2
* Copyright (c) 2003, 2007-8 Matteo Frigo
3
* Copyright (c) 2003, 2007-8 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 12 06:40:51 EDT 2009 */
24
#include "codelet-dft.h"
28
/* Generated by: ../../../genfft/gen_twiddle_c -fma -reorder-insns -schedule-for-pipeline -simd -compact -variables 4 -pipeline-latency 8 -n 2 -name t1fv_2 -include t1f.h */
31
* This function contains 3 FP additions, 2 FP multiplications,
32
* (or, 3 additions, 2 multiplications, 0 fused multiply/add),
33
* 5 stack variables, 0 constants, and 4 memory accesses
37
static void t1fv_2(R *ri, R *ii, const R *W, stride rs, INT mb, INT me, INT ms)
42
for (m = mb, W = W + (mb * ((TWVL / VL) * 2)); m < me; m = m + VL, x = x + (VL * ms), W = W + (TWVL * 2), MAKE_VOLATILE_STRIDE(rs)) {
44
T1 = LD(&(x[0]), ms, &(x[0]));
45
T2 = LD(&(x[WS(rs, 1)]), ms, &(x[WS(rs, 1)]));
46
T3 = BYTWJ(&(W[0]), T2);
47
ST(&(x[0]), VADD(T1, T3), ms, &(x[0]));
48
ST(&(x[WS(rs, 1)]), VSUB(T1, T3), ms, &(x[WS(rs, 1)]));
52
static const tw_instr twinstr[] = {
57
static const ct_desc desc = { 2, "t1fv_2", twinstr, &GENUS, {3, 2, 0, 0}, 0, 0, 0 };
59
void X(codelet_t1fv_2) (planner *p) {
60
X(kdft_dit_register) (p, t1fv_2, &desc);
64
/* Generated by: ../../../genfft/gen_twiddle_c -simd -compact -variables 4 -pipeline-latency 8 -n 2 -name t1fv_2 -include t1f.h */
67
* This function contains 3 FP additions, 2 FP multiplications,
68
* (or, 3 additions, 2 multiplications, 0 fused multiply/add),
69
* 5 stack variables, 0 constants, and 4 memory accesses
73
static void t1fv_2(R *ri, R *ii, const R *W, stride rs, INT mb, INT me, INT ms)
78
for (m = mb, W = W + (mb * ((TWVL / VL) * 2)); m < me; m = m + VL, x = x + (VL * ms), W = W + (TWVL * 2), MAKE_VOLATILE_STRIDE(rs)) {
80
T1 = LD(&(x[0]), ms, &(x[0]));
81
T2 = LD(&(x[WS(rs, 1)]), ms, &(x[WS(rs, 1)]));
82
T3 = BYTWJ(&(W[0]), T2);
83
ST(&(x[WS(rs, 1)]), VSUB(T1, T3), ms, &(x[WS(rs, 1)]));
84
ST(&(x[0]), VADD(T1, T3), ms, &(x[0]));
88
static const tw_instr twinstr[] = {
93
static const ct_desc desc = { 2, "t1fv_2", twinstr, &GENUS, {3, 2, 0, 0}, 0, 0, 0 };
95
void X(codelet_t1fv_2) (planner *p) {
96
X(kdft_dit_register) (p, t1fv_2, &desc);