Mercurial > hg > sv-dependency-builds

comparison src/fftw-3.3.8/rdft/scalar/r2cb/r2cbIII_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 08:07:44 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_r2cb.native -fma -compact -variables 4 -pipeline-latency 4 -sign 1 -n 12 -name r2cbIII_12 -dft-III -include rdft/scalar/r2cbIII.h */
29
30 /*
31 * This function contains 42 FP additions, 20 FP multiplications,
32 * (or, 30 additions, 8 multiplications, 12 fused multiply/add),
33 * 25 stack variables, 4 constants, and 24 memory accesses
34 */
35 #include "rdft/scalar/r2cbIII.h"
36
37 static void r2cbIII_12(R *R0, R *R1, R *Cr, R *Ci, stride rs, stride csr, stride csi, INT v, INT ivs, INT ovs)
38 {
39 DK(KP707106781, +0.707106781186547524400844362104849039284835938);
40 DK(KP1_732050807, +1.732050807568877293527446341505872366942805254);
41 DK(KP1_414213562, +1.414213562373095048801688724209698078569671875);
42 DK(KP2_000000000, +2.000000000000000000000000000000000000000000000);
43 {
44 INT i;
45 for (i = v; i > 0; i = i - 1, R0 = R0 + ovs, R1 = R1 + ovs, Cr = Cr + ivs, Ci = Ci + ivs, MAKE_VOLATILE_STRIDE(48, rs), MAKE_VOLATILE_STRIDE(48, csr), MAKE_VOLATILE_STRIDE(48, csi)) {
46 E T5, Tx, Tb, Te, Tw, Ts, Ta, TA, Tg, Tj, Tz, Tp, Tt, Tu;
47 {
48 E T1, T2, T3, T4;
49 T1 = Cr[WS(csr, 1)];
50 T2 = Cr[WS(csr, 5)];
51 T3 = Cr[WS(csr, 2)];
52 T4 = T2 + T3;
53 T5 = T1 + T4;
54 Tx = T2 - T3;
55 Tb = FNMS(KP2_000000000, T1, T4);
56 }
57 {
58 E Tq, Tc, Td, Tr;
59 Tq = Ci[WS(csi, 1)];
60 Tc = Ci[WS(csi, 5)];
61 Td = Ci[WS(csi, 2)];
62 Tr = Td - Tc;
63 Te = Tc + Td;
64 Tw = FMA(KP2_000000000, Tq, Tr);
65 Ts = Tq - Tr;
66 }
67 {
68 E T6, T7, T8, T9;
69 T6 = Cr[WS(csr, 4)];
70 T7 = Cr[0];
71 T8 = Cr[WS(csr, 3)];
72 T9 = T7 + T8;
73 Ta = T6 + T9;
74 TA = T7 - T8;
75 Tg = FNMS(KP2_000000000, T6, T9);
76 }
77 {
78 E To, Th, Ti, Tn;
79 To = Ci[WS(csi, 4)];
80 Th = Ci[0];
81 Ti = Ci[WS(csi, 3)];
82 Tn = Ti - Th;
83 Tj = Th + Ti;
84 Tz = FMA(KP2_000000000, To, Tn);
85 Tp = Tn - To;
86 }
87 R0[0] = KP2_000000000 * (T5 + Ta);
88 R0[WS(rs, 3)] = KP2_000000000 * (Ts + Tp);
89 Tt = Tp - Ts;
90 Tu = T5 - Ta;
91 R1[WS(rs, 1)] = KP1_414213562 * (Tt - Tu);
92 R1[WS(rs, 4)] = KP1_414213562 * (Tu + Tt);
93 {
94 E Tf, Tk, Tv, Ty, TB, TC;
95 Tf = FMA(KP1_732050807, Te, Tb);
96 Tk = FNMS(KP1_732050807, Tj, Tg);
97 Tv = Tk - Tf;
98 Ty = FMA(KP1_732050807, Tx, Tw);
99 TB = FNMS(KP1_732050807, TA, Tz);
100 TC = Ty + TB;
101 R0[WS(rs, 2)] = Tf + Tk;
102 R0[WS(rs, 5)] = TB - Ty;
103 R1[0] = KP707106781 * (Tv - TC);
104 R1[WS(rs, 3)] = KP707106781 * (Tv + TC);
105 }
106 {
107 E Tl, Tm, TF, TD, TE, TG;
108 Tl = FNMS(KP1_732050807, Te, Tb);
109 Tm = FMA(KP1_732050807, Tj, Tg);
110 TF = Tl - Tm;
111 TD = FMA(KP1_732050807, TA, Tz);
112 TE = FNMS(KP1_732050807, Tx, Tw);
113 TG = TE + TD;
114 R0[WS(rs, 4)] = -(Tl + Tm);
115 R1[WS(rs, 2)] = KP707106781 * (TF + TG);
116 R0[WS(rs, 1)] = TD - TE;
117 R1[WS(rs, 5)] = KP707106781 * (TF - TG);
118 }
119 }
120 }
121 }
122
123 static const kr2c_desc desc = { 12, "r2cbIII_12", {30, 8, 12, 0}, &GENUS };
124
125 void X(codelet_r2cbIII_12) (planner *p) {
126 X(kr2c_register) (p, r2cbIII_12, &desc);
127 }
128
129 #else
130
131 /* Generated by: ../../../genfft/gen_r2cb.native -compact -variables 4 -pipeline-latency 4 -sign 1 -n 12 -name r2cbIII_12 -dft-III -include rdft/scalar/r2cbIII.h */
132
133 /*
134 * This function contains 42 FP additions, 20 FP multiplications,
135 * (or, 38 additions, 16 multiplications, 4 fused multiply/add),
136 * 25 stack variables, 4 constants, and 24 memory accesses
137 */
138 #include "rdft/scalar/r2cbIII.h"
139
140 static void r2cbIII_12(R *R0, R *R1, R *Cr, R *Ci, stride rs, stride csr, stride csi, INT v, INT ivs, INT ovs)
141 {
142 DK(KP1_414213562, +1.414213562373095048801688724209698078569671875);
143 DK(KP2_000000000, +2.000000000000000000000000000000000000000000000);
144 DK(KP500000000, +0.500000000000000000000000000000000000000000000);
145 DK(KP866025403, +0.866025403784438646763723170752936183471402627);
146 {
147 INT i;
148 for (i = v; i > 0; i = i - 1, R0 = R0 + ovs, R1 = R1 + ovs, Cr = Cr + ivs, Ci = Ci + ivs, MAKE_VOLATILE_STRIDE(48, rs), MAKE_VOLATILE_STRIDE(48, csr), MAKE_VOLATILE_STRIDE(48, csi)) {
149 E T5, Tw, Tb, Te, Tx, Ts, Ta, TA, Tg, Tj, Tz, Tp, Tt, Tu;
150 {
151 E T1, T2, T3, T4;
152 T1 = Cr[WS(csr, 1)];
153 T2 = Cr[WS(csr, 5)];
154 T3 = Cr[WS(csr, 2)];
155 T4 = T2 + T3;
156 T5 = T1 + T4;
157 Tw = KP866025403 * (T2 - T3);
158 Tb = FNMS(KP500000000, T4, T1);
159 }
160 {
161 E Tq, Tc, Td, Tr;
162 Tq = Ci[WS(csi, 1)];
163 Tc = Ci[WS(csi, 5)];
164 Td = Ci[WS(csi, 2)];
165 Tr = Td - Tc;
166 Te = KP866025403 * (Tc + Td);
167 Tx = FMA(KP500000000, Tr, Tq);
168 Ts = Tq - Tr;
169 }
170 {
171 E T6, T7, T8, T9;
172 T6 = Cr[WS(csr, 4)];
173 T7 = Cr[0];
174 T8 = Cr[WS(csr, 3)];
175 T9 = T7 + T8;
176 Ta = T6 + T9;
177 TA = KP866025403 * (T7 - T8);
178 Tg = FNMS(KP500000000, T9, T6);
179 }
180 {
181 E To, Th, Ti, Tn;
182 To = Ci[WS(csi, 4)];
183 Th = Ci[0];
184 Ti = Ci[WS(csi, 3)];
185 Tn = Ti - Th;
186 Tj = KP866025403 * (Th + Ti);
187 Tz = FMA(KP500000000, Tn, To);
188 Tp = Tn - To;
189 }
190 R0[0] = KP2_000000000 * (T5 + Ta);
191 R0[WS(rs, 3)] = KP2_000000000 * (Ts + Tp);
192 Tt = Tp - Ts;
193 Tu = T5 - Ta;
194 R1[WS(rs, 1)] = KP1_414213562 * (Tt - Tu);
195 R1[WS(rs, 4)] = KP1_414213562 * (Tu + Tt);
196 {
197 E Tf, Tk, Tv, Ty, TB, TC;
198 Tf = Tb - Te;
199 Tk = Tg + Tj;
200 Tv = Tf - Tk;
201 Ty = Tw + Tx;
202 TB = Tz - TA;
203 TC = Ty + TB;
204 R0[WS(rs, 2)] = -(KP2_000000000 * (Tf + Tk));
205 R0[WS(rs, 5)] = KP2_000000000 * (TB - Ty);
206 R1[0] = KP1_414213562 * (Tv - TC);
207 R1[WS(rs, 3)] = KP1_414213562 * (Tv + TC);
208 }
209 {
210 E Tl, Tm, TF, TD, TE, TG;
211 Tl = Tb + Te;
212 Tm = Tg - Tj;
213 TF = Tm - Tl;
214 TD = TA + Tz;
215 TE = Tx - Tw;
216 TG = TE + TD;
217 R0[WS(rs, 4)] = KP2_000000000 * (Tl + Tm);
218 R1[WS(rs, 2)] = KP1_414213562 * (TF + TG);
219 R0[WS(rs, 1)] = KP2_000000000 * (TD - TE);
220 R1[WS(rs, 5)] = KP1_414213562 * (TF - TG);
221 }
222 }
223 }
224 }
225
226 static const kr2c_desc desc = { 12, "r2cbIII_12", {38, 16, 4, 0}, &GENUS };
227
228 void X(codelet_r2cbIII_12) (planner *p) {
229 X(kr2c_register) (p, r2cbIII_12, &desc);
230 }
231
232 #endif