2
* Copyright (c) 2003, 2007-11 Matteo Frigo
3
* Copyright (c) 2003, 2007-11 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 Wed Jul 27 06:13:36 EDT 2011 */
24
#include "codelet-dft.h"
28
/* Generated by: ../../../genfft/gen_notw_c.native -fma -reorder-insns -schedule-for-pipeline -simd -compact -variables 4 -pipeline-latency 8 -sign 1 -n 3 -name n1bv_3 -include n1b.h */
31
* This function contains 6 FP additions, 4 FP multiplications,
32
* (or, 3 additions, 1 multiplications, 3 fused multiply/add),
33
* 11 stack variables, 2 constants, and 6 memory accesses
37
static void n1bv_3(const R *ri, const R *ii, R *ro, R *io, stride is, stride os, INT v, INT ivs, INT ovs)
39
DVK(KP500000000, +0.500000000000000000000000000000000000000000000);
40
DVK(KP866025403, +0.866025403784438646763723170752936183471402627);
47
for (i = v; i > 0; i = i - VL, xi = xi + (VL * ivs), xo = xo + (VL * ovs), MAKE_VOLATILE_STRIDE(is), MAKE_VOLATILE_STRIDE(os)) {
48
V T1, T2, T3, T6, T4, T5;
49
T1 = LD(&(xi[0]), ivs, &(xi[0]));
50
T2 = LD(&(xi[WS(is, 1)]), ivs, &(xi[WS(is, 1)]));
51
T3 = LD(&(xi[WS(is, 2)]), ivs, &(xi[0]));
52
T6 = VMUL(LDK(KP866025403), VSUB(T2, T3));
54
T5 = VFNMS(LDK(KP500000000), T4, T1);
55
ST(&(xo[0]), VADD(T1, T4), ovs, &(xo[0]));
56
ST(&(xo[WS(os, 2)]), VFNMSI(T6, T5), ovs, &(xo[0]));
57
ST(&(xo[WS(os, 1)]), VFMAI(T6, T5), ovs, &(xo[WS(os, 1)]));
63
static const kdft_desc desc = { 3, XSIMD_STRING("n1bv_3"), {3, 1, 3, 0}, &GENUS, 0, 0, 0, 0 };
65
void XSIMD(codelet_n1bv_3) (planner *p) {
66
X(kdft_register) (p, n1bv_3, &desc);
71
/* Generated by: ../../../genfft/gen_notw_c.native -simd -compact -variables 4 -pipeline-latency 8 -sign 1 -n 3 -name n1bv_3 -include n1b.h */
74
* This function contains 6 FP additions, 2 FP multiplications,
75
* (or, 5 additions, 1 multiplications, 1 fused multiply/add),
76
* 11 stack variables, 2 constants, and 6 memory accesses
80
static void n1bv_3(const R *ri, const R *ii, R *ro, R *io, stride is, stride os, INT v, INT ivs, INT ovs)
82
DVK(KP500000000, +0.500000000000000000000000000000000000000000000);
83
DVK(KP866025403, +0.866025403784438646763723170752936183471402627);
90
for (i = v; i > 0; i = i - VL, xi = xi + (VL * ivs), xo = xo + (VL * ovs), MAKE_VOLATILE_STRIDE(is), MAKE_VOLATILE_STRIDE(os)) {
91
V T4, T3, T5, T1, T2, T6;
92
T4 = LD(&(xi[0]), ivs, &(xi[0]));
93
T1 = LD(&(xi[WS(is, 1)]), ivs, &(xi[WS(is, 1)]));
94
T2 = LD(&(xi[WS(is, 2)]), ivs, &(xi[0]));
95
T3 = VBYI(VMUL(LDK(KP866025403), VSUB(T1, T2)));
97
ST(&(xo[0]), VADD(T4, T5), ovs, &(xo[0]));
98
T6 = VFNMS(LDK(KP500000000), T5, T4);
99
ST(&(xo[WS(os, 1)]), VADD(T3, T6), ovs, &(xo[WS(os, 1)]));
100
ST(&(xo[WS(os, 2)]), VSUB(T6, T3), ovs, &(xo[0]));
106
static const kdft_desc desc = { 3, XSIMD_STRING("n1bv_3"), {5, 1, 1, 0}, &GENUS, 0, 0, 0, 0 };
108
void XSIMD(codelet_n1bv_3) (planner *p) {
109
X(kdft_register) (p, n1bv_3, &desc);
112
#endif /* HAVE_FMA */