Mercurial > hg > sv-dependency-builds

comparison src/fftw-3.3.8/rdft/scalar/r2cf/r2cf_20.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 08:06:27 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 20 -name r2cf_20 -include rdft/scalar/r2cf.h */
29
30 /*
31 * This function contains 86 FP additions, 32 FP multiplications,
32 * (or, 58 additions, 4 multiplications, 28 fused multiply/add),
33 * 51 stack variables, 4 constants, and 40 memory accesses
34 */
35 #include "rdft/scalar/r2cf.h"
36
37 static void r2cf_20(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(80, rs), MAKE_VOLATILE_STRIDE(80, csr), MAKE_VOLATILE_STRIDE(80, csi)) {
46 E T3, T1d, TJ, TV, T16, T1k, T1l, T19, Ta, Th, Ti, T1e, T1f, T1g, TP;
47 E TQ, TX, Tn, Ts, TK, TS, TT, TW, Ty, TD, TL;
48 {
49 E T1, T2, TF, TG, TH, TI;
50 T1 = R0[0];
51 T2 = R0[WS(rs, 5)];
52 TF = T1 + T2;
53 TG = R1[WS(rs, 2)];
54 TH = R1[WS(rs, 7)];
55 TI = TG + TH;
56 T3 = T1 - T2;
57 T1d = TG - TH;
58 TJ = TF - TI;
59 TV = TF + TI;
60 }
61 {
62 E T6, To, Tx, T17, TC, T18, T9, Tj, Td, Tu, Tm, T15, Tr, T14, Tg;
63 E Tz;
64 {
65 E T4, T5, Tv, Tw;
66 T4 = R0[WS(rs, 2)];
67 T5 = R0[WS(rs, 7)];
68 T6 = T4 - T5;
69 To = T4 + T5;
70 Tv = R1[WS(rs, 6)];
71 Tw = R1[WS(rs, 1)];
72 Tx = Tv + Tw;
73 T17 = Tw - Tv;
74 }
75 {
76 E TA, TB, T7, T8;
77 TA = R1[WS(rs, 8)];
78 TB = R1[WS(rs, 3)];
79 TC = TA + TB;
80 T18 = TB - TA;
81 T7 = R0[WS(rs, 8)];
82 T8 = R0[WS(rs, 3)];
83 T9 = T7 - T8;
84 Tj = T7 + T8;
85 }
86 {
87 E Tb, Tc, Tk, Tl;
88 Tb = R0[WS(rs, 4)];
89 Tc = R0[WS(rs, 9)];
90 Td = Tb - Tc;
91 Tu = Tb + Tc;
92 Tk = R1[0];
93 Tl = R1[WS(rs, 5)];
94 Tm = Tk + Tl;
95 T15 = Tl - Tk;
96 }
97 {
98 E Tp, Tq, Te, Tf;
99 Tp = R1[WS(rs, 4)];
100 Tq = R1[WS(rs, 9)];
101 Tr = Tp + Tq;
102 T14 = Tq - Tp;
103 Te = R0[WS(rs, 6)];
104 Tf = R0[WS(rs, 1)];
105 Tg = Te - Tf;
106 Tz = Te + Tf;
107 }
108 T16 = T14 - T15;
109 T1k = T6 - T9;
110 T1l = Td - Tg;
111 T19 = T17 - T18;
112 Ta = T6 + T9;
113 Th = Td + Tg;
114 Ti = Ta + Th;
115 T1e = T14 + T15;
116 T1f = T17 + T18;
117 T1g = T1e + T1f;
118 TP = Tu + Tx;
119 TQ = Tz + TC;
120 TX = TP + TQ;
121 Tn = Tj - Tm;
122 Ts = To - Tr;
123 TK = Ts + Tn;
124 TS = To + Tr;
125 TT = Tj + Tm;
126 TW = TS + TT;
127 Ty = Tu - Tx;
128 TD = Tz - TC;
129 TL = Ty + TD;
130 }
131 Cr[WS(csr, 5)] = T3 + Ti;
132 Ci[WS(csi, 5)] = T1g - T1d;
133 {
134 E Tt, TE, TR, TU;
135 Tt = Tn - Ts;
136 TE = Ty - TD;
137 Ci[WS(csi, 6)] = KP951056516 * (FNMS(KP618033988, TE, Tt));
138 Ci[WS(csi, 2)] = KP951056516 * (FMA(KP618033988, Tt, TE));
139 TR = TP - TQ;
140 TU = TS - TT;
141 Ci[WS(csi, 8)] = -(KP951056516 * (FNMS(KP618033988, TU, TR)));
142 Ci[WS(csi, 4)] = KP951056516 * (FMA(KP618033988, TR, TU));
143 }
144 {
145 E T10, TY, TZ, TO, TM, TN;
146 T10 = TW - TX;
147 TY = TW + TX;
148 TZ = FNMS(KP250000000, TY, TV);
149 Cr[WS(csr, 4)] = FMA(KP559016994, T10, TZ);
150 Cr[0] = TV + TY;
151 Cr[WS(csr, 8)] = FNMS(KP559016994, T10, TZ);
152 TO = TK - TL;
153 TM = TK + TL;
154 TN = FNMS(KP250000000, TM, TJ);
155 Cr[WS(csr, 2)] = FNMS(KP559016994, TO, TN);
156 Cr[WS(csr, 10)] = TJ + TM;
157 Cr[WS(csr, 6)] = FMA(KP559016994, TO, TN);
158 }
159 {
160 E T1a, T1c, T13, T1b, T11, T12;
161 T1a = FMA(KP618033988, T19, T16);
162 T1c = FNMS(KP618033988, T16, T19);
163 T11 = FNMS(KP250000000, Ti, T3);
164 T12 = Ta - Th;
165 T13 = FMA(KP559016994, T12, T11);
166 T1b = FNMS(KP559016994, T12, T11);
167 Cr[WS(csr, 9)] = FNMS(KP951056516, T1a, T13);
168 Cr[WS(csr, 7)] = FMA(KP951056516, T1c, T1b);
169 Cr[WS(csr, 1)] = FMA(KP951056516, T1a, T13);
170 Cr[WS(csr, 3)] = FNMS(KP951056516, T1c, T1b);
171 }
172 {
173 E T1m, T1o, T1j, T1n, T1h, T1i;
174 T1m = FMA(KP618033988, T1l, T1k);
175 T1o = FNMS(KP618033988, T1k, T1l);
176 T1h = FMA(KP250000000, T1g, T1d);
177 T1i = T1e - T1f;
178 T1j = FNMS(KP559016994, T1i, T1h);
179 T1n = FMA(KP559016994, T1i, T1h);
180 Ci[WS(csi, 1)] = -(FMA(KP951056516, T1m, T1j));
181 Ci[WS(csi, 7)] = FMA(KP951056516, T1o, T1n);
182 Ci[WS(csi, 9)] = FMS(KP951056516, T1m, T1j);
183 Ci[WS(csi, 3)] = FNMS(KP951056516, T1o, T1n);
184 }
185 }
186 }
187 }
188
189 static const kr2c_desc desc = { 20, "r2cf_20", {58, 4, 28, 0}, &GENUS };
190
191 void X(codelet_r2cf_20) (planner *p) {
192 X(kr2c_register) (p, r2cf_20, &desc);
193 }
194
195 #else
196
197 /* Generated by: ../../../genfft/gen_r2cf.native -compact -variables 4 -pipeline-latency 4 -n 20 -name r2cf_20 -include rdft/scalar/r2cf.h */
198
199 /*
200 * This function contains 86 FP additions, 24 FP multiplications,
201 * (or, 74 additions, 12 multiplications, 12 fused multiply/add),
202 * 51 stack variables, 4 constants, and 40 memory accesses
203 */
204 #include "rdft/scalar/r2cf.h"
205
206 static void r2cf_20(R *R0, R *R1, R *Cr, R *Ci, stride rs, stride csr, stride csi, INT v, INT ivs, INT ovs)
207 {
208 DK(KP250000000, +0.250000000000000000000000000000000000000000000);
209 DK(KP559016994, +0.559016994374947424102293417182819058860154590);
210 DK(KP587785252, +0.587785252292473129168705954639072768597652438);
211 DK(KP951056516, +0.951056516295153572116439333379382143405698634);
212 {
213 INT i;
214 for (i = v; i > 0; i = i - 1, R0 = R0 + ivs, R1 = R1 + ivs, Cr = Cr + ovs, Ci = Ci + ovs, MAKE_VOLATILE_STRIDE(80, rs), MAKE_VOLATILE_STRIDE(80, csr), MAKE_VOLATILE_STRIDE(80, csi)) {
215 E T3, T1m, TF, T17, Ts, TM, TN, Tz, Ta, Th, Ti, T1g, T1h, T1k, T10;
216 E T13, T19, TG, TH, TI, T1d, T1e, T1j, TT, TW, T18;
217 {
218 E T1, T2, T15, TD, TE, T16;
219 T1 = R0[0];
220 T2 = R0[WS(rs, 5)];
221 T15 = T1 + T2;
222 TD = R1[WS(rs, 7)];
223 TE = R1[WS(rs, 2)];
224 T16 = TE + TD;
225 T3 = T1 - T2;
226 T1m = T15 + T16;
227 TF = TD - TE;
228 T17 = T15 - T16;
229 }
230 {
231 E T6, TU, Tv, T12, Ty, TZ, T9, TR, Td, TY, To, TS, Tr, TV, Tg;
232 E T11;
233 {
234 E T4, T5, Tt, Tu;
235 T4 = R0[WS(rs, 2)];
236 T5 = R0[WS(rs, 7)];
237 T6 = T4 - T5;
238 TU = T4 + T5;
239 Tt = R1[WS(rs, 8)];
240 Tu = R1[WS(rs, 3)];
241 Tv = Tt - Tu;
242 T12 = Tt + Tu;
243 }
244 {
245 E Tw, Tx, T7, T8;
246 Tw = R1[WS(rs, 6)];
247 Tx = R1[WS(rs, 1)];
248 Ty = Tw - Tx;
249 TZ = Tw + Tx;
250 T7 = R0[WS(rs, 8)];
251 T8 = R0[WS(rs, 3)];
252 T9 = T7 - T8;
253 TR = T7 + T8;
254 }
255 {
256 E Tb, Tc, Tm, Tn;
257 Tb = R0[WS(rs, 4)];
258 Tc = R0[WS(rs, 9)];
259 Td = Tb - Tc;
260 TY = Tb + Tc;
261 Tm = R1[0];
262 Tn = R1[WS(rs, 5)];
263 To = Tm - Tn;
264 TS = Tm + Tn;
265 }
266 {
267 E Tp, Tq, Te, Tf;
268 Tp = R1[WS(rs, 4)];
269 Tq = R1[WS(rs, 9)];
270 Tr = Tp - Tq;
271 TV = Tp + Tq;
272 Te = R0[WS(rs, 6)];
273 Tf = R0[WS(rs, 1)];
274 Tg = Te - Tf;
275 T11 = Te + Tf;
276 }
277 Ts = To - Tr;
278 TM = T6 - T9;
279 TN = Td - Tg;
280 Tz = Tv - Ty;
281 Ta = T6 + T9;
282 Th = Td + Tg;
283 Ti = Ta + Th;
284 T1g = TY + TZ;
285 T1h = T11 + T12;
286 T1k = T1g + T1h;
287 T10 = TY - TZ;
288 T13 = T11 - T12;
289 T19 = T10 + T13;
290 TG = Tr + To;
291 TH = Ty + Tv;
292 TI = TG + TH;
293 T1d = TU + TV;
294 T1e = TR + TS;
295 T1j = T1d + T1e;
296 TT = TR - TS;
297 TW = TU - TV;
298 T18 = TW + TT;
299 }
300 Cr[WS(csr, 5)] = T3 + Ti;
301 Ci[WS(csi, 5)] = TF - TI;
302 {
303 E TX, T14, T1f, T1i;
304 TX = TT - TW;
305 T14 = T10 - T13;
306 Ci[WS(csi, 6)] = FNMS(KP587785252, T14, KP951056516 * TX);
307 Ci[WS(csi, 2)] = FMA(KP587785252, TX, KP951056516 * T14);
308 T1f = T1d - T1e;
309 T1i = T1g - T1h;
310 Ci[WS(csi, 8)] = FNMS(KP951056516, T1i, KP587785252 * T1f);
311 Ci[WS(csi, 4)] = FMA(KP951056516, T1f, KP587785252 * T1i);
312 }
313 {
314 E T1l, T1n, T1o, T1c, T1a, T1b;
315 T1l = KP559016994 * (T1j - T1k);
316 T1n = T1j + T1k;
317 T1o = FNMS(KP250000000, T1n, T1m);
318 Cr[WS(csr, 4)] = T1l + T1o;
319 Cr[0] = T1m + T1n;
320 Cr[WS(csr, 8)] = T1o - T1l;
321 T1c = KP559016994 * (T18 - T19);
322 T1a = T18 + T19;
323 T1b = FNMS(KP250000000, T1a, T17);
324 Cr[WS(csr, 2)] = T1b - T1c;
325 Cr[WS(csr, 10)] = T17 + T1a;
326 Cr[WS(csr, 6)] = T1c + T1b;
327 }
328 {
329 E TA, TC, Tl, TB, Tj, Tk;
330 TA = FMA(KP951056516, Ts, KP587785252 * Tz);
331 TC = FNMS(KP587785252, Ts, KP951056516 * Tz);
332 Tj = KP559016994 * (Ta - Th);
333 Tk = FNMS(KP250000000, Ti, T3);
334 Tl = Tj + Tk;
335 TB = Tk - Tj;
336 Cr[WS(csr, 9)] = Tl - TA;
337 Cr[WS(csr, 7)] = TB + TC;
338 Cr[WS(csr, 1)] = Tl + TA;
339 Cr[WS(csr, 3)] = TB - TC;
340 }
341 {
342 E TO, TQ, TL, TP, TJ, TK;
343 TO = FMA(KP951056516, TM, KP587785252 * TN);
344 TQ = FNMS(KP587785252, TM, KP951056516 * TN);
345 TJ = FMA(KP250000000, TI, TF);
346 TK = KP559016994 * (TH - TG);
347 TL = TJ + TK;
348 TP = TK - TJ;
349 Ci[WS(csi, 1)] = TL - TO;
350 Ci[WS(csi, 7)] = TQ + TP;
351 Ci[WS(csi, 9)] = TO + TL;
352 Ci[WS(csi, 3)] = TP - TQ;
353 }
354 }
355 }
356 }
357
358 static const kr2c_desc desc = { 20, "r2cf_20", {74, 12, 12, 0}, &GENUS };
359
360 void X(codelet_r2cf_20) (planner *p) {
361 X(kr2c_register) (p, r2cf_20, &desc);
362 }
363
364 #endif