Mercurial > hg > sv-dependency-builds

annotate src/fftw-3.3.8/dft/simd/common/n1bv_15.c @ 82:d0c2a83c1364

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
date Tue, 19 Nov 2019 14:52:55 +0000
parents
children
rev   line source
Chris@82 1 /*
Chris@82 2 * Copyright (c) 2003, 2007-14 Matteo Frigo
Chris@82 3 * Copyright (c) 2003, 2007-14 Massachusetts Institute of Technology
Chris@82 4 *
Chris@82 5 * This program is free software; you can redistribute it and/or modify
Chris@82 6 * it under the terms of the GNU General Public License as published by
Chris@82 7 * the Free Software Foundation; either version 2 of the License, or
Chris@82 8 * (at your option) any later version.
Chris@82 9 *
Chris@82 10 * This program is distributed in the hope that it will be useful,
Chris@82 11 * but WITHOUT ANY WARRANTY; without even the implied warranty of
Chris@82 12 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
Chris@82 13 * GNU General Public License for more details.
Chris@82 14 *
Chris@82 15 * You should have received a copy of the GNU General Public License
Chris@82 16 * along with this program; if not, write to the Free Software
Chris@82 17 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
Chris@82 18 *
Chris@82 19 */
Chris@82 20
Chris@82 21 /* This file was automatically generated --- DO NOT EDIT */
Chris@82 22 /* Generated on Thu May 24 08:04:58 EDT 2018 */
Chris@82 23
Chris@82 24 #include "dft/codelet-dft.h"
Chris@82 25
Chris@82 26 #if defined(ARCH_PREFERS_FMA) || defined(ISA_EXTENSION_PREFERS_FMA)
Chris@82 27
Chris@82 28 /* Generated by: ../../../genfft/gen_notw_c.native -fma -simd -compact -variables 4 -pipeline-latency 8 -sign 1 -n 15 -name n1bv_15 -include dft/simd/n1b.h */
Chris@82 29
Chris@82 30 /*
Chris@82 31 * This function contains 78 FP additions, 49 FP multiplications,
Chris@82 32 * (or, 36 additions, 7 multiplications, 42 fused multiply/add),
Chris@82 33 * 53 stack variables, 8 constants, and 30 memory accesses
Chris@82 34 */
Chris@82 35 #include "dft/simd/n1b.h"
Chris@82 36
Chris@82 37 static void n1bv_15(const R *ri, const R *ii, R *ro, R *io, stride is, stride os, INT v, INT ivs, INT ovs)
Chris@82 38 {
Chris@82 39 DVK(KP910592997, +0.910592997310029334643087372129977886038870291);
Chris@82 40 DVK(KP823639103, +0.823639103546331925877420039278190003029660514);
Chris@82 41 DVK(KP559016994, +0.559016994374947424102293417182819058860154590);
Chris@82 42 DVK(KP618033988, +0.618033988749894848204586834365638117720309180);
Chris@82 43 DVK(KP951056516, +0.951056516295153572116439333379382143405698634);
Chris@82 44 DVK(KP250000000, +0.250000000000000000000000000000000000000000000);
Chris@82 45 DVK(KP866025403, +0.866025403784438646763723170752936183471402627);
Chris@82 46 DVK(KP500000000, +0.500000000000000000000000000000000000000000000);
Chris@82 47 {
Chris@82 48 INT i;
Chris@82 49 const R *xi;
Chris@82 50 R *xo;
Chris@82 51 xi = ii;
Chris@82 52 xo = io;
Chris@82 53 for (i = v; i > 0; i = i - VL, xi = xi + (VL * ivs), xo = xo + (VL * ovs), MAKE_VOLATILE_STRIDE(30, is), MAKE_VOLATILE_STRIDE(30, os)) {
Chris@82 54 V T5, T11, TH, Ty, TE, TF, TB, Tg, Tr, Ts, T12, T13, T14, T15, T16;
Chris@82 55 V T17, TK, TM, TZ, T10;
Chris@82 56 {
Chris@82 57 V T1, T2, T3, T4;
Chris@82 58 T1 = LD(&(xi[0]), ivs, &(xi[0]));
Chris@82 59 T2 = LD(&(xi[WS(is, 5)]), ivs, &(xi[WS(is, 1)]));
Chris@82 60 T3 = LD(&(xi[WS(is, 10)]), ivs, &(xi[0]));
Chris@82 61 T4 = VADD(T2, T3);
Chris@82 62 T5 = VFNMS(LDK(KP500000000), T4, T1);
Chris@82 63 T11 = VADD(T1, T4);
Chris@82 64 TH = VSUB(T2, T3);
Chris@82 65 }
Chris@82 66 {
Chris@82 67 V T6, T9, Ta, Tw, Tm, Tp, Tq, TA, Tb, Te, Tf, Tx, Th, Tk, Tl;
Chris@82 68 V Tz, TI, TJ;
Chris@82 69 {
Chris@82 70 V T7, T8, Tn, To;
Chris@82 71 T6 = LD(&(xi[WS(is, 3)]), ivs, &(xi[WS(is, 1)]));
Chris@82 72 T7 = LD(&(xi[WS(is, 8)]), ivs, &(xi[0]));
Chris@82 73 T8 = LD(&(xi[WS(is, 13)]), ivs, &(xi[WS(is, 1)]));
Chris@82 74 T9 = VADD(T7, T8);
Chris@82 75 Ta = VFNMS(LDK(KP500000000), T9, T6);
Chris@82 76 Tw = VSUB(T7, T8);
Chris@82 77 Tm = LD(&(xi[WS(is, 9)]), ivs, &(xi[WS(is, 1)]));
Chris@82 78 Tn = LD(&(xi[WS(is, 14)]), ivs, &(xi[0]));
Chris@82 79 To = LD(&(xi[WS(is, 4)]), ivs, &(xi[0]));
Chris@82 80 Tp = VADD(Tn, To);
Chris@82 81 Tq = VFNMS(LDK(KP500000000), Tp, Tm);
Chris@82 82 TA = VSUB(Tn, To);
Chris@82 83 }
Chris@82 84 {
Chris@82 85 V Tc, Td, Ti, Tj;
Chris@82 86 Tb = LD(&(xi[WS(is, 12)]), ivs, &(xi[0]));
Chris@82 87 Tc = LD(&(xi[WS(is, 2)]), ivs, &(xi[0]));
Chris@82 88 Td = LD(&(xi[WS(is, 7)]), ivs, &(xi[WS(is, 1)]));
Chris@82 89 Te = VADD(Tc, Td);
Chris@82 90 Tf = VFNMS(LDK(KP500000000), Te, Tb);
Chris@82 91 Tx = VSUB(Tc, Td);
Chris@82 92 Th = LD(&(xi[WS(is, 6)]), ivs, &(xi[0]));
Chris@82 93 Ti = LD(&(xi[WS(is, 11)]), ivs, &(xi[WS(is, 1)]));
Chris@82 94 Tj = LD(&(xi[WS(is, 1)]), ivs, &(xi[WS(is, 1)]));
Chris@82 95 Tk = VADD(Ti, Tj);
Chris@82 96 Tl = VFNMS(LDK(KP500000000), Tk, Th);
Chris@82 97 Tz = VSUB(Ti, Tj);
Chris@82 98 }
Chris@82 99 Ty = VSUB(Tw, Tx);
Chris@82 100 TE = VSUB(Ta, Tf);
Chris@82 101 TF = VSUB(Tl, Tq);
Chris@82 102 TB = VSUB(Tz, TA);
Chris@82 103 Tg = VADD(Ta, Tf);
Chris@82 104 Tr = VADD(Tl, Tq);
Chris@82 105 Ts = VADD(Tg, Tr);
Chris@82 106 T12 = VADD(T6, T9);
Chris@82 107 T13 = VADD(Tb, Te);
Chris@82 108 T14 = VADD(T12, T13);
Chris@82 109 T15 = VADD(Th, Tk);
Chris@82 110 T16 = VADD(Tm, Tp);
Chris@82 111 T17 = VADD(T15, T16);
Chris@82 112 TI = VADD(Tw, Tx);
Chris@82 113 TJ = VADD(Tz, TA);
Chris@82 114 TK = VADD(TI, TJ);
Chris@82 115 TM = VSUB(TI, TJ);
Chris@82 116 }
Chris@82 117 TZ = VADD(T5, Ts);
Chris@82 118 T10 = VMUL(LDK(KP866025403), VADD(TH, TK));
Chris@82 119 ST(&(xo[WS(os, 5)]), VFNMSI(T10, TZ), ovs, &(xo[WS(os, 1)]));
Chris@82 120 ST(&(xo[WS(os, 10)]), VFMAI(T10, TZ), ovs, &(xo[0]));
Chris@82 121 {
Chris@82 122 V T1a, T18, T19, T1e, T1g, T1c, T1d, T1f, T1b;
Chris@82 123 T1a = VSUB(T14, T17);
Chris@82 124 T18 = VADD(T14, T17);
Chris@82 125 T19 = VFNMS(LDK(KP250000000), T18, T11);
Chris@82 126 T1c = VSUB(T15, T16);
Chris@82 127 T1d = VSUB(T12, T13);
Chris@82 128 T1e = VMUL(LDK(KP951056516), VFNMS(LDK(KP618033988), T1d, T1c));
Chris@82 129 T1g = VMUL(LDK(KP951056516), VFMA(LDK(KP618033988), T1c, T1d));
Chris@82 130 ST(&(xo[0]), VADD(T11, T18), ovs, &(xo[0]));
Chris@82 131 T1f = VFMA(LDK(KP559016994), T1a, T19);
Chris@82 132 ST(&(xo[WS(os, 6)]), VFMAI(T1g, T1f), ovs, &(xo[0]));
Chris@82 133 ST(&(xo[WS(os, 9)]), VFNMSI(T1g, T1f), ovs, &(xo[WS(os, 1)]));
Chris@82 134 T1b = VFNMS(LDK(KP559016994), T1a, T19);
Chris@82 135 ST(&(xo[WS(os, 3)]), VFMAI(T1e, T1b), ovs, &(xo[WS(os, 1)]));
Chris@82 136 ST(&(xo[WS(os, 12)]), VFNMSI(T1e, T1b), ovs, &(xo[0]));
Chris@82 137 }
Chris@82 138 {
Chris@82 139 V TC, TG, TU, TS, TN, TV, Tv, TR, TL, Tt, Tu;
Chris@82 140 TC = VFMA(LDK(KP618033988), TB, Ty);
Chris@82 141 TG = VFMA(LDK(KP618033988), TF, TE);
Chris@82 142 TU = VFNMS(LDK(KP618033988), TE, TF);
Chris@82 143 TS = VFNMS(LDK(KP618033988), Ty, TB);
Chris@82 144 TL = VFNMS(LDK(KP250000000), TK, TH);
Chris@82 145 TN = VFMA(LDK(KP559016994), TM, TL);
Chris@82 146 TV = VFNMS(LDK(KP559016994), TM, TL);
Chris@82 147 Tt = VFNMS(LDK(KP250000000), Ts, T5);
Chris@82 148 Tu = VSUB(Tg, Tr);
Chris@82 149 Tv = VFMA(LDK(KP559016994), Tu, Tt);
Chris@82 150 TR = VFNMS(LDK(KP559016994), Tu, Tt);
Chris@82 151 {
Chris@82 152 V TD, TO, TX, TY;
Chris@82 153 TD = VFNMS(LDK(KP823639103), TC, Tv);
Chris@82 154 TO = VMUL(LDK(KP951056516), VFMA(LDK(KP910592997), TN, TG));
Chris@82 155 ST(&(xo[WS(os, 1)]), VFMAI(TO, TD), ovs, &(xo[WS(os, 1)]));
Chris@82 156 ST(&(xo[WS(os, 14)]), VFNMSI(TO, TD), ovs, &(xo[0]));
Chris@82 157 TX = VFMA(LDK(KP823639103), TS, TR);
Chris@82 158 TY = VMUL(LDK(KP951056516), VFNMS(LDK(KP910592997), TV, TU));
Chris@82 159 ST(&(xo[WS(os, 7)]), VFNMSI(TY, TX), ovs, &(xo[WS(os, 1)]));
Chris@82 160 ST(&(xo[WS(os, 8)]), VFMAI(TY, TX), ovs, &(xo[0]));
Chris@82 161 }
Chris@82 162 {
Chris@82 163 V TP, TQ, TT, TW;
Chris@82 164 TP = VFMA(LDK(KP823639103), TC, Tv);
Chris@82 165 TQ = VMUL(LDK(KP951056516), VFNMS(LDK(KP910592997), TN, TG));
Chris@82 166 ST(&(xo[WS(os, 4)]), VFNMSI(TQ, TP), ovs, &(xo[0]));
Chris@82 167 ST(&(xo[WS(os, 11)]), VFMAI(TQ, TP), ovs, &(xo[WS(os, 1)]));
Chris@82 168 TT = VFNMS(LDK(KP823639103), TS, TR);
Chris@82 169 TW = VMUL(LDK(KP951056516), VFMA(LDK(KP910592997), TV, TU));
Chris@82 170 ST(&(xo[WS(os, 2)]), VFNMSI(TW, TT), ovs, &(xo[0]));
Chris@82 171 ST(&(xo[WS(os, 13)]), VFMAI(TW, TT), ovs, &(xo[WS(os, 1)]));
Chris@82 172 }
Chris@82 173 }
Chris@82 174 }
Chris@82 175 }
Chris@82 176 VLEAVE();
Chris@82 177 }
Chris@82 178
Chris@82 179 static const kdft_desc desc = { 15, XSIMD_STRING("n1bv_15"), {36, 7, 42, 0}, &GENUS, 0, 0, 0, 0 };
Chris@82 180
Chris@82 181 void XSIMD(codelet_n1bv_15) (planner *p) {
Chris@82 182 X(kdft_register) (p, n1bv_15, &desc);
Chris@82 183 }
Chris@82 184
Chris@82 185 #else
Chris@82 186
Chris@82 187 /* Generated by: ../../../genfft/gen_notw_c.native -simd -compact -variables 4 -pipeline-latency 8 -sign 1 -n 15 -name n1bv_15 -include dft/simd/n1b.h */
Chris@82 188
Chris@82 189 /*
Chris@82 190 * This function contains 78 FP additions, 25 FP multiplications,
Chris@82 191 * (or, 64 additions, 11 multiplications, 14 fused multiply/add),
Chris@82 192 * 55 stack variables, 10 constants, and 30 memory accesses
Chris@82 193 */
Chris@82 194 #include "dft/simd/n1b.h"
Chris@82 195
Chris@82 196 static void n1bv_15(const R *ri, const R *ii, R *ro, R *io, stride is, stride os, INT v, INT ivs, INT ovs)
Chris@82 197 {
Chris@82 198 DVK(KP216506350, +0.216506350946109661690930792688234045867850657);
Chris@82 199 DVK(KP509036960, +0.509036960455127183450980863393907648510733164);
Chris@82 200 DVK(KP823639103, +0.823639103546331925877420039278190003029660514);
Chris@82 201 DVK(KP951056516, +0.951056516295153572116439333379382143405698634);
Chris@82 202 DVK(KP587785252, +0.587785252292473129168705954639072768597652438);
Chris@82 203 DVK(KP250000000, +0.250000000000000000000000000000000000000000000);
Chris@82 204 DVK(KP559016994, +0.559016994374947424102293417182819058860154590);
Chris@82 205 DVK(KP866025403, +0.866025403784438646763723170752936183471402627);
Chris@82 206 DVK(KP484122918, +0.484122918275927110647408174972799951354115213);
Chris@82 207 DVK(KP500000000, +0.500000000000000000000000000000000000000000000);
Chris@82 208 {
Chris@82 209 INT i;
Chris@82 210 const R *xi;
Chris@82 211 R *xo;
Chris@82 212 xi = ii;
Chris@82 213 xo = io;
Chris@82 214 for (i = v; i > 0; i = i - VL, xi = xi + (VL * ivs), xo = xo + (VL * ovs), MAKE_VOLATILE_STRIDE(30, is), MAKE_VOLATILE_STRIDE(30, os)) {
Chris@82 215 V Ti, T11, TH, Ts, TL, TM, Tz, TC, TD, TI, T12, T13, T14, T15, T16;
Chris@82 216 V T17, Tf, Tj, TZ, T10;
Chris@82 217 {
Chris@82 218 V TF, Tg, Th, TG;
Chris@82 219 TF = LD(&(xi[0]), ivs, &(xi[0]));
Chris@82 220 Tg = LD(&(xi[WS(is, 5)]), ivs, &(xi[WS(is, 1)]));
Chris@82 221 Th = LD(&(xi[WS(is, 10)]), ivs, &(xi[0]));
Chris@82 222 TG = VADD(Tg, Th);
Chris@82 223 Ti = VSUB(Tg, Th);
Chris@82 224 T11 = VADD(TF, TG);
Chris@82 225 TH = VFNMS(LDK(KP500000000), TG, TF);
Chris@82 226 }
Chris@82 227 {
Chris@82 228 V Tm, Tn, T3, To, Tw, Tx, Td, Ty, Tp, Tq, T6, Tr, Tt, Tu, Ta;
Chris@82 229 V Tv, T7, Te;
Chris@82 230 {
Chris@82 231 V T1, T2, Tb, Tc;
Chris@82 232 Tm = LD(&(xi[WS(is, 3)]), ivs, &(xi[WS(is, 1)]));
Chris@82 233 T1 = LD(&(xi[WS(is, 8)]), ivs, &(xi[0]));
Chris@82 234 T2 = LD(&(xi[WS(is, 13)]), ivs, &(xi[WS(is, 1)]));
Chris@82 235 Tn = VADD(T1, T2);
Chris@82 236 T3 = VSUB(T1, T2);
Chris@82 237 To = VFNMS(LDK(KP500000000), Tn, Tm);
Chris@82 238 Tw = LD(&(xi[WS(is, 9)]), ivs, &(xi[WS(is, 1)]));
Chris@82 239 Tb = LD(&(xi[WS(is, 14)]), ivs, &(xi[0]));
Chris@82 240 Tc = LD(&(xi[WS(is, 4)]), ivs, &(xi[0]));
Chris@82 241 Tx = VADD(Tb, Tc);
Chris@82 242 Td = VSUB(Tb, Tc);
Chris@82 243 Ty = VFNMS(LDK(KP500000000), Tx, Tw);
Chris@82 244 }
Chris@82 245 {
Chris@82 246 V T4, T5, T8, T9;
Chris@82 247 Tp = LD(&(xi[WS(is, 12)]), ivs, &(xi[0]));
Chris@82 248 T4 = LD(&(xi[WS(is, 2)]), ivs, &(xi[0]));
Chris@82 249 T5 = LD(&(xi[WS(is, 7)]), ivs, &(xi[WS(is, 1)]));
Chris@82 250 Tq = VADD(T4, T5);
Chris@82 251 T6 = VSUB(T4, T5);
Chris@82 252 Tr = VFNMS(LDK(KP500000000), Tq, Tp);
Chris@82 253 Tt = LD(&(xi[WS(is, 6)]), ivs, &(xi[0]));
Chris@82 254 T8 = LD(&(xi[WS(is, 11)]), ivs, &(xi[WS(is, 1)]));
Chris@82 255 T9 = LD(&(xi[WS(is, 1)]), ivs, &(xi[WS(is, 1)]));
Chris@82 256 Tu = VADD(T8, T9);
Chris@82 257 Ta = VSUB(T8, T9);
Chris@82 258 Tv = VFNMS(LDK(KP500000000), Tu, Tt);
Chris@82 259 }
Chris@82 260 Ts = VSUB(To, Tr);
Chris@82 261 TL = VSUB(T3, T6);
Chris@82 262 TM = VSUB(Ta, Td);
Chris@82 263 Tz = VSUB(Tv, Ty);
Chris@82 264 TC = VADD(To, Tr);
Chris@82 265 TD = VADD(Tv, Ty);
Chris@82 266 TI = VADD(TC, TD);
Chris@82 267 T12 = VADD(Tm, Tn);
Chris@82 268 T13 = VADD(Tp, Tq);
Chris@82 269 T14 = VADD(T12, T13);
Chris@82 270 T15 = VADD(Tt, Tu);
Chris@82 271 T16 = VADD(Tw, Tx);
Chris@82 272 T17 = VADD(T15, T16);
Chris@82 273 T7 = VADD(T3, T6);
Chris@82 274 Te = VADD(Ta, Td);
Chris@82 275 Tf = VMUL(LDK(KP484122918), VSUB(T7, Te));
Chris@82 276 Tj = VADD(T7, Te);
Chris@82 277 }
Chris@82 278 TZ = VADD(TH, TI);
Chris@82 279 T10 = VBYI(VMUL(LDK(KP866025403), VADD(Ti, Tj)));
Chris@82 280 ST(&(xo[WS(os, 5)]), VSUB(TZ, T10), ovs, &(xo[WS(os, 1)]));
Chris@82 281 ST(&(xo[WS(os, 10)]), VADD(T10, TZ), ovs, &(xo[0]));
Chris@82 282 {
Chris@82 283 V T1a, T18, T19, T1e, T1f, T1c, T1d, T1g, T1b;
Chris@82 284 T1a = VMUL(LDK(KP559016994), VSUB(T14, T17));
Chris@82 285 T18 = VADD(T14, T17);
Chris@82 286 T19 = VFNMS(LDK(KP250000000), T18, T11);
Chris@82 287 T1c = VSUB(T12, T13);
Chris@82 288 T1d = VSUB(T15, T16);
Chris@82 289 T1e = VBYI(VFNMS(LDK(KP951056516), T1d, VMUL(LDK(KP587785252), T1c)));
Chris@82 290 T1f = VBYI(VFMA(LDK(KP951056516), T1c, VMUL(LDK(KP587785252), T1d)));
Chris@82 291 ST(&(xo[0]), VADD(T11, T18), ovs, &(xo[0]));
Chris@82 292 T1g = VADD(T1a, T19);
Chris@82 293 ST(&(xo[WS(os, 6)]), VADD(T1f, T1g), ovs, &(xo[0]));
Chris@82 294 ST(&(xo[WS(os, 9)]), VSUB(T1g, T1f), ovs, &(xo[WS(os, 1)]));
Chris@82 295 T1b = VSUB(T19, T1a);
Chris@82 296 ST(&(xo[WS(os, 3)]), VSUB(T1b, T1e), ovs, &(xo[WS(os, 1)]));
Chris@82 297 ST(&(xo[WS(os, 12)]), VADD(T1e, T1b), ovs, &(xo[0]));
Chris@82 298 }
Chris@82 299 {
Chris@82 300 V TA, TN, TU, TS, Tl, TR, TK, TV, Tk, TE, TJ;
Chris@82 301 TA = VFMA(LDK(KP951056516), Ts, VMUL(LDK(KP587785252), Tz));
Chris@82 302 TN = VFMA(LDK(KP823639103), TL, VMUL(LDK(KP509036960), TM));
Chris@82 303 TU = VFNMS(LDK(KP823639103), TM, VMUL(LDK(KP509036960), TL));
Chris@82 304 TS = VFNMS(LDK(KP951056516), Tz, VMUL(LDK(KP587785252), Ts));
Chris@82 305 Tk = VFNMS(LDK(KP216506350), Tj, VMUL(LDK(KP866025403), Ti));
Chris@82 306 Tl = VADD(Tf, Tk);
Chris@82 307 TR = VSUB(Tf, Tk);
Chris@82 308 TE = VMUL(LDK(KP559016994), VSUB(TC, TD));
Chris@82 309 TJ = VFNMS(LDK(KP250000000), TI, TH);
Chris@82 310 TK = VADD(TE, TJ);
Chris@82 311 TV = VSUB(TJ, TE);
Chris@82 312 {
Chris@82 313 V TB, TO, TX, TY;
Chris@82 314 TB = VBYI(VADD(Tl, TA));
Chris@82 315 TO = VSUB(TK, TN);
Chris@82 316 ST(&(xo[WS(os, 1)]), VADD(TB, TO), ovs, &(xo[WS(os, 1)]));
Chris@82 317 ST(&(xo[WS(os, 14)]), VSUB(TO, TB), ovs, &(xo[0]));
Chris@82 318 TX = VBYI(VSUB(TS, TR));
Chris@82 319 TY = VSUB(TV, TU);
Chris@82 320 ST(&(xo[WS(os, 7)]), VADD(TX, TY), ovs, &(xo[WS(os, 1)]));
Chris@82 321 ST(&(xo[WS(os, 8)]), VSUB(TY, TX), ovs, &(xo[0]));
Chris@82 322 }
Chris@82 323 {
Chris@82 324 V TP, TQ, TT, TW;
Chris@82 325 TP = VBYI(VSUB(Tl, TA));
Chris@82 326 TQ = VADD(TN, TK);
Chris@82 327 ST(&(xo[WS(os, 4)]), VADD(TP, TQ), ovs, &(xo[0]));
Chris@82 328 ST(&(xo[WS(os, 11)]), VSUB(TQ, TP), ovs, &(xo[WS(os, 1)]));
Chris@82 329 TT = VBYI(VADD(TR, TS));
Chris@82 330 TW = VADD(TU, TV);
Chris@82 331 ST(&(xo[WS(os, 2)]), VADD(TT, TW), ovs, &(xo[0]));
Chris@82 332 ST(&(xo[WS(os, 13)]), VSUB(TW, TT), ovs, &(xo[WS(os, 1)]));
Chris@82 333 }
Chris@82 334 }
Chris@82 335 }
Chris@82 336 }
Chris@82 337 VLEAVE();
Chris@82 338 }
Chris@82 339
Chris@82 340 static const kdft_desc desc = { 15, XSIMD_STRING("n1bv_15"), {64, 11, 14, 0}, &GENUS, 0, 0, 0, 0 };
Chris@82 341
Chris@82 342 void XSIMD(codelet_n1bv_15) (planner *p) {
Chris@82 343 X(kdft_register) (p, n1bv_15, &desc);
Chris@82 344 }
Chris@82 345
Chris@82 346 #endif