Mercurial > hg > sv-dependency-builds

comparison src/fftw-3.3.8/rdft/scalar/r2cf/r2cf_10.c @ 167:bd3cc4d1df30

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
Add FFTW 3.3.8 source, and a Linux build
author Chris Cannam <cannam@all-day-breakfast.com>
date Tue, 19 Nov 2019 14:52:55 +0000
parents
children
comparison
equal deleted inserted replaced
166:cbd6d7e562c7 167:bd3cc4d1df30
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 May 24 08:06:26 EDT 2018 */
23
24 #include "rdft/codelet-rdft.h"
25
26 #if defined(ARCH_PREFERS_FMA) || defined(ISA_EXTENSION_PREFERS_FMA)
27
28 /* Generated by: ../../../genfft/gen_r2cf.native -fma -compact -variables 4 -pipeline-latency 4 -n 10 -name r2cf_10 -include rdft/scalar/r2cf.h */
29
30 /*
31 * This function contains 34 FP additions, 14 FP multiplications,
32 * (or, 24 additions, 4 multiplications, 10 fused multiply/add),
33 * 26 stack variables, 4 constants, and 20 memory accesses
34 */
35 #include "rdft/scalar/r2cf.h"
36
37 static void r2cf_10(R *R0, R *R1, R *Cr, R *Ci, stride rs, stride csr, stride csi, INT v, INT ivs, INT ovs)
38 {
39 DK(KP559016994, +0.559016994374947424102293417182819058860154590);
40 DK(KP250000000, +0.250000000000000000000000000000000000000000000);
41 DK(KP618033988, +0.618033988749894848204586834365638117720309180);
42 DK(KP951056516, +0.951056516295153572116439333379382143405698634);
43 {
44 INT i;
45 for (i = v; i > 0; i = i - 1, R0 = R0 + ivs, R1 = R1 + ivs, Cr = Cr + ovs, Ci = Ci + ovs, MAKE_VOLATILE_STRIDE(40, rs), MAKE_VOLATILE_STRIDE(40, csr), MAKE_VOLATILE_STRIDE(40, csi)) {
46 E T3, Tt, Td, Tn, Tg, To, Th, Tv, T6, Tq, T9, Tr, Ta, Tu, T1;
47 E T2;
48 T1 = R0[0];
49 T2 = R1[WS(rs, 2)];
50 T3 = T1 - T2;
51 Tt = T1 + T2;
52 {
53 E Tb, Tc, Te, Tf;
54 Tb = R0[WS(rs, 2)];
55 Tc = R1[WS(rs, 4)];
56 Td = Tb - Tc;
57 Tn = Tb + Tc;
58 Te = R0[WS(rs, 3)];
59 Tf = R1[0];
60 Tg = Te - Tf;
61 To = Te + Tf;
62 }
63 Th = Td + Tg;
64 Tv = Tn + To;
65 {
66 E T4, T5, T7, T8;
67 T4 = R0[WS(rs, 1)];
68 T5 = R1[WS(rs, 3)];
69 T6 = T4 - T5;
70 Tq = T4 + T5;
71 T7 = R0[WS(rs, 4)];
72 T8 = R1[WS(rs, 1)];
73 T9 = T7 - T8;
74 Tr = T7 + T8;
75 }
76 Ta = T6 + T9;
77 Tu = Tq + Tr;
78 {
79 E Tl, Tm, Tk, Ti, Tj;
80 Tl = T6 - T9;
81 Tm = Tg - Td;
82 Ci[WS(csi, 1)] = -(KP951056516 * (FNMS(KP618033988, Tm, Tl)));
83 Ci[WS(csi, 3)] = KP951056516 * (FMA(KP618033988, Tl, Tm));
84 Tk = Ta - Th;
85 Ti = Ta + Th;
86 Tj = FNMS(KP250000000, Ti, T3);
87 Cr[WS(csr, 1)] = FMA(KP559016994, Tk, Tj);
88 Cr[WS(csr, 5)] = T3 + Ti;
89 Cr[WS(csr, 3)] = FNMS(KP559016994, Tk, Tj);
90 }
91 {
92 E Tp, Ts, Ty, Tw, Tx;
93 Tp = Tn - To;
94 Ts = Tq - Tr;
95 Ci[WS(csi, 2)] = KP951056516 * (FNMS(KP618033988, Ts, Tp));
96 Ci[WS(csi, 4)] = KP951056516 * (FMA(KP618033988, Tp, Ts));
97 Ty = Tu - Tv;
98 Tw = Tu + Tv;
99 Tx = FNMS(KP250000000, Tw, Tt);
100 Cr[WS(csr, 2)] = FNMS(KP559016994, Ty, Tx);
101 Cr[0] = Tt + Tw;
102 Cr[WS(csr, 4)] = FMA(KP559016994, Ty, Tx);
103 }
104 }
105 }
106 }
107
108 static const kr2c_desc desc = { 10, "r2cf_10", {24, 4, 10, 0}, &GENUS };
109
110 void X(codelet_r2cf_10) (planner *p) {
111 X(kr2c_register) (p, r2cf_10, &desc);
112 }
113
114 #else
115
116 /* Generated by: ../../../genfft/gen_r2cf.native -compact -variables 4 -pipeline-latency 4 -n 10 -name r2cf_10 -include rdft/scalar/r2cf.h */
117
118 /*
119 * This function contains 34 FP additions, 12 FP multiplications,
120 * (or, 28 additions, 6 multiplications, 6 fused multiply/add),
121 * 26 stack variables, 4 constants, and 20 memory accesses
122 */
123 #include "rdft/scalar/r2cf.h"
124
125 static void r2cf_10(R *R0, R *R1, R *Cr, R *Ci, stride rs, stride csr, stride csi, INT v, INT ivs, INT ovs)
126 {
127 DK(KP250000000, +0.250000000000000000000000000000000000000000000);
128 DK(KP559016994, +0.559016994374947424102293417182819058860154590);
129 DK(KP951056516, +0.951056516295153572116439333379382143405698634);
130 DK(KP587785252, +0.587785252292473129168705954639072768597652438);
131 {
132 INT i;
133 for (i = v; i > 0; i = i - 1, R0 = R0 + ivs, R1 = R1 + ivs, Cr = Cr + ovs, Ci = Ci + ovs, MAKE_VOLATILE_STRIDE(40, rs), MAKE_VOLATILE_STRIDE(40, csr), MAKE_VOLATILE_STRIDE(40, csi)) {
134 E Ti, Tt, Ta, Tn, Td, To, Te, Tv, T3, Tq, T6, Tr, T7, Tu, Tg;
135 E Th;
136 Tg = R0[0];
137 Th = R1[WS(rs, 2)];
138 Ti = Tg - Th;
139 Tt = Tg + Th;
140 {
141 E T8, T9, Tb, Tc;
142 T8 = R0[WS(rs, 2)];
143 T9 = R1[WS(rs, 4)];
144 Ta = T8 - T9;
145 Tn = T8 + T9;
146 Tb = R0[WS(rs, 3)];
147 Tc = R1[0];
148 Td = Tb - Tc;
149 To = Tb + Tc;
150 }
151 Te = Ta + Td;
152 Tv = Tn + To;
153 {
154 E T1, T2, T4, T5;
155 T1 = R0[WS(rs, 1)];
156 T2 = R1[WS(rs, 3)];
157 T3 = T1 - T2;
158 Tq = T1 + T2;
159 T4 = R0[WS(rs, 4)];
160 T5 = R1[WS(rs, 1)];
161 T6 = T4 - T5;
162 Tr = T4 + T5;
163 }
164 T7 = T3 + T6;
165 Tu = Tq + Tr;
166 {
167 E Tl, Tm, Tf, Tj, Tk;
168 Tl = Td - Ta;
169 Tm = T3 - T6;
170 Ci[WS(csi, 1)] = FNMS(KP951056516, Tm, KP587785252 * Tl);
171 Ci[WS(csi, 3)] = FMA(KP587785252, Tm, KP951056516 * Tl);
172 Tf = KP559016994 * (T7 - Te);
173 Tj = T7 + Te;
174 Tk = FNMS(KP250000000, Tj, Ti);
175 Cr[WS(csr, 1)] = Tf + Tk;
176 Cr[WS(csr, 5)] = Ti + Tj;
177 Cr[WS(csr, 3)] = Tk - Tf;
178 }
179 {
180 E Tp, Ts, Ty, Tw, Tx;
181 Tp = Tn - To;
182 Ts = Tq - Tr;
183 Ci[WS(csi, 2)] = FNMS(KP587785252, Ts, KP951056516 * Tp);
184 Ci[WS(csi, 4)] = FMA(KP951056516, Ts, KP587785252 * Tp);
185 Ty = KP559016994 * (Tu - Tv);
186 Tw = Tu + Tv;
187 Tx = FNMS(KP250000000, Tw, Tt);
188 Cr[WS(csr, 2)] = Tx - Ty;
189 Cr[0] = Tt + Tw;
190 Cr[WS(csr, 4)] = Ty + Tx;
191 }
192 }
193 }
194 }
195
196 static const kr2c_desc desc = { 10, "r2cf_10", {28, 6, 6, 0}, &GENUS };
197
198 void X(codelet_r2cf_10) (planner *p) {
199 X(kr2c_register) (p, r2cf_10, &desc);
200 }
201
202 #endif