mirror of
https://github.com/tildearrow/furnace.git
synced 2024-11-09 14:25:06 +00:00
239 lines
6.6 KiB
C
239 lines
6.6 KiB
C
|
/*
|
||
|
* Copyright (c) 2003, 2007-14 Matteo Frigo
|
||
|
* Copyright (c) 2003, 2007-14 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., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
|
||
|
*
|
||
|
*/
|
||
|
|
||
|
|
||
|
#include "rdft/rdft.h"
|
||
|
#include <stddef.h>
|
||
|
|
||
|
static void destroy(problem *ego_)
|
||
|
{
|
||
|
problem_rdft *ego = (problem_rdft *) ego_;
|
||
|
#if !defined(STRUCT_HACK_C99) && !defined(STRUCT_HACK_KR)
|
||
|
X(ifree0)(ego->kind);
|
||
|
#endif
|
||
|
X(tensor_destroy2)(ego->vecsz, ego->sz);
|
||
|
X(ifree)(ego_);
|
||
|
}
|
||
|
|
||
|
static void kind_hash(md5 *m, const rdft_kind *kind, int rnk)
|
||
|
{
|
||
|
int i;
|
||
|
for (i = 0; i < rnk; ++i)
|
||
|
X(md5int)(m, kind[i]);
|
||
|
}
|
||
|
|
||
|
static void hash(const problem *p_, md5 *m)
|
||
|
{
|
||
|
const problem_rdft *p = (const problem_rdft *) p_;
|
||
|
X(md5puts)(m, "rdft");
|
||
|
X(md5int)(m, p->I == p->O);
|
||
|
kind_hash(m, p->kind, p->sz->rnk);
|
||
|
X(md5int)(m, X(ialignment_of)(p->I));
|
||
|
X(md5int)(m, X(ialignment_of)(p->O));
|
||
|
X(tensor_md5)(m, p->sz);
|
||
|
X(tensor_md5)(m, p->vecsz);
|
||
|
}
|
||
|
|
||
|
static void recur(const iodim *dims, int rnk, R *I)
|
||
|
{
|
||
|
if (rnk == RNK_MINFTY)
|
||
|
return;
|
||
|
else if (rnk == 0)
|
||
|
I[0] = K(0.0);
|
||
|
else if (rnk > 0) {
|
||
|
INT i, n = dims[0].n, is = dims[0].is;
|
||
|
|
||
|
if (rnk == 1) {
|
||
|
/* this case is redundant but faster */
|
||
|
for (i = 0; i < n; ++i)
|
||
|
I[i * is] = K(0.0);
|
||
|
} else {
|
||
|
for (i = 0; i < n; ++i)
|
||
|
recur(dims + 1, rnk - 1, I + i * is);
|
||
|
}
|
||
|
}
|
||
|
}
|
||
|
|
||
|
void X(rdft_zerotens)(tensor *sz, R *I)
|
||
|
{
|
||
|
recur(sz->dims, sz->rnk, I);
|
||
|
}
|
||
|
|
||
|
#define KSTR_LEN 8
|
||
|
|
||
|
const char *X(rdft_kind_str)(rdft_kind kind)
|
||
|
{
|
||
|
static const char kstr[][KSTR_LEN] = {
|
||
|
"r2hc", "r2hc01", "r2hc10", "r2hc11",
|
||
|
"hc2r", "hc2r01", "hc2r10", "hc2r11",
|
||
|
"dht",
|
||
|
"redft00", "redft01", "redft10", "redft11",
|
||
|
"rodft00", "rodft01", "rodft10", "rodft11"
|
||
|
};
|
||
|
A(kind >= 0 && kind < sizeof(kstr) / KSTR_LEN);
|
||
|
return kstr[kind];
|
||
|
}
|
||
|
|
||
|
static void print(const problem *ego_, printer *p)
|
||
|
{
|
||
|
const problem_rdft *ego = (const problem_rdft *) ego_;
|
||
|
int i;
|
||
|
p->print(p, "(rdft %d %D %T %T",
|
||
|
X(ialignment_of)(ego->I),
|
||
|
(INT)(ego->O - ego->I),
|
||
|
ego->sz,
|
||
|
ego->vecsz);
|
||
|
for (i = 0; i < ego->sz->rnk; ++i)
|
||
|
p->print(p, " %d", (int)ego->kind[i]);
|
||
|
p->print(p, ")");
|
||
|
}
|
||
|
|
||
|
static void zero(const problem *ego_)
|
||
|
{
|
||
|
const problem_rdft *ego = (const problem_rdft *) ego_;
|
||
|
tensor *sz = X(tensor_append)(ego->vecsz, ego->sz);
|
||
|
X(rdft_zerotens)(sz, UNTAINT(ego->I));
|
||
|
X(tensor_destroy)(sz);
|
||
|
}
|
||
|
|
||
|
static const problem_adt padt =
|
||
|
{
|
||
|
PROBLEM_RDFT,
|
||
|
hash,
|
||
|
zero,
|
||
|
print,
|
||
|
destroy
|
||
|
};
|
||
|
|
||
|
/* Dimensions of size 1 that are not REDFT/RODFT are no-ops and can be
|
||
|
eliminated. REDFT/RODFT unit dimensions often have factors of 2.0
|
||
|
and suchlike from normalization and phases, although in principle
|
||
|
these constant factors from different dimensions could be combined. */
|
||
|
static int nontrivial(const iodim *d, rdft_kind kind)
|
||
|
{
|
||
|
return (d->n > 1 || kind == R2HC11 || kind == HC2R11
|
||
|
|| (REODFT_KINDP(kind) && kind != REDFT01 && kind != RODFT01));
|
||
|
}
|
||
|
|
||
|
problem *X(mkproblem_rdft)(const tensor *sz, const tensor *vecsz,
|
||
|
R *I, R *O, const rdft_kind *kind)
|
||
|
{
|
||
|
problem_rdft *ego;
|
||
|
int rnk = sz->rnk;
|
||
|
int i;
|
||
|
|
||
|
A(X(tensor_kosherp)(sz));
|
||
|
A(X(tensor_kosherp)(vecsz));
|
||
|
A(FINITE_RNK(sz->rnk));
|
||
|
|
||
|
if (UNTAINT(I) == UNTAINT(O))
|
||
|
I = O = JOIN_TAINT(I, O);
|
||
|
|
||
|
if (I == O && !X(tensor_inplace_locations)(sz, vecsz))
|
||
|
return X(mkproblem_unsolvable)();
|
||
|
|
||
|
for (i = rnk = 0; i < sz->rnk; ++i) {
|
||
|
A(sz->dims[i].n > 0);
|
||
|
if (nontrivial(sz->dims + i, kind[i]))
|
||
|
++rnk;
|
||
|
}
|
||
|
|
||
|
#if defined(STRUCT_HACK_KR)
|
||
|
ego = (problem_rdft *) X(mkproblem)(sizeof(problem_rdft)
|
||
|
+ sizeof(rdft_kind)
|
||
|
* (rnk > 0 ? rnk - 1u : 0u), &padt);
|
||
|
#elif defined(STRUCT_HACK_C99)
|
||
|
ego = (problem_rdft *) X(mkproblem)(sizeof(problem_rdft)
|
||
|
+ sizeof(rdft_kind) * (unsigned)rnk, &padt);
|
||
|
#else
|
||
|
ego = (problem_rdft *) X(mkproblem)(sizeof(problem_rdft), &padt);
|
||
|
ego->kind = (rdft_kind *) MALLOC(sizeof(rdft_kind) * (unsigned)rnk, PROBLEMS);
|
||
|
#endif
|
||
|
|
||
|
/* do compression and sorting as in X(tensor_compress), but take
|
||
|
transform kind into account (sigh) */
|
||
|
ego->sz = X(mktensor)(rnk);
|
||
|
for (i = rnk = 0; i < sz->rnk; ++i) {
|
||
|
if (nontrivial(sz->dims + i, kind[i])) {
|
||
|
ego->kind[rnk] = kind[i];
|
||
|
ego->sz->dims[rnk++] = sz->dims[i];
|
||
|
}
|
||
|
}
|
||
|
for (i = 0; i + 1 < rnk; ++i) {
|
||
|
int j;
|
||
|
for (j = i + 1; j < rnk; ++j)
|
||
|
if (X(dimcmp)(ego->sz->dims + i, ego->sz->dims + j) > 0) {
|
||
|
iodim dswap;
|
||
|
rdft_kind kswap;
|
||
|
dswap = ego->sz->dims[i];
|
||
|
ego->sz->dims[i] = ego->sz->dims[j];
|
||
|
ego->sz->dims[j] = dswap;
|
||
|
kswap = ego->kind[i];
|
||
|
ego->kind[i] = ego->kind[j];
|
||
|
ego->kind[j] = kswap;
|
||
|
}
|
||
|
}
|
||
|
|
||
|
for (i = 0; i < rnk; ++i)
|
||
|
if (ego->sz->dims[i].n == 2 && (ego->kind[i] == REDFT00
|
||
|
|| ego->kind[i] == DHT
|
||
|
|| ego->kind[i] == HC2R))
|
||
|
ego->kind[i] = R2HC; /* size-2 transforms are equivalent */
|
||
|
|
||
|
ego->vecsz = X(tensor_compress_contiguous)(vecsz);
|
||
|
ego->I = I;
|
||
|
ego->O = O;
|
||
|
|
||
|
A(FINITE_RNK(ego->sz->rnk));
|
||
|
|
||
|
return &(ego->super);
|
||
|
}
|
||
|
|
||
|
/* Same as X(mkproblem_rdft), but also destroy input tensors. */
|
||
|
problem *X(mkproblem_rdft_d)(tensor *sz, tensor *vecsz,
|
||
|
R *I, R *O, const rdft_kind *kind)
|
||
|
{
|
||
|
problem *p = X(mkproblem_rdft)(sz, vecsz, I, O, kind);
|
||
|
X(tensor_destroy2)(vecsz, sz);
|
||
|
return p;
|
||
|
}
|
||
|
|
||
|
/* As above, but for rnk <= 1 only and takes a scalar kind parameter */
|
||
|
problem *X(mkproblem_rdft_1)(const tensor *sz, const tensor *vecsz,
|
||
|
R *I, R *O, rdft_kind kind)
|
||
|
{
|
||
|
A(sz->rnk <= 1);
|
||
|
return X(mkproblem_rdft)(sz, vecsz, I, O, &kind);
|
||
|
}
|
||
|
|
||
|
problem *X(mkproblem_rdft_1_d)(tensor *sz, tensor *vecsz,
|
||
|
R *I, R *O, rdft_kind kind)
|
||
|
{
|
||
|
A(sz->rnk <= 1);
|
||
|
return X(mkproblem_rdft_d)(sz, vecsz, I, O, &kind);
|
||
|
}
|
||
|
|
||
|
/* create a zero-dimensional problem */
|
||
|
problem *X(mkproblem_rdft_0_d)(tensor *vecsz, R *I, R *O)
|
||
|
{
|
||
|
return X(mkproblem_rdft_d)(X(mktensor_0d)(), vecsz, I, O,
|
||
|
(const rdft_kind *)0);
|
||
|
}
|