Mercurial > hg > sv-dependency-builds

comparison src/fftw-3.3.8/rdft/scalar/r2cf/r2cfII_15.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 08:06:43 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 15 -name r2cfII_15 -dft-II -include rdft/scalar/r2cfII.h */
29
30 /*
31 * This function contains 72 FP additions, 41 FP multiplications,
32 * (or, 38 additions, 7 multiplications, 34 fused multiply/add),
33 * 42 stack variables, 12 constants, and 30 memory accesses
34 */
35 #include "rdft/scalar/r2cfII.h"
36
37 static void r2cfII_15(R *R0, R *R1, R *Cr, R *Ci, stride rs, stride csr, stride csi, INT v, INT ivs, INT ovs)
38 {
39 DK(KP823639103, +0.823639103546331925877420039278190003029660514);
40 DK(KP910592997, +0.910592997310029334643087372129977886038870291);
41 DK(KP951056516, +0.951056516295153572116439333379382143405698634);
42 DK(KP559016994, +0.559016994374947424102293417182819058860154590);
43 DK(KP866025403, +0.866025403784438646763723170752936183471402627);
44 DK(KP500000000, +0.500000000000000000000000000000000000000000000);
45 DK(KP690983005, +0.690983005625052575897706582817180941139845410);
46 DK(KP447213595, +0.447213595499957939281834733746255247088123672);
47 DK(KP552786404, +0.552786404500042060718165266253744752911876328);
48 DK(KP809016994, +0.809016994374947424102293417182819058860154590);
49 DK(KP618033988, +0.618033988749894848204586834365638117720309180);
50 DK(KP250000000, +0.250000000000000000000000000000000000000000000);
51 {
52 INT i;
53 for (i = v; i > 0; i = i - 1, R0 = R0 + ivs, R1 = R1 + ivs, Cr = Cr + ovs, Ci = Ci + ovs, MAKE_VOLATILE_STRIDE(60, rs), MAKE_VOLATILE_STRIDE(60, csr), MAKE_VOLATILE_STRIDE(60, csi)) {
54 E Ta, Tl, T1, T6, T7, TX, TT, T8, Tg, Th, TM, TZ, Tj, Tz, Tr;
55 E Ts, TP, TY, Tu, TC;
56 Ta = R0[WS(rs, 5)];
57 Tl = R1[WS(rs, 2)];
58 {
59 E T2, T5, T3, T4, TR, TS;
60 T1 = R0[0];
61 T2 = R0[WS(rs, 3)];
62 T5 = R1[WS(rs, 4)];
63 T3 = R0[WS(rs, 6)];
64 T4 = R1[WS(rs, 1)];
65 TR = T2 + T5;
66 TS = T3 + T4;
67 T6 = T2 + T3 - T4 - T5;
68 T7 = FNMS(KP250000000, T6, T1);
69 TX = FNMS(KP618033988, TR, TS);
70 TT = FMA(KP618033988, TS, TR);
71 T8 = (T3 + T5 - T2) - T4;
72 }
73 {
74 E Tf, TL, TK, Ti, Ty;
75 {
76 E Tb, Tc, Td, Te;
77 Tb = R1[0];
78 Tg = R0[WS(rs, 2)];
79 Tc = R1[WS(rs, 3)];
80 Td = R1[WS(rs, 6)];
81 Te = Tc + Td;
82 Tf = Tb - Te;
83 TL = Tc - Td;
84 Th = Tb + Te;
85 TK = Tg + Tb;
86 }
87 TM = FMA(KP618033988, TL, TK);
88 TZ = FNMS(KP618033988, TK, TL);
89 Ti = FMA(KP809016994, Th, Tg);
90 Tj = FNMS(KP552786404, Ti, Tf);
91 Ty = FMA(KP447213595, Th, Tf);
92 Tz = FNMS(KP690983005, Ty, Tg);
93 }
94 {
95 E Tq, TO, TN, Tt, TB;
96 {
97 E Tm, Tn, To, Tp;
98 Tm = R0[WS(rs, 7)];
99 Tr = R1[WS(rs, 5)];
100 Tn = R0[WS(rs, 1)];
101 To = R0[WS(rs, 4)];
102 Tp = Tn + To;
103 Tq = Tm - Tp;
104 TO = To - Tn;
105 Ts = Tm + Tp;
106 TN = Tr + Tm;
107 }
108 TP = FMA(KP618033988, TO, TN);
109 TY = FNMS(KP618033988, TN, TO);
110 Tt = FMA(KP809016994, Ts, Tr);
111 Tu = FNMS(KP552786404, Tt, Tq);
112 TB = FMA(KP447213595, Ts, Tq);
113 TC = FNMS(KP690983005, TB, Tr);
114 }
115 {
116 E TF, TG, TH, TI;
117 TF = T1 + T6;
118 TG = Ts - Tr - Tl;
119 TH = Ta + Tg - Th;
120 TI = TG + TH;
121 Cr[WS(csr, 2)] = FNMS(KP500000000, TI, TF);
122 Ci[WS(csi, 2)] = KP866025403 * (TH - TG);
123 Cr[WS(csr, 7)] = TF + TI;
124 }
125 {
126 E Tx, T14, T10, T11, TE, T12, TA, TD, T13;
127 Tx = FMA(KP559016994, T8, T7);
128 T14 = TZ - TY;
129 T10 = TY + TZ;
130 T11 = FMA(KP500000000, T10, TX);
131 TA = FNMS(KP809016994, Tz, Ta);
132 TD = FNMS(KP809016994, TC, Tl);
133 TE = TA - TD;
134 T12 = TD + TA;
135 Cr[WS(csr, 1)] = Tx + TE;
136 Ci[WS(csi, 1)] = KP951056516 * (T10 - TX);
137 Ci[WS(csi, 3)] = KP951056516 * (FNMS(KP910592997, T12, T11));
138 Ci[WS(csi, 6)] = -(KP951056516 * (FMA(KP910592997, T12, T11)));
139 T13 = FNMS(KP500000000, TE, Tx);
140 Cr[WS(csr, 3)] = FNMS(KP823639103, T14, T13);
141 Cr[WS(csr, 6)] = FMA(KP823639103, T14, T13);
142 }
143 {
144 E T9, TQ, TU, TV, Tw, TW, Tk, Tv, TJ;
145 T9 = FNMS(KP559016994, T8, T7);
146 TQ = TM - TP;
147 TU = TP + TM;
148 TV = FMA(KP500000000, TU, TT);
149 Tk = FNMS(KP559016994, Tj, Ta);
150 Tv = FNMS(KP559016994, Tu, Tl);
151 Tw = Tk - Tv;
152 TW = Tv + Tk;
153 Cr[WS(csr, 4)] = T9 + Tw;
154 Ci[WS(csi, 4)] = KP951056516 * (TT - TU);
155 Ci[0] = -(KP951056516 * (FMA(KP910592997, TW, TV)));
156 Ci[WS(csi, 5)] = -(KP951056516 * (FNMS(KP910592997, TW, TV)));
157 TJ = FNMS(KP500000000, Tw, T9);
158 Cr[WS(csr, 5)] = FNMS(KP823639103, TQ, TJ);
159 Cr[0] = FMA(KP823639103, TQ, TJ);
160 }
161 }
162 }
163 }
164
165 static const kr2c_desc desc = { 15, "r2cfII_15", {38, 7, 34, 0}, &GENUS };
166
167 void X(codelet_r2cfII_15) (planner *p) {
168 X(kr2c_register) (p, r2cfII_15, &desc);
169 }
170
171 #else
172
173 /* Generated by: ../../../genfft/gen_r2cf.native -compact -variables 4 -pipeline-latency 4 -n 15 -name r2cfII_15 -dft-II -include rdft/scalar/r2cfII.h */
174
175 /*
176 * This function contains 72 FP additions, 33 FP multiplications,
177 * (or, 54 additions, 15 multiplications, 18 fused multiply/add),
178 * 37 stack variables, 8 constants, and 30 memory accesses
179 */
180 #include "rdft/scalar/r2cfII.h"
181
182 static void r2cfII_15(R *R0, R *R1, R *Cr, R *Ci, stride rs, stride csr, stride csi, INT v, INT ivs, INT ovs)
183 {
184 DK(KP500000000, +0.500000000000000000000000000000000000000000000);
185 DK(KP866025403, +0.866025403784438646763723170752936183471402627);
186 DK(KP809016994, +0.809016994374947424102293417182819058860154590);
187 DK(KP309016994, +0.309016994374947424102293417182819058860154590);
188 DK(KP250000000, +0.250000000000000000000000000000000000000000000);
189 DK(KP559016994, +0.559016994374947424102293417182819058860154590);
190 DK(KP587785252, +0.587785252292473129168705954639072768597652438);
191 DK(KP951056516, +0.951056516295153572116439333379382143405698634);
192 {
193 INT i;
194 for (i = v; i > 0; i = i - 1, R0 = R0 + ivs, R1 = R1 + ivs, Cr = Cr + ovs, Ci = Ci + ovs, MAKE_VOLATILE_STRIDE(60, rs), MAKE_VOLATILE_STRIDE(60, csr), MAKE_VOLATILE_STRIDE(60, csi)) {
195 E T1, T2, Tx, TR, TE, T7, TD, Th, Tm, Tr, TQ, TA, TB, Tf, Te;
196 E Tu, TS, Td, TH, TO;
197 T1 = R0[WS(rs, 5)];
198 {
199 E T3, Tv, T6, Tw, T4, T5;
200 T2 = R0[WS(rs, 2)];
201 T3 = R1[0];
202 Tv = T2 + T3;
203 T4 = R1[WS(rs, 3)];
204 T5 = R1[WS(rs, 6)];
205 T6 = T4 + T5;
206 Tw = T4 - T5;
207 Tx = FMA(KP951056516, Tv, KP587785252 * Tw);
208 TR = FNMS(KP587785252, Tv, KP951056516 * Tw);
209 TE = KP559016994 * (T3 - T6);
210 T7 = T3 + T6;
211 TD = KP250000000 * T7;
212 }
213 {
214 E Ti, Tl, Tj, Tk, Tp, Tq;
215 Th = R0[0];
216 Ti = R1[WS(rs, 4)];
217 Tl = R0[WS(rs, 6)];
218 Tj = R1[WS(rs, 1)];
219 Tk = R0[WS(rs, 3)];
220 Tp = Tk + Ti;
221 Tq = Tl + Tj;
222 Tm = Ti + Tj - (Tk + Tl);
223 Tr = FMA(KP951056516, Tp, KP587785252 * Tq);
224 TQ = FNMS(KP951056516, Tq, KP587785252 * Tp);
225 TA = FMA(KP250000000, Tm, Th);
226 TB = KP559016994 * (Tl + Ti - (Tk + Tj));
227 }
228 {
229 E T9, Tt, Tc, Ts, Ta, Tb, TG;
230 Tf = R1[WS(rs, 2)];
231 T9 = R0[WS(rs, 7)];
232 Te = R1[WS(rs, 5)];
233 Tt = T9 + Te;
234 Ta = R0[WS(rs, 1)];
235 Tb = R0[WS(rs, 4)];
236 Tc = Ta + Tb;
237 Ts = Ta - Tb;
238 Tu = FNMS(KP951056516, Tt, KP587785252 * Ts);
239 TS = FMA(KP951056516, Ts, KP587785252 * Tt);
240 Td = T9 + Tc;
241 TG = KP559016994 * (T9 - Tc);
242 TH = FNMS(KP309016994, Te, TG) + FNMA(KP250000000, Td, Tf);
243 TO = FMS(KP809016994, Te, Tf) + FNMA(KP250000000, Td, TG);
244 }
245 {
246 E Tn, T8, Tg, To;
247 Tn = Th - Tm;
248 T8 = T1 + T2 - T7;
249 Tg = Td - Te - Tf;
250 To = T8 + Tg;
251 Ci[WS(csi, 2)] = KP866025403 * (T8 - Tg);
252 Cr[WS(csr, 2)] = FNMS(KP500000000, To, Tn);
253 Cr[WS(csr, 7)] = Tn + To;
254 }
255 {
256 E TM, TX, TT, TV, TP, TU, TN, TW;
257 TM = TB + TA;
258 TX = KP866025403 * (TR + TS);
259 TT = TR - TS;
260 TV = FMS(KP500000000, TT, TQ);
261 TN = T1 + TE + FNMS(KP809016994, T2, TD);
262 TP = TN + TO;
263 TU = KP866025403 * (TO - TN);
264 Cr[WS(csr, 1)] = TM + TP;
265 Ci[WS(csi, 1)] = TQ + TT;
266 Ci[WS(csi, 6)] = TU - TV;
267 Ci[WS(csi, 3)] = TU + TV;
268 TW = FNMS(KP500000000, TP, TM);
269 Cr[WS(csr, 3)] = TW - TX;
270 Cr[WS(csr, 6)] = TW + TX;
271 }
272 {
273 E Tz, TC, Ty, TK, TI, TL, TF, TJ;
274 Tz = KP866025403 * (Tx + Tu);
275 TC = TA - TB;
276 Ty = Tu - Tx;
277 TK = FMS(KP500000000, Ty, Tr);
278 TF = FMA(KP309016994, T2, T1) + TD - TE;
279 TI = TF + TH;
280 TL = KP866025403 * (TH - TF);
281 Ci[WS(csi, 4)] = Tr + Ty;
282 Cr[WS(csr, 4)] = TC + TI;
283 Ci[WS(csi, 5)] = TK - TL;
284 Ci[0] = TK + TL;
285 TJ = FNMS(KP500000000, TI, TC);
286 Cr[0] = Tz + TJ;
287 Cr[WS(csr, 5)] = TJ - Tz;
288 }
289 }
290 }
291 }
292
293 static const kr2c_desc desc = { 15, "r2cfII_15", {54, 15, 18, 0}, &GENUS };
294
295 void X(codelet_r2cfII_15) (planner *p) {
296 X(kr2c_register) (p, r2cfII_15, &desc);
297 }
298
299 #endif