line |
stmt |
bran |
cond |
sub |
pod |
time |
code |
1
|
|
|
|
|
|
|
#include |
2
|
|
|
|
|
|
|
/* |
3
|
|
|
|
|
|
|
Copyright © 2002, University of Tennessee Research Foundation. |
4
|
|
|
|
|
|
|
All rights reserved. |
5
|
|
|
|
|
|
|
|
6
|
|
|
|
|
|
|
Redistribution and use in source and binary forms, with or without |
7
|
|
|
|
|
|
|
modification, are permitted provided that the following conditions are met: |
8
|
|
|
|
|
|
|
|
9
|
|
|
|
|
|
|
* Redistributions of source code must retain the above copyright notice, this |
10
|
|
|
|
|
|
|
list of conditions and the following disclaimer. |
11
|
|
|
|
|
|
|
|
12
|
|
|
|
|
|
|
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
|
|
|
|
|
|
|
* Neither the name of the University of Tennessee nor the names of its |
17
|
|
|
|
|
|
|
contributors may be used to endorse or promote products derived from this |
18
|
|
|
|
|
|
|
software without specific prior written permission. |
19
|
|
|
|
|
|
|
|
20
|
|
|
|
|
|
|
THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" |
21
|
|
|
|
|
|
|
AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE |
22
|
|
|
|
|
|
|
IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE |
23
|
|
|
|
|
|
|
ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE |
24
|
|
|
|
|
|
|
LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR |
25
|
|
|
|
|
|
|
CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF |
26
|
|
|
|
|
|
|
SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS |
27
|
|
|
|
|
|
|
INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN |
28
|
|
|
|
|
|
|
CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) |
29
|
|
|
|
|
|
|
ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE |
30
|
|
|
|
|
|
|
POSSIBILITY OF SUCH DAMAGE. |
31
|
|
|
|
|
|
|
*/ |
32
|
|
|
|
|
|
|
/* |
33
|
|
|
|
|
|
|
* Modified by moocow for PDL::SVDLIBC distribution |
34
|
|
|
|
|
|
|
*/ |
35
|
|
|
|
|
|
|
|
36
|
|
|
|
|
|
|
#include |
37
|
|
|
|
|
|
|
#include |
38
|
|
|
|
|
|
|
#include |
39
|
|
|
|
|
|
|
#include |
40
|
|
|
|
|
|
|
#include |
41
|
|
|
|
|
|
|
#include |
42
|
|
|
|
|
|
|
#include |
43
|
|
|
|
|
|
|
#include |
44
|
|
|
|
|
|
|
#if 0 |
45
|
|
|
|
|
|
|
/* Mon, Tue, 24 Nov 2015 09:42:30 +0100 moocow |
46
|
|
|
|
|
|
|
* + disable netinet/in.h for win32 (why is it even here anyways?) |
47
|
|
|
|
|
|
|
* - probably for ntohl() and htonl(); gnu libc docs say: |
48
|
|
|
|
|
|
|
* uint32_t ntohl (uint32_t NETLONG) |
49
|
|
|
|
|
|
|
* converts the 'uint32_t' integer NETLONG from network byte order to host byte order; |
50
|
|
|
|
|
|
|
* This is used for IPv4 Internet addresses. |
51
|
|
|
|
|
|
|
* * this is only used by svd_(read|write)Bin(Int|Float)() functions, which |
52
|
|
|
|
|
|
|
* we don't need here; replacing calls with dummy macros |
53
|
|
|
|
|
|
|
*/ |
54
|
|
|
|
|
|
|
#include |
55
|
|
|
|
|
|
|
#else |
56
|
|
|
|
|
|
|
#define svdlibc_ntohl(x) x |
57
|
|
|
|
|
|
|
#define svdlibc_htonl(x) x |
58
|
|
|
|
|
|
|
#endif |
59
|
|
|
|
|
|
|
#include "svdlib.h" |
60
|
|
|
|
|
|
|
#include "svdutil.h" |
61
|
|
|
|
|
|
|
|
62
|
|
|
|
|
|
|
#define BUNZIP2 "bzip2 -d" |
63
|
|
|
|
|
|
|
#define BZIP2 "bzip2 -1" |
64
|
|
|
|
|
|
|
#define UNZIP "gzip -d" |
65
|
|
|
|
|
|
|
#define ZIP "gzip -1" |
66
|
|
|
|
|
|
|
#define COMPRESS "compress" |
67
|
|
|
|
|
|
|
|
68
|
|
|
|
|
|
|
#define MAX_FILENAME 512 |
69
|
|
|
|
|
|
|
#define MAX_PIPES 64 |
70
|
|
|
|
|
|
|
static FILE *Pipe[MAX_PIPES]; |
71
|
|
|
|
|
|
|
static int numPipes = 0; |
72
|
|
|
|
|
|
|
|
73
|
6
|
|
|
|
|
|
__SVDLIBC_LONG *svd_longArray(__SVDLIBC_LONG size, char empty, char *name) { |
74
|
|
|
|
|
|
|
__SVDLIBC_LONG *a; |
75
|
6
|
100
|
|
|
|
|
if (empty) a = (__SVDLIBC_LONG *) calloc(size, sizeof(__SVDLIBC_LONG)); |
76
|
3
|
|
|
|
|
|
else a = (__SVDLIBC_LONG *) malloc(size * sizeof(__SVDLIBC_LONG)); |
77
|
6
|
50
|
|
|
|
|
if (!a) { |
78
|
0
|
|
|
|
|
|
perror(name); |
79
|
|
|
|
|
|
|
/* exit(errno); */ |
80
|
|
|
|
|
|
|
} |
81
|
6
|
|
|
|
|
|
return a; |
82
|
|
|
|
|
|
|
} |
83
|
|
|
|
|
|
|
|
84
|
243
|
|
|
|
|
|
double *svd_doubleArray(__SVDLIBC_LONG size, char empty, char *name) { |
85
|
|
|
|
|
|
|
double *a; |
86
|
243
|
100
|
|
|
|
|
if (empty) a = (double *) calloc(size, sizeof(double)); |
87
|
193
|
|
|
|
|
|
else a = (double *) malloc(size * sizeof(double)); |
88
|
243
|
50
|
|
|
|
|
if (!a) { |
89
|
0
|
|
|
|
|
|
perror(name); |
90
|
|
|
|
|
|
|
/* exit(errno); */ |
91
|
|
|
|
|
|
|
} |
92
|
243
|
|
|
|
|
|
return a; |
93
|
|
|
|
|
|
|
} |
94
|
|
|
|
|
|
|
|
95
|
0
|
|
|
|
|
|
void svd_beep(void) { |
96
|
0
|
|
|
|
|
|
fputc('\a', stderr); |
97
|
0
|
|
|
|
|
|
fflush(stderr); |
98
|
0
|
|
|
|
|
|
} |
99
|
|
|
|
|
|
|
|
100
|
0
|
|
|
|
|
|
void svd_debug(char *fmt, ...) { |
101
|
|
|
|
|
|
|
va_list args; |
102
|
0
|
|
|
|
|
|
va_start(args, fmt); |
103
|
0
|
|
|
|
|
|
vfprintf(stderr, fmt, args); |
104
|
0
|
|
|
|
|
|
va_end(args); |
105
|
0
|
|
|
|
|
|
} |
106
|
|
|
|
|
|
|
|
107
|
0
|
|
|
|
|
|
void svd_error(char *fmt, ...) { |
108
|
|
|
|
|
|
|
va_list args; |
109
|
0
|
|
|
|
|
|
va_start(args, fmt); |
110
|
0
|
|
|
|
|
|
svd_beep(); |
111
|
0
|
|
|
|
|
|
fprintf(stderr, "ERROR: "); |
112
|
0
|
|
|
|
|
|
vfprintf(stderr, fmt, args); |
113
|
0
|
|
|
|
|
|
fprintf(stderr, "\n"); |
114
|
0
|
|
|
|
|
|
va_end(args); |
115
|
0
|
|
|
|
|
|
} |
116
|
|
|
|
|
|
|
|
117
|
0
|
|
|
|
|
|
void svd_fatalError(char *fmt, ...) { |
118
|
|
|
|
|
|
|
va_list args; |
119
|
0
|
|
|
|
|
|
va_start(args, fmt); |
120
|
0
|
|
|
|
|
|
svd_beep(); |
121
|
0
|
|
|
|
|
|
fprintf(stderr, "ERROR: "); |
122
|
0
|
|
|
|
|
|
vfprintf(stderr, fmt, args); |
123
|
0
|
|
|
|
|
|
fprintf(stderr, "\a\n"); |
124
|
0
|
|
|
|
|
|
va_end(args); |
125
|
0
|
|
|
|
|
|
exit(1); |
126
|
|
|
|
|
|
|
} |
127
|
|
|
|
|
|
|
|
128
|
0
|
|
|
|
|
|
static void registerPipe(FILE *p) { |
129
|
0
|
0
|
|
|
|
|
if (numPipes >= MAX_PIPES) svd_error("Too many pipes open"); |
130
|
0
|
|
|
|
|
|
Pipe[numPipes++] = p; |
131
|
0
|
|
|
|
|
|
} |
132
|
|
|
|
|
|
|
|
133
|
0
|
|
|
|
|
|
static char isPipe(FILE *p) { |
134
|
|
|
|
|
|
|
int i; |
135
|
0
|
0
|
|
|
|
|
for (i = 0; i < numPipes && Pipe[i] != p; i++); |
|
|
0
|
|
|
|
|
|
136
|
0
|
0
|
|
|
|
|
if (i == numPipes) return FALSE; |
137
|
0
|
|
|
|
|
|
Pipe[i] = Pipe[--numPipes]; |
138
|
0
|
|
|
|
|
|
return TRUE; |
139
|
|
|
|
|
|
|
} |
140
|
|
|
|
|
|
|
|
141
|
0
|
|
|
|
|
|
static FILE *openPipe(char *pipeName, char *mode) { |
142
|
|
|
|
|
|
|
FILE *pipe; |
143
|
0
|
|
|
|
|
|
fflush(stdout); |
144
|
0
|
0
|
|
|
|
|
if ((pipe = popen(pipeName, mode))) registerPipe(pipe); |
145
|
0
|
|
|
|
|
|
return pipe; |
146
|
|
|
|
|
|
|
} |
147
|
|
|
|
|
|
|
|
148
|
0
|
|
|
|
|
|
static FILE *readZippedFile(char *command, char *fileName) { |
149
|
|
|
|
|
|
|
char buf[MAX_FILENAME]; |
150
|
0
|
|
|
|
|
|
sprintf(buf, "%s < %s 2>/dev/null", command, fileName); |
151
|
0
|
|
|
|
|
|
return openPipe(buf, "r"); |
152
|
|
|
|
|
|
|
} |
153
|
|
|
|
|
|
|
|
154
|
0
|
|
|
|
|
|
FILE *svd_fatalReadFile(char *filename) { |
155
|
|
|
|
|
|
|
FILE *file; |
156
|
0
|
0
|
|
|
|
|
if (!(file = svd_readFile(filename))) |
157
|
0
|
|
|
|
|
|
svd_fatalError("couldn't read the file %s", filename); |
158
|
0
|
|
|
|
|
|
return file; |
159
|
|
|
|
|
|
|
} |
160
|
|
|
|
|
|
|
|
161
|
0
|
|
|
|
|
|
static int stringEndsIn(char *s, char *t) { |
162
|
0
|
|
|
|
|
|
int ls = strlen(s); |
163
|
0
|
|
|
|
|
|
int lt = strlen(t); |
164
|
0
|
0
|
|
|
|
|
if (ls < lt) return FALSE; |
165
|
0
|
|
|
|
|
|
return (strcmp(s + ls - lt, t)) ? FALSE : TRUE; |
166
|
|
|
|
|
|
|
} |
167
|
|
|
|
|
|
|
|
168
|
|
|
|
|
|
|
/* Will silently return NULL if file couldn't be opened */ |
169
|
0
|
|
|
|
|
|
FILE *svd_readFile(char *fileName) { |
170
|
|
|
|
|
|
|
char fileBuf[MAX_FILENAME]; |
171
|
|
|
|
|
|
|
struct stat statbuf; |
172
|
|
|
|
|
|
|
|
173
|
|
|
|
|
|
|
/* Special file name */ |
174
|
0
|
0
|
|
|
|
|
if (!strcmp(fileName, "-")) |
175
|
0
|
|
|
|
|
|
return stdin; |
176
|
|
|
|
|
|
|
|
177
|
|
|
|
|
|
|
/* If it is a pipe */ |
178
|
0
|
0
|
|
|
|
|
if (fileName[0] == '|') |
179
|
0
|
|
|
|
|
|
return openPipe(fileName + 1, "r"); |
180
|
|
|
|
|
|
|
|
181
|
|
|
|
|
|
|
/* Check if already ends in .gz or .Z and assume compressed */ |
182
|
0
|
0
|
|
|
|
|
if (stringEndsIn(fileName, ".gz") || stringEndsIn(fileName, ".Z")) { |
|
|
0
|
|
|
|
|
|
183
|
0
|
0
|
|
|
|
|
if (!stat(fileName, &statbuf)) |
184
|
0
|
|
|
|
|
|
return readZippedFile(UNZIP, fileName); |
185
|
0
|
|
|
|
|
|
return NULL; |
186
|
|
|
|
|
|
|
} |
187
|
|
|
|
|
|
|
/* Check if already ends in .bz or .bz2 and assume compressed */ |
188
|
0
|
0
|
|
|
|
|
if (stringEndsIn(fileName, ".bz") || stringEndsIn(fileName, ".bz2")) { |
|
|
0
|
|
|
|
|
|
189
|
0
|
0
|
|
|
|
|
if (!stat(fileName, &statbuf)) |
190
|
0
|
|
|
|
|
|
return readZippedFile(BUNZIP2, fileName); |
191
|
0
|
|
|
|
|
|
return NULL; |
192
|
|
|
|
|
|
|
} |
193
|
|
|
|
|
|
|
/* Try just opening normally */ |
194
|
0
|
0
|
|
|
|
|
if (!stat(fileName, &statbuf)) |
195
|
0
|
|
|
|
|
|
return fopen(fileName, "r"); |
196
|
|
|
|
|
|
|
/* Try adding .gz */ |
197
|
0
|
|
|
|
|
|
sprintf(fileBuf, "%s.gz", fileName); |
198
|
0
|
0
|
|
|
|
|
if (!stat(fileBuf, &statbuf)) |
199
|
0
|
|
|
|
|
|
return readZippedFile(UNZIP, fileBuf); |
200
|
|
|
|
|
|
|
/* Try adding .Z */ |
201
|
0
|
|
|
|
|
|
sprintf(fileBuf, "%s.Z", fileName); |
202
|
0
|
0
|
|
|
|
|
if (!stat(fileBuf, &statbuf)) |
203
|
0
|
|
|
|
|
|
return readZippedFile(UNZIP, fileBuf); |
204
|
|
|
|
|
|
|
/* Try adding .bz2 */ |
205
|
0
|
|
|
|
|
|
sprintf(fileBuf, "%s.bz2", fileName); |
206
|
0
|
0
|
|
|
|
|
if (!stat(fileBuf, &statbuf)) |
207
|
0
|
|
|
|
|
|
return readZippedFile(BUNZIP2, fileBuf); |
208
|
|
|
|
|
|
|
/* Try adding .bz */ |
209
|
0
|
|
|
|
|
|
sprintf(fileBuf, "%s.bz", fileName); |
210
|
0
|
0
|
|
|
|
|
if (!stat(fileBuf, &statbuf)) |
211
|
0
|
|
|
|
|
|
return readZippedFile(BUNZIP2, fileBuf); |
212
|
|
|
|
|
|
|
|
213
|
0
|
|
|
|
|
|
return NULL; |
214
|
|
|
|
|
|
|
} |
215
|
|
|
|
|
|
|
|
216
|
0
|
|
|
|
|
|
static FILE *writeZippedFile(char *fileName, char append) { |
217
|
|
|
|
|
|
|
char buf[MAX_FILENAME]; |
218
|
0
|
0
|
|
|
|
|
const char *op = (append) ? ">>" : ">"; |
219
|
0
|
0
|
|
|
|
|
if (stringEndsIn(fileName, ".bz2") || stringEndsIn(fileName, ".bz")) |
|
|
0
|
|
|
|
|
|
220
|
0
|
|
|
|
|
|
sprintf(buf, "%s %s \"%s\"", BZIP2, op, fileName); |
221
|
0
|
0
|
|
|
|
|
else if (stringEndsIn(fileName, ".Z")) |
222
|
0
|
|
|
|
|
|
sprintf(buf, "%s %s \"%s\"", COMPRESS, op, fileName); |
223
|
|
|
|
|
|
|
else |
224
|
0
|
|
|
|
|
|
sprintf(buf, "%s %s \"%s\"", ZIP, op, fileName); |
225
|
0
|
|
|
|
|
|
return openPipe(buf, "w"); |
226
|
|
|
|
|
|
|
} |
227
|
|
|
|
|
|
|
|
228
|
0
|
|
|
|
|
|
FILE *svd_writeFile(char *fileName, char append) { |
229
|
|
|
|
|
|
|
/* Special file name */ |
230
|
0
|
0
|
|
|
|
|
if (!strcmp(fileName, "-")) |
231
|
0
|
|
|
|
|
|
return stdout; |
232
|
|
|
|
|
|
|
|
233
|
|
|
|
|
|
|
/* If it is a pipe */ |
234
|
0
|
0
|
|
|
|
|
if (fileName[0] == '|') |
235
|
0
|
|
|
|
|
|
return openPipe(fileName + 1, "w"); |
236
|
|
|
|
|
|
|
|
237
|
|
|
|
|
|
|
/* Check if ends in .gz, .Z, .bz, .bz2 */ |
238
|
0
|
0
|
|
|
|
|
if (stringEndsIn(fileName, ".gz") || stringEndsIn(fileName, ".Z") || |
239
|
0
|
0
|
|
|
|
|
stringEndsIn(fileName, ".bz") || stringEndsIn(fileName, ".bz2")) |
240
|
0
|
|
|
|
|
|
return writeZippedFile(fileName, append); |
241
|
0
|
0
|
|
|
|
|
return (append) ? fopen(fileName, "a") : fopen(fileName, "w"); |
242
|
|
|
|
|
|
|
} |
243
|
|
|
|
|
|
|
|
244
|
|
|
|
|
|
|
/* Could be a file or a stream. */ |
245
|
0
|
|
|
|
|
|
void svd_closeFile(FILE *file) { |
246
|
0
|
0
|
|
|
|
|
if (file == stdin || file == stdout) return; |
|
|
0
|
|
|
|
|
|
247
|
0
|
0
|
|
|
|
|
if (isPipe(file)) pclose(file); |
248
|
0
|
|
|
|
|
|
else fclose(file); |
249
|
|
|
|
|
|
|
} |
250
|
|
|
|
|
|
|
|
251
|
|
|
|
|
|
|
|
252
|
0
|
|
|
|
|
|
char svd_readBinInt(FILE *file, int *val) { |
253
|
|
|
|
|
|
|
int x; |
254
|
0
|
0
|
|
|
|
|
if (fread(&x, sizeof(int), 1, file) == 1) { |
255
|
0
|
|
|
|
|
|
*val = svdlibc_ntohl(x); //-- moo: disable ntohl() for PDL::SVDLIBC |
256
|
0
|
|
|
|
|
|
return FALSE; |
257
|
|
|
|
|
|
|
} |
258
|
0
|
|
|
|
|
|
return TRUE; |
259
|
|
|
|
|
|
|
} |
260
|
|
|
|
|
|
|
|
261
|
|
|
|
|
|
|
/* This reads a float in network order and converts to a real in host order. */ |
262
|
0
|
|
|
|
|
|
char svd_readBinFloat(FILE *file, float *val) { |
263
|
|
|
|
|
|
|
int x; |
264
|
|
|
|
|
|
|
float y; |
265
|
0
|
0
|
|
|
|
|
if (fread(&x, sizeof(int), 1, file) == 1) { |
266
|
0
|
|
|
|
|
|
x = svdlibc_ntohl(x); //-- moo: disable ntohl() for PDL::SVDLIBC |
267
|
0
|
|
|
|
|
|
y = *((float *) &x); |
268
|
0
|
|
|
|
|
|
*val = y; |
269
|
0
|
|
|
|
|
|
return FALSE; |
270
|
|
|
|
|
|
|
} |
271
|
0
|
|
|
|
|
|
return TRUE; |
272
|
|
|
|
|
|
|
} |
273
|
|
|
|
|
|
|
|
274
|
0
|
|
|
|
|
|
char svd_writeBinInt(FILE *file, int x) { |
275
|
0
|
|
|
|
|
|
int y = svdlibc_htonl(x); //-- moo: disable htonl() for PDL::SVDLIBC |
276
|
0
|
0
|
|
|
|
|
if (fwrite(&y, sizeof(int), 1, file) != 1) return TRUE; |
277
|
0
|
|
|
|
|
|
return FALSE; |
278
|
|
|
|
|
|
|
} |
279
|
|
|
|
|
|
|
|
280
|
|
|
|
|
|
|
/* This takes a real in host order and writes a float in network order. */ |
281
|
0
|
|
|
|
|
|
char svd_writeBinFloat(FILE *file, float r) { |
282
|
0
|
|
|
|
|
|
int y = svdlibc_htonl(*((int *) &r)); //-- moo: disable htonl() for PDL::SVDLIBC |
283
|
0
|
0
|
|
|
|
|
if (fwrite(&y, sizeof(int), 1, file) != 1) return TRUE; |
284
|
0
|
|
|
|
|
|
return FALSE; |
285
|
|
|
|
|
|
|
} |
286
|
|
|
|
|
|
|
|
287
|
|
|
|
|
|
|
|
288
|
|
|
|
|
|
|
/************************************************************** |
289
|
|
|
|
|
|
|
* returns |a| if b is positive; else fsign returns -|a| * |
290
|
|
|
|
|
|
|
**************************************************************/ |
291
|
180
|
|
|
|
|
|
double svd_fsign(double a, double b) { |
292
|
180
|
50
|
|
|
|
|
if ((a>=0.0 && b>=0.0) || (a<0.0 && b<0.0))return(a); |
|
|
100
|
|
|
|
|
|
|
|
50
|
|
|
|
|
|
|
|
0
|
|
|
|
|
|
293
|
100
|
|
|
|
|
|
else return -a; |
294
|
|
|
|
|
|
|
} |
295
|
|
|
|
|
|
|
|
296
|
|
|
|
|
|
|
/************************************************************** |
297
|
|
|
|
|
|
|
* returns the larger of two double precision numbers * |
298
|
|
|
|
|
|
|
**************************************************************/ |
299
|
810
|
|
|
|
|
|
double svd_dmax(double a, double b) { |
300
|
810
|
100
|
|
|
|
|
return (a > b) ? a : b; |
301
|
|
|
|
|
|
|
} |
302
|
|
|
|
|
|
|
|
303
|
|
|
|
|
|
|
/************************************************************** |
304
|
|
|
|
|
|
|
* returns the smaller of two double precision numbers * |
305
|
|
|
|
|
|
|
**************************************************************/ |
306
|
860
|
|
|
|
|
|
double svd_dmin(double a, double b) { |
307
|
860
|
100
|
|
|
|
|
return (a < b) ? a : b; |
308
|
|
|
|
|
|
|
} |
309
|
|
|
|
|
|
|
|
310
|
|
|
|
|
|
|
/************************************************************** |
311
|
|
|
|
|
|
|
* returns the larger of two integers * |
312
|
|
|
|
|
|
|
**************************************************************/ |
313
|
10
|
|
|
|
|
|
__SVDLIBC_LONG svd_imax(__SVDLIBC_LONG a, __SVDLIBC_LONG b) { |
314
|
10
|
|
|
|
|
|
return (a > b) ? a : b; |
315
|
|
|
|
|
|
|
} |
316
|
|
|
|
|
|
|
|
317
|
|
|
|
|
|
|
/************************************************************** |
318
|
|
|
|
|
|
|
* returns the smaller of two integers * |
319
|
|
|
|
|
|
|
**************************************************************/ |
320
|
80
|
|
|
|
|
|
__SVDLIBC_LONG svd_imin(__SVDLIBC_LONG a, __SVDLIBC_LONG b) { |
321
|
80
|
|
|
|
|
|
return (a < b) ? a : b; |
322
|
|
|
|
|
|
|
} |
323
|
|
|
|
|
|
|
|
324
|
|
|
|
|
|
|
/************************************************************** |
325
|
|
|
|
|
|
|
* Function scales a vector by a constant. * |
326
|
|
|
|
|
|
|
* Based on Fortran-77 routine from Linpack by J. Dongarra * |
327
|
|
|
|
|
|
|
**************************************************************/ |
328
|
117
|
|
|
|
|
|
void svd_dscal(__SVDLIBC_LONG n, double da, double *dx, __SVDLIBC_LONG incx) { |
329
|
|
|
|
|
|
|
__SVDLIBC_LONG i; |
330
|
|
|
|
|
|
|
|
331
|
117
|
50
|
|
|
|
|
if (n <= 0 || incx == 0) return; |
|
|
50
|
|
|
|
|
|
332
|
117
|
50
|
|
|
|
|
if (incx < 0) dx += (-n+1) * incx; |
333
|
879
|
100
|
|
|
|
|
for (i=0; i < n; i++) { |
334
|
762
|
|
|
|
|
|
*dx *= da; |
335
|
762
|
|
|
|
|
|
dx += incx; |
336
|
|
|
|
|
|
|
} |
337
|
117
|
|
|
|
|
|
return; |
338
|
|
|
|
|
|
|
} |
339
|
|
|
|
|
|
|
|
340
|
|
|
|
|
|
|
/************************************************************** |
341
|
|
|
|
|
|
|
* function scales a vector by a constant. * |
342
|
|
|
|
|
|
|
* Based on Fortran-77 routine from Linpack by J. Dongarra * |
343
|
|
|
|
|
|
|
**************************************************************/ |
344
|
60
|
|
|
|
|
|
void svd_datx(__SVDLIBC_LONG n, double da, double *dx, __SVDLIBC_LONG incx, double *dy, __SVDLIBC_LONG incy) { |
345
|
|
|
|
|
|
|
__SVDLIBC_LONG i; |
346
|
|
|
|
|
|
|
|
347
|
60
|
50
|
|
|
|
|
if (n <= 0 || incx == 0 || incy == 0 || da == 0.0) return; |
|
|
50
|
|
|
|
|
|
|
|
50
|
|
|
|
|
|
|
|
50
|
|
|
|
|
|
348
|
120
|
50
|
|
|
|
|
if (incx == 1 && incy == 1) |
|
|
50
|
|
|
|
|
|
349
|
480
|
100
|
|
|
|
|
for (i=0; i < n; i++) *dy++ = da * (*dx++); |
350
|
|
|
|
|
|
|
|
351
|
|
|
|
|
|
|
else { |
352
|
0
|
0
|
|
|
|
|
if (incx < 0) dx += (-n+1) * incx; |
353
|
0
|
0
|
|
|
|
|
if (incy < 0) dy += (-n+1) * incy; |
354
|
0
|
0
|
|
|
|
|
for (i=0; i < n; i++) { |
355
|
0
|
|
|
|
|
|
*dy = da * (*dx); |
356
|
0
|
|
|
|
|
|
dx += incx; |
357
|
0
|
|
|
|
|
|
dy += incy; |
358
|
|
|
|
|
|
|
} |
359
|
|
|
|
|
|
|
} |
360
|
60
|
|
|
|
|
|
return; |
361
|
|
|
|
|
|
|
} |
362
|
|
|
|
|
|
|
|
363
|
|
|
|
|
|
|
/************************************************************** |
364
|
|
|
|
|
|
|
* Function copies a vector x to a vector y * |
365
|
|
|
|
|
|
|
* Based on Fortran-77 routine from Linpack by J. Dongarra * |
366
|
|
|
|
|
|
|
**************************************************************/ |
367
|
560
|
|
|
|
|
|
void svd_dcopy(__SVDLIBC_LONG n, double *dx, __SVDLIBC_LONG incx, double *dy, __SVDLIBC_LONG incy) { |
368
|
|
|
|
|
|
|
__SVDLIBC_LONG i; |
369
|
|
|
|
|
|
|
|
370
|
560
|
50
|
|
|
|
|
if (n <= 0 || incx == 0 || incy == 0) return; |
|
|
50
|
|
|
|
|
|
|
|
50
|
|
|
|
|
|
371
|
1080
|
50
|
|
|
|
|
if (incx == 1 && incy == 1) |
|
|
100
|
|
|
|
|
|
372
|
4160
|
100
|
|
|
|
|
for (i=0; i < n; i++) *dy++ = *dx++; |
373
|
|
|
|
|
|
|
|
374
|
|
|
|
|
|
|
else { |
375
|
40
|
50
|
|
|
|
|
if (incx < 0) dx += (-n+1) * incx; |
376
|
40
|
50
|
|
|
|
|
if (incy < 0) dy += (-n+1) * incy; |
377
|
260
|
100
|
|
|
|
|
for (i=0; i < n; i++) { |
378
|
220
|
|
|
|
|
|
*dy = *dx; |
379
|
220
|
|
|
|
|
|
dx += incx; |
380
|
220
|
|
|
|
|
|
dy += incy; |
381
|
|
|
|
|
|
|
} |
382
|
|
|
|
|
|
|
} |
383
|
560
|
|
|
|
|
|
return; |
384
|
|
|
|
|
|
|
} |
385
|
|
|
|
|
|
|
|
386
|
|
|
|
|
|
|
/************************************************************** |
387
|
|
|
|
|
|
|
* Function forms the dot product of two vectors. * |
388
|
|
|
|
|
|
|
* Based on Fortran-77 routine from Linpack by J. Dongarra * |
389
|
|
|
|
|
|
|
**************************************************************/ |
390
|
364
|
|
|
|
|
|
double svd_ddot(__SVDLIBC_LONG n, double *dx, __SVDLIBC_LONG incx, double *dy, __SVDLIBC_LONG incy) { |
391
|
|
|
|
|
|
|
__SVDLIBC_LONG i; |
392
|
|
|
|
|
|
|
double dot_product; |
393
|
|
|
|
|
|
|
|
394
|
364
|
50
|
|
|
|
|
if (n <= 0 || incx == 0 || incy == 0) return(0.0); |
|
|
50
|
|
|
|
|
|
|
|
50
|
|
|
|
|
|
395
|
364
|
|
|
|
|
|
dot_product = 0.0; |
396
|
728
|
50
|
|
|
|
|
if (incx == 1 && incy == 1) |
|
|
50
|
|
|
|
|
|
397
|
2912
|
100
|
|
|
|
|
for (i=0; i < n; i++) dot_product += (*dx++) * (*dy++); |
398
|
|
|
|
|
|
|
else { |
399
|
0
|
0
|
|
|
|
|
if (incx < 0) dx += (-n+1) * incx; |
400
|
0
|
0
|
|
|
|
|
if (incy < 0) dy += (-n+1) * incy; |
401
|
0
|
0
|
|
|
|
|
for (i=0; i < n; i++) { |
402
|
0
|
|
|
|
|
|
dot_product += (*dx) * (*dy); |
403
|
0
|
|
|
|
|
|
dx += incx; |
404
|
0
|
|
|
|
|
|
dy += incy; |
405
|
|
|
|
|
|
|
} |
406
|
|
|
|
|
|
|
} |
407
|
364
|
|
|
|
|
|
return(dot_product); |
408
|
|
|
|
|
|
|
} |
409
|
|
|
|
|
|
|
|
410
|
|
|
|
|
|
|
/************************************************************** |
411
|
|
|
|
|
|
|
* Constant times a vector plus a vector * |
412
|
|
|
|
|
|
|
* Based on Fortran-77 routine from Linpack by J. Dongarra * |
413
|
|
|
|
|
|
|
**************************************************************/ |
414
|
637
|
|
|
|
|
|
void svd_daxpy (__SVDLIBC_LONG n, double da, double *dx, __SVDLIBC_LONG incx, double *dy, __SVDLIBC_LONG incy) { |
415
|
|
|
|
|
|
|
__SVDLIBC_LONG i; |
416
|
|
|
|
|
|
|
|
417
|
637
|
50
|
|
|
|
|
if (n <= 0 || incx == 0 || incy == 0 || da == 0.0) return; |
|
|
50
|
|
|
|
|
|
|
|
50
|
|
|
|
|
|
|
|
50
|
|
|
|
|
|
418
|
1274
|
50
|
|
|
|
|
if (incx == 1 && incy == 1) |
|
|
50
|
|
|
|
|
|
419
|
5096
|
100
|
|
|
|
|
for (i=0; i < n; i++) { |
420
|
4459
|
|
|
|
|
|
*dy += da * (*dx++); |
421
|
4459
|
|
|
|
|
|
dy++; |
422
|
|
|
|
|
|
|
} |
423
|
|
|
|
|
|
|
else { |
424
|
0
|
0
|
|
|
|
|
if (incx < 0) dx += (-n+1) * incx; |
425
|
0
|
0
|
|
|
|
|
if (incy < 0) dy += (-n+1) * incy; |
426
|
0
|
0
|
|
|
|
|
for (i=0; i < n; i++) { |
427
|
0
|
|
|
|
|
|
*dy += da * (*dx); |
428
|
0
|
|
|
|
|
|
dx += incx; |
429
|
0
|
|
|
|
|
|
dy += incy; |
430
|
|
|
|
|
|
|
} |
431
|
|
|
|
|
|
|
} |
432
|
637
|
|
|
|
|
|
return; |
433
|
|
|
|
|
|
|
} |
434
|
|
|
|
|
|
|
|
435
|
|
|
|
|
|
|
/********************************************************************* |
436
|
|
|
|
|
|
|
* Function sorts array1 and array2 into increasing order for array1 * |
437
|
|
|
|
|
|
|
*********************************************************************/ |
438
|
30
|
|
|
|
|
|
void svd_dsort2(__SVDLIBC_LONG igap, __SVDLIBC_LONG n, double *array1, double *array2) { |
439
|
|
|
|
|
|
|
double temp; |
440
|
|
|
|
|
|
|
__SVDLIBC_LONG i, j, index; |
441
|
|
|
|
|
|
|
|
442
|
30
|
100
|
|
|
|
|
if (!igap) return; |
443
|
|
|
|
|
|
|
else { |
444
|
100
|
100
|
|
|
|
|
for (i = igap; i < n; i++) { |
445
|
80
|
|
|
|
|
|
j = i - igap; |
446
|
80
|
|
|
|
|
|
index = i; |
447
|
80
|
50
|
|
|
|
|
while (j >= 0 && array1[j] > array1[index]) { |
|
|
50
|
|
|
|
|
|
448
|
0
|
|
|
|
|
|
temp = array1[j]; |
449
|
0
|
|
|
|
|
|
array1[j] = array1[index]; |
450
|
0
|
|
|
|
|
|
array1[index] = temp; |
451
|
0
|
|
|
|
|
|
temp = array2[j]; |
452
|
0
|
|
|
|
|
|
array2[j] = array2[index]; |
453
|
0
|
|
|
|
|
|
array2[index] = temp; |
454
|
0
|
|
|
|
|
|
j -= igap; |
455
|
0
|
|
|
|
|
|
index = j + igap; |
456
|
|
|
|
|
|
|
} |
457
|
|
|
|
|
|
|
} |
458
|
|
|
|
|
|
|
} |
459
|
20
|
|
|
|
|
|
svd_dsort2(igap/2,n,array1,array2); |
460
|
|
|
|
|
|
|
} |
461
|
|
|
|
|
|
|
|
462
|
|
|
|
|
|
|
/************************************************************** |
463
|
|
|
|
|
|
|
* Function interchanges two vectors * |
464
|
|
|
|
|
|
|
* Based on Fortran-77 routine from Linpack by J. Dongarra * |
465
|
|
|
|
|
|
|
**************************************************************/ |
466
|
50
|
|
|
|
|
|
void svd_dswap(__SVDLIBC_LONG n, double *dx, __SVDLIBC_LONG incx, double *dy, __SVDLIBC_LONG incy) { |
467
|
|
|
|
|
|
|
__SVDLIBC_LONG i; |
468
|
|
|
|
|
|
|
double dtemp; |
469
|
|
|
|
|
|
|
|
470
|
50
|
50
|
|
|
|
|
if (n <= 0 || incx == 0 || incy == 0) return; |
|
|
50
|
|
|
|
|
|
|
|
50
|
|
|
|
|
|
471
|
100
|
50
|
|
|
|
|
if (incx == 1 && incy == 1) { |
|
|
50
|
|
|
|
|
|
472
|
200
|
100
|
|
|
|
|
for (i=0; i < n; i++) { |
473
|
150
|
|
|
|
|
|
dtemp = *dy; |
474
|
150
|
|
|
|
|
|
*dy++ = *dx; |
475
|
150
|
|
|
|
|
|
*dx++ = dtemp; |
476
|
|
|
|
|
|
|
} |
477
|
|
|
|
|
|
|
} |
478
|
|
|
|
|
|
|
else { |
479
|
0
|
0
|
|
|
|
|
if (incx < 0) dx += (-n+1) * incx; |
480
|
0
|
0
|
|
|
|
|
if (incy < 0) dy += (-n+1) * incy; |
481
|
0
|
0
|
|
|
|
|
for (i=0; i < n; i++) { |
482
|
0
|
|
|
|
|
|
dtemp = *dy; |
483
|
0
|
|
|
|
|
|
*dy = *dx; |
484
|
0
|
|
|
|
|
|
*dx = dtemp; |
485
|
0
|
|
|
|
|
|
dx += incx; |
486
|
0
|
|
|
|
|
|
dy += incy; |
487
|
|
|
|
|
|
|
} |
488
|
|
|
|
|
|
|
} |
489
|
|
|
|
|
|
|
} |
490
|
|
|
|
|
|
|
|
491
|
|
|
|
|
|
|
/***************************************************************** |
492
|
|
|
|
|
|
|
* Function finds the index of element having max. absolute value* |
493
|
|
|
|
|
|
|
* based on FORTRAN 77 routine from Linpack by J. Dongarra * |
494
|
|
|
|
|
|
|
*****************************************************************/ |
495
|
50
|
|
|
|
|
|
__SVDLIBC_LONG svd_idamax(__SVDLIBC_LONG n, double *dx, __SVDLIBC_LONG incx) { |
496
|
|
|
|
|
|
|
__SVDLIBC_LONG ix,i,imax; |
497
|
|
|
|
|
|
|
double dtemp, dmax; |
498
|
|
|
|
|
|
|
|
499
|
50
|
50
|
|
|
|
|
if (n < 1) return(-1); |
500
|
50
|
100
|
|
|
|
|
if (n == 1) return(0); |
501
|
40
|
50
|
|
|
|
|
if (incx == 0) return(-1); |
502
|
|
|
|
|
|
|
|
503
|
40
|
50
|
|
|
|
|
if (incx < 0) ix = (-n+1) * incx; |
504
|
40
|
|
|
|
|
|
else ix = 0; |
505
|
40
|
|
|
|
|
|
imax = ix; |
506
|
40
|
|
|
|
|
|
dx += ix; |
507
|
40
|
|
|
|
|
|
dmax = fabs(*dx); |
508
|
150
|
100
|
|
|
|
|
for (i=1; i < n; i++) { |
509
|
110
|
|
|
|
|
|
ix += incx; |
510
|
110
|
|
|
|
|
|
dx += incx; |
511
|
110
|
|
|
|
|
|
dtemp = fabs(*dx); |
512
|
110
|
50
|
|
|
|
|
if (dtemp > dmax) { |
513
|
0
|
|
|
|
|
|
dmax = dtemp; |
514
|
0
|
|
|
|
|
|
imax = ix; |
515
|
|
|
|
|
|
|
} |
516
|
|
|
|
|
|
|
} |
517
|
40
|
|
|
|
|
|
return(imax); |
518
|
|
|
|
|
|
|
} |
519
|
|
|
|
|
|
|
|
520
|
|
|
|
|
|
|
/************************************************************** |
521
|
|
|
|
|
|
|
* multiplication of matrix B by vector x, where B = A'A, * |
522
|
|
|
|
|
|
|
* and A is nrow by ncol (nrow >> ncol). Hence, B is of order * |
523
|
|
|
|
|
|
|
* n = ncol (y stores product vector). * |
524
|
|
|
|
|
|
|
**************************************************************/ |
525
|
127
|
|
|
|
|
|
void svd_opb(SMat A, double *x, double *y, double *temp) { |
526
|
|
|
|
|
|
|
__SVDLIBC_LONG i, j, end; |
527
|
127
|
|
|
|
|
|
__SVDLIBC_LONG *pointr = A->pointr, *rowind = A->rowind; |
528
|
127
|
|
|
|
|
|
double *value = A->value; |
529
|
127
|
|
|
|
|
|
__SVDLIBC_LONG n = A->cols; |
530
|
|
|
|
|
|
|
|
531
|
127
|
|
|
|
|
|
SVDCount[SVD_MXV] += 2; |
532
|
127
|
|
|
|
|
|
memset(y, 0, n * sizeof(double)); |
533
|
889
|
100
|
|
|
|
|
for (i = 0; i < A->rows; i++) temp[i] = 0.0; |
534
|
|
|
|
|
|
|
|
535
|
1016
|
100
|
|
|
|
|
for (i = 0; i < A->cols; i++) { |
536
|
889
|
|
|
|
|
|
end = pointr[i+1]; |
537
|
3683
|
100
|
|
|
|
|
for (j = pointr[i]; j < end; j++) |
538
|
2794
|
|
|
|
|
|
temp[rowind[j]] += value[j] * (*x); |
539
|
889
|
|
|
|
|
|
x++; |
540
|
|
|
|
|
|
|
} |
541
|
1016
|
100
|
|
|
|
|
for (i = 0; i < A->cols; i++) { |
542
|
889
|
|
|
|
|
|
end = pointr[i+1]; |
543
|
3683
|
100
|
|
|
|
|
for (j = pointr[i]; j < end; j++) |
544
|
2794
|
|
|
|
|
|
*y += value[j] * temp[rowind[j]]; |
545
|
889
|
|
|
|
|
|
y++; |
546
|
|
|
|
|
|
|
} |
547
|
127
|
|
|
|
|
|
return; |
548
|
|
|
|
|
|
|
} |
549
|
|
|
|
|
|
|
|
550
|
|
|
|
|
|
|
/*********************************************************** |
551
|
|
|
|
|
|
|
* multiplication of matrix A by vector x, where A is * |
552
|
|
|
|
|
|
|
* nrow by ncol (nrow >> ncol). y stores product vector. * |
553
|
|
|
|
|
|
|
***********************************************************/ |
554
|
57
|
|
|
|
|
|
void svd_opa(SMat A, double *x, double *y) { |
555
|
|
|
|
|
|
|
__SVDLIBC_LONG end, i, j; |
556
|
57
|
|
|
|
|
|
__SVDLIBC_LONG *pointr = A->pointr, *rowind = A->rowind; |
557
|
57
|
|
|
|
|
|
double *value = A->value; |
558
|
|
|
|
|
|
|
|
559
|
57
|
|
|
|
|
|
SVDCount[SVD_MXV]++; |
560
|
57
|
|
|
|
|
|
memset(y, 0, A->rows * sizeof(double)); |
561
|
|
|
|
|
|
|
|
562
|
456
|
100
|
|
|
|
|
for (i = 0; i < A->cols; i++) { |
563
|
399
|
|
|
|
|
|
end = pointr[i+1]; |
564
|
1653
|
100
|
|
|
|
|
for (j = pointr[i]; j < end; j++) |
565
|
1254
|
|
|
|
|
|
y[rowind[j]] += value[j] * x[i]; |
566
|
|
|
|
|
|
|
} |
567
|
57
|
|
|
|
|
|
return; |
568
|
|
|
|
|
|
|
} |
569
|
|
|
|
|
|
|
|
570
|
|
|
|
|
|
|
|
571
|
|
|
|
|
|
|
/*********************************************************************** |
572
|
|
|
|
|
|
|
* * |
573
|
|
|
|
|
|
|
* random() * |
574
|
|
|
|
|
|
|
* (double precision) * |
575
|
|
|
|
|
|
|
***********************************************************************/ |
576
|
|
|
|
|
|
|
/*********************************************************************** |
577
|
|
|
|
|
|
|
|
578
|
|
|
|
|
|
|
Description |
579
|
|
|
|
|
|
|
----------- |
580
|
|
|
|
|
|
|
|
581
|
|
|
|
|
|
|
This is a translation of a Fortran-77 uniform random number |
582
|
|
|
|
|
|
|
generator. The code is based on theory and suggestions given in |
583
|
|
|
|
|
|
|
D. E. Knuth (1969), vol 2. The argument to the function should |
584
|
|
|
|
|
|
|
be initialized to an arbitrary integer prior to the first call to |
585
|
|
|
|
|
|
|
random. The calling program should not alter the value of the |
586
|
|
|
|
|
|
|
argument between subsequent calls to random. Random returns values |
587
|
|
|
|
|
|
|
within the interval (0,1). |
588
|
|
|
|
|
|
|
|
589
|
|
|
|
|
|
|
|
590
|
|
|
|
|
|
|
Arguments |
591
|
|
|
|
|
|
|
--------- |
592
|
|
|
|
|
|
|
|
593
|
|
|
|
|
|
|
(input) |
594
|
|
|
|
|
|
|
iy an integer seed whose value must not be altered by the caller |
595
|
|
|
|
|
|
|
between subsequent calls |
596
|
|
|
|
|
|
|
|
597
|
|
|
|
|
|
|
(output) |
598
|
|
|
|
|
|
|
random a double precision random number between (0,1) |
599
|
|
|
|
|
|
|
|
600
|
|
|
|
|
|
|
***********************************************************************/ |
601
|
70
|
|
|
|
|
|
double svd_random2(__SVDLIBC_LONG *iy) { |
602
|
|
|
|
|
|
|
static __SVDLIBC_LONG m2 = 0; |
603
|
|
|
|
|
|
|
static __SVDLIBC_LONG ia, ic, mic; |
604
|
|
|
|
|
|
|
static double halfm, s; |
605
|
|
|
|
|
|
|
|
606
|
|
|
|
|
|
|
/* If first entry, compute (max int) / 2 */ |
607
|
70
|
100
|
|
|
|
|
if (!m2) { |
608
|
1
|
|
|
|
|
|
m2 = 1 << (8 * (int)sizeof(int) - 2); |
609
|
1
|
|
|
|
|
|
halfm = m2; |
610
|
|
|
|
|
|
|
|
611
|
|
|
|
|
|
|
/* compute multiplier and increment for linear congruential |
612
|
|
|
|
|
|
|
* method */ |
613
|
1
|
|
|
|
|
|
ia = 8 * (__SVDLIBC_LONG)(halfm * atan(1.0) / 8.0) + 5; |
614
|
1
|
|
|
|
|
|
ic = 2 * (__SVDLIBC_LONG)(halfm * (0.5 - sqrt(3.0)/6.0)) + 1; |
615
|
1
|
|
|
|
|
|
mic = (m2-ic) + m2; |
616
|
|
|
|
|
|
|
|
617
|
|
|
|
|
|
|
/* s is the scale factor for converting to floating point */ |
618
|
1
|
|
|
|
|
|
s = 0.5 / halfm; |
619
|
|
|
|
|
|
|
} |
620
|
|
|
|
|
|
|
|
621
|
|
|
|
|
|
|
/* compute next random number */ |
622
|
70
|
|
|
|
|
|
*iy = *iy * ia; |
623
|
|
|
|
|
|
|
|
624
|
|
|
|
|
|
|
/* for computers which do not allow integer overflow on addition */ |
625
|
70
|
100
|
|
|
|
|
if (*iy > mic) *iy = (*iy - m2) - m2; |
626
|
|
|
|
|
|
|
|
627
|
70
|
|
|
|
|
|
*iy = *iy + ic; |
628
|
|
|
|
|
|
|
|
629
|
|
|
|
|
|
|
/* for computers whose word length for addition is greater than |
630
|
|
|
|
|
|
|
* for multiplication */ |
631
|
70
|
100
|
|
|
|
|
if (*iy / 2 > m2) *iy = (*iy - m2) - m2; |
632
|
|
|
|
|
|
|
|
633
|
|
|
|
|
|
|
/* for computers whose integer overflow affects the sign bit */ |
634
|
70
|
100
|
|
|
|
|
if (*iy < 0) *iy = (*iy + m2) + m2; |
635
|
|
|
|
|
|
|
|
636
|
70
|
|
|
|
|
|
return((double)(*iy) * s); |
637
|
|
|
|
|
|
|
} |
638
|
|
|
|
|
|
|
|
639
|
|
|
|
|
|
|
/************************************************************** |
640
|
|
|
|
|
|
|
* * |
641
|
|
|
|
|
|
|
* Function finds sqrt(a^2 + b^2) without overflow or * |
642
|
|
|
|
|
|
|
* destructive underflow. * |
643
|
|
|
|
|
|
|
* * |
644
|
|
|
|
|
|
|
**************************************************************/ |
645
|
|
|
|
|
|
|
/************************************************************** |
646
|
|
|
|
|
|
|
|
647
|
|
|
|
|
|
|
Funtions used |
648
|
|
|
|
|
|
|
------------- |
649
|
|
|
|
|
|
|
|
650
|
|
|
|
|
|
|
UTILITY dmax, dmin |
651
|
|
|
|
|
|
|
|
652
|
|
|
|
|
|
|
**************************************************************/ |
653
|
800
|
|
|
|
|
|
double svd_pythag(double a, double b) { |
654
|
|
|
|
|
|
|
double p, r, s, t, u, temp; |
655
|
|
|
|
|
|
|
|
656
|
800
|
|
|
|
|
|
p = svd_dmax(fabs(a), fabs(b)); |
657
|
800
|
50
|
|
|
|
|
if (p != 0.0) { |
658
|
800
|
|
|
|
|
|
temp = svd_dmin(fabs(a), fabs(b)) / p; |
659
|
800
|
|
|
|
|
|
r = temp * temp; |
660
|
800
|
|
|
|
|
|
t = 4.0 + r; |
661
|
2300
|
100
|
|
|
|
|
while (t != 4.0) { |
662
|
1500
|
|
|
|
|
|
s = r / t; |
663
|
1500
|
|
|
|
|
|
u = 1.0 + 2.0 * s; |
664
|
1500
|
|
|
|
|
|
p *= u; |
665
|
1500
|
|
|
|
|
|
temp = s / u; |
666
|
1500
|
|
|
|
|
|
r *= temp * temp; |
667
|
1500
|
|
|
|
|
|
t = 4.0 + r; |
668
|
|
|
|
|
|
|
} |
669
|
|
|
|
|
|
|
} |
670
|
800
|
|
|
|
|
|
return(p); |
671
|
|
|
|
|
|
|
} |
672
|
|
|
|
|
|
|
|