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 16:15:25 EDT 2006 */
24
#include "codelet-rdft.h"
28
/* Generated by: ../../../genfft/gen_hc2r -fma -reorder-insns -schedule-for-pipeline -compact -variables 4 -pipeline-latency 4 -sign 1 -n 7 -name hc2r_7 -include hc2r.h */
31
* This function contains 24 FP additions, 22 FP multiplications,
32
* (or, 2 additions, 0 multiplications, 22 fused multiply/add),
33
* 31 stack variables, and 14 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_hc2r.ml,v 1.19 2006-02-12 23:34:12 athena Exp $
44
static void hc2r_7(const R *ri, const R *ii, R *O, stride ris, stride iis, stride os, INT v, INT ivs, INT ovs)
46
DK(KP1_949855824, +1.949855824363647214036263365987862434465571601);
47
DK(KP1_801937735, +1.801937735804838252472204639014890102331838324);
48
DK(KP2_000000000, +2.000000000000000000000000000000000000000000000);
49
DK(KP692021471, +0.692021471630095869627814897002069140197260599);
50
DK(KP801937735, +0.801937735804838252472204639014890102331838324);
51
DK(KP356895867, +0.356895867892209443894399510021300583399127187);
52
DK(KP554958132, +0.554958132087371191422194871006410481067288862);
54
for (i = v; i > 0; i = i - 1, ri = ri + ivs, ii = ii + ivs, O = O + ovs, MAKE_VOLATILE_STRIDE(ris), MAKE_VOLATILE_STRIDE(iis), MAKE_VOLATILE_STRIDE(os)) {
55
E Tn, Td, Tg, Ti, Tl, T8;
57
E T1, T9, Tb, Ta, T2, T4, Th, Tm, Tc, T3, Te;
64
Th = FMA(KP554958132, T9, Tb);
65
Tm = FMS(KP554958132, Ta, T9);
66
Tc = FMA(KP554958132, Tb, Ta);
68
Te = FNMS(KP356895867, T2, T4);
69
Tn = FMA(KP801937735, Tm, Tb);
71
E Tf, Tk, T7, T5, Tj, T6;
72
Td = FMA(KP801937735, Tc, T9);
74
Tj = FNMS(KP356895867, T4, T3);
75
T6 = FNMS(KP356895867, T3, T2);
76
Tf = FNMS(KP692021471, Te, T3);
77
O[0] = FMA(KP2_000000000, T5, T1);
78
Tk = FNMS(KP692021471, Tj, T2);
79
T7 = FNMS(KP692021471, T6, T4);
80
Tg = FNMS(KP1_801937735, Tf, T1);
81
Ti = FNMS(KP801937735, Th, Ta);
82
Tl = FNMS(KP1_801937735, Tk, T1);
83
T8 = FNMS(KP1_801937735, T7, T1);
86
O[WS(os, 5)] = FMA(KP1_949855824, Ti, Tg);
87
O[WS(os, 2)] = FNMS(KP1_949855824, Ti, Tg);
88
O[WS(os, 4)] = FMA(KP1_949855824, Tn, Tl);
89
O[WS(os, 3)] = FNMS(KP1_949855824, Tn, Tl);
90
O[WS(os, 6)] = FMA(KP1_949855824, Td, T8);
91
O[WS(os, 1)] = FNMS(KP1_949855824, Td, T8);
95
static const khc2r_desc desc = { 7, "hc2r_7", {2, 0, 22, 0}, &GENUS, 0, 0, 0, 0, 0 };
97
void X(codelet_hc2r_7) (planner *p) {
98
X(khc2r_register) (p, hc2r_7, &desc);
103
/* Generated by: ../../../genfft/gen_hc2r -compact -variables 4 -pipeline-latency 4 -sign 1 -n 7 -name hc2r_7 -include hc2r.h */
106
* This function contains 24 FP additions, 19 FP multiplications,
107
* (or, 11 additions, 6 multiplications, 13 fused multiply/add),
108
* 21 stack variables, and 14 memory accesses
112
* $Id: algsimp.ml,v 1.9 2006-02-12 23:34:12 athena Exp $
113
* $Id: fft.ml,v 1.4 2006-01-05 03:04:27 stevenj Exp $
114
* $Id: gen_hc2r.ml,v 1.19 2006-02-12 23:34:12 athena Exp $
119
static void hc2r_7(const R *ri, const R *ii, R *O, stride ris, stride iis, stride os, INT v, INT ivs, INT ovs)
121
DK(KP2_000000000, +2.000000000000000000000000000000000000000000000);
122
DK(KP1_801937735, +1.801937735804838252472204639014890102331838324);
123
DK(KP445041867, +0.445041867912628808577805128993589518932711138);
124
DK(KP1_246979603, +1.246979603717467061050009768008479621264549462);
125
DK(KP867767478, +0.867767478235116240951536665696717509219981456);
126
DK(KP1_949855824, +1.949855824363647214036263365987862434465571601);
127
DK(KP1_563662964, +1.563662964936059617416889053348115500464669037);
129
for (i = v; i > 0; i = i - 1, ri = ri + ivs, ii = ii + ivs, O = O + ovs, MAKE_VOLATILE_STRIDE(ris), MAKE_VOLATILE_STRIDE(iis), MAKE_VOLATILE_STRIDE(os)) {
130
E T9, Td, Tb, T1, T4, T2, T3, T5, Tc, Ta, T6, T8, T7;
134
T9 = FNMS(KP1_949855824, T7, KP1_563662964 * T6) - (KP867767478 * T8);
135
Td = FMA(KP867767478, T6, KP1_563662964 * T7) - (KP1_949855824 * T8);
136
Tb = FMA(KP1_563662964, T8, KP1_949855824 * T6) + (KP867767478 * T7);
141
T5 = FMA(KP1_246979603, T3, T1) + FNMA(KP445041867, T4, KP1_801937735 * T2);
142
Tc = FMA(KP1_246979603, T4, T1) + FNMA(KP1_801937735, T3, KP445041867 * T2);
143
Ta = FMA(KP1_246979603, T2, T1) + FNMA(KP1_801937735, T4, KP445041867 * T3);
144
O[WS(os, 4)] = T5 - T9;
145
O[WS(os, 3)] = T5 + T9;
146
O[WS(os, 2)] = Tc + Td;
147
O[WS(os, 5)] = Tc - Td;
148
O[WS(os, 6)] = Ta + Tb;
149
O[WS(os, 1)] = Ta - Tb;
150
O[0] = FMA(KP2_000000000, T2 + T3 + T4, T1);
154
static const khc2r_desc desc = { 7, "hc2r_7", {11, 6, 13, 0}, &GENUS, 0, 0, 0, 0, 0 };
156
void X(codelet_hc2r_7) (planner *p) {
157
X(khc2r_register) (p, hc2r_7, &desc);
160
#endif /* HAVE_FMA */