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
|
/*
* Copyright (c) 2003 Matteo Frigo
* Copyright (c) 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
*
*/
/* $Id: ct-dit.c,v 1.1 2008/10/17 06:11:08 scuri Exp $ */
/* decimation in time Cooley-Tukey */
#include "dft.h"
#include "ct.h"
static void apply(const plan *ego_, R *ri, R *ii, R *ro, R *io)
{
const plan_ct *ego = (const plan_ct *) ego_;
plan *cld0 = ego->cld;
plan_dft *cld = (plan_dft *) cld0;
/* two-dimensional r x vl sub-transform: */
cld->apply(cld0, ri, ii, ro, io);
{
int i, m = ego->m, vl = ego->vl;
int os = ego->os, ovs = ego->ovs;
for (i = 0; i < vl; ++i)
ego->k.dit(ro + i * ovs, io + i * ovs, ego->td->W,
ego->ios, m, os);
}
}
static int applicable0(const solver_ct *ego, const problem *p_,
const planner *plnr)
{
UNUSED(plnr);
if (X(dft_ct_applicable)(ego, p_)) {
int ivs, ovs;
int vl;
const ct_desc *e = ego->desc;
const problem_dft *p = (const problem_dft *) p_;
iodim *d = p->sz->dims;
int m = d[0].n / e->radix;
X(tensor_tornk1)(p->vecsz, &vl, &ivs, &ovs);
return (1
&& (e->genus->okp(e, p->ro, p->io,
(int)m * d[0].os, 0, m, d[0].os, plnr))
&& (e->genus->okp(e, p->ro + ovs, p->io + ovs,
(int)m * d[0].os, 0, m, d[0].os, plnr))
);
}
return 0;
}
static int applicable(const solver_ct *ego, const problem *p_,
const planner *plnr)
{
const problem_dft *p;
if (!applicable0(ego, p_, plnr))
return 0;
p = (const problem_dft *) p_;
/* emulate fftw2 behavior */
if (NO_VRECURSEP(plnr) && (p->vecsz->rnk > 0)) return 0;
if (NO_UGLYP(plnr)) {
if (X(ct_uglyp)(16, p->sz->dims[0].n, ego->desc->radix)) return 0;
if (NONTHREADED_ICKYP(plnr))
return 0; /* prefer threaded version */
}
return 1;
}
static void finish(plan_ct *ego)
{
const ct_desc *d = ego->slv->desc;
ego->ios = X(mkstride)(ego->r, ego->m * ego->os);
X(ops_madd)(ego->vl * ego->m / d->genus->vl, &d->ops, &ego->cld->ops,
&ego->super.super.ops);
}
static plan *mkplan(const solver *ego, const problem *p, planner *plnr)
{
static const ctadt adt = {
sizeof(plan_ct), X(dft_mkcld_dit), finish, applicable, apply
};
return X(mkplan_dft_ct)((const solver_ct *) ego, p, plnr, &adt);
}
solver *X(mksolver_dft_ct_dit)(kdft_dit codelet, const ct_desc *desc)
{
static const solver_adt sadt = { mkplan };
static const char name[] = "dft-dit";
union kct k;
k.dit = codelet;
return X(mksolver_dft_ct)(k, desc, name, &sadt);
}
|
