2
* Copyright (c) 2003, 2006 Matteo Frigo
3
* Copyright (c) 2003, 2006 Massachusetts Institute of Technology
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.
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.
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., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
21
/* This file was automatically generated --- DO NOT EDIT */
22
/* Generated on Sun Jul 2 16:08:53 EDT 2006 */
24
#include "codelet-rdft.h"
28
/* Generated by: ../../../genfft/gen_r2hc -fma -reorder-insns -schedule-for-pipeline -compact -variables 4 -pipeline-latency 4 -n 8 -name r2hcII_8 -dft-II -include r2hcII.h */
31
* This function contains 22 FP additions, 16 FP multiplications,
32
* (or, 6 additions, 0 multiplications, 16 fused multiply/add),
33
* 22 stack variables, and 16 memory accesses
37
* $Id: algsimp.ml,v 1.9 2006-02-12 23:34:12 athena Exp $
38
* $Id: fft.ml,v 1.4 2006-01-05 03:04:27 stevenj Exp $
39
* $Id: gen_r2hc.ml,v 1.18 2006-02-12 23:34:12 athena Exp $
44
static void r2hcII_8(const R *I, R *ro, R *io, stride is, stride ros, stride ios, INT v, INT ivs, INT ovs)
46
DK(KP923879532, +0.923879532511286756128183189396788286822416626);
47
DK(KP707106781, +0.707106781186547524400844362104849039284835938);
48
DK(KP414213562, +0.414213562373095048801688724209698078569671875);
50
for (i = v; i > 0; i = i - 1, I = I + ivs, ro = ro + ovs, io = io + ovs, MAKE_VOLATILE_STRIDE(is), MAKE_VOLATILE_STRIDE(ros), MAKE_VOLATILE_STRIDE(ios)) {
51
E Te, T8, Td, T5, Tj, Tl, Tf, Tb;
53
E T1, Th, T9, Ti, T4, Ta;
65
Te = FMA(KP414213562, T6, T7);
66
T8 = FNMS(KP414213562, T7, T6);
69
Td = FNMS(KP707106781, T4, T1);
70
T5 = FMA(KP707106781, T4, T1);
71
Tj = FMA(KP707106781, Ti, Th);
72
Tl = FNMS(KP707106781, Ti, Th);
73
Tf = FMA(KP414213562, T9, Ta);
74
Tb = FMS(KP414213562, Ta, T9);
82
ro[WS(ros, 1)] = FMA(KP923879532, Tg, Td);
83
ro[WS(ros, 2)] = FNMS(KP923879532, Tg, Td);
84
io[WS(ios, 3)] = FNMS(KP923879532, Tk, Tj);
85
io[0] = -(FMA(KP923879532, Tk, Tj));
86
io[WS(ios, 1)] = FMA(KP923879532, Tm, Tl);
87
io[WS(ios, 2)] = FMS(KP923879532, Tm, Tl);
88
ro[0] = FMA(KP923879532, Tc, T5);
89
ro[WS(ros, 3)] = FNMS(KP923879532, Tc, T5);
94
static const kr2hc_desc desc = { 8, "r2hcII_8", {6, 0, 16, 0}, &GENUS, 0, 0, 0, 0, 0 };
96
void X(codelet_r2hcII_8) (planner *p) {
97
X(kr2hcII_register) (p, r2hcII_8, &desc);
102
/* Generated by: ../../../genfft/gen_r2hc -compact -variables 4 -pipeline-latency 4 -n 8 -name r2hcII_8 -dft-II -include r2hcII.h */
105
* This function contains 22 FP additions, 10 FP multiplications,
106
* (or, 18 additions, 6 multiplications, 4 fused multiply/add),
107
* 18 stack variables, and 16 memory accesses
111
* $Id: algsimp.ml,v 1.9 2006-02-12 23:34:12 athena Exp $
112
* $Id: fft.ml,v 1.4 2006-01-05 03:04:27 stevenj Exp $
113
* $Id: gen_r2hc.ml,v 1.18 2006-02-12 23:34:12 athena Exp $
118
static void r2hcII_8(const R *I, R *ro, R *io, stride is, stride ros, stride ios, INT v, INT ivs, INT ovs)
120
DK(KP382683432, +0.382683432365089771728459984030398866761344562);
121
DK(KP923879532, +0.923879532511286756128183189396788286822416626);
122
DK(KP707106781, +0.707106781186547524400844362104849039284835938);
124
for (i = v; i > 0; i = i - 1, I = I + ivs, ro = ro + ovs, io = io + ovs, MAKE_VOLATILE_STRIDE(is), MAKE_VOLATILE_STRIDE(ros), MAKE_VOLATILE_STRIDE(ios)) {
125
E T1, Tj, T4, Ti, T8, Te, Tb, Tf, T2, T3;
130
T4 = KP707106781 * (T2 - T3);
131
Ti = KP707106781 * (T2 + T3);
136
T8 = FNMS(KP382683432, T7, KP923879532 * T6);
137
Te = FMA(KP382683432, T6, KP923879532 * T7);
140
Tb = FNMS(KP923879532, Ta, KP382683432 * T9);
141
Tf = FMA(KP923879532, T9, KP382683432 * Ta);
147
ro[WS(ros, 3)] = T5 - Tc;
152
io[WS(ios, 3)] = Tk - Th;
158
ro[WS(ros, 2)] = Td - Tg;
159
ro[WS(ros, 1)] = Td + Tg;
162
io[WS(ios, 2)] = Tl - Tm;
163
io[WS(ios, 1)] = Tl + Tm;
168
static const kr2hc_desc desc = { 8, "r2hcII_8", {18, 6, 4, 0}, &GENUS, 0, 0, 0, 0, 0 };
170
void X(codelet_r2hcII_8) (planner *p) {
171
X(kr2hcII_register) (p, r2hcII_8, &desc);
174
#endif /* HAVE_FMA */