Mercurial > hg > sv-dependency-builds

annotate src/fftw-3.3.5/rdft/simd/common/hc2cfdftv_12.c @ 56:af97cad61ff0

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
Add updated build of PortAudio for OSX
author Chris Cannam <cannam@all-day-breakfast.com>
date Tue, 03 Jan 2017 15:10:52 +0000
parents 2cd0e3b3e1fd
children
rev   line source
Chris@42 1 /*
Chris@42 2 * Copyright (c) 2003, 2007-14 Matteo Frigo
Chris@42 3 * Copyright (c) 2003, 2007-14 Massachusetts Institute of Technology
Chris@42 4 *
Chris@42 5 * This program is free software; you can redistribute it and/or modify
Chris@42 6 * it under the terms of the GNU General Public License as published by
Chris@42 7 * the Free Software Foundation; either version 2 of the License, or
Chris@42 8 * (at your option) any later version.
Chris@42 9 *
Chris@42 10 * This program is distributed in the hope that it will be useful,
Chris@42 11 * but WITHOUT ANY WARRANTY; without even the implied warranty of
Chris@42 12 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
Chris@42 13 * GNU General Public License for more details.
Chris@42 14 *
Chris@42 15 * You should have received a copy of the GNU General Public License
Chris@42 16 * along with this program; if not, write to the Free Software
Chris@42 17 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
Chris@42 18 *
Chris@42 19 */
Chris@42 20
Chris@42 21 /* This file was automatically generated --- DO NOT EDIT */
Chris@42 22 /* Generated on Sat Jul 30 16:52:40 EDT 2016 */
Chris@42 23
Chris@42 24 #include "codelet-rdft.h"
Chris@42 25
Chris@42 26 #ifdef HAVE_FMA
Chris@42 27
Chris@42 28 /* Generated by: ../../../genfft/gen_hc2cdft_c.native -fma -reorder-insns -schedule-for-pipeline -simd -compact -variables 4 -pipeline-latency 8 -trivial-stores -variables 32 -no-generate-bytw -n 12 -dit -name hc2cfdftv_12 -include hc2cfv.h */
Chris@42 29
Chris@42 30 /*
Chris@42 31 * This function contains 71 FP additions, 66 FP multiplications,
Chris@42 32 * (or, 41 additions, 36 multiplications, 30 fused multiply/add),
Chris@42 33 * 86 stack variables, 2 constants, and 24 memory accesses
Chris@42 34 */
Chris@42 35 #include "hc2cfv.h"
Chris@42 36
Chris@42 37 static void hc2cfdftv_12(R *Rp, R *Ip, R *Rm, R *Im, const R *W, stride rs, INT mb, INT me, INT ms)
Chris@42 38 {
Chris@42 39 DVK(KP866025403, +0.866025403784438646763723170752936183471402627);
Chris@42 40 DVK(KP500000000, +0.500000000000000000000000000000000000000000000);
Chris@42 41 {
Chris@42 42 INT m;
Chris@42 43 for (m = mb, W = W + ((mb - 1) * ((TWVL / VL) * 22)); m < me; m = m + VL, Rp = Rp + (VL * ms), Ip = Ip + (VL * ms), Rm = Rm - (VL * ms), Im = Im - (VL * ms), W = W + (TWVL * 22), MAKE_VOLATILE_STRIDE(48, rs)) {
Chris@42 44 V T3, T7, TH, TE, Th, TC, Tq, T11, TU, Tx, Tb, Tz, Tu, Tw, Tp;
Chris@42 45 V Tl, T9, Ta, T8, Ty, Tn, To, Tm, TG, T1, T2, Tt, T5, T6, T4;
Chris@42 46 V Tv, Tj, Tk, Ti, TD, Tf, Tg, Te, TB, TT, TF, TR, Tr;
Chris@42 47 T1 = LD(&(Rp[0]), ms, &(Rp[0]));
Chris@42 48 T2 = LD(&(Rm[0]), -ms, &(Rm[0]));
Chris@42 49 Tt = LDW(&(W[0]));
Chris@42 50 T5 = LD(&(Rp[WS(rs, 2)]), ms, &(Rp[0]));
Chris@42 51 T6 = LD(&(Rm[WS(rs, 2)]), -ms, &(Rm[0]));
Chris@42 52 T4 = LDW(&(W[TWVL * 6]));
Chris@42 53 Tv = LDW(&(W[TWVL * 8]));
Chris@42 54 Tn = LD(&(Rp[WS(rs, 1)]), ms, &(Rp[WS(rs, 1)]));
Chris@42 55 To = LD(&(Rm[WS(rs, 1)]), -ms, &(Rm[WS(rs, 1)]));
Chris@42 56 T3 = VFMACONJ(T2, T1);
Chris@42 57 Tu = VZMULIJ(Tt, VFNMSCONJ(T2, T1));
Chris@42 58 Tm = LDW(&(W[TWVL * 2]));
Chris@42 59 TG = LDW(&(W[TWVL * 4]));
Chris@42 60 T7 = VZMULJ(T4, VFMACONJ(T6, T5));
Chris@42 61 Tw = VZMULIJ(Tv, VFNMSCONJ(T6, T5));
Chris@42 62 Tj = LD(&(Rp[WS(rs, 5)]), ms, &(Rp[WS(rs, 1)]));
Chris@42 63 Tk = LD(&(Rm[WS(rs, 5)]), -ms, &(Rm[WS(rs, 1)]));
Chris@42 64 Ti = LDW(&(W[TWVL * 18]));
Chris@42 65 TD = LDW(&(W[TWVL * 20]));
Chris@42 66 Tp = VZMULJ(Tm, VFMACONJ(To, Tn));
Chris@42 67 TH = VZMULIJ(TG, VFNMSCONJ(To, Tn));
Chris@42 68 Tf = LD(&(Rp[WS(rs, 3)]), ms, &(Rp[WS(rs, 1)]));
Chris@42 69 Tg = LD(&(Rm[WS(rs, 3)]), -ms, &(Rm[WS(rs, 1)]));
Chris@42 70 Te = LDW(&(W[TWVL * 10]));
Chris@42 71 TB = LDW(&(W[TWVL * 12]));
Chris@42 72 Tl = VZMULJ(Ti, VFMACONJ(Tk, Tj));
Chris@42 73 TE = VZMULIJ(TD, VFNMSCONJ(Tk, Tj));
Chris@42 74 T9 = LD(&(Rp[WS(rs, 4)]), ms, &(Rp[0]));
Chris@42 75 Ta = LD(&(Rm[WS(rs, 4)]), -ms, &(Rm[0]));
Chris@42 76 T8 = LDW(&(W[TWVL * 14]));
Chris@42 77 Ty = LDW(&(W[TWVL * 16]));
Chris@42 78 Th = VZMULJ(Te, VFMACONJ(Tg, Tf));
Chris@42 79 TC = VZMULIJ(TB, VFNMSCONJ(Tg, Tf));
Chris@42 80 Tq = VADD(Tl, Tp);
Chris@42 81 T11 = VSUB(Tp, Tl);
Chris@42 82 TU = VSUB(Tu, Tw);
Chris@42 83 Tx = VADD(Tu, Tw);
Chris@42 84 Tb = VZMULJ(T8, VFMACONJ(Ta, T9));
Chris@42 85 Tz = VZMULIJ(Ty, VFNMSCONJ(Ta, T9));
Chris@42 86 TT = VSUB(TC, TE);
Chris@42 87 TF = VADD(TC, TE);
Chris@42 88 TR = VFNMS(LDK(KP500000000), Tq, Th);
Chris@42 89 Tr = VADD(Th, Tq);
Chris@42 90 {
Chris@42 91 V TX, TA, T1d, TV, TY, TI, T1e, T12, TQ, Td, T10, Tc, T1a, TN, TJ;
Chris@42 92 V T1j, T1f, T1b, TS, TM, Ts, T17, T13, TZ, T1i, T1c, T16, TW, TP, TO;
Chris@42 93 V TL, TK, T1k, T1l, T1h, T1g, T18, T19, T15, T14;
Chris@42 94 T10 = VSUB(Tb, T7);
Chris@42 95 Tc = VADD(T7, Tb);
Chris@42 96 TX = VFNMS(LDK(KP500000000), Tx, Tz);
Chris@42 97 TA = VADD(Tx, Tz);
Chris@42 98 T1d = VADD(TU, TT);
Chris@42 99 TV = VSUB(TT, TU);
Chris@42 100 TY = VFNMS(LDK(KP500000000), TF, TH);
Chris@42 101 TI = VADD(TF, TH);
Chris@42 102 T1e = VADD(T10, T11);
Chris@42 103 T12 = VSUB(T10, T11);
Chris@42 104 TQ = VFNMS(LDK(KP500000000), Tc, T3);
Chris@42 105 Td = VADD(T3, Tc);
Chris@42 106 T1a = VADD(TX, TY);
Chris@42 107 TZ = VSUB(TX, TY);
Chris@42 108 TN = VADD(TA, TI);
Chris@42 109 TJ = VSUB(TA, TI);
Chris@42 110 T1j = VMUL(LDK(KP866025403), VADD(T1d, T1e));
Chris@42 111 T1f = VMUL(LDK(KP866025403), VSUB(T1d, T1e));
Chris@42 112 T1b = VADD(TQ, TR);
Chris@42 113 TS = VSUB(TQ, TR);
Chris@42 114 TM = VADD(Td, Tr);
Chris@42 115 Ts = VSUB(Td, Tr);
Chris@42 116 T17 = VFMA(LDK(KP866025403), T12, TZ);
Chris@42 117 T13 = VFNMS(LDK(KP866025403), T12, TZ);
Chris@42 118 T1i = VSUB(T1b, T1a);
Chris@42 119 T1c = VADD(T1a, T1b);
Chris@42 120 T16 = VFNMS(LDK(KP866025403), TV, TS);
Chris@42 121 TW = VFMA(LDK(KP866025403), TV, TS);
Chris@42 122 TP = VCONJ(VMUL(LDK(KP500000000), VADD(TN, TM)));
Chris@42 123 TO = VMUL(LDK(KP500000000), VSUB(TM, TN));
Chris@42 124 TL = VCONJ(VMUL(LDK(KP500000000), VFNMSI(TJ, Ts)));
Chris@42 125 TK = VMUL(LDK(KP500000000), VFMAI(TJ, Ts));
Chris@42 126 T1k = VCONJ(VMUL(LDK(KP500000000), VFNMSI(T1j, T1i)));
Chris@42 127 T1l = VMUL(LDK(KP500000000), VFMAI(T1j, T1i));
Chris@42 128 T1h = VMUL(LDK(KP500000000), VFMAI(T1f, T1c));
Chris@42 129 T1g = VCONJ(VMUL(LDK(KP500000000), VFNMSI(T1f, T1c)));
Chris@42 130 T18 = VMUL(LDK(KP500000000), VFNMSI(T17, T16));
Chris@42 131 T19 = VCONJ(VMUL(LDK(KP500000000), VFMAI(T17, T16)));
Chris@42 132 T15 = VCONJ(VMUL(LDK(KP500000000), VFMAI(T13, TW)));
Chris@42 133 T14 = VMUL(LDK(KP500000000), VFNMSI(T13, TW));
Chris@42 134 ST(&(Rm[WS(rs, 5)]), TP, -ms, &(Rm[WS(rs, 1)]));
Chris@42 135 ST(&(Rp[0]), TO, ms, &(Rp[0]));
Chris@42 136 ST(&(Rm[WS(rs, 2)]), TL, -ms, &(Rm[0]));
Chris@42 137 ST(&(Rp[WS(rs, 3)]), TK, ms, &(Rp[WS(rs, 1)]));
Chris@42 138 ST(&(Rm[WS(rs, 3)]), T1k, -ms, &(Rm[WS(rs, 1)]));
Chris@42 139 ST(&(Rp[WS(rs, 4)]), T1l, ms, &(Rp[0]));
Chris@42 140 ST(&(Rp[WS(rs, 2)]), T1h, ms, &(Rp[0]));
Chris@42 141 ST(&(Rm[WS(rs, 1)]), T1g, -ms, &(Rm[WS(rs, 1)]));
Chris@42 142 ST(&(Rp[WS(rs, 5)]), T18, ms, &(Rp[WS(rs, 1)]));
Chris@42 143 ST(&(Rm[WS(rs, 4)]), T19, -ms, &(Rm[0]));
Chris@42 144 ST(&(Rm[0]), T15, -ms, &(Rm[0]));
Chris@42 145 ST(&(Rp[WS(rs, 1)]), T14, ms, &(Rp[WS(rs, 1)]));
Chris@42 146 }
Chris@42 147 }
Chris@42 148 }
Chris@42 149 VLEAVE();
Chris@42 150 }
Chris@42 151
Chris@42 152 static const tw_instr twinstr[] = {
Chris@42 153 VTW(1, 1),
Chris@42 154 VTW(1, 2),
Chris@42 155 VTW(1, 3),
Chris@42 156 VTW(1, 4),
Chris@42 157 VTW(1, 5),
Chris@42 158 VTW(1, 6),
Chris@42 159 VTW(1, 7),
Chris@42 160 VTW(1, 8),
Chris@42 161 VTW(1, 9),
Chris@42 162 VTW(1, 10),
Chris@42 163 VTW(1, 11),
Chris@42 164 {TW_NEXT, VL, 0}
Chris@42 165 };
Chris@42 166
Chris@42 167 static const hc2c_desc desc = { 12, XSIMD_STRING("hc2cfdftv_12"), twinstr, &GENUS, {41, 36, 30, 0} };
Chris@42 168
Chris@42 169 void XSIMD(codelet_hc2cfdftv_12) (planner *p) {
Chris@42 170 X(khc2c_register) (p, hc2cfdftv_12, &desc, HC2C_VIA_DFT);
Chris@42 171 }
Chris@42 172 #else /* HAVE_FMA */
Chris@42 173
Chris@42 174 /* Generated by: ../../../genfft/gen_hc2cdft_c.native -simd -compact -variables 4 -pipeline-latency 8 -trivial-stores -variables 32 -no-generate-bytw -n 12 -dit -name hc2cfdftv_12 -include hc2cfv.h */
Chris@42 175
Chris@42 176 /*
Chris@42 177 * This function contains 71 FP additions, 41 FP multiplications,
Chris@42 178 * (or, 67 additions, 37 multiplications, 4 fused multiply/add),
Chris@42 179 * 58 stack variables, 4 constants, and 24 memory accesses
Chris@42 180 */
Chris@42 181 #include "hc2cfv.h"
Chris@42 182
Chris@42 183 static void hc2cfdftv_12(R *Rp, R *Ip, R *Rm, R *Im, const R *W, stride rs, INT mb, INT me, INT ms)
Chris@42 184 {
Chris@42 185 DVK(KP433012701, +0.433012701892219323381861585376468091735701313);
Chris@42 186 DVK(KP866025403, +0.866025403784438646763723170752936183471402627);
Chris@42 187 DVK(KP250000000, +0.250000000000000000000000000000000000000000000);
Chris@42 188 DVK(KP500000000, +0.500000000000000000000000000000000000000000000);
Chris@42 189 {
Chris@42 190 INT m;
Chris@42 191 for (m = mb, W = W + ((mb - 1) * ((TWVL / VL) * 22)); m < me; m = m + VL, Rp = Rp + (VL * ms), Ip = Ip + (VL * ms), Rm = Rm - (VL * ms), Im = Im - (VL * ms), W = W + (TWVL * 22), MAKE_VOLATILE_STRIDE(48, rs)) {
Chris@42 192 V TX, T13, T4, Tf, TZ, TD, TF, T17, TW, T14, Tw, Tl, T10, TL, TN;
Chris@42 193 V T16;
Chris@42 194 {
Chris@42 195 V T1, T3, TA, Tb, Td, Te, T9, TC, T2, Tz, Tc, Ta, T6, T8, T7;
Chris@42 196 V T5, TB, TE, Ti, Tk, TI, Ts, Tu, Tv, Tq, TK, Tj, TH, Tt, Tr;
Chris@42 197 V Tn, Tp, To, Tm, TJ, Th, TM;
Chris@42 198 T1 = LD(&(Rp[0]), ms, &(Rp[0]));
Chris@42 199 T2 = LD(&(Rm[0]), -ms, &(Rm[0]));
Chris@42 200 T3 = VCONJ(T2);
Chris@42 201 Tz = LDW(&(W[0]));
Chris@42 202 TA = VZMULIJ(Tz, VSUB(T3, T1));
Chris@42 203 Tb = LD(&(Rp[WS(rs, 4)]), ms, &(Rp[0]));
Chris@42 204 Tc = LD(&(Rm[WS(rs, 4)]), -ms, &(Rm[0]));
Chris@42 205 Td = VCONJ(Tc);
Chris@42 206 Ta = LDW(&(W[TWVL * 14]));
Chris@42 207 Te = VZMULJ(Ta, VADD(Tb, Td));
Chris@42 208 T6 = LD(&(Rp[WS(rs, 2)]), ms, &(Rp[0]));
Chris@42 209 T7 = LD(&(Rm[WS(rs, 2)]), -ms, &(Rm[0]));
Chris@42 210 T8 = VCONJ(T7);
Chris@42 211 T5 = LDW(&(W[TWVL * 6]));
Chris@42 212 T9 = VZMULJ(T5, VADD(T6, T8));
Chris@42 213 TB = LDW(&(W[TWVL * 8]));
Chris@42 214 TC = VZMULIJ(TB, VSUB(T8, T6));
Chris@42 215 TX = VSUB(TC, TA);
Chris@42 216 T13 = VSUB(Te, T9);
Chris@42 217 T4 = VADD(T1, T3);
Chris@42 218 Tf = VADD(T9, Te);
Chris@42 219 TZ = VFNMS(LDK(KP250000000), Tf, VMUL(LDK(KP500000000), T4));
Chris@42 220 TD = VADD(TA, TC);
Chris@42 221 TE = LDW(&(W[TWVL * 16]));
Chris@42 222 TF = VZMULIJ(TE, VSUB(Td, Tb));
Chris@42 223 T17 = VFNMS(LDK(KP500000000), TD, TF);
Chris@42 224 Ti = LD(&(Rp[WS(rs, 3)]), ms, &(Rp[WS(rs, 1)]));
Chris@42 225 Tj = LD(&(Rm[WS(rs, 3)]), -ms, &(Rm[WS(rs, 1)]));
Chris@42 226 Tk = VCONJ(Tj);
Chris@42 227 TH = LDW(&(W[TWVL * 12]));
Chris@42 228 TI = VZMULIJ(TH, VSUB(Tk, Ti));
Chris@42 229 Ts = LD(&(Rp[WS(rs, 1)]), ms, &(Rp[WS(rs, 1)]));
Chris@42 230 Tt = LD(&(Rm[WS(rs, 1)]), -ms, &(Rm[WS(rs, 1)]));
Chris@42 231 Tu = VCONJ(Tt);
Chris@42 232 Tr = LDW(&(W[TWVL * 2]));
Chris@42 233 Tv = VZMULJ(Tr, VADD(Ts, Tu));
Chris@42 234 Tn = LD(&(Rp[WS(rs, 5)]), ms, &(Rp[WS(rs, 1)]));
Chris@42 235 To = LD(&(Rm[WS(rs, 5)]), -ms, &(Rm[WS(rs, 1)]));
Chris@42 236 Tp = VCONJ(To);
Chris@42 237 Tm = LDW(&(W[TWVL * 18]));
Chris@42 238 Tq = VZMULJ(Tm, VADD(Tn, Tp));
Chris@42 239 TJ = LDW(&(W[TWVL * 20]));
Chris@42 240 TK = VZMULIJ(TJ, VSUB(Tp, Tn));
Chris@42 241 TW = VSUB(TK, TI);
Chris@42 242 T14 = VSUB(Tv, Tq);
Chris@42 243 Tw = VADD(Tq, Tv);
Chris@42 244 Th = LDW(&(W[TWVL * 10]));
Chris@42 245 Tl = VZMULJ(Th, VADD(Ti, Tk));
Chris@42 246 T10 = VFNMS(LDK(KP250000000), Tw, VMUL(LDK(KP500000000), Tl));
Chris@42 247 TL = VADD(TI, TK);
Chris@42 248 TM = LDW(&(W[TWVL * 4]));
Chris@42 249 TN = VZMULIJ(TM, VSUB(Tu, Ts));
Chris@42 250 T16 = VFNMS(LDK(KP500000000), TL, TN);
Chris@42 251 }
Chris@42 252 {
Chris@42 253 V Ty, TS, TP, TT, Tg, Tx, TG, TO, TQ, TV, TR, TU, T1i, T1o, T1l;
Chris@42 254 V T1p, T1g, T1h, T1j, T1k, T1m, T1r, T1n, T1q, T12, T1c, T19, T1d, TY, T11;
Chris@42 255 V T15, T18, T1a, T1f, T1b, T1e;
Chris@42 256 Tg = VADD(T4, Tf);
Chris@42 257 Tx = VADD(Tl, Tw);
Chris@42 258 Ty = VADD(Tg, Tx);
Chris@42 259 TS = VSUB(Tg, Tx);
Chris@42 260 TG = VADD(TD, TF);
Chris@42 261 TO = VADD(TL, TN);
Chris@42 262 TP = VADD(TG, TO);
Chris@42 263 TT = VBYI(VSUB(TO, TG));
Chris@42 264 TQ = VCONJ(VMUL(LDK(KP500000000), VSUB(Ty, TP)));
Chris@42 265 ST(&(Rm[WS(rs, 5)]), TQ, -ms, &(Rm[WS(rs, 1)]));
Chris@42 266 TV = VMUL(LDK(KP500000000), VADD(TS, TT));
Chris@42 267 ST(&(Rp[WS(rs, 3)]), TV, ms, &(Rp[WS(rs, 1)]));
Chris@42 268 TR = VMUL(LDK(KP500000000), VADD(Ty, TP));
Chris@42 269 ST(&(Rp[0]), TR, ms, &(Rp[0]));
Chris@42 270 TU = VCONJ(VMUL(LDK(KP500000000), VSUB(TS, TT)));
Chris@42 271 ST(&(Rm[WS(rs, 2)]), TU, -ms, &(Rm[0]));
Chris@42 272 T1g = VADD(TX, TW);
Chris@42 273 T1h = VADD(T13, T14);
Chris@42 274 T1i = VMUL(LDK(KP500000000), VBYI(VMUL(LDK(KP866025403), VSUB(T1g, T1h))));
Chris@42 275 T1o = VMUL(LDK(KP500000000), VBYI(VMUL(LDK(KP866025403), VADD(T1g, T1h))));
Chris@42 276 T1j = VADD(TZ, T10);
Chris@42 277 T1k = VMUL(LDK(KP500000000), VADD(T17, T16));
Chris@42 278 T1l = VSUB(T1j, T1k);
Chris@42 279 T1p = VADD(T1j, T1k);
Chris@42 280 T1m = VADD(T1i, T1l);
Chris@42 281 ST(&(Rp[WS(rs, 2)]), T1m, ms, &(Rp[0]));
Chris@42 282 T1r = VCONJ(VSUB(T1p, T1o));
Chris@42 283 ST(&(Rm[WS(rs, 3)]), T1r, -ms, &(Rm[WS(rs, 1)]));
Chris@42 284 T1n = VCONJ(VSUB(T1l, T1i));
Chris@42 285 ST(&(Rm[WS(rs, 1)]), T1n, -ms, &(Rm[WS(rs, 1)]));
Chris@42 286 T1q = VADD(T1o, T1p);
Chris@42 287 ST(&(Rp[WS(rs, 4)]), T1q, ms, &(Rp[0]));
Chris@42 288 TY = VMUL(LDK(KP433012701), VSUB(TW, TX));
Chris@42 289 T11 = VSUB(TZ, T10);
Chris@42 290 T12 = VADD(TY, T11);
Chris@42 291 T1c = VSUB(T11, TY);
Chris@42 292 T15 = VMUL(LDK(KP866025403), VSUB(T13, T14));
Chris@42 293 T18 = VSUB(T16, T17);
Chris@42 294 T19 = VMUL(LDK(KP500000000), VBYI(VSUB(T15, T18)));
Chris@42 295 T1d = VMUL(LDK(KP500000000), VBYI(VADD(T15, T18)));
Chris@42 296 T1a = VCONJ(VSUB(T12, T19));
Chris@42 297 ST(&(Rm[0]), T1a, -ms, &(Rm[0]));
Chris@42 298 T1f = VCONJ(VADD(T1c, T1d));
Chris@42 299 ST(&(Rm[WS(rs, 4)]), T1f, -ms, &(Rm[0]));
Chris@42 300 T1b = VADD(T12, T19);
Chris@42 301 ST(&(Rp[WS(rs, 1)]), T1b, ms, &(Rp[WS(rs, 1)]));
Chris@42 302 T1e = VSUB(T1c, T1d);
Chris@42 303 ST(&(Rp[WS(rs, 5)]), T1e, ms, &(Rp[WS(rs, 1)]));
Chris@42 304 }
Chris@42 305 }
Chris@42 306 }
Chris@42 307 VLEAVE();
Chris@42 308 }
Chris@42 309
Chris@42 310 static const tw_instr twinstr[] = {
Chris@42 311 VTW(1, 1),
Chris@42 312 VTW(1, 2),
Chris@42 313 VTW(1, 3),
Chris@42 314 VTW(1, 4),
Chris@42 315 VTW(1, 5),
Chris@42 316 VTW(1, 6),
Chris@42 317 VTW(1, 7),
Chris@42 318 VTW(1, 8),
Chris@42 319 VTW(1, 9),
Chris@42 320 VTW(1, 10),
Chris@42 321 VTW(1, 11),
Chris@42 322 {TW_NEXT, VL, 0}
Chris@42 323 };
Chris@42 324
Chris@42 325 static const hc2c_desc desc = { 12, XSIMD_STRING("hc2cfdftv_12"), twinstr, &GENUS, {67, 37, 4, 0} };
Chris@42 326
Chris@42 327 void XSIMD(codelet_hc2cfdftv_12) (planner *p) {
Chris@42 328 X(khc2c_register) (p, hc2cfdftv_12, &desc, HC2C_VIA_DFT);
Chris@42 329 }
Chris@42 330 #endif /* HAVE_FMA */