + ldt*(j+3)] = h[j+3 + ld*(i )];
177 vec[i + ldt*(j+4)] = h[j+4 + ld*(i )];
178 vec[i + ldt*(j+5)] = h[j+5 + ld*(i )];
179 vec[i + ldt*(j+6)] = h[j+6 + ld*(i )];
180 vec[i + ldt*(j+7)] = h[j+7 + ld*(i )];
181 vec[i + ldt*(j+8)] = h[j+8 + ld*(i )];
182 vec[i + ldt*(j+9)] = h[j+9 + ld*(i )];
183 }
184 for (int j = mlim; j != m; ++j) {
185 vec[i + ldt*(j )] = h[j + ld*(i )];
186 }
187 }
188 #endif
189 }
190
191
192 void transpose_conj(const int m, const int n, const complex<double>* h, const int ld, complex<double>* vec, const int ldt) {
193 #ifdef HAVE_MKL_H
194 complex<double> one = 1.0;
195 mkl_zomatcopy_("c", "c", &m, &n, &one, h, &ld, vec, &ldt);
196 #else
197 const int mresidual = m % 10;
198 const int nresidual = n % 10;
199 const int mlim = m - mresidual;
200 const int nlim = n - nresidual;
201 for (int i = 0; i < nlim; i += 10) {
202 for (int j = 0; j < mlim; j += 10) {
203 vec[i + ldt*(j )] = conj(h[j + ld*(i )]);
204 vec[i+1 + ldt*(j )] = conj(h[j + ld*(i+1)]);
205 vec[i+2 + ldt*(j )] = conj(h[j + ld*(i+2)]);
206 vec[i+3 + ldt*(j )] = conj(h[j + ld*(i+3)]);
207 vec[i+4 + ldt*(j )] = conj(h[j + ld*(i+4)]);
208 vec[i+5 + ldt*(j )] = conj(h[j + ld*(i+5)]);
209 vec[i+6 + ldt*(j )] = conj(h[j + ld*(i+6)]);
210 vec[i+7 + ldt*(j )] = conj(h[j + ld*(i+7)]);
211 vec[i+8 + ldt*(j )] = conj(h[j + ld*(i+8)]);
212 vec[i+9 + ldt*(j )] = conj(h[j + ld*(i+9)]);
213 vec[i + ldt*(j+1)] = conj(h[j+1 + ld*(i )]);
214 vec[i+1 + ldt*(j+1)] = conj(h[j+1 + ld*(i+1)]);
215 vec[i+2 + ldt*(j+1)] = conj(h[j+1 + ld*(i+2)]);
216 vec[i+3 + ldt*(j+1)] = conj(h[j+1 + ld*(i+3)]);
217 vec[i+4 + ldt*(j+1)] = conj(h[j+1 + ld*(i+4)]);
218 vec[i+5 + ldt*(j+1)] = conj(h[j+1 + ld*(i+5)]);
219 vec[i+6 + ldt*(j+1)] = conj(h[j+1 + ld*(i+6)]);
220 vec[i+7 + ldt*(j+1)] = conj(h[j+1 + ld*(i+7)]);
221 vec[i+8 + ldt*(j+1)] = conj(h[j+1 + ld*(i+8)]);
222 vec[i+9 + ldt*(j+1)] = conj(h[j+1 + ld*(i+9)]);
223 vec[i + ldt*(j+2)] = conj(h[j+2 + ld*(i )]);
224 vec[i+1 + ldt*(j+2)] = conj(h[j+2 + ld*(i+1)]);
225 vec[i+2 + ldt*(j+2)] = conj(h[j+2 + ld*(i+2)]);
226 vec[i+3 + ldt*(j+2)] = conj(h[j+2 + ld*(i+3)]);
227 vec[i+4 + ldt*(j+2)] = conj(h[j+2 + ld*(i+4)]);
228 vec[i+5 + ldt*(j+2)] = conj(h[j+2 + ld*(i+5)]);
229 vec[i+6 + ldt*(j+2)] = conj(h[j+2 + ld*(i+6)]);
230 vec[i+7 + ldt*(j+2)] = conj(h[j+2 + ld*(i+7)]);
231 vec[i+8 + ldt*(j+2)] = conj(h[j+2 + ld*(i+8)]);
232 vec[i+9 + ldt*(j+2)] = conj(h[j+2 + ld*(i+9)]);
233 vec[i + ldt*(j+3)] = conj(h[j+3 + ld*(i )]);
234 vec[i+1 + ldt*(j+3)] = conj(h[j+3 + ld*(i+1)]);
235 vec[i+2 + ldt*(j+3)] = conj(h[j+3 + ld*(i+2)]);
236 vec[i+3 + ldt*(j+3)] = conj(h[j+3 + ld*(i+3)]);
237 vec[i+4 + ldt*(j+3)] = conj(h[j+3 + ld*(i+4)]);
238 vec[i+5 + ldt*(j+3)] = conj(h[j+3 + ld*(i+5)]);
239 vec[i+6 + ldt*(j+3)] = conj(h[j+3 + ld*(i+6)]);
240 vec[i+7 + ldt*(j+3)] = conj(h[j+3 + ld*(i+7)]);
241 vec[i+8 + ldt*(j+3)] = conj(h[j+3 + ld*(i+8)]);
242 vec[i+9 + ldt*(j+3)] = conj(h[j+3 + ld*(i+9)]);
243 vec[i + ldt*(j+4)] = conj(h[j+4 + ld*(i )]);
244 vec[i+1 + ldt*(j+4)] = conj(h[j+4 + ld*(i+1)]);
245 vec[i+2 + ldt*(j+4)] = conj(h[j+4 + ld*(i+2)]);
246 vec[i+3 + ldt*(j+4)] = conj(h[j+4 + ld*(i+3)]);
247 vec[i+4 + ldt*(j+4)] = conj(h[j+4 + ld*(i+4)]);
248 vec[i+5 + ldt*(j+4)] = conj(h[j+4 + ld*(i+5)]);
249 vec[i+6 + ldt*(j+4)] = conj(h[j+4 + ld*(i+6)]);
250 vec[i+7 + ldt*(j+4)] = conj(h[j+4 + ld*(i+7)]);
251 vec[i+8 + ldt*(j+4)] = conj(h[j+4 + ld*(i+8)]);
252 vec[i+9 + ldt*(j+4)] = conj(h[j+4 + ld*(i+9)]);
253 vec[i + ldt*(j+5)] = conj(h[j+5 + ld*(i )]);
254 vec[i+1 + ldt*(j+5)] = conj(h[j+5 + ld*(i+1)]);
255 vec[i+2 + ldt*(j+5)] = conj(h[j+5 + ld*(i+2)]);
256 vec[i+3 + ldt*(j+5)] = conj(h[j+5 + ld*(i+3)]);
257 vec[i+4 + ldt*(j+5)] = conj(h[j+5 + ld*(i+4)]);
258 vec[i+5 + ldt*(j+5)] = conj(h[j+5 + ld*(i+5)]);
259 vec[i+6 + ldt*(j+5)] = conj(h[j+5 + ld*(i+6)]);
260 vec[i+7 + ldt*(j+5)] = conj(h[j+5 + ld*(i+7)]);
261 vec[i+8 + ldt*(j+5)] = conj(h[j+5 + ld*(i+8)]);
262 vec[i+9 + ldt*(j+5)] = conj(h[j+5 + ld*(i+9)]);
263 vec[i + ldt*(j+6)] = conj(h[j+6 + ld*(i )]);
264 vec[i+1 + ldt*(j+6)] = conj(h[j+6 + ld*(i+1)]);
265 vec[i+2 + ldt*(j+6)] = conj(h[j+6 + ld*(i+2)]);
266 vec[i+3 + ldt*(j+6)] = conj(h[j+6 + ld*(i+3)]);
267 vec[i+4 + ldt*(j+6)] = conj(h[j+6 + ld*(i+4)]);
268 vec[i+5 + ldt*(j+6)] = conj(h[j+6 + ld*(i+5)]);
269 vec[i+6 + ldt*(j+6)] = conj(h[j+6 + ld*(i+6)]);
270 vec[i+7 + ldt*(j+6)] = conj(h[j+6 + ld*(i+7)]);
271 vec[i+8 + ldt*(j+6)] = conj(h[j+6 + ld*(i+8)]);
272 vec[i+9 + ldt*(j+6)] = conj(h[j+6 + ld*(i+9)]);
273 vec[i + ldt*(j+7)] = conj(h[j+7 + ld*(i )]);
274 vec[i+1 + ldt*(j+7)] = conj(h[j+7 + ld*(i+1)]);
275 vec[i+2 + ldt*(j+7)] = conj(h[j+7 + ld*(i+2)]);
276 vec[i+3 + ldt*(j+7)] = conj(h[j+7 + ld*(i+3)]);
277 vec[i+4 + ldt*(j+7)] = conj(h[j+7 + ld*(i+4)]);
278 vec[i+5 + ldt*(j+7)] = conj(h[j+7 + ld*(i+5)]);
279 vec[i+6 + ldt*(j+7)] = conj(h[j+7 + ld*(i+6)]);
280 vec[i+7 + ldt*(j+7)] = conj(h[j+7 + ld*(i+7)]);
281 vec[i+8 + ldt*(j+7)] = conj(h[j+7 + ld*(i+8)]);
282 vec[i+9 + ldt*(j+7)] = conj(h[j+7 + ld*(i+9)]);
283 vec[i + ldt*(j+8)] = conj(h[j+8 + ld*(i )]);
284 vec[i+1 + ldt*(j+8)] = conj(h[j+8 + ld*(i+1)]);
285 vec[i+2 + ldt*(j+8)] = conj(h[j+8 + ld*(i+2)]);
286 vec[i+3 + ldt*(j+8)] = conj(h[j+8 + ld*(i+3)]);
287 vec[i+4 + ldt*(j+8)] = conj(h[j+8 + ld*(i+4)]);
288 vec[i+5 + ldt*(j+8)] = conj(h[j+8 + ld*(i+5)]);
289 vec[i+6 + ldt*(j+8)] = conj(h[j+8 + ld*(i+6)]);
290 vec[i+7 + ldt*(j+8)] = conj(h[j+8 + ld*(i+7)]);
291 vec[i+8 + ldt*(j+8)] = conj(h[j+8 + ld*(i+8)]);
292 vec[i+9 + ldt*(j+8)] = conj(h[j+8 + ld*(i+9)]);
293 vec[i + ldt*(j+9)] = conj(h[j+9 + ld*(i )]);
294 vec[i+1 + ldt*(j+9)] = conj(h[j+9 + ld*(i+1)]);
295 vec[i+2 + ldt*(j+9)] = conj(h[j+9 + ld*(i+2)]);
296 vec[i+3 + ldt*(j+9)] = conj(h[j+9 + ld*(i+3)]);
297 vec[i+4 + ldt*(j+9)] = conj(h[j+9 + ld*(i+4)]);
298 vec[i+5 + ldt*(j+9)] = conj(h[j+9 + ld*(i+5)]);
299 vec[i+6 + ldt*(j+9)] = conj(h[j+9 + ld*(i+6)]);
300 vec[i+7 + ldt*(j+9)] = conj(h[j+9 + ld*(i+7)]);
301 vec[i+8 + ldt*(j+9)] = conj(h[j+9 + ld*(i+8)]);
302 vec[i+9 + ldt*(j+9)] = conj(h[j+9 + ld*(i+9)]);
303 }
304 for (int j = mlim; j != m; ++j) {
305 vec[i + ldt*j] = conj(h[j + ld*(i )]);
306 vec[i+1 + ldt*j] = conj(h[j + ld*(i+1)]);
307 vec[i+2 + ldt*j] = conj(h[j + ld*(i+2)]);
308 vec[i+3 + ldt*j] = conj(h[j + ld*(i+3)]);
309 vec[i+4 + ldt*j] = conj(h[j + ld*(i+4)]);
310 vec[i+5 + ldt*j] = conj(h[j + ld*(i+5)]);
311 vec[i+6 + ldt*j] = conj(h[j + ld*(i+6)]);
312 vec[i+7 + ldt*j] = conj(h[j + ld*(i+7)]);
313 vec[i+8 + ldt*j] = conj(h[j + ld*(i+8)]);
314 vec[i+9 + ldt*j] = conj(h[j + ld*(i+9)]);
315 }
316 }
317 for (int i = nlim; i != n; ++i) {
318 for (int j = 0; j < mlim; j += 10) {
319 vec[i + ldt*(j )] = conj(h[j + ld*(i )]);
320 vec[i + ldt*(j+1)] = conj(h[j+1 + ld*(i )]);
321 vec[i + ldt*(j+2)] = conj(h[j+2 + ld*(i )]);
322 vec[i + ldt*(j+3)] = conj(h[j+3 + ld*(i )]);
323 vec[i + ldt*(j+4)] = conj(h[j+4 + ld*(i )]);
324 vec[i + ldt*(j+5)] = conj(h[j+5 + ld*(i )]);
325 vec[i + ldt*(j+6)] = conj(h[j+6 + ld*(i )]);
326 vec[i + ldt*(j+7)] = conj(h[j+7 + ld*(i )]);
327 vec[i + ldt*(j+8)] = conj(h[j+8 + ld*(i )]);
328 vec[i + ldt*(j+9)] = conj(h[j+9 + ld*(i )]);
329 }
330 for (int j = mlim; j != m; ++j) {
331 vec[i + ldt*(j )] = conj(h[j + ld*(i )]);
332 }
333 }
334 #endif
335 }
336
337 }}