1 /*
2 * Copyright (c) 2003, 2007-14 Matteo Frigo
3 * Copyright (c) 2003, 2007-14 Massachusetts Institute of Technology
4 *
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.
9 *
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.
14 *
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., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
18 *
19 */
20
21 /* This file was automatically generated --- DO NOT EDIT */
22 /* Generated on Thu Dec 10 07:04:10 EST 2020 */
23
24 #include "dft/codelet-dft.h"
25
26 #if defined(ARCH_PREFERS_FMA) || defined(ISA_EXTENSION_PREFERS_FMA)
27
28 /* Generated by: ../../../genfft/gen_twiddle.native -fma -compact -variables 4 -pipeline-latency 4 -n 2 -name t1_2 -include dft/scalar/t.h */
29
30 /*
31 * This function contains 6 FP additions, 4 FP multiplications,
32 * (or, 4 additions, 2 multiplications, 2 fused multiply/add),
33 * 11 stack variables, 0 constants, and 8 memory accesses
34 */
35 #include "dft/scalar/t.h"
36
t1_2(R * ri,R * ii,const R * W,stride rs,INT mb,INT me,INT ms)37 static void t1_2(R *ri, R *ii, const R *W, stride rs, INT mb, INT me, INT ms)
38 {
39 {
40 INT m;
41 for (m = mb, W = W + (mb * 2); m < me; m = m + 1, ri = ri + ms, ii = ii + ms, W = W + 2, MAKE_VOLATILE_STRIDE(4, rs)) {
42 E T1, Ta, T3, T6, T4, T8, T2, T7, T9, T5;
43 T1 = ri[0];
44 Ta = ii[0];
45 T3 = ri[WS(rs, 1)];
46 T6 = ii[WS(rs, 1)];
47 T2 = W[0];
48 T4 = T2 * T3;
49 T8 = T2 * T6;
50 T5 = W[1];
51 T7 = FMA(T5, T6, T4);
52 T9 = FNMS(T5, T3, T8);
53 ri[WS(rs, 1)] = T1 - T7;
54 ii[WS(rs, 1)] = Ta - T9;
55 ri[0] = T1 + T7;
56 ii[0] = T9 + Ta;
57 }
58 }
59 }
60
61 static const tw_instr twinstr[] = {
62 { TW_FULL, 0, 2 },
63 { TW_NEXT, 1, 0 }
64 };
65
66 static const ct_desc desc = { 2, "t1_2", twinstr, &GENUS, { 4, 2, 2, 0 }, 0, 0, 0 };
67
X(codelet_t1_2)68 void X(codelet_t1_2) (planner *p) {
69 X(kdft_dit_register) (p, t1_2, &desc);
70 }
71 #else
72
73 /* Generated by: ../../../genfft/gen_twiddle.native -compact -variables 4 -pipeline-latency 4 -n 2 -name t1_2 -include dft/scalar/t.h */
74
75 /*
76 * This function contains 6 FP additions, 4 FP multiplications,
77 * (or, 4 additions, 2 multiplications, 2 fused multiply/add),
78 * 9 stack variables, 0 constants, and 8 memory accesses
79 */
80 #include "dft/scalar/t.h"
81
t1_2(R * ri,R * ii,const R * W,stride rs,INT mb,INT me,INT ms)82 static void t1_2(R *ri, R *ii, const R *W, stride rs, INT mb, INT me, INT ms)
83 {
84 {
85 INT m;
86 for (m = mb, W = W + (mb * 2); m < me; m = m + 1, ri = ri + ms, ii = ii + ms, W = W + 2, MAKE_VOLATILE_STRIDE(4, rs)) {
87 E T1, T8, T6, T7;
88 T1 = ri[0];
89 T8 = ii[0];
90 {
91 E T3, T5, T2, T4;
92 T3 = ri[WS(rs, 1)];
93 T5 = ii[WS(rs, 1)];
94 T2 = W[0];
95 T4 = W[1];
96 T6 = FMA(T2, T3, T4 * T5);
97 T7 = FNMS(T4, T3, T2 * T5);
98 }
99 ri[WS(rs, 1)] = T1 - T6;
100 ii[WS(rs, 1)] = T8 - T7;
101 ri[0] = T1 + T6;
102 ii[0] = T7 + T8;
103 }
104 }
105 }
106
107 static const tw_instr twinstr[] = {
108 { TW_FULL, 0, 2 },
109 { TW_NEXT, 1, 0 }
110 };
111
112 static const ct_desc desc = { 2, "t1_2", twinstr, &GENUS, { 4, 2, 2, 0 }, 0, 0, 0 };
113
X(codelet_t1_2)114 void X(codelet_t1_2) (planner *p) {
115 X(kdft_dit_register) (p, t1_2, &desc);
116 }
117 #endif
118