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
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
|
/*
*
* ====================================================
* Copyright (C) 1993 by Sun Microsystems, Inc. All rights reserved.
*
* Developed at SunPro, a Sun Microsystems, Inc. business.
* Permission to use, copy, modify, and distribute this
* software is freely granted, provided that this notice
* is preserved.
* ====================================================
*/
/*
* from: @(#)fdlibm.h 5.1 93/09/24
* $FreeBSD$
*/
#ifndef _NPY_MATH_PRIVATE_H_
#define _NPY_MATH_PRIVATE_H_
#include <Python.h>
#ifdef __cplusplus
#include <cmath>
using std::isgreater;
using std::isless;
#else
#include <math.h>
#endif
#include "npy_config.h"
#include "npy_fpmath.h"
#include "numpy/npy_math.h"
#include "numpy/npy_endian.h"
#include "numpy/npy_common.h"
/*
* The original fdlibm code used statements like:
* n0 = ((*(int*)&one)>>29)^1; * index of high word *
* ix0 = *(n0+(int*)&x); * high word of x *
* ix1 = *((1-n0)+(int*)&x); * low word of x *
* to dig two 32 bit words out of the 64 bit IEEE floating point
* value. That is non-ANSI, and, moreover, the gcc instruction
* scheduler gets it wrong. We instead use the following macros.
* Unlike the original code, we determine the endianness at compile
* time, not at run time; I don't see much benefit to selecting
* endianness at run time.
*/
/*
* A union which permits us to convert between a double and two 32 bit
* ints.
*/
/* XXX: not really, but we already make this assumption elsewhere. Will have to
* fix this at some point */
#define IEEE_WORD_ORDER NPY_BYTE_ORDER
#if IEEE_WORD_ORDER == NPY_BIG_ENDIAN
typedef union
{
double value;
struct
{
npy_uint32 msw;
npy_uint32 lsw;
} parts;
} ieee_double_shape_type;
#endif
#if IEEE_WORD_ORDER == NPY_LITTLE_ENDIAN
typedef union
{
double value;
struct
{
npy_uint32 lsw;
npy_uint32 msw;
} parts;
} ieee_double_shape_type;
#endif
/* Get two 32 bit ints from a double. */
#define EXTRACT_WORDS(ix0,ix1,d) \
do { \
ieee_double_shape_type ew_u; \
ew_u.value = (d); \
(ix0) = ew_u.parts.msw; \
(ix1) = ew_u.parts.lsw; \
} while (0)
/* Get the more significant 32 bit int from a double. */
#define GET_HIGH_WORD(i,d) \
do { \
ieee_double_shape_type gh_u; \
gh_u.value = (d); \
(i) = gh_u.parts.msw; \
} while (0)
/* Get the less significant 32 bit int from a double. */
#define GET_LOW_WORD(i,d) \
do { \
ieee_double_shape_type gl_u; \
gl_u.value = (d); \
(i) = gl_u.parts.lsw; \
} while (0)
/* Set the more significant 32 bits of a double from an int. */
#define SET_HIGH_WORD(d,v) \
do { \
ieee_double_shape_type sh_u; \
sh_u.value = (d); \
sh_u.parts.msw = (v); \
(d) = sh_u.value; \
} while (0)
/* Set the less significant 32 bits of a double from an int. */
#define SET_LOW_WORD(d,v) \
do { \
ieee_double_shape_type sl_u; \
sl_u.value = (d); \
sl_u.parts.lsw = (v); \
(d) = sl_u.value; \
} while (0)
/* Set a double from two 32 bit ints. */
#define INSERT_WORDS(d,ix0,ix1) \
do { \
ieee_double_shape_type iw_u; \
iw_u.parts.msw = (ix0); \
iw_u.parts.lsw = (ix1); \
(d) = iw_u.value; \
} while (0)
/*
* A union which permits us to convert between a float and a 32 bit
* int.
*/
typedef union
{
float value;
/* FIXME: Assumes 32 bit int. */
npy_uint32 word;
} ieee_float_shape_type;
/* Get a 32 bit int from a float. */
#define GET_FLOAT_WORD(i,d) \
do { \
ieee_float_shape_type gf_u; \
gf_u.value = (d); \
(i) = gf_u.word; \
} while (0)
/* Set a float from a 32 bit int. */
#define SET_FLOAT_WORD(d,i) \
do { \
ieee_float_shape_type sf_u; \
sf_u.word = (i); \
(d) = sf_u.value; \
} while (0)
#ifdef NPY_USE_C99_COMPLEX
#include <complex.h>
#endif
/*
* Long double support
*/
#if defined(HAVE_LDOUBLE_INTEL_EXTENDED_12_BYTES_LE)
/*
* Intel extended 80 bits precision. Bit representation is
* | junk | s |eeeeeeeeeeeeeee|mmmmmmmm................mmmmmmm|
* | 16 bits| 1 bit | 15 bits | 64 bits |
* | a[2] | a[1] | a[0] |
*
* 16 stronger bits of a[2] are junk
*/
typedef npy_uint32 IEEEl2bitsrep_part;
union IEEEl2bitsrep {
npy_longdouble e;
IEEEl2bitsrep_part a[3];
};
#define LDBL_MANL_INDEX 0
#define LDBL_MANL_MASK 0xFFFFFFFF
#define LDBL_MANL_SHIFT 0
#define LDBL_MANH_INDEX 1
#define LDBL_MANH_MASK 0xFFFFFFFF
#define LDBL_MANH_SHIFT 0
#define LDBL_EXP_INDEX 2
#define LDBL_EXP_MASK 0x7FFF
#define LDBL_EXP_SHIFT 0
#define LDBL_SIGN_INDEX 2
#define LDBL_SIGN_MASK 0x8000
#define LDBL_SIGN_SHIFT 15
#define LDBL_NBIT 0x80000000
typedef npy_uint32 ldouble_man_t;
typedef npy_uint32 ldouble_exp_t;
typedef npy_uint32 ldouble_sign_t;
#elif defined(HAVE_LDOUBLE_INTEL_EXTENDED_16_BYTES_LE)
/*
* Intel extended 80 bits precision, 16 bytes alignment.. Bit representation is
* | junk | s |eeeeeeeeeeeeeee|mmmmmmmm................mmmmmmm|
* | 16 bits| 1 bit | 15 bits | 64 bits |
* | a[2] | a[1] | a[0] |
*
* a[3] and 16 stronger bits of a[2] are junk
*/
typedef npy_uint32 IEEEl2bitsrep_part;
union IEEEl2bitsrep {
npy_longdouble e;
IEEEl2bitsrep_part a[4];
};
#define LDBL_MANL_INDEX 0
#define LDBL_MANL_MASK 0xFFFFFFFF
#define LDBL_MANL_SHIFT 0
#define LDBL_MANH_INDEX 1
#define LDBL_MANH_MASK 0xFFFFFFFF
#define LDBL_MANH_SHIFT 0
#define LDBL_EXP_INDEX 2
#define LDBL_EXP_MASK 0x7FFF
#define LDBL_EXP_SHIFT 0
#define LDBL_SIGN_INDEX 2
#define LDBL_SIGN_MASK 0x8000
#define LDBL_SIGN_SHIFT 15
#define LDBL_NBIT 0x800000000
typedef npy_uint32 ldouble_man_t;
typedef npy_uint32 ldouble_exp_t;
typedef npy_uint32 ldouble_sign_t;
#elif defined(HAVE_LDOUBLE_MOTOROLA_EXTENDED_12_BYTES_BE)
/*
* Motorola extended 80 bits precision. Bit representation is
* | s |eeeeeeeeeeeeeee| junk |mmmmmmmm................mmmmmmm|
* | 1 bit | 15 bits | 16 bits| 64 bits |
* | a[0] | a[1] | a[2] |
*
* 16 low bits of a[0] are junk
*/
typedef npy_uint32 IEEEl2bitsrep_part;
union IEEEl2bitsrep {
npy_longdouble e;
IEEEl2bitsrep_part a[3];
};
#define LDBL_MANL_INDEX 2
#define LDBL_MANL_MASK 0xFFFFFFFF
#define LDBL_MANL_SHIFT 0
#define LDBL_MANH_INDEX 1
#define LDBL_MANH_MASK 0xFFFFFFFF
#define LDBL_MANH_SHIFT 0
#define LDBL_EXP_INDEX 0
#define LDBL_EXP_MASK 0x7FFF0000
#define LDBL_EXP_SHIFT 16
#define LDBL_SIGN_INDEX 0
#define LDBL_SIGN_MASK 0x80000000
#define LDBL_SIGN_SHIFT 31
#define LDBL_NBIT 0x80000000
typedef npy_uint32 ldouble_man_t;
typedef npy_uint32 ldouble_exp_t;
typedef npy_uint32 ldouble_sign_t;
#elif defined(HAVE_LDOUBLE_IEEE_DOUBLE_BE)
/* 64 bits IEEE double precision aligned on 16 bytes: used by ppc arch on
* Mac OS X */
/*
* IEEE double precision. Bit representation is
* | s |eeeeeeeeeee|mmmmmmmm................mmmmmmm|
* |1 bit| 11 bits | 52 bits |
* | a[0] | a[1] |
*/
typedef npy_uint32 IEEEl2bitsrep_part;
union IEEEl2bitsrep {
npy_longdouble e;
IEEEl2bitsrep_part a[2];
};
#define LDBL_MANL_INDEX 1
#define LDBL_MANL_MASK 0xFFFFFFFF
#define LDBL_MANL_SHIFT 0
#define LDBL_MANH_INDEX 0
#define LDBL_MANH_MASK 0x000FFFFF
#define LDBL_MANH_SHIFT 0
#define LDBL_EXP_INDEX 0
#define LDBL_EXP_MASK 0x7FF00000
#define LDBL_EXP_SHIFT 20
#define LDBL_SIGN_INDEX 0
#define LDBL_SIGN_MASK 0x80000000
#define LDBL_SIGN_SHIFT 31
#define LDBL_NBIT 0
typedef npy_uint32 ldouble_man_t;
typedef npy_uint32 ldouble_exp_t;
typedef npy_uint32 ldouble_sign_t;
#elif defined(HAVE_LDOUBLE_IEEE_DOUBLE_LE)
/* 64 bits IEEE double precision, Little Endian. */
/*
* IEEE double precision. Bit representation is
* | s |eeeeeeeeeee|mmmmmmmm................mmmmmmm|
* |1 bit| 11 bits | 52 bits |
* | a[1] | a[0] |
*/
typedef npy_uint32 IEEEl2bitsrep_part;
union IEEEl2bitsrep {
npy_longdouble e;
IEEEl2bitsrep_part a[2];
};
#define LDBL_MANL_INDEX 0
#define LDBL_MANL_MASK 0xFFFFFFFF
#define LDBL_MANL_SHIFT 0
#define LDBL_MANH_INDEX 1
#define LDBL_MANH_MASK 0x000FFFFF
#define LDBL_MANH_SHIFT 0
#define LDBL_EXP_INDEX 1
#define LDBL_EXP_MASK 0x7FF00000
#define LDBL_EXP_SHIFT 20
#define LDBL_SIGN_INDEX 1
#define LDBL_SIGN_MASK 0x80000000
#define LDBL_SIGN_SHIFT 31
#define LDBL_NBIT 0x00000080
typedef npy_uint32 ldouble_man_t;
typedef npy_uint32 ldouble_exp_t;
typedef npy_uint32 ldouble_sign_t;
#elif defined(HAVE_LDOUBLE_IEEE_QUAD_BE)
/*
* IEEE quad precision, Big Endian. Bit representation is
* | s |eeeeeeeeeee|mmmmmmmm................mmmmmmm|
* |1 bit| 15 bits | 112 bits |
* | a[0] | a[1] |
*/
typedef npy_uint64 IEEEl2bitsrep_part;
union IEEEl2bitsrep {
npy_longdouble e;
IEEEl2bitsrep_part a[2];
};
#define LDBL_MANL_INDEX 1
#define LDBL_MANL_MASK 0xFFFFFFFFFFFFFFFF
#define LDBL_MANL_SHIFT 0
#define LDBL_MANH_INDEX 0
#define LDBL_MANH_MASK 0x0000FFFFFFFFFFFF
#define LDBL_MANH_SHIFT 0
#define LDBL_EXP_INDEX 0
#define LDBL_EXP_MASK 0x7FFF000000000000
#define LDBL_EXP_SHIFT 48
#define LDBL_SIGN_INDEX 0
#define LDBL_SIGN_MASK 0x8000000000000000
#define LDBL_SIGN_SHIFT 63
#define LDBL_NBIT 0
typedef npy_uint64 ldouble_man_t;
typedef npy_uint64 ldouble_exp_t;
typedef npy_uint32 ldouble_sign_t;
#elif defined(HAVE_LDOUBLE_IEEE_QUAD_LE)
/*
* IEEE quad precision, Little Endian. Bit representation is
* | s |eeeeeeeeeee|mmmmmmmm................mmmmmmm|
* |1 bit| 15 bits | 112 bits |
* | a[1] | a[0] |
*/
typedef npy_uint64 IEEEl2bitsrep_part;
union IEEEl2bitsrep {
npy_longdouble e;
IEEEl2bitsrep_part a[2];
};
#define LDBL_MANL_INDEX 0
#define LDBL_MANL_MASK 0xFFFFFFFFFFFFFFFF
#define LDBL_MANL_SHIFT 0
#define LDBL_MANH_INDEX 1
#define LDBL_MANH_MASK 0x0000FFFFFFFFFFFF
#define LDBL_MANH_SHIFT 0
#define LDBL_EXP_INDEX 1
#define LDBL_EXP_MASK 0x7FFF000000000000
#define LDBL_EXP_SHIFT 48
#define LDBL_SIGN_INDEX 1
#define LDBL_SIGN_MASK 0x8000000000000000
#define LDBL_SIGN_SHIFT 63
#define LDBL_NBIT 0
typedef npy_uint64 ldouble_man_t;
typedef npy_uint64 ldouble_exp_t;
typedef npy_uint32 ldouble_sign_t;
#endif
#if !defined(HAVE_LDOUBLE_IBM_DOUBLE_DOUBLE_BE) && \
!defined(HAVE_LDOUBLE_IBM_DOUBLE_DOUBLE_LE)
/* Get the sign bit of x. x should be of type IEEEl2bitsrep */
#define GET_LDOUBLE_SIGN(x) \
(((x).a[LDBL_SIGN_INDEX] & LDBL_SIGN_MASK) >> LDBL_SIGN_SHIFT)
/* Set the sign bit of x to v. x should be of type IEEEl2bitsrep */
#define SET_LDOUBLE_SIGN(x, v) \
((x).a[LDBL_SIGN_INDEX] = \
((x).a[LDBL_SIGN_INDEX] & ~LDBL_SIGN_MASK) | \
(((IEEEl2bitsrep_part)(v) << LDBL_SIGN_SHIFT) & LDBL_SIGN_MASK))
/* Get the exp bits of x. x should be of type IEEEl2bitsrep */
#define GET_LDOUBLE_EXP(x) \
(((x).a[LDBL_EXP_INDEX] & LDBL_EXP_MASK) >> LDBL_EXP_SHIFT)
/* Set the exp bit of x to v. x should be of type IEEEl2bitsrep */
#define SET_LDOUBLE_EXP(x, v) \
((x).a[LDBL_EXP_INDEX] = \
((x).a[LDBL_EXP_INDEX] & ~LDBL_EXP_MASK) | \
(((IEEEl2bitsrep_part)(v) << LDBL_EXP_SHIFT) & LDBL_EXP_MASK))
/* Get the manl bits of x. x should be of type IEEEl2bitsrep */
#define GET_LDOUBLE_MANL(x) \
(((x).a[LDBL_MANL_INDEX] & LDBL_MANL_MASK) >> LDBL_MANL_SHIFT)
/* Set the manl bit of x to v. x should be of type IEEEl2bitsrep */
#define SET_LDOUBLE_MANL(x, v) \
((x).a[LDBL_MANL_INDEX] = \
((x).a[LDBL_MANL_INDEX] & ~LDBL_MANL_MASK) | \
(((IEEEl2bitsrep_part)(v) << LDBL_MANL_SHIFT) & LDBL_MANL_MASK))
/* Get the manh bits of x. x should be of type IEEEl2bitsrep */
#define GET_LDOUBLE_MANH(x) \
(((x).a[LDBL_MANH_INDEX] & LDBL_MANH_MASK) >> LDBL_MANH_SHIFT)
/* Set the manh bit of x to v. x should be of type IEEEl2bitsrep */
#define SET_LDOUBLE_MANH(x, v) \
((x).a[LDBL_MANH_INDEX] = \
((x).a[LDBL_MANH_INDEX] & ~LDBL_MANH_MASK) | \
(((IEEEl2bitsrep_part)(v) << LDBL_MANH_SHIFT) & LDBL_MANH_MASK))
#endif /* !HAVE_LDOUBLE_DOUBLE_DOUBLE_* */
/*
* Those unions are used to convert a pointer of npy_cdouble to native C99
* complex or our own complex type independently on whether C99 complex
* support is available
*/
#ifdef NPY_USE_C99_COMPLEX
/*
* Microsoft C defines _MSC_VER
* Intel compiler does not use MSVC complex types, but defines _MSC_VER by
* default.
*/
#if defined(_MSC_VER) && !defined(__INTEL_COMPILER)
typedef union {
npy_cdouble npy_z;
_Dcomplex c99_z;
} __npy_cdouble_to_c99_cast;
typedef union {
npy_cfloat npy_z;
_Fcomplex c99_z;
} __npy_cfloat_to_c99_cast;
typedef union {
npy_clongdouble npy_z;
_Lcomplex c99_z;
} __npy_clongdouble_to_c99_cast;
#else /* !_MSC_VER */
typedef union {
npy_cdouble npy_z;
#ifdef __cplusplus
std::complex<double> c99_z;
#else
complex double c99_z;
#endif
} __npy_cdouble_to_c99_cast;
typedef union {
npy_cfloat npy_z;
#ifdef __cplusplus
std::complex<float> c99_z;
#else
complex float c99_z;
#endif
} __npy_cfloat_to_c99_cast;
typedef union {
npy_clongdouble npy_z;
#ifdef __cplusplus
std::complex<long double> c99_z;
#else
complex long double c99_z;
#endif
} __npy_clongdouble_to_c99_cast;
#endif /* !_MSC_VER */
#else /* !NPY_USE_C99_COMPLEX */
typedef union {
npy_cdouble npy_z;
npy_cdouble c99_z;
} __npy_cdouble_to_c99_cast;
typedef union {
npy_cfloat npy_z;
npy_cfloat c99_z;
} __npy_cfloat_to_c99_cast;
typedef union {
npy_clongdouble npy_z;
npy_clongdouble c99_z;
} __npy_clongdouble_to_c99_cast;
#endif /* !NPY_USE_C99_COMPLEX */
#endif /* !_NPY_MATH_PRIVATE_H_ */
|
