Coverage Report

Created: 2026-09-01 06:11

next uncovered line (L), next uncovered region (R), next uncovered branch (B)
/src/fftw3/dft/buffered.c
Line
Count
Source
1
/*
2
 * Copyright (c) 2003, 2007-14 Matteo Frigo
3
 * Copyright (c) 2003, 2007-14 Massachusetts Institute of Technology
4
 *
5
 * This program is free software; you can redistribute it and/or modify
6
 * it under the terms of the GNU General Public License as published by
7
 * the Free Software Foundation; either version 2 of the License, or
8
 * (at your option) any later version.
9
 *
10
 * This program is distributed in the hope that it will be useful,
11
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
12
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
13
 * GNU General Public License for more details.
14
 *
15
 * You should have received a copy of the GNU General Public License
16
 * along with this program; if not, write to the Free Software
17
 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA  02110-1301  USA
18
 *
19
 */
20
21
22
#include "dft/dft.h"
23
24
typedef struct {
25
     solver super;
26
     size_t maxnbuf_ndx;
27
} S;
28
29
static const INT maxnbufs[] = { 8, 256 };
30
31
typedef struct {
32
     plan_dft super;
33
34
     plan *cld, *cldcpy, *cldrest;
35
     INT n, vl, nbuf, bufdist;
36
     INT ivs_by_nbuf, ovs_by_nbuf;
37
     INT roffset, ioffset;
38
} P;
39
40
/* transform a vector input with the help of bufs */
41
static void apply(const plan *ego_, R *ri, R *ii, R *ro, R *io)
42
67
{
43
67
     const P *ego = (const P *) ego_;
44
67
     INT nbuf = ego->nbuf;
45
67
     R *bufs = (R *)MALLOC(sizeof(R) * nbuf * ego->bufdist * 2, BUFFERS);
46
47
67
     plan_dft *cld = (plan_dft *) ego->cld;
48
67
     plan_dft *cldcpy = (plan_dft *) ego->cldcpy;
49
67
     plan_dft *cldrest;
50
67
     INT i, vl = ego->vl;
51
67
     INT ivs_by_nbuf = ego->ivs_by_nbuf, ovs_by_nbuf = ego->ovs_by_nbuf;
52
67
     INT roffset = ego->roffset, ioffset = ego->ioffset;
53
54
134
     for (i = nbuf; i <= vl; i += nbuf) {
55
          /* transform to bufs: */
56
67
          cld->apply((plan *) cld, ri, ii, bufs + roffset, bufs + ioffset);
57
67
    ri += ivs_by_nbuf; ii += ivs_by_nbuf;
58
59
          /* copy back */
60
67
          cldcpy->apply((plan *) cldcpy, bufs+roffset, bufs+ioffset, ro, io);
61
67
    ro += ovs_by_nbuf; io += ovs_by_nbuf;
62
67
     }
63
64
67
     X(ifree)(bufs);
65
66
     /* Do the remaining transforms, if any: */
67
67
     cldrest = (plan_dft *) ego->cldrest;
68
67
     cldrest->apply((plan *) cldrest, ri, ii, ro, io);
69
67
}
70
71
72
static void awake(plan *ego_, enum wakefulness wakefulness)
73
54
{
74
54
     P *ego = (P *) ego_;
75
76
54
     X(plan_awake)(ego->cld, wakefulness);
77
54
     X(plan_awake)(ego->cldcpy, wakefulness);
78
54
     X(plan_awake)(ego->cldrest, wakefulness);
79
54
}
80
81
static void destroy(plan *ego_)
82
201
{
83
201
     P *ego = (P *) ego_;
84
201
     X(plan_destroy_internal)(ego->cldrest);
85
201
     X(plan_destroy_internal)(ego->cldcpy);
86
201
     X(plan_destroy_internal)(ego->cld);
87
201
}
88
89
static void print(const plan *ego_, printer *p)
90
0
{
91
0
     const P *ego = (const P *) ego_;
92
0
     p->print(p, "(dft-buffered-%D%v/%D-%D%(%p%)%(%p%)%(%p%))",
93
0
              ego->n, ego->nbuf,
94
0
              ego->vl, ego->bufdist % ego->n,
95
0
              ego->cld, ego->cldcpy, ego->cldrest);
96
0
}
97
98
static int applicable0(const S *ego, const problem *p_, const planner *plnr)
99
2.23k
{
100
2.23k
     const problem_dft *p = (const problem_dft *) p_;
101
2.23k
     const iodim *d = p->sz->dims;
102
103
2.23k
     if (1
104
2.23k
   && p->vecsz->rnk <= 1
105
1.21k
   && p->sz->rnk == 1
106
2.23k
    ) {
107
1.18k
    INT vl, ivs, ovs;
108
1.18k
    X(tensor_tornk1)(p->vecsz, &vl, &ivs, &ovs);
109
110
1.18k
    if (X(toobig)(p->sz->dims[0].n) && CONSERVE_MEMORYP(plnr))
111
0
         return 0;
112
113
    /* if this solver is redundant, in the sense that a solver
114
       of lower index generates the same plan, then prune this
115
       solver */
116
1.18k
    if (X(nbuf_redundant)(d[0].n, vl, 
117
1.18k
        ego->maxnbuf_ndx,
118
1.18k
        maxnbufs, NELEM(maxnbufs)))
119
432
         return 0;
120
121
    /*
122
      In principle, the buffered transforms might be useful
123
      when working out of place.  However, in order to
124
      prevent infinite loops in the planner, we require
125
      that the output stride of the buffered transforms be
126
      greater than 2.
127
    */
128
748
    if (p->ri != p->ro)
129
440
         return (d[0].os > 2);
130
131
    /*
132
     * If the problem is in place, the input/output strides must
133
     * be the same or the whole thing must fit in the buffer.
134
     */
135
308
    if (X(tensor_inplace_strides2)(p->sz, p->vecsz))
136
0
         return 1;
137
138
308
    if (/* fits into buffer: */
139
308
         ((p->vecsz->rnk == 0)
140
308
    ||
141
308
    (X(nbuf)(d[0].n, p->vecsz->dims[0].n, 
142
308
       maxnbufs[ego->maxnbuf_ndx]) 
143
308
     == p->vecsz->dims[0].n)))
144
201
         return 1;
145
308
     }
146
147
1.15k
     return 0;
148
2.23k
}
149
150
static int applicable(const S *ego, const problem *p_, const planner *plnr)
151
2.23k
{
152
2.23k
     if (NO_BUFFERINGP(plnr)) return 0;
153
2.23k
     if (!applicable0(ego, p_, plnr)) return 0;
154
155
201
     if (NO_UGLYP(plnr)) {
156
201
    const problem_dft *p = (const problem_dft *) p_;
157
201
    if (p->ri != p->ro) return 0;
158
201
    if (X(toobig)(p->sz->dims[0].n)) return 0;
159
201
     }
160
201
     return 1;
161
201
}
162
163
static plan *mkplan(const solver *ego_, const problem *p_, planner *plnr)
164
2.23k
{
165
2.23k
     P *pln;
166
2.23k
     const S *ego = (const S *)ego_;
167
2.23k
     plan *cld = (plan *) 0;
168
2.23k
     plan *cldcpy = (plan *) 0;
169
2.23k
     plan *cldrest = (plan *) 0;
170
2.23k
     const problem_dft *p = (const problem_dft *) p_;
171
2.23k
     R *bufs = (R *) 0;
172
2.23k
     INT nbuf = 0, bufdist, n, vl;
173
2.23k
     INT ivs, ovs, roffset, ioffset;
174
175
2.23k
     static const plan_adt padt = {
176
2.23k
    X(dft_solve), awake, print, destroy
177
2.23k
     };
178
179
2.23k
     if (!applicable(ego, p_, plnr))
180
2.03k
          goto nada;
181
182
201
     n = X(tensor_sz)(p->sz);
183
184
201
     X(tensor_tornk1)(p->vecsz, &vl, &ivs, &ovs);
185
186
201
     nbuf = X(nbuf)(n, vl, maxnbufs[ego->maxnbuf_ndx]);
187
201
     bufdist = X(bufdist)(n, vl);
188
201
     A(nbuf > 0);
189
190
     /* attempt to keep real and imaginary part in the same order,
191
  so as to allow optimizations in the the copy plan */
192
201
     roffset = (p->ri - p->ii > 0) ? (INT)1 : (INT)0;
193
201
     ioffset = 1 - roffset;
194
195
     /* initial allocation for the purpose of planning */
196
201
     bufs = (R *) MALLOC(sizeof(R) * nbuf * bufdist * 2, BUFFERS);
197
198
     /* allow destruction of input if problem is in place */
199
201
     cld = X(mkplan_f_d)(plnr,
200
201
       X(mkproblem_dft_d)(
201
201
            X(mktensor_1d)(n, p->sz->dims[0].is, 2),
202
201
            X(mktensor_1d)(nbuf, ivs, bufdist * 2),
203
201
            TAINT(p->ri, ivs * nbuf),
204
201
            TAINT(p->ii, ivs * nbuf),
205
201
            bufs + roffset, 
206
201
            bufs + ioffset),
207
201
       0, 0, (p->ri == p->ro) ? NO_DESTROY_INPUT : 0);
208
201
     if (!cld)
209
0
          goto nada;
210
211
     /* copying back from the buffer is a rank-0 transform: */
212
201
     cldcpy = X(mkplan_d)(plnr,
213
201
        X(mkproblem_dft_d)(
214
201
             X(mktensor_0d)(),
215
201
             X(mktensor_2d)(nbuf, bufdist * 2, ovs,
216
201
                n, 2, p->sz->dims[0].os),
217
201
             bufs + roffset, 
218
201
             bufs + ioffset, 
219
201
             TAINT(p->ro, ovs * nbuf), 
220
201
             TAINT(p->io, ovs * nbuf)));
221
201
     if (!cldcpy)
222
0
          goto nada;
223
224
     /* deallocate buffers, let apply() allocate them for real */
225
201
     X(ifree)(bufs);
226
201
     bufs = 0;
227
228
     /* plan the leftover transforms (cldrest): */
229
201
     {
230
201
    INT id = ivs * (nbuf * (vl / nbuf));
231
201
    INT od = ovs * (nbuf * (vl / nbuf));
232
201
    cldrest = X(mkplan_d)(plnr, 
233
201
        X(mkproblem_dft_d)(
234
201
             X(tensor_copy)(p->sz),
235
201
             X(mktensor_1d)(vl % nbuf, ivs, ovs),
236
201
             p->ri+id, p->ii+id, p->ro+od, p->io+od));
237
201
     }
238
201
     if (!cldrest)
239
0
          goto nada;
240
241
201
     pln = MKPLAN_DFT(P, &padt, apply);
242
201
     pln->cld = cld;
243
201
     pln->cldcpy = cldcpy;
244
201
     pln->cldrest = cldrest;
245
201
     pln->n = n;
246
201
     pln->vl = vl;
247
201
     pln->ivs_by_nbuf = ivs * nbuf;
248
201
     pln->ovs_by_nbuf = ovs * nbuf;
249
201
     pln->roffset = roffset;
250
201
     pln->ioffset = ioffset;
251
252
201
     pln->nbuf = nbuf;
253
201
     pln->bufdist = bufdist;
254
255
201
     {
256
201
    opcnt t;
257
201
    X(ops_add)(&cld->ops, &cldcpy->ops, &t);
258
201
    X(ops_madd)(vl / nbuf, &t, &cldrest->ops, &pln->super.super.ops);
259
201
     }
260
261
201
     return &(pln->super.super);
262
263
2.03k
 nada:
264
2.03k
     X(ifree0)(bufs);
265
2.03k
     X(plan_destroy_internal)(cldrest);
266
2.03k
     X(plan_destroy_internal)(cldcpy);
267
2.03k
     X(plan_destroy_internal)(cld);
268
2.03k
     return (plan *) 0;
269
201
}
270
271
static solver *mksolver(size_t maxnbuf_ndx)
272
4
{
273
4
     static const solver_adt sadt = { PROBLEM_DFT, mkplan, 0 };
274
4
     S *slv = MKSOLVER(S, &sadt);
275
4
     slv->maxnbuf_ndx = maxnbuf_ndx;
276
4
     return &(slv->super);
277
4
}
278
279
void X(dft_buffered_register)(planner *p)
280
1
{
281
1
     size_t i;
282
3
     for (i = 0; i < NELEM(maxnbufs); ++i)
283
2
    REGISTER_SOLVER(p, mksolver(i));
284
1
}