Mercurial > hg > sv-dependency-builds

comparison src/fftw-3.3.8/dft/simd/common/n1fv_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 08:04:53 EDT 2018 */
23
24 #include "dft/codelet-dft.h"
25
26 #if defined(ARCH_PREFERS_FMA) || defined(ISA_EXTENSION_PREFERS_FMA)
27
28 /* Generated by: ../../../genfft/gen_notw_c.native -fma -simd -compact -variables 4 -pipeline-latency 8 -n 20 -name n1fv_20 -include dft/simd/n1f.h */
29
30 /*
31 * This function contains 104 FP additions, 50 FP multiplications,
32 * (or, 58 additions, 4 multiplications, 46 fused multiply/add),
33 * 53 stack variables, 4 constants, and 40 memory accesses
34 */
35 #include "dft/simd/n1f.h"
36
37 static void n1fv_20(const R *ri, const R *ii, R *ro, R *io, stride is, stride os, INT v, INT ivs, INT ovs)
38 {
39 DVK(KP559016994, +0.559016994374947424102293417182819058860154590);
40 DVK(KP618033988, +0.618033988749894848204586834365638117720309180);
41 DVK(KP951056516, +0.951056516295153572116439333379382143405698634);
42 DVK(KP250000000, +0.250000000000000000000000000000000000000000000);
43 {
44 INT i;
45 const R *xi;
46 R *xo;
47 xi = ri;
48 xo = ro;
49 for (i = v; i > 0; i = i - VL, xi = xi + (VL * ivs), xo = xo + (VL * ovs), MAKE_VOLATILE_STRIDE(40, is), MAKE_VOLATILE_STRIDE(40, os)) {
50 V T3, T1r, Tm, T13, TG, TN, TO, TH, T16, T19, T1a, T1v, T1w, T1x, T1s;
51 V T1t, T1u, T1d, T1g, T1h, Ti, TE, TB, TL, Tj, TC;
52 {
53 V T1, T2, T11, Tk, Tl, T12;
54 T1 = LD(&(xi[0]), ivs, &(xi[0]));
55 T2 = LD(&(xi[WS(is, 10)]), ivs, &(xi[0]));
56 T11 = VADD(T1, T2);
57 Tk = LD(&(xi[WS(is, 5)]), ivs, &(xi[WS(is, 1)]));
58 Tl = LD(&(xi[WS(is, 15)]), ivs, &(xi[WS(is, 1)]));
59 T12 = VADD(Tk, Tl);
60 T3 = VSUB(T1, T2);
61 T1r = VADD(T11, T12);
62 Tm = VSUB(Tk, Tl);
63 T13 = VSUB(T11, T12);
64 }
65 {
66 V T6, T14, Tw, T1c, Tz, T1f, T9, T17, Td, T1b, Tp, T15, Ts, T18, Tg;
67 V T1e;
68 {
69 V T4, T5, Tu, Tv;
70 T4 = LD(&(xi[WS(is, 4)]), ivs, &(xi[0]));
71 T5 = LD(&(xi[WS(is, 14)]), ivs, &(xi[0]));
72 T6 = VSUB(T4, T5);
73 T14 = VADD(T4, T5);
74 Tu = LD(&(xi[WS(is, 13)]), ivs, &(xi[WS(is, 1)]));
75 Tv = LD(&(xi[WS(is, 3)]), ivs, &(xi[WS(is, 1)]));
76 Tw = VSUB(Tu, Tv);
77 T1c = VADD(Tu, Tv);
78 }
79 {
80 V Tx, Ty, T7, T8;
81 Tx = LD(&(xi[WS(is, 17)]), ivs, &(xi[WS(is, 1)]));
82 Ty = LD(&(xi[WS(is, 7)]), ivs, &(xi[WS(is, 1)]));
83 Tz = VSUB(Tx, Ty);
84 T1f = VADD(Tx, Ty);
85 T7 = LD(&(xi[WS(is, 16)]), ivs, &(xi[0]));
86 T8 = LD(&(xi[WS(is, 6)]), ivs, &(xi[0]));
87 T9 = VSUB(T7, T8);
88 T17 = VADD(T7, T8);
89 }
90 {
91 V Tb, Tc, Tn, To;
92 Tb = LD(&(xi[WS(is, 8)]), ivs, &(xi[0]));
93 Tc = LD(&(xi[WS(is, 18)]), ivs, &(xi[0]));
94 Td = VSUB(Tb, Tc);
95 T1b = VADD(Tb, Tc);
96 Tn = LD(&(xi[WS(is, 9)]), ivs, &(xi[WS(is, 1)]));
97 To = LD(&(xi[WS(is, 19)]), ivs, &(xi[WS(is, 1)]));
98 Tp = VSUB(Tn, To);
99 T15 = VADD(Tn, To);
100 }
101 {
102 V Tq, Tr, Te, Tf;
103 Tq = LD(&(xi[WS(is, 1)]), ivs, &(xi[WS(is, 1)]));
104 Tr = LD(&(xi[WS(is, 11)]), ivs, &(xi[WS(is, 1)]));
105 Ts = VSUB(Tq, Tr);
106 T18 = VADD(Tq, Tr);
107 Te = LD(&(xi[WS(is, 12)]), ivs, &(xi[0]));
108 Tf = LD(&(xi[WS(is, 2)]), ivs, &(xi[0]));
109 Tg = VSUB(Te, Tf);
110 T1e = VADD(Te, Tf);
111 }
112 TG = VSUB(Ts, Tp);
113 TN = VSUB(T6, T9);
114 TO = VSUB(Td, Tg);
115 TH = VSUB(Tz, Tw);
116 T16 = VSUB(T14, T15);
117 T19 = VSUB(T17, T18);
118 T1a = VADD(T16, T19);
119 T1v = VADD(T1b, T1c);
120 T1w = VADD(T1e, T1f);
121 T1x = VADD(T1v, T1w);
122 T1s = VADD(T14, T15);
123 T1t = VADD(T17, T18);
124 T1u = VADD(T1s, T1t);
125 T1d = VSUB(T1b, T1c);
126 T1g = VSUB(T1e, T1f);
127 T1h = VADD(T1d, T1g);
128 {
129 V Ta, Th, Tt, TA;
130 Ta = VADD(T6, T9);
131 Th = VADD(Td, Tg);
132 Ti = VADD(Ta, Th);
133 TE = VSUB(Ta, Th);
134 Tt = VADD(Tp, Ts);
135 TA = VADD(Tw, Tz);
136 TB = VADD(Tt, TA);
137 TL = VSUB(TA, Tt);
138 }
139 }
140 Tj = VADD(T3, Ti);
141 TC = VADD(Tm, TB);
142 ST(&(xo[WS(os, 5)]), VFNMSI(TC, Tj), ovs, &(xo[WS(os, 1)]));
143 ST(&(xo[WS(os, 15)]), VFMAI(TC, Tj), ovs, &(xo[WS(os, 1)]));
144 {
145 V T1A, T1y, T1z, T1E, T1G, T1C, T1D, T1F, T1B;
146 T1A = VSUB(T1u, T1x);
147 T1y = VADD(T1u, T1x);
148 T1z = VFNMS(LDK(KP250000000), T1y, T1r);
149 T1C = VSUB(T1s, T1t);
150 T1D = VSUB(T1v, T1w);
151 T1E = VMUL(LDK(KP951056516), VFMA(LDK(KP618033988), T1D, T1C));
152 T1G = VMUL(LDK(KP951056516), VFNMS(LDK(KP618033988), T1C, T1D));
153 ST(&(xo[0]), VADD(T1r, T1y), ovs, &(xo[0]));
154 T1F = VFNMS(LDK(KP559016994), T1A, T1z);
155 ST(&(xo[WS(os, 8)]), VFNMSI(T1G, T1F), ovs, &(xo[0]));
156 ST(&(xo[WS(os, 12)]), VFMAI(T1G, T1F), ovs, &(xo[0]));
157 T1B = VFMA(LDK(KP559016994), T1A, T1z);
158 ST(&(xo[WS(os, 4)]), VFMAI(T1E, T1B), ovs, &(xo[0]));
159 ST(&(xo[WS(os, 16)]), VFNMSI(T1E, T1B), ovs, &(xo[0]));
160 }
161 {
162 V T1k, T1i, T1j, T1o, T1q, T1m, T1n, T1p, T1l;
163 T1k = VSUB(T1a, T1h);
164 T1i = VADD(T1a, T1h);
165 T1j = VFNMS(LDK(KP250000000), T1i, T13);
166 T1m = VSUB(T1d, T1g);
167 T1n = VSUB(T16, T19);
168 T1o = VMUL(LDK(KP951056516), VFNMS(LDK(KP618033988), T1n, T1m));
169 T1q = VMUL(LDK(KP951056516), VFMA(LDK(KP618033988), T1m, T1n));
170 ST(&(xo[WS(os, 10)]), VADD(T13, T1i), ovs, &(xo[0]));
171 T1p = VFMA(LDK(KP559016994), T1k, T1j);
172 ST(&(xo[WS(os, 6)]), VFNMSI(T1q, T1p), ovs, &(xo[0]));
173 ST(&(xo[WS(os, 14)]), VFMAI(T1q, T1p), ovs, &(xo[0]));
174 T1l = VFNMS(LDK(KP559016994), T1k, T1j);
175 ST(&(xo[WS(os, 2)]), VFMAI(T1o, T1l), ovs, &(xo[0]));
176 ST(&(xo[WS(os, 18)]), VFNMSI(T1o, T1l), ovs, &(xo[0]));
177 }
178 {
179 V TI, TP, TX, TU, TM, TW, TF, TT, TK, TD;
180 TI = VFMA(LDK(KP618033988), TH, TG);
181 TP = VFMA(LDK(KP618033988), TO, TN);
182 TX = VFNMS(LDK(KP618033988), TN, TO);
183 TU = VFNMS(LDK(KP618033988), TG, TH);
184 TK = VFNMS(LDK(KP250000000), TB, Tm);
185 TM = VFNMS(LDK(KP559016994), TL, TK);
186 TW = VFMA(LDK(KP559016994), TL, TK);
187 TD = VFNMS(LDK(KP250000000), Ti, T3);
188 TF = VFMA(LDK(KP559016994), TE, TD);
189 TT = VFNMS(LDK(KP559016994), TE, TD);
190 {
191 V TJ, TQ, TZ, T10;
192 TJ = VFMA(LDK(KP951056516), TI, TF);
193 TQ = VFMA(LDK(KP951056516), TP, TM);
194 ST(&(xo[WS(os, 1)]), VFNMSI(TQ, TJ), ovs, &(xo[WS(os, 1)]));
195 ST(&(xo[WS(os, 19)]), VFMAI(TQ, TJ), ovs, &(xo[WS(os, 1)]));
196 TZ = VFMA(LDK(KP951056516), TU, TT);
197 T10 = VFMA(LDK(KP951056516), TX, TW);
198 ST(&(xo[WS(os, 13)]), VFNMSI(T10, TZ), ovs, &(xo[WS(os, 1)]));
199 ST(&(xo[WS(os, 7)]), VFMAI(T10, TZ), ovs, &(xo[WS(os, 1)]));
200 }
201 {
202 V TR, TS, TV, TY;
203 TR = VFNMS(LDK(KP951056516), TI, TF);
204 TS = VFNMS(LDK(KP951056516), TP, TM);
205 ST(&(xo[WS(os, 9)]), VFNMSI(TS, TR), ovs, &(xo[WS(os, 1)]));
206 ST(&(xo[WS(os, 11)]), VFMAI(TS, TR), ovs, &(xo[WS(os, 1)]));
207 TV = VFNMS(LDK(KP951056516), TU, TT);
208 TY = VFNMS(LDK(KP951056516), TX, TW);
209 ST(&(xo[WS(os, 17)]), VFNMSI(TY, TV), ovs, &(xo[WS(os, 1)]));
210 ST(&(xo[WS(os, 3)]), VFMAI(TY, TV), ovs, &(xo[WS(os, 1)]));
211 }
212 }
213 }
214 }
215 VLEAVE();
216 }
217
218 static const kdft_desc desc = { 20, XSIMD_STRING("n1fv_20"), {58, 4, 46, 0}, &GENUS, 0, 0, 0, 0 };
219
220 void XSIMD(codelet_n1fv_20) (planner *p) {
221 X(kdft_register) (p, n1fv_20, &desc);
222 }
223
224 #else
225
226 /* Generated by: ../../../genfft/gen_notw_c.native -simd -compact -variables 4 -pipeline-latency 8 -n 20 -name n1fv_20 -include dft/simd/n1f.h */
227
228 /*
229 * This function contains 104 FP additions, 24 FP multiplications,
230 * (or, 92 additions, 12 multiplications, 12 fused multiply/add),
231 * 53 stack variables, 4 constants, and 40 memory accesses
232 */
233 #include "dft/simd/n1f.h"
234
235 static void n1fv_20(const R *ri, const R *ii, R *ro, R *io, stride is, stride os, INT v, INT ivs, INT ovs)
236 {
237 DVK(KP587785252, +0.587785252292473129168705954639072768597652438);
238 DVK(KP951056516, +0.951056516295153572116439333379382143405698634);
239 DVK(KP250000000, +0.250000000000000000000000000000000000000000000);
240 DVK(KP559016994, +0.559016994374947424102293417182819058860154590);
241 {
242 INT i;
243 const R *xi;
244 R *xo;
245 xi = ri;
246 xo = ro;
247 for (i = v; i > 0; i = i - VL, xi = xi + (VL * ivs), xo = xo + (VL * ovs), MAKE_VOLATILE_STRIDE(40, is), MAKE_VOLATILE_STRIDE(40, os)) {
248 V T3, T1B, Tm, T1i, TG, TN, TO, TH, T13, T16, T1k, T1u, T1v, T1z, T1r;
249 V T1s, T1y, T1a, T1d, T1j, Ti, TD, TB, TL, Tj, TC;
250 {
251 V T1, T2, T1g, Tk, Tl, T1h;
252 T1 = LD(&(xi[0]), ivs, &(xi[0]));
253 T2 = LD(&(xi[WS(is, 10)]), ivs, &(xi[0]));
254 T1g = VADD(T1, T2);
255 Tk = LD(&(xi[WS(is, 5)]), ivs, &(xi[WS(is, 1)]));
256 Tl = LD(&(xi[WS(is, 15)]), ivs, &(xi[WS(is, 1)]));
257 T1h = VADD(Tk, Tl);
258 T3 = VSUB(T1, T2);
259 T1B = VADD(T1g, T1h);
260 Tm = VSUB(Tk, Tl);
261 T1i = VSUB(T1g, T1h);
262 }
263 {
264 V T6, T18, Tw, T12, Tz, T15, T9, T1b, Td, T11, Tp, T19, Ts, T1c, Tg;
265 V T14;
266 {
267 V T4, T5, Tu, Tv;
268 T4 = LD(&(xi[WS(is, 4)]), ivs, &(xi[0]));
269 T5 = LD(&(xi[WS(is, 14)]), ivs, &(xi[0]));
270 T6 = VSUB(T4, T5);
271 T18 = VADD(T4, T5);
272 Tu = LD(&(xi[WS(is, 13)]), ivs, &(xi[WS(is, 1)]));
273 Tv = LD(&(xi[WS(is, 3)]), ivs, &(xi[WS(is, 1)]));
274 Tw = VSUB(Tu, Tv);
275 T12 = VADD(Tu, Tv);
276 }
277 {
278 V Tx, Ty, T7, T8;
279 Tx = LD(&(xi[WS(is, 17)]), ivs, &(xi[WS(is, 1)]));
280 Ty = LD(&(xi[WS(is, 7)]), ivs, &(xi[WS(is, 1)]));
281 Tz = VSUB(Tx, Ty);
282 T15 = VADD(Tx, Ty);
283 T7 = LD(&(xi[WS(is, 16)]), ivs, &(xi[0]));
284 T8 = LD(&(xi[WS(is, 6)]), ivs, &(xi[0]));
285 T9 = VSUB(T7, T8);
286 T1b = VADD(T7, T8);
287 }
288 {
289 V Tb, Tc, Tn, To;
290 Tb = LD(&(xi[WS(is, 8)]), ivs, &(xi[0]));
291 Tc = LD(&(xi[WS(is, 18)]), ivs, &(xi[0]));
292 Td = VSUB(Tb, Tc);
293 T11 = VADD(Tb, Tc);
294 Tn = LD(&(xi[WS(is, 9)]), ivs, &(xi[WS(is, 1)]));
295 To = LD(&(xi[WS(is, 19)]), ivs, &(xi[WS(is, 1)]));
296 Tp = VSUB(Tn, To);
297 T19 = VADD(Tn, To);
298 }
299 {
300 V Tq, Tr, Te, Tf;
301 Tq = LD(&(xi[WS(is, 1)]), ivs, &(xi[WS(is, 1)]));
302 Tr = LD(&(xi[WS(is, 11)]), ivs, &(xi[WS(is, 1)]));
303 Ts = VSUB(Tq, Tr);
304 T1c = VADD(Tq, Tr);
305 Te = LD(&(xi[WS(is, 12)]), ivs, &(xi[0]));
306 Tf = LD(&(xi[WS(is, 2)]), ivs, &(xi[0]));
307 Tg = VSUB(Te, Tf);
308 T14 = VADD(Te, Tf);
309 }
310 TG = VSUB(Ts, Tp);
311 TN = VSUB(T6, T9);
312 TO = VSUB(Td, Tg);
313 TH = VSUB(Tz, Tw);
314 T13 = VSUB(T11, T12);
315 T16 = VSUB(T14, T15);
316 T1k = VADD(T13, T16);
317 T1u = VADD(T11, T12);
318 T1v = VADD(T14, T15);
319 T1z = VADD(T1u, T1v);
320 T1r = VADD(T18, T19);
321 T1s = VADD(T1b, T1c);
322 T1y = VADD(T1r, T1s);
323 T1a = VSUB(T18, T19);
324 T1d = VSUB(T1b, T1c);
325 T1j = VADD(T1a, T1d);
326 {
327 V Ta, Th, Tt, TA;
328 Ta = VADD(T6, T9);
329 Th = VADD(Td, Tg);
330 Ti = VADD(Ta, Th);
331 TD = VMUL(LDK(KP559016994), VSUB(Ta, Th));
332 Tt = VADD(Tp, Ts);
333 TA = VADD(Tw, Tz);
334 TB = VADD(Tt, TA);
335 TL = VMUL(LDK(KP559016994), VSUB(TA, Tt));
336 }
337 }
338 Tj = VADD(T3, Ti);
339 TC = VBYI(VADD(Tm, TB));
340 ST(&(xo[WS(os, 5)]), VSUB(Tj, TC), ovs, &(xo[WS(os, 1)]));
341 ST(&(xo[WS(os, 15)]), VADD(Tj, TC), ovs, &(xo[WS(os, 1)]));
342 {
343 V T1A, T1C, T1D, T1x, T1G, T1t, T1w, T1F, T1E;
344 T1A = VMUL(LDK(KP559016994), VSUB(T1y, T1z));
345 T1C = VADD(T1y, T1z);
346 T1D = VFNMS(LDK(KP250000000), T1C, T1B);
347 T1t = VSUB(T1r, T1s);
348 T1w = VSUB(T1u, T1v);
349 T1x = VBYI(VFMA(LDK(KP951056516), T1t, VMUL(LDK(KP587785252), T1w)));
350 T1G = VBYI(VFNMS(LDK(KP587785252), T1t, VMUL(LDK(KP951056516), T1w)));
351 ST(&(xo[0]), VADD(T1B, T1C), ovs, &(xo[0]));
352 T1F = VSUB(T1D, T1A);
353 ST(&(xo[WS(os, 8)]), VSUB(T1F, T1G), ovs, &(xo[0]));
354 ST(&(xo[WS(os, 12)]), VADD(T1G, T1F), ovs, &(xo[0]));
355 T1E = VADD(T1A, T1D);
356 ST(&(xo[WS(os, 4)]), VADD(T1x, T1E), ovs, &(xo[0]));
357 ST(&(xo[WS(os, 16)]), VSUB(T1E, T1x), ovs, &(xo[0]));
358 }
359 {
360 V T1n, T1l, T1m, T1f, T1q, T17, T1e, T1p, T1o;
361 T1n = VMUL(LDK(KP559016994), VSUB(T1j, T1k));
362 T1l = VADD(T1j, T1k);
363 T1m = VFNMS(LDK(KP250000000), T1l, T1i);
364 T17 = VSUB(T13, T16);
365 T1e = VSUB(T1a, T1d);
366 T1f = VBYI(VFNMS(LDK(KP587785252), T1e, VMUL(LDK(KP951056516), T17)));
367 T1q = VBYI(VFMA(LDK(KP951056516), T1e, VMUL(LDK(KP587785252), T17)));
368 ST(&(xo[WS(os, 10)]), VADD(T1i, T1l), ovs, &(xo[0]));
369 T1p = VADD(T1n, T1m);
370 ST(&(xo[WS(os, 6)]), VSUB(T1p, T1q), ovs, &(xo[0]));
371 ST(&(xo[WS(os, 14)]), VADD(T1q, T1p), ovs, &(xo[0]));
372 T1o = VSUB(T1m, T1n);
373 ST(&(xo[WS(os, 2)]), VADD(T1f, T1o), ovs, &(xo[0]));
374 ST(&(xo[WS(os, 18)]), VSUB(T1o, T1f), ovs, &(xo[0]));
375 }
376 {
377 V TI, TP, TX, TU, TM, TW, TF, TT, TK, TE;
378 TI = VFMA(LDK(KP951056516), TG, VMUL(LDK(KP587785252), TH));
379 TP = VFMA(LDK(KP951056516), TN, VMUL(LDK(KP587785252), TO));
380 TX = VFNMS(LDK(KP587785252), TN, VMUL(LDK(KP951056516), TO));
381 TU = VFNMS(LDK(KP587785252), TG, VMUL(LDK(KP951056516), TH));
382 TK = VFMS(LDK(KP250000000), TB, Tm);
383 TM = VADD(TK, TL);
384 TW = VSUB(TL, TK);
385 TE = VFNMS(LDK(KP250000000), Ti, T3);
386 TF = VADD(TD, TE);
387 TT = VSUB(TE, TD);
388 {
389 V TJ, TQ, TZ, T10;
390 TJ = VADD(TF, TI);
391 TQ = VBYI(VSUB(TM, TP));
392 ST(&(xo[WS(os, 19)]), VSUB(TJ, TQ), ovs, &(xo[WS(os, 1)]));
393 ST(&(xo[WS(os, 1)]), VADD(TJ, TQ), ovs, &(xo[WS(os, 1)]));
394 TZ = VADD(TT, TU);
395 T10 = VBYI(VADD(TX, TW));
396 ST(&(xo[WS(os, 13)]), VSUB(TZ, T10), ovs, &(xo[WS(os, 1)]));
397 ST(&(xo[WS(os, 7)]), VADD(TZ, T10), ovs, &(xo[WS(os, 1)]));
398 }
399 {
400 V TR, TS, TV, TY;
401 TR = VSUB(TF, TI);
402 TS = VBYI(VADD(TP, TM));
403 ST(&(xo[WS(os, 11)]), VSUB(TR, TS), ovs, &(xo[WS(os, 1)]));
404 ST(&(xo[WS(os, 9)]), VADD(TR, TS), ovs, &(xo[WS(os, 1)]));
405 TV = VSUB(TT, TU);
406 TY = VBYI(VSUB(TW, TX));
407 ST(&(xo[WS(os, 17)]), VSUB(TV, TY), ovs, &(xo[WS(os, 1)]));
408 ST(&(xo[WS(os, 3)]), VADD(TV, TY), ovs, &(xo[WS(os, 1)]));
409 }
410 }
411 }
412 }
413 VLEAVE();
414 }
415
416 static const kdft_desc desc = { 20, XSIMD_STRING("n1fv_20"), {92, 12, 12, 0}, &GENUS, 0, 0, 0, 0 };
417
418 void XSIMD(codelet_n1fv_20) (planner *p) {
419 X(kdft_register) (p, n1fv_20, &desc);
420 }
421
422 #endif