mirror of
https://github.com/tildearrow/furnace.git
synced 2024-11-14 00:35:06 +00:00
54e93db207
not reliable yet
341 lines
8.8 KiB
C
341 lines
8.8 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
|
|
*
|
|
*/
|
|
|
|
|
|
/* direct RDFT solver, using r2c codelets */
|
|
|
|
#include "rdft/rdft.h"
|
|
|
|
typedef struct {
|
|
solver super;
|
|
const kr2c_desc *desc;
|
|
kr2c k;
|
|
int bufferedp;
|
|
} S;
|
|
|
|
typedef struct {
|
|
plan_rdft super;
|
|
|
|
stride rs, csr, csi;
|
|
stride brs, bcsr, bcsi;
|
|
INT n, vl, rs0, ivs, ovs, ioffset, bioffset;
|
|
kr2c k;
|
|
const S *slv;
|
|
} P;
|
|
|
|
/*************************************************************
|
|
Nonbuffered code
|
|
*************************************************************/
|
|
static void apply_r2hc(const plan *ego_, R *I, R *O)
|
|
{
|
|
const P *ego = (const P *) ego_;
|
|
ASSERT_ALIGNED_DOUBLE;
|
|
ego->k(I, I + ego->rs0, O, O + ego->ioffset,
|
|
ego->rs, ego->csr, ego->csi,
|
|
ego->vl, ego->ivs, ego->ovs);
|
|
}
|
|
|
|
static void apply_hc2r(const plan *ego_, R *I, R *O)
|
|
{
|
|
const P *ego = (const P *) ego_;
|
|
ASSERT_ALIGNED_DOUBLE;
|
|
ego->k(O, O + ego->rs0, I, I + ego->ioffset,
|
|
ego->rs, ego->csr, ego->csi,
|
|
ego->vl, ego->ivs, ego->ovs);
|
|
}
|
|
|
|
/*************************************************************
|
|
Buffered code
|
|
*************************************************************/
|
|
/* should not be 2^k to avoid associativity conflicts */
|
|
static INT compute_batchsize(INT radix)
|
|
{
|
|
/* round up to multiple of 4 */
|
|
radix += 3;
|
|
radix &= -4;
|
|
|
|
return (radix + 2);
|
|
}
|
|
|
|
static void dobatch_r2hc(const P *ego, R *I, R *O, R *buf, INT batchsz)
|
|
{
|
|
X(cpy2d_ci)(I, buf,
|
|
ego->n, ego->rs0, WS(ego->bcsr /* hack */, 1),
|
|
batchsz, ego->ivs, 1, 1);
|
|
|
|
if (IABS(WS(ego->csr, 1)) < IABS(ego->ovs)) {
|
|
/* transform directly to output */
|
|
ego->k(buf, buf + WS(ego->bcsr /* hack */, 1),
|
|
O, O + ego->ioffset,
|
|
ego->brs, ego->csr, ego->csi,
|
|
batchsz, 1, ego->ovs);
|
|
} else {
|
|
/* transform to buffer and copy back */
|
|
ego->k(buf, buf + WS(ego->bcsr /* hack */, 1),
|
|
buf, buf + ego->bioffset,
|
|
ego->brs, ego->bcsr, ego->bcsi,
|
|
batchsz, 1, 1);
|
|
X(cpy2d_co)(buf, O,
|
|
ego->n, WS(ego->bcsr, 1), WS(ego->csr, 1),
|
|
batchsz, 1, ego->ovs, 1);
|
|
}
|
|
}
|
|
|
|
static void dobatch_hc2r(const P *ego, R *I, R *O, R *buf, INT batchsz)
|
|
{
|
|
if (IABS(WS(ego->csr, 1)) < IABS(ego->ivs)) {
|
|
/* transform directly from input */
|
|
ego->k(buf, buf + WS(ego->bcsr /* hack */, 1),
|
|
I, I + ego->ioffset,
|
|
ego->brs, ego->csr, ego->csi,
|
|
batchsz, ego->ivs, 1);
|
|
} else {
|
|
/* copy into buffer and transform in place */
|
|
X(cpy2d_ci)(I, buf,
|
|
ego->n, WS(ego->csr, 1), WS(ego->bcsr, 1),
|
|
batchsz, ego->ivs, 1, 1);
|
|
ego->k(buf, buf + WS(ego->bcsr /* hack */, 1),
|
|
buf, buf + ego->bioffset,
|
|
ego->brs, ego->bcsr, ego->bcsi,
|
|
batchsz, 1, 1);
|
|
}
|
|
X(cpy2d_co)(buf, O,
|
|
ego->n, WS(ego->bcsr /* hack */, 1), ego->rs0,
|
|
batchsz, 1, ego->ovs, 1);
|
|
}
|
|
|
|
static void iterate(const P *ego, R *I, R *O,
|
|
void (*dobatch)(const P *ego, R *I, R *O,
|
|
R *buf, INT batchsz))
|
|
{
|
|
R *buf;
|
|
INT vl = ego->vl;
|
|
INT n = ego->n;
|
|
INT i;
|
|
INT batchsz = compute_batchsize(n);
|
|
size_t bufsz = n * batchsz * sizeof(R);
|
|
|
|
BUF_ALLOC(R *, buf, bufsz);
|
|
|
|
for (i = 0; i < vl - batchsz; i += batchsz) {
|
|
dobatch(ego, I, O, buf, batchsz);
|
|
I += batchsz * ego->ivs;
|
|
O += batchsz * ego->ovs;
|
|
}
|
|
dobatch(ego, I, O, buf, vl - i);
|
|
|
|
BUF_FREE(buf, bufsz);
|
|
}
|
|
|
|
static void apply_buf_r2hc(const plan *ego_, R *I, R *O)
|
|
{
|
|
iterate((const P *) ego_, I, O, dobatch_r2hc);
|
|
}
|
|
|
|
static void apply_buf_hc2r(const plan *ego_, R *I, R *O)
|
|
{
|
|
iterate((const P *) ego_, I, O, dobatch_hc2r);
|
|
}
|
|
|
|
static void destroy(plan *ego_)
|
|
{
|
|
P *ego = (P *) ego_;
|
|
X(stride_destroy)(ego->rs);
|
|
X(stride_destroy)(ego->csr);
|
|
X(stride_destroy)(ego->csi);
|
|
X(stride_destroy)(ego->brs);
|
|
X(stride_destroy)(ego->bcsr);
|
|
X(stride_destroy)(ego->bcsi);
|
|
}
|
|
|
|
static void print(const plan *ego_, printer *p)
|
|
{
|
|
const P *ego = (const P *) ego_;
|
|
const S *s = ego->slv;
|
|
|
|
if (ego->slv->bufferedp)
|
|
p->print(p, "(rdft-%s-directbuf/%D-r2c-%D%v \"%s\")",
|
|
X(rdft_kind_str)(s->desc->genus->kind),
|
|
/* hack */ WS(ego->bcsr, 1), ego->n,
|
|
ego->vl, s->desc->nam);
|
|
|
|
else
|
|
p->print(p, "(rdft-%s-direct-r2c-%D%v \"%s\")",
|
|
X(rdft_kind_str)(s->desc->genus->kind), ego->n,
|
|
ego->vl, s->desc->nam);
|
|
}
|
|
|
|
static INT ioffset(rdft_kind kind, INT sz, INT s)
|
|
{
|
|
return(s * ((kind == R2HC || kind == HC2R) ? sz : (sz - 1)));
|
|
}
|
|
|
|
static int applicable(const solver *ego_, const problem *p_)
|
|
{
|
|
const S *ego = (const S *) ego_;
|
|
const kr2c_desc *desc = ego->desc;
|
|
const problem_rdft *p = (const problem_rdft *) p_;
|
|
INT vl, ivs, ovs;
|
|
|
|
return (
|
|
1
|
|
&& p->sz->rnk == 1
|
|
&& p->vecsz->rnk <= 1
|
|
&& p->sz->dims[0].n == desc->n
|
|
&& p->kind[0] == desc->genus->kind
|
|
|
|
/* check strides etc */
|
|
&& X(tensor_tornk1)(p->vecsz, &vl, &ivs, &ovs)
|
|
|
|
&& (0
|
|
/* can operate out-of-place */
|
|
|| p->I != p->O
|
|
|
|
/* computing one transform */
|
|
|| vl == 1
|
|
|
|
/* can operate in-place as long as strides are the same */
|
|
|| X(tensor_inplace_strides2)(p->sz, p->vecsz)
|
|
)
|
|
);
|
|
}
|
|
|
|
static int applicable_buf(const solver *ego_, const problem *p_)
|
|
{
|
|
const S *ego = (const S *) ego_;
|
|
const kr2c_desc *desc = ego->desc;
|
|
const problem_rdft *p = (const problem_rdft *) p_;
|
|
INT vl, ivs, ovs, batchsz;
|
|
|
|
return (
|
|
1
|
|
&& p->sz->rnk == 1
|
|
&& p->vecsz->rnk <= 1
|
|
&& p->sz->dims[0].n == desc->n
|
|
&& p->kind[0] == desc->genus->kind
|
|
|
|
/* check strides etc */
|
|
&& X(tensor_tornk1)(p->vecsz, &vl, &ivs, &ovs)
|
|
|
|
&& (batchsz = compute_batchsize(desc->n), 1)
|
|
|
|
&& (0
|
|
/* can operate out-of-place */
|
|
|| p->I != p->O
|
|
|
|
/* can operate in-place as long as strides are the same */
|
|
|| X(tensor_inplace_strides2)(p->sz, p->vecsz)
|
|
|
|
/* can do it if the problem fits in the buffer, no matter
|
|
what the strides are */
|
|
|| vl <= batchsz
|
|
)
|
|
);
|
|
}
|
|
|
|
static plan *mkplan(const solver *ego_, const problem *p_, planner *plnr)
|
|
{
|
|
const S *ego = (const S *) ego_;
|
|
P *pln;
|
|
const problem_rdft *p;
|
|
iodim *d;
|
|
INT rs, cs, b, n;
|
|
|
|
static const plan_adt padt = {
|
|
X(rdft_solve), X(null_awake), print, destroy
|
|
};
|
|
|
|
UNUSED(plnr);
|
|
|
|
if (ego->bufferedp) {
|
|
if (!applicable_buf(ego_, p_))
|
|
return (plan *)0;
|
|
} else {
|
|
if (!applicable(ego_, p_))
|
|
return (plan *)0;
|
|
}
|
|
|
|
p = (const problem_rdft *) p_;
|
|
|
|
if (R2HC_KINDP(p->kind[0])) {
|
|
rs = p->sz->dims[0].is; cs = p->sz->dims[0].os;
|
|
pln = MKPLAN_RDFT(P, &padt,
|
|
ego->bufferedp ? apply_buf_r2hc : apply_r2hc);
|
|
} else {
|
|
rs = p->sz->dims[0].os; cs = p->sz->dims[0].is;
|
|
pln = MKPLAN_RDFT(P, &padt,
|
|
ego->bufferedp ? apply_buf_hc2r : apply_hc2r);
|
|
}
|
|
|
|
d = p->sz->dims;
|
|
n = d[0].n;
|
|
|
|
pln->k = ego->k;
|
|
pln->n = n;
|
|
|
|
pln->rs0 = rs;
|
|
pln->rs = X(mkstride)(n, 2 * rs);
|
|
pln->csr = X(mkstride)(n, cs);
|
|
pln->csi = X(mkstride)(n, -cs);
|
|
pln->ioffset = ioffset(p->kind[0], n, cs);
|
|
|
|
b = compute_batchsize(n);
|
|
pln->brs = X(mkstride)(n, 2 * b);
|
|
pln->bcsr = X(mkstride)(n, b);
|
|
pln->bcsi = X(mkstride)(n, -b);
|
|
pln->bioffset = ioffset(p->kind[0], n, b);
|
|
|
|
X(tensor_tornk1)(p->vecsz, &pln->vl, &pln->ivs, &pln->ovs);
|
|
|
|
pln->slv = ego;
|
|
X(ops_zero)(&pln->super.super.ops);
|
|
|
|
X(ops_madd2)(pln->vl / ego->desc->genus->vl,
|
|
&ego->desc->ops,
|
|
&pln->super.super.ops);
|
|
|
|
if (ego->bufferedp)
|
|
pln->super.super.ops.other += 2 * n * pln->vl;
|
|
|
|
pln->super.super.could_prune_now_p = !ego->bufferedp;
|
|
|
|
return &(pln->super.super);
|
|
}
|
|
|
|
/* constructor */
|
|
static solver *mksolver(kr2c k, const kr2c_desc *desc, int bufferedp)
|
|
{
|
|
static const solver_adt sadt = { PROBLEM_RDFT, mkplan, 0 };
|
|
S *slv = MKSOLVER(S, &sadt);
|
|
slv->k = k;
|
|
slv->desc = desc;
|
|
slv->bufferedp = bufferedp;
|
|
return &(slv->super);
|
|
}
|
|
|
|
solver *X(mksolver_rdft_r2c_direct)(kr2c k, const kr2c_desc *desc)
|
|
{
|
|
return mksolver(k, desc, 0);
|
|
}
|
|
|
|
solver *X(mksolver_rdft_r2c_directbuf)(kr2c k, const kr2c_desc *desc)
|
|
{
|
|
return mksolver(k, desc, 1);
|
|
}
|