miriad 4.1.0.0
Sign up to get free protection for your applications and to get access to all the features.
- data/README +103 -0
- data/Rakefile +82 -0
- data/ext/bug.c +341 -0
- data/ext/dio.c +317 -0
- data/ext/extconf.rb +49 -0
- data/ext/headio.c +835 -0
- data/ext/hio.c +1515 -0
- data/ext/hio.h +48 -0
- data/ext/interface.c +74 -0
- data/ext/io.h +56 -0
- data/ext/key.c +934 -0
- data/ext/maskio.c +398 -0
- data/ext/maxdimc.h.in +9 -0
- data/ext/miriad.h +371 -0
- data/ext/miriad.i +464 -0
- data/ext/miriad_ruby.c +602 -0
- data/ext/miriad_ruby.i +443 -0
- data/ext/miriad_wrap.c +4210 -0
- data/ext/narray_ruby.swg +59 -0
- data/ext/pack.c +639 -0
- data/ext/scrio.c +132 -0
- data/ext/sysdep.h +185 -0
- data/ext/uvio.c +4934 -0
- data/ext/xyio.c +476 -0
- data/ext/xyzio.c +2020 -0
- data/lib/miriad.rb +564 -0
- metadata +93 -0
data/ext/narray_ruby.swg
ADDED
@@ -0,0 +1,59 @@
|
|
1
|
+
// Common setup for NArray typemaps
|
2
|
+
//
|
3
|
+
// Requires NArray Ruby extension
|
4
|
+
//
|
5
|
+
// http://narray.rubyforge.org/
|
6
|
+
//
|
7
|
+
|
8
|
+
#if defined SWIGRUBY
|
9
|
+
|
10
|
+
%{
|
11
|
+
#include "narray.h"
|
12
|
+
extern const char * na_typestring[];
|
13
|
+
|
14
|
+
// Raises TypeError is v is not an NArray
|
15
|
+
#define Check_NArray(v) \
|
16
|
+
if(!NA_IsNArray(v)) { \
|
17
|
+
rb_raise(rb_eTypeError,"requires NArray, got %s", \
|
18
|
+
rb_class2name(CLASS_OF(v))); \
|
19
|
+
}
|
20
|
+
|
21
|
+
#define Check_NArrayType(v,tc) \
|
22
|
+
Check_NArray(v) \
|
23
|
+
if(tc >= NA_NTYPES) { \
|
24
|
+
rb_raise(rb_eArgError,"invalid type code check requested (%d)", tc); \
|
25
|
+
} \
|
26
|
+
if(NA_TYPE(v) >= NA_NTYPES) { \
|
27
|
+
rb_raise(rb_eArgError,"invalid type code found in NArray object (%d)", NA_TYPE(v)); \
|
28
|
+
} \
|
29
|
+
if(NA_TYPE(v) != tc) { \
|
30
|
+
rb_raise(rb_eTypeError,"requires type code %s (not %s)", \
|
31
|
+
na_typestring[tc], \
|
32
|
+
na_typestring[NA_TYPE(v)]); \
|
33
|
+
}
|
34
|
+
|
35
|
+
#define Check_NArrayType2(v,tc1,tc2) \
|
36
|
+
Check_NArray(v) \
|
37
|
+
if(tc1 >= NA_NTYPES) { \
|
38
|
+
rb_raise(rb_eArgError,"invalid type code check requested (%d)", tc1); \
|
39
|
+
} \
|
40
|
+
if(tc2 >= NA_NTYPES) { \
|
41
|
+
rb_raise(rb_eArgError,"invalid type code check requested (%d)", tc2); \
|
42
|
+
} \
|
43
|
+
if(NA_TYPE(v) >= NA_NTYPES) { \
|
44
|
+
rb_raise(rb_eArgError,"invalid type code found in NArray object (%d)", NA_TYPE(v)); \
|
45
|
+
} \
|
46
|
+
if(NA_TYPE(v) != tc1 && NA_TYPE(v) != tc2) { \
|
47
|
+
rb_raise(rb_eTypeError,"requires type code %s or %s (not %s)", \
|
48
|
+
na_typestring[tc1], \
|
49
|
+
na_typestring[tc2], \
|
50
|
+
na_typestring[NA_TYPE(v)]); \
|
51
|
+
}
|
52
|
+
%}
|
53
|
+
|
54
|
+
%init %{
|
55
|
+
rb_require("narray");
|
56
|
+
%}
|
57
|
+
|
58
|
+
// vim: set ts=2 sw=2 et sta syntax=c cindent :
|
59
|
+
#endif // SWIGRUBY
|
data/ext/pack.c
ADDED
@@ -0,0 +1,639 @@
|
|
1
|
+
/* pack */
|
2
|
+
/* & pjt */
|
3
|
+
/* : low-level-i/o */
|
4
|
+
/* + */
|
5
|
+
/* */
|
6
|
+
/* This converts data between disk and internal */
|
7
|
+
/* format. Disk format is IEEE reals and 16 or 32 */
|
8
|
+
/* bit integers (most significant byte first). */
|
9
|
+
/* */
|
10
|
+
/* This assumes that these are the local machine format */
|
11
|
+
/* (float == IEEE real, int == 32 bit integer, */
|
12
|
+
/* short int == 16 bit integer). */
|
13
|
+
/* */
|
14
|
+
/* packx_c, unpackx_c, pack32_c and unpack32_c are */
|
15
|
+
/* implemented as macros (calling bcopy) in the */
|
16
|
+
/* system dependent include file. */
|
17
|
+
/*-- */
|
18
|
+
/* */
|
19
|
+
/* History: */
|
20
|
+
/* rjs Dark-ages Original version. */
|
21
|
+
/* bs ?????89 Improved efficiency using "register" declarations. */
|
22
|
+
/* rjs 1nov89 Incoporated Brian's changes. */
|
23
|
+
/* mjs 28feb91 Merge Sun and Cray versions. */
|
24
|
+
/* mjs 18mar91 Added convex definition. */
|
25
|
+
/* mjs 19feb93 Added mips definition. */
|
26
|
+
/* pjt 25jan95 linux kludge to include packALPHA.c */
|
27
|
+
/* pjt 14jun01 packALPHA.c now included in this source code */
|
28
|
+
/* and using the standard WORDS_BIGENDIAN macro */
|
29
|
+
/* pjt 21jun02 MIR4 prototyping */
|
30
|
+
/************************************************************************/
|
31
|
+
|
32
|
+
#include "sysdep.h"
|
33
|
+
#include "miriad.h"
|
34
|
+
|
35
|
+
#if defined(WORDS_BIGENDIAN)
|
36
|
+
|
37
|
+
static int words_bigendian = 1; /* never used actually, but handy symbol to find via nm(1) */
|
38
|
+
|
39
|
+
void pack16_c(register int *from,char *to,int n)
|
40
|
+
{
|
41
|
+
register short int *tto;
|
42
|
+
register int i;
|
43
|
+
|
44
|
+
tto = (short int *)to;
|
45
|
+
for (i=0; i < n; i++) *tto++ = *from++;
|
46
|
+
}
|
47
|
+
void unpack16_c(char *from,register int *to,int n)
|
48
|
+
{
|
49
|
+
register short int *ffrom;
|
50
|
+
register int i;
|
51
|
+
|
52
|
+
ffrom = (short int *)from;
|
53
|
+
for (i=0; i < n; i++) *to++ = *ffrom++;
|
54
|
+
}
|
55
|
+
|
56
|
+
void pack64_c(register int *from,char *to,int n)
|
57
|
+
{
|
58
|
+
register short int *tto;
|
59
|
+
register int i;
|
60
|
+
|
61
|
+
tto = (short int *)to;
|
62
|
+
for (i=0; i < n; i++) *tto++ = *from++;
|
63
|
+
}
|
64
|
+
void unpack64_c(char *from,register int *to,int n)
|
65
|
+
{
|
66
|
+
register short int *ffrom;
|
67
|
+
register int i;
|
68
|
+
|
69
|
+
ffrom = (short int *)from;
|
70
|
+
for (i=0; i < n; i++) *to++ = *ffrom++;
|
71
|
+
}
|
72
|
+
|
73
|
+
#endif
|
74
|
+
|
75
|
+
|
76
|
+
#ifndef WORDS_BIGENDIAN
|
77
|
+
#ifndef unicos
|
78
|
+
static int words_littleendian = 1; /* never used actually, but handy symbol to find via nm(1) */
|
79
|
+
/************************************************************************/
|
80
|
+
/* */
|
81
|
+
/* The pack routines -- these convert between the host format and */
|
82
|
+
/* the disk format. Disk format is IEEE 32 and 64 bit reals, and 2's */
|
83
|
+
/* complement integers. Byte order is the FITS byte order (most */
|
84
|
+
/* significant bytes first). */
|
85
|
+
/* */
|
86
|
+
/* This version is for a machine which uses IEEE internally, but which */
|
87
|
+
/* uses least significant bytes first (little endian), e.g. PCs and */
|
88
|
+
/* Alphas. */
|
89
|
+
/* */
|
90
|
+
/* History: */
|
91
|
+
/* rjs 21nov94 Original version. */
|
92
|
+
/************************************************************************/
|
93
|
+
void pack16_c(int *in,char *out,int n)
|
94
|
+
/*
|
95
|
+
Pack an integer array into 16 bit integers.
|
96
|
+
------------------------------------------------------------------------*/
|
97
|
+
{
|
98
|
+
int i;
|
99
|
+
char *s;
|
100
|
+
|
101
|
+
s = (char *)in;
|
102
|
+
for(i=0; i < n; i++){
|
103
|
+
*out++ = *(s+1);
|
104
|
+
*out++ = *s;
|
105
|
+
s += sizeof(int);
|
106
|
+
}
|
107
|
+
}
|
108
|
+
/************************************************************************/
|
109
|
+
void unpack16_c(char *in,int *out,int n)
|
110
|
+
/*
|
111
|
+
Unpack an array of 16 bit integers into integers.
|
112
|
+
------------------------------------------------------------------------*/
|
113
|
+
{
|
114
|
+
int i;
|
115
|
+
char *s;
|
116
|
+
|
117
|
+
s = (char *)out;
|
118
|
+
for(i=0; i < n; i++){
|
119
|
+
*s++ = *(in+1);
|
120
|
+
*s++ = *in;
|
121
|
+
if(0x80 & *in){
|
122
|
+
*s++ = 0xFF;
|
123
|
+
*s++ = 0xFF;
|
124
|
+
} else {
|
125
|
+
*s++ = 0;
|
126
|
+
*s++ = 0;
|
127
|
+
}
|
128
|
+
in += 2;
|
129
|
+
}
|
130
|
+
}
|
131
|
+
/************************************************************************/
|
132
|
+
void pack32_c(int *in,char *out,int n)
|
133
|
+
/*
|
134
|
+
Pack an array of integers into 32 bit integers.
|
135
|
+
------------------------------------------------------------------------*/
|
136
|
+
{
|
137
|
+
int i;
|
138
|
+
char *s;
|
139
|
+
|
140
|
+
s = (char *)in;
|
141
|
+
for(i = 0; i < n; i++){
|
142
|
+
*out++ = *(s+3);
|
143
|
+
*out++ = *(s+2);
|
144
|
+
*out++ = *(s+1);
|
145
|
+
*out++ = *s;
|
146
|
+
s += 4;
|
147
|
+
}
|
148
|
+
}
|
149
|
+
/************************************************************************/
|
150
|
+
void unpack32_c(char *in,int *out,int n)
|
151
|
+
/*
|
152
|
+
Unpack an array of 32 bit integers into integers.
|
153
|
+
------------------------------------------------------------------------*/
|
154
|
+
{
|
155
|
+
int i;
|
156
|
+
char *s;
|
157
|
+
|
158
|
+
s = (char *)out;
|
159
|
+
for(i = 0; i < n; i++){
|
160
|
+
*s++ = *(in+3);
|
161
|
+
*s++ = *(in+2);
|
162
|
+
*s++ = *(in+1);
|
163
|
+
*s++ = *in;
|
164
|
+
in += 4;
|
165
|
+
}
|
166
|
+
}
|
167
|
+
/************************************************************************/
|
168
|
+
void pack64_c(int8 *in,char *out,int n)
|
169
|
+
/*
|
170
|
+
Pack an integer array into 64 bit integers.
|
171
|
+
------------------------------------------------------------------------*/
|
172
|
+
{
|
173
|
+
int i;
|
174
|
+
char *s;
|
175
|
+
|
176
|
+
s = (char *)in;
|
177
|
+
for(i=0; i < n; i++){
|
178
|
+
*out++ = *(s+7);
|
179
|
+
*out++ = *(s+6);
|
180
|
+
*out++ = *(s+5);
|
181
|
+
*out++ = *(s+4);
|
182
|
+
*out++ = *(s+3);
|
183
|
+
*out++ = *(s+2);
|
184
|
+
*out++ = *(s+1);
|
185
|
+
*out++ = *s;
|
186
|
+
s += 8;
|
187
|
+
}
|
188
|
+
}
|
189
|
+
/************************************************************************/
|
190
|
+
void unpack64_c(char *in,int8 *out,int n)
|
191
|
+
/*
|
192
|
+
Unpack an array of 64 bit integers into integers.
|
193
|
+
------------------------------------------------------------------------*/
|
194
|
+
{
|
195
|
+
int i;
|
196
|
+
char *s;
|
197
|
+
|
198
|
+
s = (char *)out;
|
199
|
+
for(i=0; i < n; i++){
|
200
|
+
*s++ = *(in+7);
|
201
|
+
*s++ = *(in+6);
|
202
|
+
*s++ = *(in+5);
|
203
|
+
*s++ = *(in+4);
|
204
|
+
*s++ = *(in+3);
|
205
|
+
*s++ = *(in+2);
|
206
|
+
*s++ = *(in+1);
|
207
|
+
*s++ = *in;
|
208
|
+
in += 8;
|
209
|
+
}
|
210
|
+
}
|
211
|
+
/************************************************************************/
|
212
|
+
void packr_c(float *in,char *out,int n)
|
213
|
+
/*
|
214
|
+
Pack an array of reals into IEEE reals -- just do byte reversal.
|
215
|
+
------------------------------------------------------------------------*/
|
216
|
+
{
|
217
|
+
int i;
|
218
|
+
char *s;
|
219
|
+
|
220
|
+
s = (char *)in;
|
221
|
+
for(i = 0; i < n; i++){
|
222
|
+
*out++ = *(s+3);
|
223
|
+
*out++ = *(s+2);
|
224
|
+
*out++ = *(s+1);
|
225
|
+
*out++ = *s;
|
226
|
+
s += 4;
|
227
|
+
}
|
228
|
+
}
|
229
|
+
/************************************************************************/
|
230
|
+
void unpackr_c(char *in,float *out,int n)
|
231
|
+
/*
|
232
|
+
Unpack an array of IEEE reals into reals -- just do byte reversal.
|
233
|
+
------------------------------------------------------------------------*/
|
234
|
+
{
|
235
|
+
int i;
|
236
|
+
char *s;
|
237
|
+
|
238
|
+
s = (char *)out;
|
239
|
+
for(i = 0; i < n; i++){
|
240
|
+
*s++ = *(in+3);
|
241
|
+
*s++ = *(in+2);
|
242
|
+
*s++ = *(in+1);
|
243
|
+
*s++ = *in;
|
244
|
+
in += 4;
|
245
|
+
}
|
246
|
+
}
|
247
|
+
/************************************************************************/
|
248
|
+
void packd_c(double *in,char *out,int n)
|
249
|
+
/*
|
250
|
+
Pack an array of doubles -- this involves simply performing byte
|
251
|
+
reversal.
|
252
|
+
------------------------------------------------------------------------*/
|
253
|
+
{
|
254
|
+
int i;
|
255
|
+
char *s;
|
256
|
+
|
257
|
+
s = (char *)in;
|
258
|
+
for(i = 0; i < n; i++){
|
259
|
+
*out++ = *(s+7);
|
260
|
+
*out++ = *(s+6);
|
261
|
+
*out++ = *(s+5);
|
262
|
+
*out++ = *(s+4);
|
263
|
+
*out++ = *(s+3);
|
264
|
+
*out++ = *(s+2);
|
265
|
+
*out++ = *(s+1);
|
266
|
+
*out++ = *s;
|
267
|
+
s += 8;
|
268
|
+
}
|
269
|
+
}
|
270
|
+
/************************************************************************/
|
271
|
+
void unpackd_c(char *in,double *out,int n)
|
272
|
+
/*
|
273
|
+
Unpack an array of doubles -- this involves simply performing byte
|
274
|
+
reversal.
|
275
|
+
------------------------------------------------------------------------*/
|
276
|
+
{
|
277
|
+
int i;
|
278
|
+
char *s;
|
279
|
+
|
280
|
+
s = (char *)out;
|
281
|
+
for(i = 0; i < n; i++){
|
282
|
+
*s++ = *(in+7);
|
283
|
+
*s++ = *(in+6);
|
284
|
+
*s++ = *(in+5);
|
285
|
+
*s++ = *(in+4);
|
286
|
+
*s++ = *(in+3);
|
287
|
+
*s++ = *(in+2);
|
288
|
+
*s++ = *(in+1);
|
289
|
+
*s++ = *in;
|
290
|
+
in += 8;
|
291
|
+
}
|
292
|
+
}
|
293
|
+
#endif
|
294
|
+
#endif
|
295
|
+
|
296
|
+
|
297
|
+
#if defined(unicos)
|
298
|
+
static int words_unicos = 1;
|
299
|
+
#define TWO15 0x8000
|
300
|
+
#define TWO16 0x10000
|
301
|
+
#define TWO31 0x80000000
|
302
|
+
#define TWO32 0x100000000
|
303
|
+
#define HILONG 0xFFFFFFFF00000000
|
304
|
+
#define LOLONG 0x00000000FFFFFFFF
|
305
|
+
#define WORD0 0x000000000000FFFF
|
306
|
+
#define WORD1 0x00000000FFFF0000
|
307
|
+
#define WORD2 0x0000FFFF00000000
|
308
|
+
#define WORD3 0xFFFF000000000000
|
309
|
+
|
310
|
+
/* Masks for IEEE floating format (both hi and lo types). */
|
311
|
+
|
312
|
+
#define IEEE_HISIGN 0x8000000000000000
|
313
|
+
#define IEEE_HIEXPO 0x7F80000000000000
|
314
|
+
#define IEEE_HIMANT 0x007FFFFF00000000
|
315
|
+
#define IEEE_LOSIGN 0x0000000080000000
|
316
|
+
#define IEEE_LOEXPO 0x000000007F800000
|
317
|
+
#define IEEE_LOMANT 0x00000000007FFFFF
|
318
|
+
#define IEEE_DMANT 0x000FFFFFFFFFFFF0
|
319
|
+
#define IEEE_DEXPO 0x7FF0000000000000
|
320
|
+
|
321
|
+
/* Masks for Cray floating format. */
|
322
|
+
|
323
|
+
#define CRAY_MANT 0x0000FFFFFF000000 /* Including unhidden bit. */
|
324
|
+
#define CRAY_MANT1 0x00007FFFFF000000 /* No unhidden bit. */
|
325
|
+
#define CRAY_DMANT 0x0000FFFFFFFFFFFF
|
326
|
+
#define CRAY_DMANT1 0x00007FFFFFFFFFFF
|
327
|
+
#define CRAY_EXPO 0x7FFF000000000000
|
328
|
+
#define SIGN 0x8000000000000000
|
329
|
+
|
330
|
+
/* Mask of a pointer to char giving the character offset in a Cray word. */
|
331
|
+
|
332
|
+
#define CHAR_OFFSET 0xE000000000000000
|
333
|
+
|
334
|
+
/************************************************************************/
|
335
|
+
void pack16_c(int *in,char *out,int n)
|
336
|
+
/*
|
337
|
+
Pack an integer array into 16 bit integers for unicos
|
338
|
+
------------------------------------------------------------------------*/
|
339
|
+
{
|
340
|
+
int temp,offset,*outd,in1,in2,in3,in4,i;
|
341
|
+
|
342
|
+
if(n <= 0)return; /* Return if nothing to do. */
|
343
|
+
temp = (int)out;
|
344
|
+
offset = ( temp & CHAR_OFFSET ) >> 62; /* Get offset of first word. */
|
345
|
+
outd = (int *)(temp & ~CHAR_OFFSET); /* Get address of words. */
|
346
|
+
|
347
|
+
/* Handle the first few which are not aligned on a Cray word. */
|
348
|
+
|
349
|
+
switch(offset){
|
350
|
+
case 1: *outd = (*outd & ~WORD2) | ((*in++ << 32) & WORD2);
|
351
|
+
if(--n == 0)break;
|
352
|
+
case 2: *outd = (*outd & ~WORD1) | ((*in++ << 16) & WORD1);
|
353
|
+
if(--n == 0)break;
|
354
|
+
case 3: *outd = (*outd & ~WORD0) | ((*in++ ) & WORD0);
|
355
|
+
outd++;
|
356
|
+
}
|
357
|
+
|
358
|
+
/* Handle the ones which are aligned on a Cray word. */
|
359
|
+
|
360
|
+
for(i=0; i < n-3; i=i+4){
|
361
|
+
in1 = *in++ << 48;
|
362
|
+
in2 = *in++ << 32;
|
363
|
+
in3 = *in++ << 16;
|
364
|
+
in4 = *in++;
|
365
|
+
*outd++ = (in1 & WORD3) | (in2 & WORD2) | (in3 & WORD1) | (in4 & WORD0);
|
366
|
+
}
|
367
|
+
n -= i;
|
368
|
+
|
369
|
+
/* Handle the last ones which are not aligned on a Cray word. */
|
370
|
+
|
371
|
+
if(n-- > 0){
|
372
|
+
*outd = (*outd & ~WORD3) | ((*in++ << 48) & WORD3);
|
373
|
+
if(n-- > 0){
|
374
|
+
*outd = (*outd & ~WORD2) | ((*in++ << 32) & WORD2);
|
375
|
+
if(n-- > 0){
|
376
|
+
*outd = (*outd & ~WORD1) | ((*in++ << 16) & WORD1);
|
377
|
+
}
|
378
|
+
}
|
379
|
+
}
|
380
|
+
}
|
381
|
+
/************************************************************************/
|
382
|
+
void unpack16_c(char *in,int *out,int n)
|
383
|
+
/*
|
384
|
+
Unpack an array of 16 bit integers into integers for unicos
|
385
|
+
------------------------------------------------------------------------*/
|
386
|
+
{
|
387
|
+
int temp,offset,*ind,i;
|
388
|
+
|
389
|
+
if(n <= 0)return; /* Return if nothing to do. */
|
390
|
+
temp = (int)in;
|
391
|
+
offset = ( temp & CHAR_OFFSET ) >> 62; /* Get offset of first word. */
|
392
|
+
ind = (int *)(temp & ~CHAR_OFFSET); /* Get address of words. */
|
393
|
+
|
394
|
+
/* Handle the first few which are not word aligned. */
|
395
|
+
|
396
|
+
switch(offset){
|
397
|
+
case 1: temp = (*ind >> 32) & WORD0;
|
398
|
+
*out++ = (temp < TWO15 ? temp : temp - TWO16);
|
399
|
+
if(--n == 0) break;
|
400
|
+
case 2: temp = (*ind >> 16) & WORD0;
|
401
|
+
*out++ = (temp < TWO15 ? temp : temp - TWO16);
|
402
|
+
if(--n == 0) break;
|
403
|
+
case 3: temp = (*ind++ ) & WORD0;
|
404
|
+
*out++ = (temp < TWO15 ? temp : temp - TWO16);
|
405
|
+
if(--n == 0) break;
|
406
|
+
}
|
407
|
+
|
408
|
+
/* Handle those that are Cray-word-aligned. */
|
409
|
+
|
410
|
+
for(i=0; i < n-3; i=i+4){
|
411
|
+
temp = (*ind >> 48) & WORD0;
|
412
|
+
*out++ = (temp < TWO15 ? temp : temp - TWO16);
|
413
|
+
temp = (*ind >> 32) & WORD0;
|
414
|
+
*out++ = (temp < TWO15 ? temp : temp - TWO16);
|
415
|
+
temp = (*ind >> 16) & WORD0;
|
416
|
+
*out++ = (temp < TWO15 ? temp : temp - TWO16);
|
417
|
+
temp = (*ind++ ) & WORD0;
|
418
|
+
*out++ = (temp < TWO15 ? temp : temp - TWO16);
|
419
|
+
}
|
420
|
+
n -= i;
|
421
|
+
|
422
|
+
/* Handle the last few which are not Cray-word-aligned. */
|
423
|
+
|
424
|
+
if(n-- > 0){
|
425
|
+
temp = (*ind >> 48) & WORD0;
|
426
|
+
*out++ = (temp < TWO15 ? temp : temp - TWO16);
|
427
|
+
if(n-- > 0){
|
428
|
+
temp = (*ind >> 32) & WORD0;
|
429
|
+
*out++ = (temp < TWO15 ? temp : temp - TWO16);
|
430
|
+
if(n-- > 0){
|
431
|
+
temp = (*ind >> 16) & WORD0;
|
432
|
+
*out++ = (temp < TWO15 ? temp : temp - TWO16);
|
433
|
+
}
|
434
|
+
}
|
435
|
+
}
|
436
|
+
}
|
437
|
+
/************************************************************************/
|
438
|
+
void pack32_c(int *in,char *out,int n)
|
439
|
+
/*
|
440
|
+
Pack an array of integers into 32 bit integers for unicos
|
441
|
+
------------------------------------------------------------------------*/
|
442
|
+
{
|
443
|
+
int temp,offset,*outd,i,in1,in2;
|
444
|
+
|
445
|
+
if(n <= 0)return; /* Return if nothing to do. */
|
446
|
+
temp = (int)out;
|
447
|
+
offset = ( temp & CHAR_OFFSET ) >> 63; /* Get offset of first long. */
|
448
|
+
outd = (int *)(temp & ~CHAR_OFFSET); /* Get address of words. */
|
449
|
+
|
450
|
+
/* Do the first one, if it is not aligned on a Cray word. */
|
451
|
+
|
452
|
+
if(offset==1){
|
453
|
+
*outd = (*outd & ~LOLONG) | (*in++ & LOLONG);
|
454
|
+
outd++;
|
455
|
+
}
|
456
|
+
n -= offset;
|
457
|
+
|
458
|
+
/* Do those which are Cray word aligned. */
|
459
|
+
|
460
|
+
for(i=0; i < n-1; i=i+2){
|
461
|
+
in1 = *in++ << 32;
|
462
|
+
in2 = *in++;
|
463
|
+
*outd++ = (in1 & HILONG) | (in2 & LOLONG);
|
464
|
+
}
|
465
|
+
n -= i;
|
466
|
+
|
467
|
+
/* Handle the last one, if there is one. */
|
468
|
+
|
469
|
+
if(n==1)*outd = (*outd & ~HILONG) | ((*in++ << 32) & HILONG);
|
470
|
+
}
|
471
|
+
/************************************************************************/
|
472
|
+
void unpack32_c(char *in,int *out,int n)
|
473
|
+
/*
|
474
|
+
Unpack an array of 32 bit integers into integers for unicos
|
475
|
+
------------------------------------------------------------------------*/
|
476
|
+
{
|
477
|
+
int temp,offset,*ind,i;
|
478
|
+
|
479
|
+
if(n <= 0)return; /* Return if nothing to do. */
|
480
|
+
temp = (int)in;
|
481
|
+
offset = ( temp & CHAR_OFFSET ) >> 63; /* Get offset of first word. */
|
482
|
+
ind = (int *)(temp & ~CHAR_OFFSET); /* Get address of words. */
|
483
|
+
|
484
|
+
/* Handle one which is not Cray word aligned. */
|
485
|
+
|
486
|
+
if(offset==1){
|
487
|
+
temp = (*ind++ & LOLONG);
|
488
|
+
*out++ = (temp < TWO31 ? temp : temp - TWO32);
|
489
|
+
}
|
490
|
+
n -= offset;
|
491
|
+
|
492
|
+
/* Handle those which are Cray word aligned. */
|
493
|
+
|
494
|
+
for(i=0; i < n-1; i=i+2){
|
495
|
+
temp = (*ind >> 32) & LOLONG;
|
496
|
+
*out++ = (temp < TWO31 ? temp : temp - TWO32);
|
497
|
+
temp = (*ind++ ) & LOLONG;
|
498
|
+
*out++ = (temp < TWO31 ? temp : temp - TWO32);
|
499
|
+
}
|
500
|
+
n -= i;
|
501
|
+
|
502
|
+
/* Possibly handle a last one which is not Cray word aligned. */
|
503
|
+
|
504
|
+
if(n==1){
|
505
|
+
temp = (*ind >> 32) & LOLONG;
|
506
|
+
*out++ = (temp < TWO31 ? temp : temp - TWO32);
|
507
|
+
}
|
508
|
+
}
|
509
|
+
/************************************************************************/
|
510
|
+
void packr_c(float *in,char *out,int n)
|
511
|
+
/*
|
512
|
+
Pack an array of Cray reals into IEEE reals.
|
513
|
+
------------------------------------------------------------------------*/
|
514
|
+
{
|
515
|
+
int temp,offset,*outd,bias,*ind,tin,tout,i;
|
516
|
+
|
517
|
+
if(n <= 0)return; /* Return if nothing to do. */
|
518
|
+
temp = (int)out;
|
519
|
+
offset = ( temp & CHAR_OFFSET ) >> 63; /* Get offset of first long. */
|
520
|
+
outd = (int *)(temp & ~CHAR_OFFSET); /* Get address of words. */
|
521
|
+
bias = (16384 - 126) << 48;
|
522
|
+
ind = (int *)in;
|
523
|
+
|
524
|
+
/* Do the first one, if it is not aligned on a Cray word. */
|
525
|
+
|
526
|
+
if(offset==1){
|
527
|
+
tin = *ind++;
|
528
|
+
*outd = (*outd & ~LOLONG) |
|
529
|
+
(tin & CRAY_MANT ? (((tin & CRAY_EXPO)-bias) >> 25) |
|
530
|
+
((tin & CRAY_MANT1) >> 24) | ((tin & SIGN) >> 32) : 0);
|
531
|
+
outd++;
|
532
|
+
}
|
533
|
+
n -= offset;
|
534
|
+
|
535
|
+
/* Do those which are Cray word aligned. */
|
536
|
+
|
537
|
+
for(i=0; i < n-1; i=i+2){
|
538
|
+
tin = *ind++;
|
539
|
+
tout = (tin & CRAY_MANT ? (((tin & CRAY_EXPO)-bias) << 7) |
|
540
|
+
((tin & CRAY_MANT1) << 8) | (tin & SIGN) : 0);
|
541
|
+
tin = *ind++;
|
542
|
+
*outd++ = tout |
|
543
|
+
(tin & CRAY_MANT ? (((tin & CRAY_EXPO)-bias) >> 25) |
|
544
|
+
((tin & CRAY_MANT1) >> 24) | ((tin & SIGN) >> 32) : 0);
|
545
|
+
}
|
546
|
+
n -= i;
|
547
|
+
|
548
|
+
/* Handle the last one, if there is one. */
|
549
|
+
|
550
|
+
if(n==1){
|
551
|
+
tin = *ind;
|
552
|
+
*outd = (*outd & ~HILONG) |
|
553
|
+
(tin & CRAY_MANT ? (((tin & CRAY_EXPO)-bias) << 7) |
|
554
|
+
((tin & CRAY_MANT1) << 8) | (tin & SIGN) : 0);
|
555
|
+
}
|
556
|
+
}
|
557
|
+
/************************************************************************/
|
558
|
+
void unpackr_c(char *in,float *out,int n)
|
559
|
+
/*
|
560
|
+
Unpack an array of IEEE reals into Cray reals.
|
561
|
+
------------------------------------------------------------------------*/
|
562
|
+
{
|
563
|
+
int temp,tin,*ind,*outd,offset,i,bias;
|
564
|
+
|
565
|
+
if(n <= 0)return; /* Return if nothing to do. */
|
566
|
+
temp = (int)in;
|
567
|
+
offset = ( temp & CHAR_OFFSET ) >> 63; /* Get offset of first word. */
|
568
|
+
ind = (int *)(temp & ~CHAR_OFFSET); /* Get address of words. */
|
569
|
+
outd = (int *)out;
|
570
|
+
bias = ((16384-126) <<48) + (1 << 47);
|
571
|
+
|
572
|
+
/* Handle the first one if it is not aligned on a Cray word. */
|
573
|
+
|
574
|
+
if(offset==1){
|
575
|
+
tin = *ind++;
|
576
|
+
*outd++ = (tin & IEEE_LOEXPO ? (((tin & IEEE_LOEXPO) << 25)+bias) |
|
577
|
+
((tin & IEEE_LOMANT) << 24) | ((tin & IEEE_LOSIGN) << 32) : 0);
|
578
|
+
}
|
579
|
+
n -= offset;
|
580
|
+
|
581
|
+
/* Handle the bulk of them that are aligned on Cray words. */
|
582
|
+
|
583
|
+
for(i=0; i < n-1; i=i+2){
|
584
|
+
tin = *ind++;
|
585
|
+
*outd++ = (tin & IEEE_HIEXPO ? (((tin & IEEE_HIEXPO) >> 7)+bias) |
|
586
|
+
((tin & IEEE_HIMANT) >> 8 ) | (tin & IEEE_HISIGN) : 0);
|
587
|
+
*outd++ = (tin & IEEE_LOEXPO ? (((tin & IEEE_LOEXPO) << 25)+bias) |
|
588
|
+
((tin & IEEE_LOMANT) << 24) | ((tin & IEEE_LOSIGN) << 32) : 0);
|
589
|
+
}
|
590
|
+
n -= i;
|
591
|
+
|
592
|
+
/* Handle the last one, if needed, which is not aligned on a Cray word. */
|
593
|
+
|
594
|
+
if(n==1){
|
595
|
+
tin = *ind;
|
596
|
+
*outd++ = (tin & IEEE_HIEXPO ? (((tin & IEEE_HIEXPO) >> 7)+bias) |
|
597
|
+
((tin & IEEE_HIMANT) >> 8 ) | (tin & IEEE_HISIGN) : 0);
|
598
|
+
}
|
599
|
+
}
|
600
|
+
/************************************************************************/
|
601
|
+
void packd_c(double *in,char *out,int n)
|
602
|
+
/*
|
603
|
+
Pack an array of Cray reals into IEEE double precision. This assumes
|
604
|
+
that a "double" and a "float" are identical.
|
605
|
+
------------------------------------------------------------------------*/
|
606
|
+
{
|
607
|
+
int *ind,*outd,bias,i,tin;
|
608
|
+
|
609
|
+
ind = (int *)in;
|
610
|
+
outd = (int *)out;
|
611
|
+
bias = (16384 - 1022) << 48;
|
612
|
+
|
613
|
+
for(i=0; i < n; i++){
|
614
|
+
tin = *ind++;
|
615
|
+
*outd++ = (tin & CRAY_DMANT ? (tin & SIGN) |
|
616
|
+
(((tin & CRAY_EXPO)-bias) << 4) | ((tin & CRAY_DMANT1) << 5) : 0 );
|
617
|
+
}
|
618
|
+
}
|
619
|
+
/************************************************************************/
|
620
|
+
void unpackd_c(char *in,double *out,int n)
|
621
|
+
/*
|
622
|
+
Unpack an array of IEEE double precision numbers into Cray reals. This
|
623
|
+
assumes that a "double" and a "float" are identical.
|
624
|
+
------------------------------------------------------------------------*/
|
625
|
+
{
|
626
|
+
int *ind,*outd,bias,i,tin;
|
627
|
+
|
628
|
+
ind = (int *)in;
|
629
|
+
outd = (int *)out;
|
630
|
+
bias = ((16384 - 1022) << 48) | (1 << 47);
|
631
|
+
|
632
|
+
for(i=0; i < n; i++){
|
633
|
+
tin = *ind++;
|
634
|
+
*outd++ = (tin & IEEE_DEXPO ? (tin & SIGN) |
|
635
|
+
(((tin & IEEE_DEXPO) >> 4) + bias) | ((tin & IEEE_DMANT) >> 5) : 0 );
|
636
|
+
}
|
637
|
+
}
|
638
|
+
|
639
|
+
#endif
|