-
Notifications
You must be signed in to change notification settings - Fork 0
/
gmx_fft.h
593 lines (534 loc) · 23 KB
/
gmx_fft.h
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
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
/* -*- mode: c; tab-width: 4; indent-tabs-mode: nil; c-basic-offset: 4; c-file-style: "stroustrup"; -*-
*
*
* Gromacs 4.0 Copyright (c) 1991-2003
* David van der Spoel, Erik Lindahl, University of Groningen.
*
* This program is free software; you can redistribute it and/or
* modify it under the terms of the GNU General Public License
* as published by the Free Software Foundation; either version 2
* of the License, or (at your option) any later version.
*
* To help us fund GROMACS development, we humbly ask that you cite
* the research papers on the package. Check out http://www.gromacs.org
*
* And Hey:
* Gnomes, ROck Monsters And Chili Sauce
*/
#ifndef _GMX_FFT_H_
#define _GMX_FFT_H_
/*! \file gmx_fft.h
* \brief Fast Fourier Transforms.
*
* This file provides an abstract Gromacs interface to Fourier transforms,
* including multi-dimensional and real-to-complex transforms.
*
* Internally it is implemented as wrappers to external libraries such
* as FFTW or the Intel Math Kernel Library, but we also have a built-in
* version of FFTPACK in case the faster alternatives are unavailable.
*
* We also provide our own multi-dimensional transform setups even when
* the underlying library does not support it directly.
*
*/
#include <stdio.h>
#include "types/simple.h"
#include "gmxcomplex.h"
#ifdef __cplusplus
extern "C" {
#endif
#if 0
} /* fixes auto-indentation problems */
#endif
/*! \brief Datatype for FFT setup
*
* The gmx_fft_t type contains all the setup information, e.g. twiddle
* factors, necessary to perform an FFT. Internally it is mapped to
* whatever FFT library we are using, or the built-in FFTPACK if no fast
* external library is available.
*
* Since some of the libraries (e.g. MKL) store work array data in their
* handles this datatype should only be used for one thread at a time, i.e.
* they should allocate one instance each when executing in parallel.
*/
typedef struct gmx_fft *
gmx_fft_t;
/*! \brief Specifier for FFT direction.
*
* The definition of the 1D forward transform from input x[] to output y[] is
* \f[
* y_{k} = \sum_{j=0}^{N-1} x_{j} \exp{-i 2 \pi j k /N}
* \f]
*
* while the corresponding backward transform is
*
* \f[
* y_{k} = \sum_{j=0}^{N-1} x_{j} \exp{i 2 \pi j k /N}
* \f]
*
* A forward-backward transform pair will this result in data scaled by N.
*
* For complex-to-complex transforms you can only use one of
* GMX_FFT_FORWARD or GMX_FFT_BACKWARD, and for real-complex transforms you
* can only use GMX_FFT_REAL_TO_COMPLEX or GMX_FFT_COMPLEX_TO_REAL.
*/
typedef enum gmx_fft_direction
{
GMX_FFT_FORWARD, /*!< Forward complex-to-complex transform */
GMX_FFT_BACKWARD, /*!< Backward complex-to-complex transform */
GMX_FFT_REAL_TO_COMPLEX, /*!< Real-to-complex valued fft */
GMX_FFT_COMPLEX_TO_REAL /*!< Complex-to-real valued fft */
} gmx_fft_direction;
/*! \brief Specifier for FFT flags.
*
* Some FFT libraries (FFTW, in particular) can do timings and other
* tricks to try and optimize the FFT for the current architecture. However,
* this can also lead to results that differ between consecutive runs with
* identical input.
* To avoid this, the conservative flag will attempt to disable such
* optimization, but there are no guarantees since we cannot control what
* the FFT libraries do internally.
*/
typedef int gmx_fft_flag;
static const int GMX_FFT_FLAG_NONE = 0;
static const int GMX_FFT_FLAG_CONSERVATIVE = (1<<0);
/*! \brief Setup a 1-dimensional complex-to-complex transform
*
* \param fft Pointer to opaque Gromacs FFT datatype
* \param nx Length of transform
* \param flags FFT options
*
* \return status - 0 or a standard error message.
*
* \note Since some of the libraries (e.g. MKL) store work array data in their
* handles this datatype should only be used for one thread at a time,
* i.e. you should create one copy per thread when executing in parallel.
*/
int
gmx_fft_init_1d (gmx_fft_t * fft,
int nx,
gmx_fft_flag flags);
/*! \brief Setup multiple 1-dimensional complex-to-complex transform
*
* \param fft Pointer to opaque Gromacs FFT datatype
* \param nx Length of transform
* \param howmany Howmany 1D FFT
* \param flags FFT options
*
* \return status - 0 or a standard error message.
*
* \note Since some of the libraries (e.g. MKL) store work array data in their
* handles this datatype should only be used for one thread at a time,
* i.e. you should create one copy per thread when executing in parallel.
*/
int
gmx_fft_init_many_1d (gmx_fft_t * fft,
int nx,
int howmany,
gmx_fft_flag flags);
/*! \brief Setup a 1-dimensional real-to-complex transform
*
* \param fft Pointer to opaque Gromacs FFT datatype
* \param nx Length of transform in real space
* \param flags FFT options
*
* \return status - 0 or a standard error message.
*
* \note Since some of the libraries (e.g. MKL) store work array data in their
* handles this datatype should only be used for one thread at a time,
* i.e. you should create one copy per thread when executing in parallel.
*/
int
gmx_fft_init_1d_real (gmx_fft_t * fft,
int nx,
gmx_fft_flag flags);
/*! \brief Setup multiple 1-dimensional real-to-complex transform
*
* \param fft Pointer to opaque Gromacs FFT datatype
* \param nx Length of transform in real space
* \param howmany Homany 1D FFTs
* \param flags FFT options
*
* \return status - 0 or a standard error message.
*
* \note Since some of the libraries (e.g. MKL) store work array data in their
* handles this datatype should only be used for one thread at a time,
* i.e. you should create one copy per thread when executing in parallel.
*/
int
gmx_fft_init_many_1d_real (gmx_fft_t * fft,
int nx,
int howmany,
gmx_fft_flag flags);
/*! \brief Setup a 2-dimensional complex-to-complex transform
*
* \param fft Pointer to opaque Gromacs FFT datatype
* \param nx Length of transform in first dimension
* \param ny Length of transform in second dimension
* \param flags FFT options
*
* \return status - 0 or a standard error message.
*
* \note Since some of the libraries (e.g. MKL) store work array data in their
* handles this datatype should only be used for one thread at a time,
* i.e. you should create one copy per thread when executing in parallel.
*/
int
gmx_fft_init_2d (gmx_fft_t * fft,
int nx,
int ny,
gmx_fft_flag flags);
/*! \brief Setup a 2-dimensional real-to-complex transform
*
* \param fft Pointer to opaque Gromacs FFT datatype
* \param nx Length of transform in first dimension
* \param ny Length of transform in second dimension
* \param flags FFT options
*
* The normal space is assumed to be real, while the values in
* frequency space are complex.
*
* \return status - 0 or a standard error message.
*
* \note Since some of the libraries (e.g. MKL) store work array data in their
* handles this datatype should only be used for one thread at a time,
* i.e. you should create one copy per thread when executing in parallel.
*/
int
gmx_fft_init_2d_real (gmx_fft_t * fft,
int nx,
int ny,
gmx_fft_flag flags);
/*! \brief Setup a 3-dimensional complex-to-complex transform
*
* \param fft Pointer to opaque Gromacs FFT datatype
* \param nx Length of transform in first dimension
* \param ny Length of transform in second dimension
* \param nz Length of transform in third dimension
* \param flags FFT options
*
* \return status - 0 or a standard error message.
*
* \note Since some of the libraries (e.g. MKL) store work array data in their
* handles this datatype should only be used for one thread at a time,
* i.e. you should create one copy per thread when executing in parallel.
*/
int
gmx_fft_init_3d (gmx_fft_t * fft,
int nx,
int ny,
int nz,
gmx_fft_flag flags);
/*! \brief Setup a 3-dimensional real-to-complex transform
*
* \param fft Pointer to opaque Gromacs FFT datatype
* \param nx Length of transform in first dimension
* \param ny Length of transform in second dimension
* \param nz Length of transform in third dimension
* \param flags FFT options
*
* The normal space is assumed to be real, while the values in
* frequency space are complex.
*
* \return status - 0 or a standard error message.
*
* \note Since some of the libraries (e.g. MKL) store work array data in their
* handles this datatype should only be used for one thread at a time,
* i.e. you should create one copy per thread when executing in parallel.
*/
int
gmx_fft_init_3d_real (gmx_fft_t * fft,
int nx,
int ny,
int nz,
gmx_fft_flag flags);
/*! \brief Perform a 1-dimensional complex-to-complex transform
*
* Performs an instance of a transform previously initiated.
*
* \param setup Setup returned from gmx_fft_init_1d()
* \param dir Forward or Backward
* \param in_data Input grid data. This should be allocated with gmx_new()
* to make it 16-byte aligned for better performance.
* \param out_data Output grid data. This should be allocated with gmx_new()
* to make it 16-byte aligned for better performance.
* You can provide the same pointer for in_data and out_data
* to perform an in-place transform.
*
* \return 0 on success, or an error code.
*
* \note Data pointers are declared as void, to avoid casting pointers
* depending on your grid type.
*/
int
gmx_fft_1d (gmx_fft_t setup,
enum gmx_fft_direction dir,
void * in_data,
void * out_data);
/*! \brief Perform many 1-dimensional complex-to-complex transforms
*
* Performs many instances of a transform previously initiated.
*
* \param setup Setup returned from gmx_fft_init_1d()
* \param dir Forward or Backward
* \param in_data Input grid data. This should be allocated with gmx_new()
* to make it 16-byte aligned for better performance.
* \param out_data Output grid data. This should be allocated with gmx_new()
* to make it 16-byte aligned for better performance.
* You can provide the same pointer for in_data and out_data
* to perform an in-place transform.
*
* \return 0 on success, or an error code.
*
* \note Data pointers are declared as void, to avoid casting pointers
* depending on your grid type.
*/
int
gmx_fft_many_1d (gmx_fft_t setup,
enum gmx_fft_direction dir,
void * in_data,
void * out_data);
/*! \brief Perform a 1-dimensional real-to-complex transform
*
* Performs an instance of a transform previously initiated.
*
* \param setup Setup returned from gmx_fft_init_1d_real()
* \param dir Real-to-complex or complex-to-real
* \param in_data Input grid data. This should be allocated with gmx_new()
* to make it 16-byte aligned for better performance.
* \param out_data Output grid data. This should be allocated with gmx_new()
* to make it 16-byte aligned for better performance.
* You can provide the same pointer for in_data and out_data
* to perform an in-place transform.
*
* If you are doing an in-place transform, the array must be padded up to
* an even integer length so n/2 complex numbers can fit. Out-of-place arrays
* should not be padded (although it doesn't matter in 1d).
*
* \return 0 on success, or an error code.
*
* \note Data pointers are declared as void, to avoid casting pointers
* depending on transform direction.
*/
int
gmx_fft_1d_real (gmx_fft_t setup,
enum gmx_fft_direction dir,
void * in_data,
void * out_data);
/*! \brief Perform many 1-dimensional real-to-complex transforms
*
* Performs many instances of a transform previously initiated.
*
* \param setup Setup returned from gmx_fft_init_1d_real()
* \param dir Real-to-complex or complex-to-real
* \param in_data Input grid data. This should be allocated with gmx_new()
* to make it 16-byte aligned for better performance.
* \param out_data Output grid data. This should be allocated with gmx_new()
* to make it 16-byte aligned for better performance.
* You can provide the same pointer for in_data and out_data
* to perform an in-place transform.
*
* If you are doing an in-place transform, the array must be padded up to
* an even integer length so n/2 complex numbers can fit. Out-of-place arrays
* should not be padded (although it doesn't matter in 1d).
*
* \return 0 on success, or an error code.
*
* \note Data pointers are declared as void, to avoid casting pointers
* depending on transform direction.
*/
int
gmx_fft_many_1d_real (gmx_fft_t setup,
enum gmx_fft_direction dir,
void * in_data,
void * out_data);
/*! \brief Perform a 2-dimensional complex-to-complex transform
*
* Performs an instance of a transform previously initiated.
*
* \param setup Setup returned from gmx_fft_init_1d()
* \param dir Forward or Backward
* \param in_data Input grid data. This should be allocated with gmx_new()
* to make it 16-byte aligned for better performance.
* \param out_data Output grid data. This should be allocated with gmx_new()
* to make it 16-byte aligned for better performance.
* You can provide the same pointer for in_data and out_data
* to perform an in-place transform.
*
* \return 0 on success, or an error code.
*
* \note Data pointers are declared as void, to avoid casting pointers
* depending on your grid type.
*/
int
gmx_fft_2d (gmx_fft_t setup,
enum gmx_fft_direction dir,
void * in_data,
void * out_data);
/*! \brief Perform a 2-dimensional real-to-complex transform
*
* Performs an instance of a transform previously initiated.
*
* \param setup Setup returned from gmx_fft_init_1d_real()
* \param dir Real-to-complex or complex-to-real
* \param in_data Input grid data. This should be allocated with gmx_new()
* to make it 16-byte aligned for better performance.
* \param out_data Output grid data. This should be allocated with gmx_new()
* to make it 16-byte aligned for better performance.
* You can provide the same pointer for in_data and out_data
* to perform an in-place transform.
*
* \return 0 on success, or an error code.
*
* \note If you are doing an in-place transform, the last dimension of the
* array MUST be padded up to an even integer length so n/2 complex numbers can
* fit. Thus, if the real grid e.g. has dimension 5*3, you must allocate it as
* a 5*4 array, where the last element in the second dimension is padding.
* The complex data will be written to the same array, but since that dimension
* is 5*2 it will now fill the entire array. Reverse complex-to-real in-place
* transformation will produce the same sort of padded array.
*
* The padding does NOT apply to out-of-place transformation. In that case the
* input array will simply be 5*3 of real, while the output is 5*2 of complex.
*
* \note Data pointers are declared as void, to avoid casting pointers
* depending on transform direction.
*/
int
gmx_fft_2d_real (gmx_fft_t setup,
enum gmx_fft_direction dir,
void * in_data,
void * out_data);
/*! \brief Perform a 3-dimensional complex-to-complex transform
*
* Performs an instance of a transform previously initiated.
*
* \param setup Setup returned from gmx_fft_init_1d()
* \param dir Forward or Backward
* \param in_data Input grid data. This should be allocated with gmx_new()
* to make it 16-byte aligned for better performance.
* \param out_data Output grid data. This should be allocated with gmx_new()
* to make it 16-byte aligned for better performance.
* You can provide the same pointer for in_data and out_data
* to perform an in-place transform.
*
* \return 0 on success, or an error code.
*
* \note Data pointers are declared as void, to avoid casting pointers
* depending on your grid type.
*/
int
gmx_fft_3d (gmx_fft_t setup,
enum gmx_fft_direction dir,
void * in_data,
void * out_data);
/*! \brief Perform a 3-dimensional real-to-complex transform
*
* Performs an instance of a transform previously initiated.
*
* \param setup Setup returned from gmx_fft_init_1d_real()
* \param dir Real-to-complex or complex-to-real
* \param in_data Input grid data. This should be allocated with gmx_new()
* to make it 16-byte aligned for better performance.
* \param out_data Output grid data. This should be allocated with gmx_new()
* to make it 16-byte aligned for better performance.
* You can provide the same pointer for in_data and out_data
* to perform an in-place transform.
*
* \return 0 on success, or an error code.
*
* \note If you are doing an in-place transform, the last dimension of the
* array MUST be padded up to an even integer length so n/2 complex numbers can
* fit. Thus, if the real grid e.g. has dimension 7*5*3, you must allocate it as
* a 7*5*4 array, where the last element in the second dimension is padding.
* The complex data will be written to the same array, but since that dimension
* is 7*5*2 it will now fill the entire array. Reverse complex-to-real in-place
* transformation will produce the same sort of padded array.
*
* The padding does NOT apply to out-of-place transformation. In that case the
* input will simply be 7*5*3 of real, while the output is 7*5*2 of complex.
*
* \note Data pointers are declared as void, to avoid casting pointers
* depending on transform direction.
*/
int
gmx_fft_3d_real (gmx_fft_t setup,
enum gmx_fft_direction dir,
void * in_data,
void * out_data);
/*! \brief Release an FFT setup structure
*
* Destroy setup and release all allocated memory.
*
* \param setup Setup returned from gmx_fft_init_1d(), or one
* of the other initializers.
*
*/
void
gmx_fft_destroy (gmx_fft_t setup);
/*! \brief Release a many FFT setup structure
*
* Destroy setup and release all allocated memory.
*
* \param setup Setup returned from gmx_fft_init_1d(), or one
* of the other initializers.
*
*/
void
gmx_many_fft_destroy (gmx_fft_t setup);
/*! \brief Transpose 2d complex matrix, in-place or out-of-place.
*
* This routines works when the matrix is non-square, i.e. nx!=ny too,
* without allocating an entire matrix of work memory, which is important
* for huge FFT grids.
*
* \param in_data Input data, to be transposed
* \param out_data Output, transposed data. If this is identical to
* in_data, an in-place transpose is performed.
* \param nx Number of rows before transpose
* \param ny Number of columns before transpose
*
* \return GMX_SUCCESS, or an error code from gmx_errno.h
*/
int
gmx_fft_transpose_2d (t_complex * in_data,
t_complex * out_data,
int nx,
int ny);
/*! \brief Transpose 2d multi-element matrix
*
* This routine is very similar to gmx_fft_transpose_2d(), but it
* supports matrices with more than one data value for each position.
* It is extremely useful when transposing the x/y dimensions of a 3d
* matrix - in that case you just set nelem to nz, and the routine will do
* and x/y transpose where it moves entire columns of z data
*
* This routines works when the matrix is non-square, i.e. nx!=ny too,
* without allocating an entire matrix of work memory, which is important
* for huge FFT grid.
*
* For performance reasons you need to provide a \a small workarray
* with length at least 2*nelem (note that the type is char, not t_complex).
*
* \param in_data Input data, to be transposed
* \param out_data Output, transposed data. If this is identical to
* in_data, an in-place transpose is performed.
* \param nx Number of rows before transpose
* \param ny Number of columns before transpose
* \param nelem Number of t_complex values in each position. If this
* is 1 it is faster to use gmx_fft_transpose_2d() directly.
* \param work Work array of length 2*nelem, type t_complex.
*
* \return GMX_SUCCESS, or an error code from gmx_errno.h
*/
int
gmx_fft_transpose_2d_nelem(t_complex * in_data,
t_complex * out_data,
int nx,
int ny,
int nelem,
t_complex * work);
#ifdef __cplusplus
}
#endif
#endif /* _GMX_FFT_H_ */