Mercurial > hg > sv-dependency-builds

comparison src/fftw-3.3.8/rdft/scalar/r2cf/r2cf_12.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 12 -name r2cf_12 -include rdft/scalar/r2cf.h */
29
30 /*
31 * This function contains 38 FP additions, 10 FP multiplications,
32 * (or, 30 additions, 2 multiplications, 8 fused multiply/add),
33 * 21 stack variables, 2 constants, and 24 memory accesses
34 */
35 #include "rdft/scalar/r2cf.h"
36
37 static void r2cf_12(R *R0, R *R1, R *Cr, R *Ci, stride rs, stride csr, stride csi, INT v, INT ivs, INT ovs)
38 {
39 DK(KP866025403, +0.866025403784438646763723170752936183471402627);
40 DK(KP500000000, +0.500000000000000000000000000000000000000000000);
41 {
42 INT i;
43 for (i = v; i > 0; i = i - 1, R0 = R0 + ivs, R1 = R1 + ivs, Cr = Cr + ovs, Ci = Ci + ovs, MAKE_VOLATILE_STRIDE(48, rs), MAKE_VOLATILE_STRIDE(48, csr), MAKE_VOLATILE_STRIDE(48, csi)) {
44 E T5, Tp, Tm, Tk, Ty, Tt, Ta, Tq, Tn, Tf, Tz, Tu, Tl, To;
45 {
46 E T1, T2, T3, T4;
47 T1 = R0[0];
48 T2 = R0[WS(rs, 2)];
49 T3 = R0[WS(rs, 4)];
50 T4 = T2 + T3;
51 T5 = T1 + T4;
52 Tp = FNMS(KP500000000, T4, T1);
53 Tm = T3 - T2;
54 }
55 {
56 E Tg, Th, Ti, Tj;
57 Tg = R1[WS(rs, 1)];
58 Th = R1[WS(rs, 3)];
59 Ti = R1[WS(rs, 5)];
60 Tj = Th + Ti;
61 Tk = FNMS(KP500000000, Tj, Tg);
62 Ty = Ti - Th;
63 Tt = Tg + Tj;
64 }
65 {
66 E T6, T7, T8, T9;
67 T6 = R0[WS(rs, 3)];
68 T7 = R0[WS(rs, 5)];
69 T8 = R0[WS(rs, 1)];
70 T9 = T7 + T8;
71 Ta = T6 + T9;
72 Tq = FNMS(KP500000000, T9, T6);
73 Tn = T8 - T7;
74 }
75 {
76 E Tb, Tc, Td, Te;
77 Tb = R1[WS(rs, 4)];
78 Tc = R1[0];
79 Td = R1[WS(rs, 2)];
80 Te = Tc + Td;
81 Tf = FNMS(KP500000000, Te, Tb);
82 Tz = Td - Tc;
83 Tu = Tb + Te;
84 }
85 Cr[WS(csr, 3)] = T5 - Ta;
86 Ci[WS(csi, 3)] = Tt - Tu;
87 Tl = Tf - Tk;
88 To = Tm - Tn;
89 Ci[WS(csi, 1)] = FMA(KP866025403, To, Tl);
90 Ci[WS(csi, 5)] = FNMS(KP866025403, To, Tl);
91 {
92 E Tx, TA, Tv, Tw;
93 Tx = Tp - Tq;
94 TA = Ty - Tz;
95 Cr[WS(csr, 5)] = FNMS(KP866025403, TA, Tx);
96 Cr[WS(csr, 1)] = FMA(KP866025403, TA, Tx);
97 Tv = T5 + Ta;
98 Tw = Tt + Tu;
99 Cr[WS(csr, 6)] = Tv - Tw;
100 Cr[0] = Tv + Tw;
101 }
102 {
103 E Tr, Ts, TB, TC;
104 Tr = Tp + Tq;
105 Ts = Tk + Tf;
106 Cr[WS(csr, 2)] = Tr - Ts;
107 Cr[WS(csr, 4)] = Tr + Ts;
108 TB = Ty + Tz;
109 TC = Tm + Tn;
110 Ci[WS(csi, 2)] = KP866025403 * (TB - TC);
111 Ci[WS(csi, 4)] = KP866025403 * (TC + TB);
112 }
113 }
114 }
115 }
116
117 static const kr2c_desc desc = { 12, "r2cf_12", {30, 2, 8, 0}, &GENUS };
118
119 void X(codelet_r2cf_12) (planner *p) {
120 X(kr2c_register) (p, r2cf_12, &desc);
121 }
122
123 #else
124
125 /* Generated by: ../../../genfft/gen_r2cf.native -compact -variables 4 -pipeline-latency 4 -n 12 -name r2cf_12 -include rdft/scalar/r2cf.h */
126
127 /*
128 * This function contains 38 FP additions, 8 FP multiplications,
129 * (or, 34 additions, 4 multiplications, 4 fused multiply/add),
130 * 21 stack variables, 2 constants, and 24 memory accesses
131 */
132 #include "rdft/scalar/r2cf.h"
133
134 static void r2cf_12(R *R0, R *R1, R *Cr, R *Ci, stride rs, stride csr, stride csi, INT v, INT ivs, INT ovs)
135 {
136 DK(KP866025403, +0.866025403784438646763723170752936183471402627);
137 DK(KP500000000, +0.500000000000000000000000000000000000000000000);
138 {
139 INT i;
140 for (i = v; i > 0; i = i - 1, R0 = R0 + ivs, R1 = R1 + ivs, Cr = Cr + ovs, Ci = Ci + ovs, MAKE_VOLATILE_STRIDE(48, rs), MAKE_VOLATILE_STRIDE(48, csr), MAKE_VOLATILE_STRIDE(48, csi)) {
141 E T5, Tp, Tb, Tn, Ty, Tt, Ta, Tq, Tc, Ti, Tz, Tu, Td, To;
142 {
143 E T1, T2, T3, T4;
144 T1 = R0[0];
145 T2 = R0[WS(rs, 2)];
146 T3 = R0[WS(rs, 4)];
147 T4 = T2 + T3;
148 T5 = T1 + T4;
149 Tp = FNMS(KP500000000, T4, T1);
150 Tb = T3 - T2;
151 }
152 {
153 E Tj, Tk, Tl, Tm;
154 Tj = R1[WS(rs, 1)];
155 Tk = R1[WS(rs, 3)];
156 Tl = R1[WS(rs, 5)];
157 Tm = Tk + Tl;
158 Tn = FNMS(KP500000000, Tm, Tj);
159 Ty = Tl - Tk;
160 Tt = Tj + Tm;
161 }
162 {
163 E T6, T7, T8, T9;
164 T6 = R0[WS(rs, 3)];
165 T7 = R0[WS(rs, 5)];
166 T8 = R0[WS(rs, 1)];
167 T9 = T7 + T8;
168 Ta = T6 + T9;
169 Tq = FNMS(KP500000000, T9, T6);
170 Tc = T8 - T7;
171 }
172 {
173 E Te, Tf, Tg, Th;
174 Te = R1[WS(rs, 4)];
175 Tf = R1[0];
176 Tg = R1[WS(rs, 2)];
177 Th = Tf + Tg;
178 Ti = FNMS(KP500000000, Th, Te);
179 Tz = Tg - Tf;
180 Tu = Te + Th;
181 }
182 Cr[WS(csr, 3)] = T5 - Ta;
183 Ci[WS(csi, 3)] = Tt - Tu;
184 Td = KP866025403 * (Tb - Tc);
185 To = Ti - Tn;
186 Ci[WS(csi, 1)] = Td + To;
187 Ci[WS(csi, 5)] = To - Td;
188 {
189 E Tx, TA, Tv, Tw;
190 Tx = Tp - Tq;
191 TA = KP866025403 * (Ty - Tz);
192 Cr[WS(csr, 5)] = Tx - TA;
193 Cr[WS(csr, 1)] = Tx + TA;
194 Tv = T5 + Ta;
195 Tw = Tt + Tu;
196 Cr[WS(csr, 6)] = Tv - Tw;
197 Cr[0] = Tv + Tw;
198 }
199 {
200 E Tr, Ts, TB, TC;
201 Tr = Tp + Tq;
202 Ts = Tn + Ti;
203 Cr[WS(csr, 2)] = Tr - Ts;
204 Cr[WS(csr, 4)] = Tr + Ts;
205 TB = Ty + Tz;
206 TC = Tb + Tc;
207 Ci[WS(csi, 2)] = KP866025403 * (TB - TC);
208 Ci[WS(csi, 4)] = KP866025403 * (TC + TB);
209 }
210 }
211 }
212 }
213
214 static const kr2c_desc desc = { 12, "r2cf_12", {34, 4, 4, 0}, &GENUS };
215
216 void X(codelet_r2cf_12) (planner *p) {
217 X(kr2c_register) (p, r2cf_12, &desc);
218 }
219
220 #endif