Mercurial > hg > sv-dependency-builds

comparison src/fftw-3.3.3/dft/simd/common/n1bv_20.c @ 10:37bf6b4a2645

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
Add FFTW3
author Chris Cannam
date Wed, 20 Mar 2013 15:35:50 +0000
parents
children
comparison
equal deleted inserted replaced
9:c0fb53affa76 10:37bf6b4a2645
1 /*
2 * Copyright (c) 2003, 2007-11 Matteo Frigo
3 * Copyright (c) 2003, 2007-11 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 Sun Nov 25 07:37:14 EST 2012 */
23
24 #include "codelet-dft.h"
25
26 #ifdef HAVE_FMA
27
28 /* Generated by: ../../../genfft/gen_notw_c.native -fma -reorder-insns -schedule-for-pipeline -simd -compact -variables 4 -pipeline-latency 8 -sign 1 -n 20 -name n1bv_20 -include n1b.h */
29
30 /*
31 * This function contains 104 FP additions, 50 FP multiplications,
32 * (or, 58 additions, 4 multiplications, 46 fused multiply/add),
33 * 71 stack variables, 4 constants, and 40 memory accesses
34 */
35 #include "n1b.h"
36
37 static void n1bv_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 = ii;
48 xo = io;
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 TS, TA, TN, TV, TK, TU, TR, Tl;
51 {
52 V T3, TE, T1r, T13, Ta, TL, Tz, TG, Ts, TF, Th, TM, T1u, T1C, T1n;
53 V T1a, T1m, T1h, T1x, T1D, Tk, Ti;
54 {
55 V T1, T2, TC, TD;
56 T1 = LD(&(xi[0]), ivs, &(xi[0]));
57 T2 = LD(&(xi[WS(is, 10)]), ivs, &(xi[0]));
58 TC = LD(&(xi[WS(is, 5)]), ivs, &(xi[WS(is, 1)]));
59 TD = LD(&(xi[WS(is, 15)]), ivs, &(xi[WS(is, 1)]));
60 {
61 V T14, T6, T1c, Tv, Tm, T1f, Ty, T17, T9, Tn, Tp, T1b, Td, Tq, Te;
62 V Tf, T15, To;
63 {
64 V Tw, Tx, T7, T8, Tb, Tc;
65 {
66 V T4, T5, Tt, Tu, T11, T12;
67 T4 = LD(&(xi[WS(is, 4)]), ivs, &(xi[0]));
68 T5 = LD(&(xi[WS(is, 14)]), ivs, &(xi[0]));
69 Tt = LD(&(xi[WS(is, 13)]), ivs, &(xi[WS(is, 1)]));
70 Tu = LD(&(xi[WS(is, 3)]), ivs, &(xi[WS(is, 1)]));
71 Tw = LD(&(xi[WS(is, 17)]), ivs, &(xi[WS(is, 1)]));
72 T3 = VSUB(T1, T2);
73 T11 = VADD(T1, T2);
74 TE = VSUB(TC, TD);
75 T12 = VADD(TC, TD);
76 T14 = VADD(T4, T5);
77 T6 = VSUB(T4, T5);
78 T1c = VADD(Tt, Tu);
79 Tv = VSUB(Tt, Tu);
80 Tx = LD(&(xi[WS(is, 7)]), ivs, &(xi[WS(is, 1)]));
81 T7 = LD(&(xi[WS(is, 16)]), ivs, &(xi[0]));
82 T8 = LD(&(xi[WS(is, 6)]), ivs, &(xi[0]));
83 T1r = VADD(T11, T12);
84 T13 = VSUB(T11, T12);
85 }
86 Tb = LD(&(xi[WS(is, 8)]), ivs, &(xi[0]));
87 Tc = LD(&(xi[WS(is, 18)]), ivs, &(xi[0]));
88 Tm = LD(&(xi[WS(is, 9)]), ivs, &(xi[WS(is, 1)]));
89 T1f = VADD(Tw, Tx);
90 Ty = VSUB(Tw, Tx);
91 T17 = VADD(T7, T8);
92 T9 = VSUB(T7, T8);
93 Tn = LD(&(xi[WS(is, 19)]), ivs, &(xi[WS(is, 1)]));
94 Tp = LD(&(xi[WS(is, 1)]), ivs, &(xi[WS(is, 1)]));
95 T1b = VADD(Tb, Tc);
96 Td = VSUB(Tb, Tc);
97 Tq = LD(&(xi[WS(is, 11)]), ivs, &(xi[WS(is, 1)]));
98 Te = LD(&(xi[WS(is, 12)]), ivs, &(xi[0]));
99 Tf = LD(&(xi[WS(is, 2)]), ivs, &(xi[0]));
100 }
101 Ta = VADD(T6, T9);
102 TL = VSUB(T6, T9);
103 T15 = VADD(Tm, Tn);
104 To = VSUB(Tm, Tn);
105 Tz = VSUB(Tv, Ty);
106 TG = VADD(Tv, Ty);
107 {
108 V T1d, T1v, T18, Tr, T1e, Tg, T16, T1s;
109 T1d = VSUB(T1b, T1c);
110 T1v = VADD(T1b, T1c);
111 T18 = VADD(Tp, Tq);
112 Tr = VSUB(Tp, Tq);
113 T1e = VADD(Te, Tf);
114 Tg = VSUB(Te, Tf);
115 T16 = VSUB(T14, T15);
116 T1s = VADD(T14, T15);
117 {
118 V T1t, T19, T1w, T1g;
119 T1t = VADD(T17, T18);
120 T19 = VSUB(T17, T18);
121 Ts = VSUB(To, Tr);
122 TF = VADD(To, Tr);
123 T1w = VADD(T1e, T1f);
124 T1g = VSUB(T1e, T1f);
125 Th = VADD(Td, Tg);
126 TM = VSUB(Td, Tg);
127 T1u = VADD(T1s, T1t);
128 T1C = VSUB(T1s, T1t);
129 T1n = VSUB(T16, T19);
130 T1a = VADD(T16, T19);
131 T1m = VSUB(T1d, T1g);
132 T1h = VADD(T1d, T1g);
133 T1x = VADD(T1v, T1w);
134 T1D = VSUB(T1v, T1w);
135 }
136 }
137 }
138 }
139 Tk = VSUB(Ta, Th);
140 Ti = VADD(Ta, Th);
141 {
142 V TJ, T1k, T1A, TZ, Tj, T1E, T1G, TI, T10, T1j, T1z, T1i, T1y, TH;
143 TJ = VSUB(TF, TG);
144 TH = VADD(TF, TG);
145 T1i = VADD(T1a, T1h);
146 T1k = VSUB(T1a, T1h);
147 T1y = VADD(T1u, T1x);
148 T1A = VSUB(T1u, T1x);
149 TZ = VADD(T3, Ti);
150 Tj = VFNMS(LDK(KP250000000), Ti, T3);
151 T1E = VMUL(LDK(KP951056516), VFMA(LDK(KP618033988), T1D, T1C));
152 T1G = VMUL(LDK(KP951056516), VFNMS(LDK(KP618033988), T1C, T1D));
153 TI = VFNMS(LDK(KP250000000), TH, TE);
154 T10 = VADD(TE, TH);
155 T1j = VFNMS(LDK(KP250000000), T1i, T13);
156 ST(&(xo[0]), VADD(T1r, T1y), ovs, &(xo[0]));
157 T1z = VFNMS(LDK(KP250000000), T1y, T1r);
158 ST(&(xo[WS(os, 10)]), VADD(T13, T1i), ovs, &(xo[0]));
159 {
160 V T1p, T1l, T1o, T1q, T1F, T1B;
161 TS = VFNMS(LDK(KP618033988), Ts, Tz);
162 TA = VFMA(LDK(KP618033988), Tz, Ts);
163 TN = VFMA(LDK(KP618033988), TM, TL);
164 TV = VFNMS(LDK(KP618033988), TL, TM);
165 ST(&(xo[WS(os, 5)]), VFMAI(T10, TZ), ovs, &(xo[WS(os, 1)]));
166 ST(&(xo[WS(os, 15)]), VFNMSI(T10, TZ), ovs, &(xo[WS(os, 1)]));
167 T1p = VFMA(LDK(KP559016994), T1k, T1j);
168 T1l = VFNMS(LDK(KP559016994), T1k, T1j);
169 T1o = VMUL(LDK(KP951056516), VFNMS(LDK(KP618033988), T1n, T1m));
170 T1q = VMUL(LDK(KP951056516), VFMA(LDK(KP618033988), T1m, T1n));
171 T1F = VFNMS(LDK(KP559016994), T1A, T1z);
172 T1B = VFMA(LDK(KP559016994), T1A, T1z);
173 ST(&(xo[WS(os, 14)]), VFNMSI(T1q, T1p), ovs, &(xo[0]));
174 ST(&(xo[WS(os, 6)]), VFMAI(T1q, T1p), ovs, &(xo[0]));
175 ST(&(xo[WS(os, 18)]), VFMAI(T1o, T1l), ovs, &(xo[0]));
176 ST(&(xo[WS(os, 2)]), VFNMSI(T1o, T1l), ovs, &(xo[0]));
177 ST(&(xo[WS(os, 16)]), VFMAI(T1E, T1B), ovs, &(xo[0]));
178 ST(&(xo[WS(os, 4)]), VFNMSI(T1E, T1B), ovs, &(xo[0]));
179 ST(&(xo[WS(os, 12)]), VFNMSI(T1G, T1F), ovs, &(xo[0]));
180 ST(&(xo[WS(os, 8)]), VFMAI(T1G, T1F), ovs, &(xo[0]));
181 TK = VFMA(LDK(KP559016994), TJ, TI);
182 TU = VFNMS(LDK(KP559016994), TJ, TI);
183 TR = VFNMS(LDK(KP559016994), Tk, Tj);
184 Tl = VFMA(LDK(KP559016994), Tk, Tj);
185 }
186 }
187 }
188 {
189 V TY, TW, TO, TQ, TB, TP, TX, TT;
190 TY = VFMA(LDK(KP951056516), TV, TU);
191 TW = VFNMS(LDK(KP951056516), TV, TU);
192 TO = VFMA(LDK(KP951056516), TN, TK);
193 TQ = VFNMS(LDK(KP951056516), TN, TK);
194 TB = VFNMS(LDK(KP951056516), TA, Tl);
195 TP = VFMA(LDK(KP951056516), TA, Tl);
196 TX = VFNMS(LDK(KP951056516), TS, TR);
197 TT = VFMA(LDK(KP951056516), TS, TR);
198 ST(&(xo[WS(os, 9)]), VFMAI(TQ, TP), ovs, &(xo[WS(os, 1)]));
199 ST(&(xo[WS(os, 11)]), VFNMSI(TQ, TP), ovs, &(xo[WS(os, 1)]));
200 ST(&(xo[WS(os, 1)]), VFMAI(TO, TB), ovs, &(xo[WS(os, 1)]));
201 ST(&(xo[WS(os, 19)]), VFNMSI(TO, TB), ovs, &(xo[WS(os, 1)]));
202 ST(&(xo[WS(os, 17)]), VFMAI(TW, TT), ovs, &(xo[WS(os, 1)]));
203 ST(&(xo[WS(os, 3)]), VFNMSI(TW, TT), ovs, &(xo[WS(os, 1)]));
204 ST(&(xo[WS(os, 13)]), VFMAI(TY, TX), ovs, &(xo[WS(os, 1)]));
205 ST(&(xo[WS(os, 7)]), VFNMSI(TY, TX), ovs, &(xo[WS(os, 1)]));
206 }
207 }
208 }
209 VLEAVE();
210 }
211
212 static const kdft_desc desc = { 20, XSIMD_STRING("n1bv_20"), {58, 4, 46, 0}, &GENUS, 0, 0, 0, 0 };
213
214 void XSIMD(codelet_n1bv_20) (planner *p) {
215 X(kdft_register) (p, n1bv_20, &desc);
216 }
217
218 #else /* HAVE_FMA */
219
220 /* Generated by: ../../../genfft/gen_notw_c.native -simd -compact -variables 4 -pipeline-latency 8 -sign 1 -n 20 -name n1bv_20 -include n1b.h */
221
222 /*
223 * This function contains 104 FP additions, 24 FP multiplications,
224 * (or, 92 additions, 12 multiplications, 12 fused multiply/add),
225 * 53 stack variables, 4 constants, and 40 memory accesses
226 */
227 #include "n1b.h"
228
229 static void n1bv_20(const R *ri, const R *ii, R *ro, R *io, stride is, stride os, INT v, INT ivs, INT ovs)
230 {
231 DVK(KP587785252, +0.587785252292473129168705954639072768597652438);
232 DVK(KP951056516, +0.951056516295153572116439333379382143405698634);
233 DVK(KP250000000, +0.250000000000000000000000000000000000000000000);
234 DVK(KP559016994, +0.559016994374947424102293417182819058860154590);
235 {
236 INT i;
237 const R *xi;
238 R *xo;
239 xi = ii;
240 xo = io;
241 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)) {
242 V T3, T1y, TH, T1i, Ts, TL, TM, Tz, T13, T16, T1j, T1u, T1v, T1w, T1r;
243 V T1s, T1t, T1a, T1d, T1k, Ti, Tk, TE, TI, TZ, T10;
244 {
245 V T1, T2, T1g, TF, TG, T1h;
246 T1 = LD(&(xi[0]), ivs, &(xi[0]));
247 T2 = LD(&(xi[WS(is, 10)]), ivs, &(xi[0]));
248 T1g = VADD(T1, T2);
249 TF = LD(&(xi[WS(is, 5)]), ivs, &(xi[WS(is, 1)]));
250 TG = LD(&(xi[WS(is, 15)]), ivs, &(xi[WS(is, 1)]));
251 T1h = VADD(TF, TG);
252 T3 = VSUB(T1, T2);
253 T1y = VADD(T1g, T1h);
254 TH = VSUB(TF, TG);
255 T1i = VSUB(T1g, T1h);
256 }
257 {
258 V T6, T11, Tv, T19, Ty, T1c, T9, T14, Td, T18, To, T12, Tr, T15, Tg;
259 V T1b;
260 {
261 V T4, T5, Tt, Tu;
262 T4 = LD(&(xi[WS(is, 4)]), ivs, &(xi[0]));
263 T5 = LD(&(xi[WS(is, 14)]), ivs, &(xi[0]));
264 T6 = VSUB(T4, T5);
265 T11 = VADD(T4, T5);
266 Tt = LD(&(xi[WS(is, 13)]), ivs, &(xi[WS(is, 1)]));
267 Tu = LD(&(xi[WS(is, 3)]), ivs, &(xi[WS(is, 1)]));
268 Tv = VSUB(Tt, Tu);
269 T19 = VADD(Tt, Tu);
270 }
271 {
272 V Tw, Tx, T7, T8;
273 Tw = LD(&(xi[WS(is, 17)]), ivs, &(xi[WS(is, 1)]));
274 Tx = LD(&(xi[WS(is, 7)]), ivs, &(xi[WS(is, 1)]));
275 Ty = VSUB(Tw, Tx);
276 T1c = VADD(Tw, Tx);
277 T7 = LD(&(xi[WS(is, 16)]), ivs, &(xi[0]));
278 T8 = LD(&(xi[WS(is, 6)]), ivs, &(xi[0]));
279 T9 = VSUB(T7, T8);
280 T14 = VADD(T7, T8);
281 }
282 {
283 V Tb, Tc, Tm, Tn;
284 Tb = LD(&(xi[WS(is, 8)]), ivs, &(xi[0]));
285 Tc = LD(&(xi[WS(is, 18)]), ivs, &(xi[0]));
286 Td = VSUB(Tb, Tc);
287 T18 = VADD(Tb, Tc);
288 Tm = LD(&(xi[WS(is, 9)]), ivs, &(xi[WS(is, 1)]));
289 Tn = LD(&(xi[WS(is, 19)]), ivs, &(xi[WS(is, 1)]));
290 To = VSUB(Tm, Tn);
291 T12 = VADD(Tm, Tn);
292 }
293 {
294 V Tp, Tq, Te, Tf;
295 Tp = LD(&(xi[WS(is, 1)]), ivs, &(xi[WS(is, 1)]));
296 Tq = LD(&(xi[WS(is, 11)]), ivs, &(xi[WS(is, 1)]));
297 Tr = VSUB(Tp, Tq);
298 T15 = VADD(Tp, Tq);
299 Te = LD(&(xi[WS(is, 12)]), ivs, &(xi[0]));
300 Tf = LD(&(xi[WS(is, 2)]), ivs, &(xi[0]));
301 Tg = VSUB(Te, Tf);
302 T1b = VADD(Te, Tf);
303 }
304 Ts = VSUB(To, Tr);
305 TL = VSUB(T6, T9);
306 TM = VSUB(Td, Tg);
307 Tz = VSUB(Tv, Ty);
308 T13 = VSUB(T11, T12);
309 T16 = VSUB(T14, T15);
310 T1j = VADD(T13, T16);
311 T1u = VADD(T18, T19);
312 T1v = VADD(T1b, T1c);
313 T1w = VADD(T1u, T1v);
314 T1r = VADD(T11, T12);
315 T1s = VADD(T14, T15);
316 T1t = VADD(T1r, T1s);
317 T1a = VSUB(T18, T19);
318 T1d = VSUB(T1b, T1c);
319 T1k = VADD(T1a, T1d);
320 {
321 V Ta, Th, TC, TD;
322 Ta = VADD(T6, T9);
323 Th = VADD(Td, Tg);
324 Ti = VADD(Ta, Th);
325 Tk = VMUL(LDK(KP559016994), VSUB(Ta, Th));
326 TC = VADD(To, Tr);
327 TD = VADD(Tv, Ty);
328 TE = VMUL(LDK(KP559016994), VSUB(TC, TD));
329 TI = VADD(TC, TD);
330 }
331 }
332 TZ = VADD(T3, Ti);
333 T10 = VBYI(VADD(TH, TI));
334 ST(&(xo[WS(os, 15)]), VSUB(TZ, T10), ovs, &(xo[WS(os, 1)]));
335 ST(&(xo[WS(os, 5)]), VADD(TZ, T10), ovs, &(xo[WS(os, 1)]));
336 {
337 V T1x, T1z, T1A, T1E, T1G, T1C, T1D, T1F, T1B;
338 T1x = VMUL(LDK(KP559016994), VSUB(T1t, T1w));
339 T1z = VADD(T1t, T1w);
340 T1A = VFNMS(LDK(KP250000000), T1z, T1y);
341 T1C = VSUB(T1r, T1s);
342 T1D = VSUB(T1u, T1v);
343 T1E = VBYI(VFMA(LDK(KP951056516), T1C, VMUL(LDK(KP587785252), T1D)));
344 T1G = VBYI(VFNMS(LDK(KP951056516), T1D, VMUL(LDK(KP587785252), T1C)));
345 ST(&(xo[0]), VADD(T1y, T1z), ovs, &(xo[0]));
346 T1F = VSUB(T1A, T1x);
347 ST(&(xo[WS(os, 8)]), VSUB(T1F, T1G), ovs, &(xo[0]));
348 ST(&(xo[WS(os, 12)]), VADD(T1G, T1F), ovs, &(xo[0]));
349 T1B = VADD(T1x, T1A);
350 ST(&(xo[WS(os, 4)]), VSUB(T1B, T1E), ovs, &(xo[0]));
351 ST(&(xo[WS(os, 16)]), VADD(T1E, T1B), ovs, &(xo[0]));
352 }
353 {
354 V T1n, T1l, T1m, T1f, T1p, T17, T1e, T1q, T1o;
355 T1n = VMUL(LDK(KP559016994), VSUB(T1j, T1k));
356 T1l = VADD(T1j, T1k);
357 T1m = VFNMS(LDK(KP250000000), T1l, T1i);
358 T17 = VSUB(T13, T16);
359 T1e = VSUB(T1a, T1d);
360 T1f = VBYI(VFNMS(LDK(KP951056516), T1e, VMUL(LDK(KP587785252), T17)));
361 T1p = VBYI(VFMA(LDK(KP951056516), T17, VMUL(LDK(KP587785252), T1e)));
362 ST(&(xo[WS(os, 10)]), VADD(T1i, T1l), ovs, &(xo[0]));
363 T1q = VADD(T1n, T1m);
364 ST(&(xo[WS(os, 6)]), VADD(T1p, T1q), ovs, &(xo[0]));
365 ST(&(xo[WS(os, 14)]), VSUB(T1q, T1p), ovs, &(xo[0]));
366 T1o = VSUB(T1m, T1n);
367 ST(&(xo[WS(os, 2)]), VADD(T1f, T1o), ovs, &(xo[0]));
368 ST(&(xo[WS(os, 18)]), VSUB(T1o, T1f), ovs, &(xo[0]));
369 }
370 {
371 V TA, TN, TU, TS, TK, TV, Tl, TR, TJ, Tj;
372 TA = VFNMS(LDK(KP951056516), Tz, VMUL(LDK(KP587785252), Ts));
373 TN = VFNMS(LDK(KP951056516), TM, VMUL(LDK(KP587785252), TL));
374 TU = VFMA(LDK(KP951056516), TL, VMUL(LDK(KP587785252), TM));
375 TS = VFMA(LDK(KP951056516), Ts, VMUL(LDK(KP587785252), Tz));
376 TJ = VFNMS(LDK(KP250000000), TI, TH);
377 TK = VSUB(TE, TJ);
378 TV = VADD(TE, TJ);
379 Tj = VFNMS(LDK(KP250000000), Ti, T3);
380 Tl = VSUB(Tj, Tk);
381 TR = VADD(Tk, Tj);
382 {
383 V TB, TO, TX, TY;
384 TB = VSUB(Tl, TA);
385 TO = VBYI(VSUB(TK, TN));
386 ST(&(xo[WS(os, 17)]), VSUB(TB, TO), ovs, &(xo[WS(os, 1)]));
387 ST(&(xo[WS(os, 3)]), VADD(TB, TO), ovs, &(xo[WS(os, 1)]));
388 TX = VADD(TR, TS);
389 TY = VBYI(VSUB(TV, TU));
390 ST(&(xo[WS(os, 11)]), VSUB(TX, TY), ovs, &(xo[WS(os, 1)]));
391 ST(&(xo[WS(os, 9)]), VADD(TX, TY), ovs, &(xo[WS(os, 1)]));
392 }
393 {
394 V TP, TQ, TT, TW;
395 TP = VADD(Tl, TA);
396 TQ = VBYI(VADD(TN, TK));
397 ST(&(xo[WS(os, 13)]), VSUB(TP, TQ), ovs, &(xo[WS(os, 1)]));
398 ST(&(xo[WS(os, 7)]), VADD(TP, TQ), ovs, &(xo[WS(os, 1)]));
399 TT = VSUB(TR, TS);
400 TW = VBYI(VADD(TU, TV));
401 ST(&(xo[WS(os, 19)]), VSUB(TT, TW), ovs, &(xo[WS(os, 1)]));
402 ST(&(xo[WS(os, 1)]), VADD(TT, TW), ovs, &(xo[WS(os, 1)]));
403 }
404 }
405 }
406 }
407 VLEAVE();
408 }
409
410 static const kdft_desc desc = { 20, XSIMD_STRING("n1bv_20"), {92, 12, 12, 0}, &GENUS, 0, 0, 0, 0 };
411
412 void XSIMD(codelet_n1bv_20) (planner *p) {
413 X(kdft_register) (p, n1bv_20, &desc);
414 }
415
416 #endif /* HAVE_FMA */