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
|
/*
* Copyright (c) 1997-1999, 2003 Massachusetts Institute of Technology
*
* 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.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
*
*/
#include <stdlib.h>
#include "fftw.h"
#include "mex.h"
/**************************************************************************/
/* MEX programs need to use special memory allocation routines,
so we use the hooks provided by FFTW to ensure that MEX
allocation is used: */
void *fftw_mex_malloc_hook(size_t n)
{
void *buf;
buf = mxMalloc(n);
/* Call this routine so that we can retain allocations and
data between calls to the FFTW MEX: */
mexMakeMemoryPersistent(buf);
return buf;
}
void fftw_mex_free_hook(void *buf)
{
mxFree(buf);
}
void install_fftw_hooks(void)
{
fftw_malloc_hook = fftw_mex_malloc_hook;
fftw_free_hook = fftw_mex_free_hook;
}
/**************************************************************************/
/* We retain various information between calls to the FFTW MEX in
order to maximize performance. (Reusing plans, data, and
allocated blocks where possible.) This information is referenced
by the following variables, which are initialized in the function
initialize_fftw_mex_data. */
#define MAX_RANK 10
int first_call = 1; /* 1 if this is the first call to the FFTW MEX,
and nothing has been initialized yet. 0 otherwise. */
/* Keep track of the array dimensions that stored data (below) is for.
When these dimensions changed, we have to recompute the plans,
work arrays, etc. */
int cur_rank = 0, /* rank of the array */
cur_dims[MAX_RANK], /* dimensions */
cur_N; /* product of the dimensions */
/* Work arrays. MATLAB stores complex numbers as separate real/imag.
arrays, so we have to translate into our format before the FFT.
In case allocation is slow, we retain these work arrays between
calls so that they can be reused. */
fftw_complex *input_work = 0, *output_work = 0;
/* The number of floating point operations required for the FFT.
(Starting with FFTW 1.3, an exact count is computed by the planner.)
This is used to update MATLAB's flops count. */
int fftw_mex_flops = 0, ifftw_mex_flops = 0;
/* Plans. These are computed once and then reused as long as the
dimensions of the array don't changed. At any point in time,
at most two plans are cached: a forward and backwards plan,
either for one- or multi-dimensional transforms. */
fftw_plan p = 0, ip = 0;
fftwnd_plan pnd = 0, ipnd = 0;
/**************************************************************************/
int compute_fftw_mex_flops(fftw_direction dir)
{
#ifdef FFTW_HAS_COUNT_PLAN_OPS /* this feature will be in FFTW 1.3 */
fftw_op_count ops;
if (dir == FFTW_FORWARD) {
if (cur_rank == 1)
fftw_count_plan_ops(p,&ops);
else
fftwnd_count_plan_ops(pnd,&ops);
}
else {
if (cur_rank == 1)
fftw_count_plan_ops(ip,&ops);
else
fftwnd_count_plan_ops(ipnd,&ops);
}
return (ops.fp_additions + ops.fp_multiplications);
#else
return 0;
#endif
}
/**************************************************************************/
/* The following functions destroy and/or initialize the data that
FFTW-MEX caches between calls. */
void destroy_fftw_mex_data(void) {
if (output_work != input_work)
fftw_free(output_work);
if (input_work)
fftw_free(input_work);
if (p)
fftw_destroy_plan(p);
if (pnd)
fftwnd_destroy_plan(pnd);
if (ip)
fftw_destroy_plan(ip);
if (ipnd)
fftwnd_destroy_plan(ipnd);
cur_rank = 0;
input_work = output_work = 0;
ip = p = 0;
ipnd = pnd = 0;
}
/* This function is called when MATLAB exits or the MEX file is
cleared, in which case we want to dispose of all data and
free any allocated blocks. */
void fftw_mex_exit_function(void)
{
if (!first_call) {
destroy_fftw_mex_data();
fftw_forget_wisdom();
first_call = 1;
}
}
#define MAGIC(x) #x
#define STRINGIZE(x) MAGIC(x)
/* Initialize the cached data each time the MEX file is called. First,
we check if we have previously computed plans and data for these
array dimensions. Only if the dimensions have changed since the
last call must we recompute the plans, etc. */
void initialize_fftw_mex_data(int rank, const int *dims, fftw_direction dir)
{
int new_plan = 0;
if (first_call) {
/* The following things need only be done once: */
install_fftw_hooks();
mexAtExit(fftw_mex_exit_function);
first_call = 0;
}
if (rank == 1) {
if (cur_rank != 1 || cur_dims[0] != dims[0]) {
destroy_fftw_mex_data();
cur_rank = 1;
cur_dims[0] = cur_N = dims[0];
input_work = (fftw_complex*)fftw_malloc(sizeof(fftw_complex) * cur_N);
output_work = (fftw_complex*)fftw_malloc(sizeof(fftw_complex) * cur_N);
new_plan = 1;
}
else if (dir == FFTW_FORWARD && !p ||
dir == FFTW_BACKWARD && !ip)
new_plan = 1;
if (new_plan) {
if (dir == FFTW_FORWARD) {
p = fftw_create_plan(cur_N,dir,
FFTW_MEASURE | FFTW_USE_WISDOM);
fftw_mex_flops = compute_fftw_mex_flops(dir);
}
else {
ip = fftw_create_plan(cur_N,dir,
FFTW_MEASURE | FFTW_USE_WISDOM);
ifftw_mex_flops = compute_fftw_mex_flops(dir);
}
}
}
else {
int same_dims = 1, dim;
if (cur_rank == rank)
for (dim = 0; dim < rank && same_dims; ++dim)
same_dims = (cur_dims[dim] == dims[rank-1-dim]);
else
same_dims = 0;
if (!same_dims) {
if (rank > MAX_RANK)
mexErrMsgTxt("Sorry, dimensionality > " STRINGIZE(MAX_RANK)
" is not supported.");
destroy_fftw_mex_data();
cur_rank = rank;
cur_N = 1;
for (dim = 0; dim < rank; ++dim)
cur_N *= (cur_dims[dim] = dims[rank-1-dim]);
input_work = (fftw_complex*)fftw_malloc(sizeof(fftw_complex) * cur_N);
output_work = input_work;
new_plan = 1;
}
else if (dir == FFTW_FORWARD && !pnd ||
dir == FFTW_BACKWARD && !ipnd)
new_plan = 1;
if (new_plan) {
if (dir == FFTW_FORWARD) {
pnd = fftwnd_create_plan(rank,cur_dims,dir,
FFTW_IN_PLACE |
FFTW_MEASURE | FFTW_USE_WISDOM);
fftw_mex_flops = compute_fftw_mex_flops(dir);
}
else {
ipnd = fftwnd_create_plan(rank,cur_dims,dir,
FFTW_IN_PLACE |
FFTW_MEASURE | FFTW_USE_WISDOM);
ifftw_mex_flops = compute_fftw_mex_flops(dir);
}
}
}
}
/**************************************************************************/
/* MATLAB stores complex numbers as separate arrays for real and
imaginary parts. The following functions take the data in
this format and pack it into a fftw_complex work array, or
unpack it, respectively. The globals input_work and output_work
are used as the arrays to pack to/unpack from.*/
void pack_input_work(double *input_re, double *input_im)
{
int i;
if (input_im)
for (i = 0; i < cur_N; ++i) {
c_re(input_work[i]) = input_re[i];
c_im(input_work[i]) = input_im[i];
}
else
for (i = 0; i < cur_N; ++i) {
c_re(input_work[i]) = input_re[i];
c_im(input_work[i]) = 0.0;
}
}
void unpack_output_work(double *output_re, double *output_im)
{
int i;
for (i = 0; i < cur_N; ++i) {
output_re[i] = c_re(output_work[i]);
output_im[i] = c_im(output_work[i]);
}
}
/**************************************************************************/
/* The following function is called by MATLAB when the FFTW
MEX is invoked from within the program.
The rhs parameters are the list of arrays on the right-hand-side
(rhs) of the MATLAB command--the arguments to FFTW. The lhs
parameters are the list of arrays on the left-hand-side (lhs) of
the MATLAB command--these are what the output(s) of FFTW are
assigned to.
The syntax for the FFTW call in MATLAB is fftw(array,sign),
as described in fftw.m */
void mexFunction(int nlhs, mxArray *plhs[],
int nrhs, const mxArray *prhs[])
{
int rank;
const int *dims;
int m, n; /* Array is m x n, C-ordered */
fftw_direction dir;
if (nrhs != 2)
mexErrMsgTxt("Two input arguments are expected.");
if (!mxIsDouble(prhs[0]))
mexErrMsgTxt("First input must be a double precision matrix.");
if (mxIsSparse(prhs[0]))
mexErrMsgTxt("Sorry, sparse matrices are not currently supported.");
if (mxGetM(prhs[1]) * mxGetN(prhs[1]) != 1)
mexErrMsgTxt("Second input must be a scalar (+/- 1).");
if (mxGetScalar(prhs[1]) > 0.0)
dir = FFTW_BACKWARD;
else
dir = FFTW_FORWARD;
if ((rank = mxGetNumberOfDimensions(prhs[0])) == 2) {
int dims2[2];
m = mxGetM(prhs[0]);
n = mxGetN(prhs[0]);
if (m == 1 || n == 1) {
dims2[0] = m * n;
initialize_fftw_mex_data(1,dims2,dir);
}
else {
dims2[0] = m;
dims2[1] = n;
initialize_fftw_mex_data(2,dims2,dir);
}
}
else
initialize_fftw_mex_data(rank,dims = mxGetDimensions(prhs[0]),dir);
pack_input_work(mxGetPr(prhs[0]),mxGetPi(prhs[0]));
if (dir == FFTW_FORWARD) {
if (cur_rank == 1)
fftw(p,1, input_work,1,0, output_work,1,0);
else
fftwnd(pnd,1, input_work,1,0, 0,0,0);
mexAddFlops(fftw_mex_flops);
}
else {
if (cur_rank == 1)
fftw(ip,1, input_work,1,0, output_work,1,0);
else
fftwnd(ipnd,1, input_work,1,0, 0,0,0);
mexAddFlops(ifftw_mex_flops);
}
/* Create a matrix for the return argument. */
if (cur_rank <= 2)
plhs[0] = mxCreateDoubleMatrix(m, n, mxCOMPLEX);
else
plhs[0] = mxCreateNumericArray(rank,dims,
mxDOUBLE_CLASS,mxCOMPLEX);
unpack_output_work(mxGetPr(plhs[0]),mxGetPi(plhs[0]));
}
|