Mercurial > hg > sv-dependency-builds

annotate src/fftw-3.3.5/dft/simd/common/t1sv_32.c @ 42:2cd0e3b3e1fd

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
Current fftw source
author Chris Cannam
date Tue, 18 Oct 2016 13:40:26 +0100
parents
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 16:45:10 EDT 2016 */
Chris@42 23
Chris@42 24 #include "codelet-dft.h"
Chris@42 25
Chris@42 26 #ifdef HAVE_FMA
Chris@42 27
Chris@42 28 /* Generated by: ../../../genfft/gen_twiddle.native -fma -reorder-insns -schedule-for-pipeline -simd -compact -variables 4 -pipeline-latency 8 -n 32 -name t1sv_32 -include ts.h */
Chris@42 29
Chris@42 30 /*
Chris@42 31 * This function contains 434 FP additions, 260 FP multiplications,
Chris@42 32 * (or, 236 additions, 62 multiplications, 198 fused multiply/add),
Chris@42 33 * 158 stack variables, 7 constants, and 128 memory accesses
Chris@42 34 */
Chris@42 35 #include "ts.h"
Chris@42 36
Chris@42 37 static void t1sv_32(R *ri, R *ii, const R *W, stride rs, INT mb, INT me, INT ms)
Chris@42 38 {
Chris@42 39 DVK(KP831469612, +0.831469612302545237078788377617905756738560812);
Chris@42 40 DVK(KP980785280, +0.980785280403230449126182236134239036973933731);
Chris@42 41 DVK(KP668178637, +0.668178637919298919997757686523080761552472251);
Chris@42 42 DVK(KP198912367, +0.198912367379658006911597622644676228597850501);
Chris@42 43 DVK(KP923879532, +0.923879532511286756128183189396788286822416626);
Chris@42 44 DVK(KP414213562, +0.414213562373095048801688724209698078569671875);
Chris@42 45 DVK(KP707106781, +0.707106781186547524400844362104849039284835938);
Chris@42 46 {
Chris@42 47 INT m;
Chris@42 48 for (m = mb, W = W + (mb * 62); m < me; m = m + (2 * VL), ri = ri + ((2 * VL) * ms), ii = ii + ((2 * VL) * ms), W = W + ((2 * VL) * 62), MAKE_VOLATILE_STRIDE(64, rs)) {
Chris@42 49 V T8Z, T90;
Chris@42 50 {
Chris@42 51 V T87, T8x, T3w, T8, T3B, T83, Tl, T8y, T6F, Tz, T3J, T5T, T6G, TM, T3Q;
Chris@42 52 V T5U, T46, T5Y, T7D, T6L, T5X, T3Z, T6M, T1f, T4l, T61, T7E, T6R, T60, T4e;
Chris@42 53 V T6O, T1G, T5r, T6c, T78, T7N, T54, T6f, T32, T7b, T4S, T65, T6X, T7I, T4v;
Chris@42 54 V T68, T29, T70, T4x, T2f, T5b, T5s, T7O, T7e, T5t, T5i, T79, T3t, T2h, T2k;
Chris@42 55 V T2j, T2o, T2r, T4H, T2y, T2n, T2q, T4y, T2i;
Chris@42 56 {
Chris@42 57 V T3U, TU, TW, TZ, TY, T13, T16, T12, T15, T3V, TX, T44, T1d;
Chris@42 58 {
Chris@42 59 V T1, T86, T3, T6, T5, Ta, Td, Tg, Tj, Tf, T84, T4, Tc, Ti, T3x;
Chris@42 60 V Tb, T2, T9;
Chris@42 61 T1 = LD(&(ri[0]), ms, &(ri[0]));
Chris@42 62 T86 = LD(&(ii[0]), ms, &(ii[0]));
Chris@42 63 T3 = LD(&(ri[WS(rs, 16)]), ms, &(ri[0]));
Chris@42 64 T6 = LD(&(ii[WS(rs, 16)]), ms, &(ii[0]));
Chris@42 65 T2 = LDW(&(W[TWVL * 30]));
Chris@42 66 T5 = LDW(&(W[TWVL * 31]));
Chris@42 67 Ta = LD(&(ri[WS(rs, 8)]), ms, &(ri[0]));
Chris@42 68 Td = LD(&(ii[WS(rs, 8)]), ms, &(ii[0]));
Chris@42 69 T9 = LDW(&(W[TWVL * 14]));
Chris@42 70 Tg = LD(&(ri[WS(rs, 24)]), ms, &(ri[0]));
Chris@42 71 Tj = LD(&(ii[WS(rs, 24)]), ms, &(ii[0]));
Chris@42 72 Tf = LDW(&(W[TWVL * 46]));
Chris@42 73 T84 = VMUL(T2, T6);
Chris@42 74 T4 = VMUL(T2, T3);
Chris@42 75 Tc = LDW(&(W[TWVL * 15]));
Chris@42 76 Ti = LDW(&(W[TWVL * 47]));
Chris@42 77 T3x = VMUL(T9, Td);
Chris@42 78 Tb = VMUL(T9, Ta);
Chris@42 79 {
Chris@42 80 V Tu, Tx, T3F, Ts, Tt, Tw;
Chris@42 81 {
Chris@42 82 V To, Tr, Tq, T3E, Tp;
Chris@42 83 {
Chris@42 84 V T3y, Te, Tn, T3A, Tk;
Chris@42 85 {
Chris@42 86 V T3z, Th, T85, T7;
Chris@42 87 To = LD(&(ri[WS(rs, 4)]), ms, &(ri[0]));
Chris@42 88 T3z = VMUL(Tf, Tj);
Chris@42 89 Th = VMUL(Tf, Tg);
Chris@42 90 T85 = VFNMS(T5, T3, T84);
Chris@42 91 T7 = VFMA(T5, T6, T4);
Chris@42 92 Tr = LD(&(ii[WS(rs, 4)]), ms, &(ii[0]));
Chris@42 93 T3y = VFNMS(Tc, Ta, T3x);
Chris@42 94 Te = VFMA(Tc, Td, Tb);
Chris@42 95 Tn = LDW(&(W[TWVL * 6]));
Chris@42 96 T3A = VFNMS(Ti, Tg, T3z);
Chris@42 97 Tk = VFMA(Ti, Tj, Th);
Chris@42 98 T87 = VADD(T85, T86);
Chris@42 99 T8x = VSUB(T86, T85);
Chris@42 100 T3w = VSUB(T1, T7);
Chris@42 101 T8 = VADD(T1, T7);
Chris@42 102 }
Chris@42 103 Tq = LDW(&(W[TWVL * 7]));
Chris@42 104 T3E = VMUL(Tn, Tr);
Chris@42 105 Tp = VMUL(Tn, To);
Chris@42 106 T3B = VSUB(T3y, T3A);
Chris@42 107 T83 = VADD(T3y, T3A);
Chris@42 108 Tl = VADD(Te, Tk);
Chris@42 109 T8y = VSUB(Te, Tk);
Chris@42 110 }
Chris@42 111 Tu = LD(&(ri[WS(rs, 20)]), ms, &(ri[0]));
Chris@42 112 Tx = LD(&(ii[WS(rs, 20)]), ms, &(ii[0]));
Chris@42 113 T3F = VFNMS(Tq, To, T3E);
Chris@42 114 Ts = VFMA(Tq, Tr, Tp);
Chris@42 115 Tt = LDW(&(W[TWVL * 38]));
Chris@42 116 Tw = LDW(&(W[TWVL * 39]));
Chris@42 117 }
Chris@42 118 {
Chris@42 119 V TB, TE, TD, TH, TK, T3G, Tv, TG, TJ, T3L, TC, TA;
Chris@42 120 TB = LD(&(ri[WS(rs, 28)]), ms, &(ri[0]));
Chris@42 121 TE = LD(&(ii[WS(rs, 28)]), ms, &(ii[0]));
Chris@42 122 TA = LDW(&(W[TWVL * 54]));
Chris@42 123 TD = LDW(&(W[TWVL * 55]));
Chris@42 124 TH = LD(&(ri[WS(rs, 12)]), ms, &(ri[0]));
Chris@42 125 TK = LD(&(ii[WS(rs, 12)]), ms, &(ii[0]));
Chris@42 126 T3G = VMUL(Tt, Tx);
Chris@42 127 Tv = VMUL(Tt, Tu);
Chris@42 128 TG = LDW(&(W[TWVL * 22]));
Chris@42 129 TJ = LDW(&(W[TWVL * 23]));
Chris@42 130 T3L = VMUL(TA, TE);
Chris@42 131 TC = VMUL(TA, TB);
Chris@42 132 {
Chris@42 133 V T19, T1c, T3P, T3K, T18, T1b, TV, T43, T1a;
Chris@42 134 {
Chris@42 135 V TQ, TT, T3M, TF, TS, T3I, T3D, T3O, TL, T3T, TR;
Chris@42 136 {
Chris@42 137 V T3H, Ty, T3N, TI, TP;
Chris@42 138 TQ = LD(&(ri[WS(rs, 2)]), ms, &(ri[0]));
Chris@42 139 TT = LD(&(ii[WS(rs, 2)]), ms, &(ii[0]));
Chris@42 140 T3H = VFNMS(Tw, Tu, T3G);
Chris@42 141 Ty = VFMA(Tw, Tx, Tv);
Chris@42 142 T3N = VMUL(TG, TK);
Chris@42 143 TI = VMUL(TG, TH);
Chris@42 144 T3M = VFNMS(TD, TB, T3L);
Chris@42 145 TF = VFMA(TD, TE, TC);
Chris@42 146 TP = LDW(&(W[TWVL * 2]));
Chris@42 147 TS = LDW(&(W[TWVL * 3]));
Chris@42 148 T6F = VADD(T3F, T3H);
Chris@42 149 T3I = VSUB(T3F, T3H);
Chris@42 150 Tz = VADD(Ts, Ty);
Chris@42 151 T3D = VSUB(Ts, Ty);
Chris@42 152 T3O = VFNMS(TJ, TH, T3N);
Chris@42 153 TL = VFMA(TJ, TK, TI);
Chris@42 154 T3T = VMUL(TP, TT);
Chris@42 155 TR = VMUL(TP, TQ);
Chris@42 156 }
Chris@42 157 T19 = LD(&(ri[WS(rs, 26)]), ms, &(ri[0]));
Chris@42 158 T1c = LD(&(ii[WS(rs, 26)]), ms, &(ii[0]));
Chris@42 159 T3J = VADD(T3D, T3I);
Chris@42 160 T5T = VSUB(T3I, T3D);
Chris@42 161 T6G = VADD(T3M, T3O);
Chris@42 162 T3P = VSUB(T3M, T3O);
Chris@42 163 TM = VADD(TF, TL);
Chris@42 164 T3K = VSUB(TF, TL);
Chris@42 165 T3U = VFNMS(TS, TQ, T3T);
Chris@42 166 TU = VFMA(TS, TT, TR);
Chris@42 167 T18 = LDW(&(W[TWVL * 50]));
Chris@42 168 T1b = LDW(&(W[TWVL * 51]));
Chris@42 169 }
Chris@42 170 TW = LD(&(ri[WS(rs, 18)]), ms, &(ri[0]));
Chris@42 171 TZ = LD(&(ii[WS(rs, 18)]), ms, &(ii[0]));
Chris@42 172 T3Q = VSUB(T3K, T3P);
Chris@42 173 T5U = VADD(T3K, T3P);
Chris@42 174 TV = LDW(&(W[TWVL * 34]));
Chris@42 175 TY = LDW(&(W[TWVL * 35]));
Chris@42 176 T43 = VMUL(T18, T1c);
Chris@42 177 T1a = VMUL(T18, T19);
Chris@42 178 T13 = LD(&(ri[WS(rs, 10)]), ms, &(ri[0]));
Chris@42 179 T16 = LD(&(ii[WS(rs, 10)]), ms, &(ii[0]));
Chris@42 180 T12 = LDW(&(W[TWVL * 18]));
Chris@42 181 T15 = LDW(&(W[TWVL * 19]));
Chris@42 182 T3V = VMUL(TV, TZ);
Chris@42 183 TX = VMUL(TV, TW);
Chris@42 184 T44 = VFNMS(T1b, T19, T43);
Chris@42 185 T1d = VFMA(T1b, T1c, T1a);
Chris@42 186 }
Chris@42 187 }
Chris@42 188 }
Chris@42 189 }
Chris@42 190 {
Chris@42 191 V T4Z, T2H, T2J, T2M, T2L, T2Q, T2T, T2P, T2S, T5p, T30, T50, T2K;
Chris@42 192 {
Chris@42 193 V T49, T1l, T1n, T1q, T1p, T1u, T1x, T4j, T1E, T1t, T1w, T4a, T1o;
Chris@42 194 {
Chris@42 195 V T1A, T1D, T1C, T4i, T1B, T1m;
Chris@42 196 {
Chris@42 197 V T1h, T1k, T41, T14, T3W, T10, T1g, T1j;
Chris@42 198 T1h = LD(&(ri[WS(rs, 30)]), ms, &(ri[0]));
Chris@42 199 T1k = LD(&(ii[WS(rs, 30)]), ms, &(ii[0]));
Chris@42 200 T41 = VMUL(T12, T16);
Chris@42 201 T14 = VMUL(T12, T13);
Chris@42 202 T3W = VFNMS(TY, TW, T3V);
Chris@42 203 T10 = VFMA(TY, TZ, TX);
Chris@42 204 T1g = LDW(&(W[TWVL * 58]));
Chris@42 205 T1j = LDW(&(W[TWVL * 59]));
Chris@42 206 {
Chris@42 207 V T6J, T3X, T11, T40, T48, T1i, T6K, T45, T1e, T3Y, T1z, T42, T17;
Chris@42 208 T1A = LD(&(ri[WS(rs, 22)]), ms, &(ri[0]));
Chris@42 209 T1D = LD(&(ii[WS(rs, 22)]), ms, &(ii[0]));
Chris@42 210 T42 = VFNMS(T15, T13, T41);
Chris@42 211 T17 = VFMA(T15, T16, T14);
Chris@42 212 T6J = VADD(T3U, T3W);
Chris@42 213 T3X = VSUB(T3U, T3W);
Chris@42 214 T11 = VADD(TU, T10);
Chris@42 215 T40 = VSUB(TU, T10);
Chris@42 216 T48 = VMUL(T1g, T1k);
Chris@42 217 T1i = VMUL(T1g, T1h);
Chris@42 218 T6K = VADD(T42, T44);
Chris@42 219 T45 = VSUB(T42, T44);
Chris@42 220 T1e = VADD(T17, T1d);
Chris@42 221 T3Y = VSUB(T17, T1d);
Chris@42 222 T1z = LDW(&(W[TWVL * 42]));
Chris@42 223 T1C = LDW(&(W[TWVL * 43]));
Chris@42 224 T49 = VFNMS(T1j, T1h, T48);
Chris@42 225 T1l = VFMA(T1j, T1k, T1i);
Chris@42 226 T46 = VADD(T40, T45);
Chris@42 227 T5Y = VSUB(T40, T45);
Chris@42 228 T7D = VADD(T6J, T6K);
Chris@42 229 T6L = VSUB(T6J, T6K);
Chris@42 230 T5X = VADD(T3X, T3Y);
Chris@42 231 T3Z = VSUB(T3X, T3Y);
Chris@42 232 T6M = VSUB(T11, T1e);
Chris@42 233 T1f = VADD(T11, T1e);
Chris@42 234 T4i = VMUL(T1z, T1D);
Chris@42 235 T1B = VMUL(T1z, T1A);
Chris@42 236 }
Chris@42 237 }
Chris@42 238 T1n = LD(&(ri[WS(rs, 14)]), ms, &(ri[0]));
Chris@42 239 T1q = LD(&(ii[WS(rs, 14)]), ms, &(ii[0]));
Chris@42 240 T1m = LDW(&(W[TWVL * 26]));
Chris@42 241 T1p = LDW(&(W[TWVL * 27]));
Chris@42 242 T1u = LD(&(ri[WS(rs, 6)]), ms, &(ri[0]));
Chris@42 243 T1x = LD(&(ii[WS(rs, 6)]), ms, &(ii[0]));
Chris@42 244 T4j = VFNMS(T1C, T1A, T4i);
Chris@42 245 T1E = VFMA(T1C, T1D, T1B);
Chris@42 246 T1t = LDW(&(W[TWVL * 10]));
Chris@42 247 T1w = LDW(&(W[TWVL * 11]));
Chris@42 248 T4a = VMUL(T1m, T1q);
Chris@42 249 T1o = VMUL(T1m, T1n);
Chris@42 250 }
Chris@42 251 {
Chris@42 252 V T2W, T2Z, T6P, T4c, T1s, T4f, T6Q, T4k, T1F, T4d, T2V, T2Y, T5o, T2X, T2I;
Chris@42 253 {
Chris@42 254 V T2D, T2G, T2C, T2F, T4g, T1v, T4b, T1r;
Chris@42 255 T2D = LD(&(ri[WS(rs, 31)]), ms, &(ri[WS(rs, 1)]));
Chris@42 256 T2G = LD(&(ii[WS(rs, 31)]), ms, &(ii[WS(rs, 1)]));
Chris@42 257 T2C = LDW(&(W[TWVL * 60]));
Chris@42 258 T2F = LDW(&(W[TWVL * 61]));
Chris@42 259 T4g = VMUL(T1t, T1x);
Chris@42 260 T1v = VMUL(T1t, T1u);
Chris@42 261 T4b = VFNMS(T1p, T1n, T4a);
Chris@42 262 T1r = VFMA(T1p, T1q, T1o);
Chris@42 263 T2W = LD(&(ri[WS(rs, 23)]), ms, &(ri[WS(rs, 1)]));
Chris@42 264 T2Z = LD(&(ii[WS(rs, 23)]), ms, &(ii[WS(rs, 1)]));
Chris@42 265 {
Chris@42 266 V T4Y, T2E, T4h, T1y;
Chris@42 267 T4Y = VMUL(T2C, T2G);
Chris@42 268 T2E = VMUL(T2C, T2D);
Chris@42 269 T4h = VFNMS(T1w, T1u, T4g);
Chris@42 270 T1y = VFMA(T1w, T1x, T1v);
Chris@42 271 T6P = VADD(T49, T4b);
Chris@42 272 T4c = VSUB(T49, T4b);
Chris@42 273 T1s = VADD(T1l, T1r);
Chris@42 274 T4f = VSUB(T1l, T1r);
Chris@42 275 T4Z = VFNMS(T2F, T2D, T4Y);
Chris@42 276 T2H = VFMA(T2F, T2G, T2E);
Chris@42 277 T6Q = VADD(T4h, T4j);
Chris@42 278 T4k = VSUB(T4h, T4j);
Chris@42 279 T1F = VADD(T1y, T1E);
Chris@42 280 T4d = VSUB(T1y, T1E);
Chris@42 281 T2V = LDW(&(W[TWVL * 44]));
Chris@42 282 }
Chris@42 283 T2Y = LDW(&(W[TWVL * 45]));
Chris@42 284 }
Chris@42 285 T2J = LD(&(ri[WS(rs, 15)]), ms, &(ri[WS(rs, 1)]));
Chris@42 286 T2M = LD(&(ii[WS(rs, 15)]), ms, &(ii[WS(rs, 1)]));
Chris@42 287 T4l = VADD(T4f, T4k);
Chris@42 288 T61 = VSUB(T4f, T4k);
Chris@42 289 T7E = VADD(T6P, T6Q);
Chris@42 290 T6R = VSUB(T6P, T6Q);
Chris@42 291 T60 = VADD(T4c, T4d);
Chris@42 292 T4e = VSUB(T4c, T4d);
Chris@42 293 T6O = VSUB(T1s, T1F);
Chris@42 294 T1G = VADD(T1s, T1F);
Chris@42 295 T5o = VMUL(T2V, T2Z);
Chris@42 296 T2X = VMUL(T2V, T2W);
Chris@42 297 T2I = LDW(&(W[TWVL * 28]));
Chris@42 298 T2L = LDW(&(W[TWVL * 29]));
Chris@42 299 T2Q = LD(&(ri[WS(rs, 7)]), ms, &(ri[WS(rs, 1)]));
Chris@42 300 T2T = LD(&(ii[WS(rs, 7)]), ms, &(ii[WS(rs, 1)]));
Chris@42 301 T2P = LDW(&(W[TWVL * 12]));
Chris@42 302 T2S = LDW(&(W[TWVL * 13]));
Chris@42 303 T5p = VFNMS(T2Y, T2W, T5o);
Chris@42 304 T30 = VFMA(T2Y, T2Z, T2X);
Chris@42 305 T50 = VMUL(T2I, T2M);
Chris@42 306 T2K = VMUL(T2I, T2J);
Chris@42 307 }
Chris@42 308 }
Chris@42 309 {
Chris@42 310 V T4q, T1O, T1Q, T1T, T1S, T1X, T20, T4Q, T27, T1W, T1Z, T4r, T1R;
Chris@42 311 {
Chris@42 312 V T23, T26, T25, T4P, T24, T1P;
Chris@42 313 {
Chris@42 314 V T1K, T1N, T5m, T2R, T1J, T1M, T51, T2N;
Chris@42 315 T1K = LD(&(ri[WS(rs, 1)]), ms, &(ri[WS(rs, 1)]));
Chris@42 316 T1N = LD(&(ii[WS(rs, 1)]), ms, &(ii[WS(rs, 1)]));
Chris@42 317 T5m = VMUL(T2P, T2T);
Chris@42 318 T2R = VMUL(T2P, T2Q);
Chris@42 319 T1J = LDW(&(W[0]));
Chris@42 320 T1M = LDW(&(W[TWVL * 1]));
Chris@42 321 T51 = VFNMS(T2L, T2J, T50);
Chris@42 322 T2N = VFMA(T2L, T2M, T2K);
Chris@42 323 {
Chris@42 324 V T76, T52, T2O, T5l, T77, T5q, T31, T53, T22;
Chris@42 325 T23 = LD(&(ri[WS(rs, 25)]), ms, &(ri[WS(rs, 1)]));
Chris@42 326 T26 = LD(&(ii[WS(rs, 25)]), ms, &(ii[WS(rs, 1)]));
Chris@42 327 {
Chris@42 328 V T5n, T2U, T4p, T1L;
Chris@42 329 T5n = VFNMS(T2S, T2Q, T5m);
Chris@42 330 T2U = VFMA(T2S, T2T, T2R);
Chris@42 331 T4p = VMUL(T1J, T1N);
Chris@42 332 T1L = VMUL(T1J, T1K);
Chris@42 333 T76 = VADD(T4Z, T51);
Chris@42 334 T52 = VSUB(T4Z, T51);
Chris@42 335 T2O = VADD(T2H, T2N);
Chris@42 336 T5l = VSUB(T2H, T2N);
Chris@42 337 T77 = VADD(T5n, T5p);
Chris@42 338 T5q = VSUB(T5n, T5p);
Chris@42 339 T31 = VADD(T2U, T30);
Chris@42 340 T53 = VSUB(T2U, T30);
Chris@42 341 T4q = VFNMS(T1M, T1K, T4p);
Chris@42 342 T1O = VFMA(T1M, T1N, T1L);
Chris@42 343 T22 = LDW(&(W[TWVL * 48]));
Chris@42 344 }
Chris@42 345 T25 = LDW(&(W[TWVL * 49]));
Chris@42 346 T5r = VADD(T5l, T5q);
Chris@42 347 T6c = VSUB(T5l, T5q);
Chris@42 348 T78 = VSUB(T76, T77);
Chris@42 349 T7N = VADD(T76, T77);
Chris@42 350 T54 = VSUB(T52, T53);
Chris@42 351 T6f = VADD(T52, T53);
Chris@42 352 T32 = VADD(T2O, T31);
Chris@42 353 T7b = VSUB(T2O, T31);
Chris@42 354 T4P = VMUL(T22, T26);
Chris@42 355 T24 = VMUL(T22, T23);
Chris@42 356 }
Chris@42 357 }
Chris@42 358 T1Q = LD(&(ri[WS(rs, 17)]), ms, &(ri[WS(rs, 1)]));
Chris@42 359 T1T = LD(&(ii[WS(rs, 17)]), ms, &(ii[WS(rs, 1)]));
Chris@42 360 T1P = LDW(&(W[TWVL * 32]));
Chris@42 361 T1S = LDW(&(W[TWVL * 33]));
Chris@42 362 T1X = LD(&(ri[WS(rs, 9)]), ms, &(ri[WS(rs, 1)]));
Chris@42 363 T20 = LD(&(ii[WS(rs, 9)]), ms, &(ii[WS(rs, 1)]));
Chris@42 364 T4Q = VFNMS(T25, T23, T4P);
Chris@42 365 T27 = VFMA(T25, T26, T24);
Chris@42 366 T1W = LDW(&(W[TWVL * 16]));
Chris@42 367 T1Z = LDW(&(W[TWVL * 17]));
Chris@42 368 T4r = VMUL(T1P, T1T);
Chris@42 369 T1R = VMUL(T1P, T1Q);
Chris@42 370 }
Chris@42 371 {
Chris@42 372 V T56, T38, T3a, T3d, T3c, T3h, T3k, T3g, T3j, T5g, T3r, T57, T3b;
Chris@42 373 {
Chris@42 374 V T3n, T3q, T6V, T4t, T1V, T4M, T6W, T4R, T28, T4u, T3m, T3p, T5f, T3o, T39;
Chris@42 375 {
Chris@42 376 V T34, T37, T33, T36, T4N, T1Y, T4s, T1U;
Chris@42 377 T34 = LD(&(ri[WS(rs, 3)]), ms, &(ri[WS(rs, 1)]));
Chris@42 378 T37 = LD(&(ii[WS(rs, 3)]), ms, &(ii[WS(rs, 1)]));
Chris@42 379 T33 = LDW(&(W[TWVL * 4]));
Chris@42 380 T36 = LDW(&(W[TWVL * 5]));
Chris@42 381 T4N = VMUL(T1W, T20);
Chris@42 382 T1Y = VMUL(T1W, T1X);
Chris@42 383 T4s = VFNMS(T1S, T1Q, T4r);
Chris@42 384 T1U = VFMA(T1S, T1T, T1R);
Chris@42 385 T3n = LD(&(ri[WS(rs, 11)]), ms, &(ri[WS(rs, 1)]));
Chris@42 386 T3q = LD(&(ii[WS(rs, 11)]), ms, &(ii[WS(rs, 1)]));
Chris@42 387 {
Chris@42 388 V T55, T35, T4O, T21;
Chris@42 389 T55 = VMUL(T33, T37);
Chris@42 390 T35 = VMUL(T33, T34);
Chris@42 391 T4O = VFNMS(T1Z, T1X, T4N);
Chris@42 392 T21 = VFMA(T1Z, T20, T1Y);
Chris@42 393 T6V = VADD(T4q, T4s);
Chris@42 394 T4t = VSUB(T4q, T4s);
Chris@42 395 T1V = VADD(T1O, T1U);
Chris@42 396 T4M = VSUB(T1O, T1U);
Chris@42 397 T56 = VFNMS(T36, T34, T55);
Chris@42 398 T38 = VFMA(T36, T37, T35);
Chris@42 399 T6W = VADD(T4O, T4Q);
Chris@42 400 T4R = VSUB(T4O, T4Q);
Chris@42 401 T28 = VADD(T21, T27);
Chris@42 402 T4u = VSUB(T21, T27);
Chris@42 403 T3m = LDW(&(W[TWVL * 20]));
Chris@42 404 }
Chris@42 405 T3p = LDW(&(W[TWVL * 21]));
Chris@42 406 }
Chris@42 407 T3a = LD(&(ri[WS(rs, 19)]), ms, &(ri[WS(rs, 1)]));
Chris@42 408 T3d = LD(&(ii[WS(rs, 19)]), ms, &(ii[WS(rs, 1)]));
Chris@42 409 T4S = VADD(T4M, T4R);
Chris@42 410 T65 = VSUB(T4M, T4R);
Chris@42 411 T6X = VSUB(T6V, T6W);
Chris@42 412 T7I = VADD(T6V, T6W);
Chris@42 413 T4v = VSUB(T4t, T4u);
Chris@42 414 T68 = VADD(T4t, T4u);
Chris@42 415 T29 = VADD(T1V, T28);
Chris@42 416 T70 = VSUB(T1V, T28);
Chris@42 417 T5f = VMUL(T3m, T3q);
Chris@42 418 T3o = VMUL(T3m, T3n);
Chris@42 419 T39 = LDW(&(W[TWVL * 36]));
Chris@42 420 T3c = LDW(&(W[TWVL * 37]));
Chris@42 421 T3h = LD(&(ri[WS(rs, 27)]), ms, &(ri[WS(rs, 1)]));
Chris@42 422 T3k = LD(&(ii[WS(rs, 27)]), ms, &(ii[WS(rs, 1)]));
Chris@42 423 T3g = LDW(&(W[TWVL * 52]));
Chris@42 424 T3j = LDW(&(W[TWVL * 53]));
Chris@42 425 T5g = VFNMS(T3p, T3n, T5f);
Chris@42 426 T3r = VFMA(T3p, T3q, T3o);
Chris@42 427 T57 = VMUL(T39, T3d);
Chris@42 428 T3b = VMUL(T39, T3a);
Chris@42 429 }
Chris@42 430 {
Chris@42 431 V T2u, T2x, T2w, T4G, T2v, T2g;
Chris@42 432 {
Chris@42 433 V T2b, T2e, T5d, T3i, T2a, T2d, T58, T3e, T2t;
Chris@42 434 T2b = LD(&(ri[WS(rs, 5)]), ms, &(ri[WS(rs, 1)]));
Chris@42 435 T2e = LD(&(ii[WS(rs, 5)]), ms, &(ii[WS(rs, 1)]));
Chris@42 436 T5d = VMUL(T3g, T3k);
Chris@42 437 T3i = VMUL(T3g, T3h);
Chris@42 438 T2a = LDW(&(W[TWVL * 8]));
Chris@42 439 T2d = LDW(&(W[TWVL * 9]));
Chris@42 440 T58 = VFNMS(T3c, T3a, T57);
Chris@42 441 T3e = VFMA(T3c, T3d, T3b);
Chris@42 442 T2u = LD(&(ri[WS(rs, 13)]), ms, &(ri[WS(rs, 1)]));
Chris@42 443 T2x = LD(&(ii[WS(rs, 13)]), ms, &(ii[WS(rs, 1)]));
Chris@42 444 {
Chris@42 445 V T5e, T3l, T4w, T2c;
Chris@42 446 T5e = VFNMS(T3j, T3h, T5d);
Chris@42 447 T3l = VFMA(T3j, T3k, T3i);
Chris@42 448 T4w = VMUL(T2a, T2e);
Chris@42 449 T2c = VMUL(T2a, T2b);
Chris@42 450 {
Chris@42 451 V T7c, T59, T3f, T5a;
Chris@42 452 T7c = VADD(T56, T58);
Chris@42 453 T59 = VSUB(T56, T58);
Chris@42 454 T3f = VADD(T38, T3e);
Chris@42 455 T5a = VSUB(T38, T3e);
Chris@42 456 {
Chris@42 457 V T7d, T5h, T3s, T5c;
Chris@42 458 T7d = VADD(T5e, T5g);
Chris@42 459 T5h = VSUB(T5e, T5g);
Chris@42 460 T3s = VADD(T3l, T3r);
Chris@42 461 T5c = VSUB(T3l, T3r);
Chris@42 462 T4x = VFNMS(T2d, T2b, T4w);
Chris@42 463 T2f = VFMA(T2d, T2e, T2c);
Chris@42 464 T5b = VSUB(T59, T5a);
Chris@42 465 T5s = VADD(T5a, T59);
Chris@42 466 T2t = LDW(&(W[TWVL * 24]));
Chris@42 467 T7O = VADD(T7c, T7d);
Chris@42 468 T7e = VSUB(T7c, T7d);
Chris@42 469 T5t = VSUB(T5c, T5h);
Chris@42 470 T5i = VADD(T5c, T5h);
Chris@42 471 T79 = VSUB(T3s, T3f);
Chris@42 472 T3t = VADD(T3f, T3s);
Chris@42 473 }
Chris@42 474 }
Chris@42 475 }
Chris@42 476 T2w = LDW(&(W[TWVL * 25]));
Chris@42 477 T4G = VMUL(T2t, T2x);
Chris@42 478 T2v = VMUL(T2t, T2u);
Chris@42 479 }
Chris@42 480 T2h = LD(&(ri[WS(rs, 21)]), ms, &(ri[WS(rs, 1)]));
Chris@42 481 T2k = LD(&(ii[WS(rs, 21)]), ms, &(ii[WS(rs, 1)]));
Chris@42 482 T2g = LDW(&(W[TWVL * 40]));
Chris@42 483 T2j = LDW(&(W[TWVL * 41]));
Chris@42 484 T2o = LD(&(ri[WS(rs, 29)]), ms, &(ri[WS(rs, 1)]));
Chris@42 485 T2r = LD(&(ii[WS(rs, 29)]), ms, &(ii[WS(rs, 1)]));
Chris@42 486 T4H = VFNMS(T2w, T2u, T4G);
Chris@42 487 T2y = VFMA(T2w, T2x, T2v);
Chris@42 488 T2n = LDW(&(W[TWVL * 56]));
Chris@42 489 T2q = LDW(&(W[TWVL * 57]));
Chris@42 490 T4y = VMUL(T2g, T2k);
Chris@42 491 T2i = VMUL(T2g, T2h);
Chris@42 492 }
Chris@42 493 }
Chris@42 494 }
Chris@42 495 }
Chris@42 496 }
Chris@42 497 {
Chris@42 498 V T4C, T4T, T4U, T4J, T7A, T7w, T7j, T75, T7i, T6U, T8p, T8n, T8v, T8t, T7q;
Chris@42 499 V T7y, T7t, T7z, T7g, T7k;
Chris@42 500 {
Chris@42 501 V T6E, T8j, T6H, T8k, T73, T6Y, T7S, T8i, T8h, T7V;
Chris@42 502 {
Chris@42 503 V T7P, T7Y, T7C, TO, T89, T8e, T3u, T7M, T8d, T1H, T7K, T7X, T2B, T7H;
Chris@42 504 {
Chris@42 505 V T71, T2m, T72, T4I, T2z, T4D, Tm, TN, T2A, T7J;
Chris@42 506 T6E = VSUB(T8, Tl);
Chris@42 507 Tm = VADD(T8, Tl);
Chris@42 508 TN = VADD(Tz, TM);
Chris@42 509 T8j = VSUB(TM, Tz);
Chris@42 510 T7P = VSUB(T7N, T7O);
Chris@42 511 T7Y = VADD(T7N, T7O);
Chris@42 512 {
Chris@42 513 V T82, T4E, T2p, T4z, T2l, T88;
Chris@42 514 T82 = VADD(T6F, T6G);
Chris@42 515 T6H = VSUB(T6F, T6G);
Chris@42 516 T4E = VMUL(T2n, T2r);
Chris@42 517 T2p = VMUL(T2n, T2o);
Chris@42 518 T4z = VFNMS(T2j, T2h, T4y);
Chris@42 519 T2l = VFMA(T2j, T2k, T2i);
Chris@42 520 T8k = VSUB(T87, T83);
Chris@42 521 T88 = VADD(T83, T87);
Chris@42 522 T7C = VSUB(Tm, TN);
Chris@42 523 TO = VADD(Tm, TN);
Chris@42 524 {
Chris@42 525 V T4F, T2s, T4A, T4B;
Chris@42 526 T4F = VFNMS(T2q, T2o, T4E);
Chris@42 527 T2s = VFMA(T2q, T2r, T2p);
Chris@42 528 T71 = VADD(T4x, T4z);
Chris@42 529 T4A = VSUB(T4x, T4z);
Chris@42 530 T2m = VADD(T2f, T2l);
Chris@42 531 T4B = VSUB(T2f, T2l);
Chris@42 532 T89 = VADD(T82, T88);
Chris@42 533 T8e = VSUB(T88, T82);
Chris@42 534 T72 = VADD(T4F, T4H);
Chris@42 535 T4I = VSUB(T4F, T4H);
Chris@42 536 T2z = VADD(T2s, T2y);
Chris@42 537 T4D = VSUB(T2s, T2y);
Chris@42 538 T4C = VSUB(T4A, T4B);
Chris@42 539 T4T = VADD(T4B, T4A);
Chris@42 540 }
Chris@42 541 }
Chris@42 542 T3u = VADD(T32, T3t);
Chris@42 543 T7M = VSUB(T32, T3t);
Chris@42 544 T7J = VADD(T71, T72);
Chris@42 545 T73 = VSUB(T71, T72);
Chris@42 546 T4U = VSUB(T4D, T4I);
Chris@42 547 T4J = VADD(T4D, T4I);
Chris@42 548 T6Y = VSUB(T2z, T2m);
Chris@42 549 T2A = VADD(T2m, T2z);
Chris@42 550 T8d = VSUB(T1G, T1f);
Chris@42 551 T1H = VADD(T1f, T1G);
Chris@42 552 T7K = VSUB(T7I, T7J);
Chris@42 553 T7X = VADD(T7I, T7J);
Chris@42 554 T2B = VADD(T29, T2A);
Chris@42 555 T7H = VSUB(T29, T2A);
Chris@42 556 }
Chris@42 557 {
Chris@42 558 V T1I, T80, T7Q, T7U, T7F, T7L, T7T, T3v, T8b, T8c, T8a, T7W, T81, T7Z;
Chris@42 559 T7W = VSUB(TO, T1H);
Chris@42 560 T1I = VADD(TO, T1H);
Chris@42 561 T7Z = VSUB(T7X, T7Y);
Chris@42 562 T80 = VADD(T7X, T7Y);
Chris@42 563 T7Q = VSUB(T7M, T7P);
Chris@42 564 T7U = VADD(T7M, T7P);
Chris@42 565 T7F = VSUB(T7D, T7E);
Chris@42 566 T81 = VADD(T7D, T7E);
Chris@42 567 T7L = VADD(T7H, T7K);
Chris@42 568 T7T = VSUB(T7K, T7H);
Chris@42 569 T3v = VADD(T2B, T3u);
Chris@42 570 T8b = VSUB(T3u, T2B);
Chris@42 571 ST(&(ri[WS(rs, 24)]), VSUB(T7W, T7Z), ms, &(ri[0]));
Chris@42 572 ST(&(ri[WS(rs, 8)]), VADD(T7W, T7Z), ms, &(ri[0]));
Chris@42 573 T8c = VSUB(T89, T81);
Chris@42 574 T8a = VADD(T81, T89);
Chris@42 575 {
Chris@42 576 V T8f, T8g, T7G, T7R;
Chris@42 577 T7S = VSUB(T7C, T7F);
Chris@42 578 T7G = VADD(T7C, T7F);
Chris@42 579 T7R = VADD(T7L, T7Q);
Chris@42 580 T8i = VSUB(T7Q, T7L);
Chris@42 581 T8h = VSUB(T8e, T8d);
Chris@42 582 T8f = VADD(T8d, T8e);
Chris@42 583 ST(&(ri[0]), VADD(T1I, T3v), ms, &(ri[0]));
Chris@42 584 ST(&(ri[WS(rs, 16)]), VSUB(T1I, T3v), ms, &(ri[0]));
Chris@42 585 T8g = VADD(T7T, T7U);
Chris@42 586 T7V = VSUB(T7T, T7U);
Chris@42 587 ST(&(ii[WS(rs, 16)]), VSUB(T8a, T80), ms, &(ii[0]));
Chris@42 588 ST(&(ii[0]), VADD(T80, T8a), ms, &(ii[0]));
Chris@42 589 ST(&(ii[WS(rs, 24)]), VSUB(T8c, T8b), ms, &(ii[0]));
Chris@42 590 ST(&(ii[WS(rs, 8)]), VADD(T8b, T8c), ms, &(ii[0]));
Chris@42 591 ST(&(ri[WS(rs, 4)]), VFMA(LDK(KP707106781), T7R, T7G), ms, &(ri[0]));
Chris@42 592 ST(&(ri[WS(rs, 20)]), VFNMS(LDK(KP707106781), T7R, T7G), ms, &(ri[0]));
Chris@42 593 ST(&(ii[WS(rs, 20)]), VFNMS(LDK(KP707106781), T8g, T8f), ms, &(ii[0]));
Chris@42 594 ST(&(ii[WS(rs, 4)]), VFMA(LDK(KP707106781), T8g, T8f), ms, &(ii[0]));
Chris@42 595 }
Chris@42 596 }
Chris@42 597 }
Chris@42 598 {
Chris@42 599 V T7f, T7a, T7m, T6I, T7s, T7r, T8r, T8l, T8m, T6T, T8s, T7p;
Chris@42 600 {
Chris@42 601 V T7n, T6N, T6S, T7o, T7u, T7v, T6Z, T74;
Chris@42 602 T7f = VSUB(T7b, T7e);
Chris@42 603 T7u = VADD(T7b, T7e);
Chris@42 604 T7v = VADD(T78, T79);
Chris@42 605 T7a = VSUB(T78, T79);
Chris@42 606 ST(&(ri[WS(rs, 12)]), VFMA(LDK(KP707106781), T7V, T7S), ms, &(ri[0]));
Chris@42 607 ST(&(ri[WS(rs, 28)]), VFNMS(LDK(KP707106781), T7V, T7S), ms, &(ri[0]));
Chris@42 608 ST(&(ii[WS(rs, 28)]), VFNMS(LDK(KP707106781), T8i, T8h), ms, &(ii[0]));
Chris@42 609 ST(&(ii[WS(rs, 12)]), VFMA(LDK(KP707106781), T8i, T8h), ms, &(ii[0]));
Chris@42 610 T7m = VADD(T6E, T6H);
Chris@42 611 T6I = VSUB(T6E, T6H);
Chris@42 612 T7A = VFMA(LDK(KP414213562), T7u, T7v);
Chris@42 613 T7w = VFNMS(LDK(KP414213562), T7v, T7u);
Chris@42 614 T7n = VADD(T6M, T6L);
Chris@42 615 T6N = VSUB(T6L, T6M);
Chris@42 616 T6S = VADD(T6O, T6R);
Chris@42 617 T7o = VSUB(T6O, T6R);
Chris@42 618 T7s = VADD(T6X, T6Y);
Chris@42 619 T6Z = VSUB(T6X, T6Y);
Chris@42 620 T74 = VSUB(T70, T73);
Chris@42 621 T7r = VADD(T70, T73);
Chris@42 622 T8r = VSUB(T8k, T8j);
Chris@42 623 T8l = VADD(T8j, T8k);
Chris@42 624 T8m = VADD(T6N, T6S);
Chris@42 625 T6T = VSUB(T6N, T6S);
Chris@42 626 T7j = VFNMS(LDK(KP414213562), T6Z, T74);
Chris@42 627 T75 = VFMA(LDK(KP414213562), T74, T6Z);
Chris@42 628 T8s = VSUB(T7o, T7n);
Chris@42 629 T7p = VADD(T7n, T7o);
Chris@42 630 }
Chris@42 631 T7i = VFNMS(LDK(KP707106781), T6T, T6I);
Chris@42 632 T6U = VFMA(LDK(KP707106781), T6T, T6I);
Chris@42 633 T8p = VFNMS(LDK(KP707106781), T8m, T8l);
Chris@42 634 T8n = VFMA(LDK(KP707106781), T8m, T8l);
Chris@42 635 T8v = VFNMS(LDK(KP707106781), T8s, T8r);
Chris@42 636 T8t = VFMA(LDK(KP707106781), T8s, T8r);
Chris@42 637 T7q = VFMA(LDK(KP707106781), T7p, T7m);
Chris@42 638 T7y = VFNMS(LDK(KP707106781), T7p, T7m);
Chris@42 639 T7t = VFMA(LDK(KP414213562), T7s, T7r);
Chris@42 640 T7z = VFNMS(LDK(KP414213562), T7r, T7s);
Chris@42 641 T7g = VFNMS(LDK(KP414213562), T7f, T7a);
Chris@42 642 T7k = VFMA(LDK(KP414213562), T7a, T7f);
Chris@42 643 }
Chris@42 644 }
Chris@42 645 {
Chris@42 646 V T5S, T8O, T8N, T5V, T6d, T6g, T66, T4L, T5I, T69, T5y, T4o, T8J, T8L, T5M;
Chris@42 647 V T5Q, T5A, T5w, T5H, T4W, T5O, T5G, T8D, T8F;
Chris@42 648 {
Chris@42 649 V T5C, T3S, T8C, T4n, T8H, T8B, T8I, T5F, T5L, T5k, T5K, T5v, T4V;
Chris@42 650 {
Chris@42 651 V T5D, T47, T4m, T5E, T8z, T8A, T3C, T3R, T5j, T5u, T4K;
Chris@42 652 T5S = VSUB(T3w, T3B);
Chris@42 653 T3C = VADD(T3w, T3B);
Chris@42 654 T3R = VADD(T3J, T3Q);
Chris@42 655 T8O = VSUB(T3Q, T3J);
Chris@42 656 {
Chris@42 657 V T8o, T7B, T7x, T8q;
Chris@42 658 T8o = VADD(T7z, T7A);
Chris@42 659 T7B = VSUB(T7z, T7A);
Chris@42 660 T7x = VADD(T7t, T7w);
Chris@42 661 T8q = VSUB(T7w, T7t);
Chris@42 662 {
Chris@42 663 V T8u, T7l, T7h, T8w;
Chris@42 664 T8u = VSUB(T7k, T7j);
Chris@42 665 T7l = VADD(T7j, T7k);
Chris@42 666 T7h = VSUB(T75, T7g);
Chris@42 667 T8w = VADD(T75, T7g);
Chris@42 668 ST(&(ri[WS(rs, 10)]), VFMA(LDK(KP923879532), T7B, T7y), ms, &(ri[0]));
Chris@42 669 ST(&(ri[WS(rs, 26)]), VFNMS(LDK(KP923879532), T7B, T7y), ms, &(ri[0]));
Chris@42 670 ST(&(ii[WS(rs, 18)]), VFNMS(LDK(KP923879532), T8o, T8n), ms, &(ii[0]));
Chris@42 671 ST(&(ii[WS(rs, 2)]), VFMA(LDK(KP923879532), T8o, T8n), ms, &(ii[0]));
Chris@42 672 ST(&(ii[WS(rs, 26)]), VFNMS(LDK(KP923879532), T8q, T8p), ms, &(ii[0]));
Chris@42 673 ST(&(ii[WS(rs, 10)]), VFMA(LDK(KP923879532), T8q, T8p), ms, &(ii[0]));
Chris@42 674 ST(&(ri[WS(rs, 2)]), VFMA(LDK(KP923879532), T7x, T7q), ms, &(ri[0]));
Chris@42 675 ST(&(ri[WS(rs, 18)]), VFNMS(LDK(KP923879532), T7x, T7q), ms, &(ri[0]));
Chris@42 676 ST(&(ri[WS(rs, 30)]), VFMA(LDK(KP923879532), T7l, T7i), ms, &(ri[0]));
Chris@42 677 ST(&(ri[WS(rs, 14)]), VFNMS(LDK(KP923879532), T7l, T7i), ms, &(ri[0]));
Chris@42 678 ST(&(ii[WS(rs, 22)]), VFNMS(LDK(KP923879532), T8u, T8t), ms, &(ii[0]));
Chris@42 679 ST(&(ii[WS(rs, 6)]), VFMA(LDK(KP923879532), T8u, T8t), ms, &(ii[0]));
Chris@42 680 ST(&(ii[WS(rs, 30)]), VFMA(LDK(KP923879532), T8w, T8v), ms, &(ii[0]));
Chris@42 681 ST(&(ii[WS(rs, 14)]), VFNMS(LDK(KP923879532), T8w, T8v), ms, &(ii[0]));
Chris@42 682 ST(&(ri[WS(rs, 6)]), VFMA(LDK(KP923879532), T7h, T6U), ms, &(ri[0]));
Chris@42 683 ST(&(ri[WS(rs, 22)]), VFNMS(LDK(KP923879532), T7h, T6U), ms, &(ri[0]));
Chris@42 684 T5C = VFMA(LDK(KP707106781), T3R, T3C);
Chris@42 685 T3S = VFNMS(LDK(KP707106781), T3R, T3C);
Chris@42 686 }
Chris@42 687 }
Chris@42 688 T5D = VFMA(LDK(KP414213562), T3Z, T46);
Chris@42 689 T47 = VFNMS(LDK(KP414213562), T46, T3Z);
Chris@42 690 T4m = VFMA(LDK(KP414213562), T4l, T4e);
Chris@42 691 T5E = VFNMS(LDK(KP414213562), T4e, T4l);
Chris@42 692 T8N = VADD(T8y, T8x);
Chris@42 693 T8z = VSUB(T8x, T8y);
Chris@42 694 T8A = VADD(T5T, T5U);
Chris@42 695 T5V = VSUB(T5T, T5U);
Chris@42 696 T6d = VSUB(T5i, T5b);
Chris@42 697 T5j = VADD(T5b, T5i);
Chris@42 698 T5u = VADD(T5s, T5t);
Chris@42 699 T6g = VSUB(T5s, T5t);
Chris@42 700 T66 = VSUB(T4J, T4C);
Chris@42 701 T4K = VADD(T4C, T4J);
Chris@42 702 T8C = VADD(T47, T4m);
Chris@42 703 T4n = VSUB(T47, T4m);
Chris@42 704 T8H = VFNMS(LDK(KP707106781), T8A, T8z);
Chris@42 705 T8B = VFMA(LDK(KP707106781), T8A, T8z);
Chris@42 706 T8I = VSUB(T5E, T5D);
Chris@42 707 T5F = VADD(T5D, T5E);
Chris@42 708 T5L = VFMA(LDK(KP707106781), T5j, T54);
Chris@42 709 T5k = VFNMS(LDK(KP707106781), T5j, T54);
Chris@42 710 T5K = VFMA(LDK(KP707106781), T5u, T5r);
Chris@42 711 T5v = VFNMS(LDK(KP707106781), T5u, T5r);
Chris@42 712 T4L = VFNMS(LDK(KP707106781), T4K, T4v);
Chris@42 713 T5I = VFMA(LDK(KP707106781), T4K, T4v);
Chris@42 714 T4V = VADD(T4T, T4U);
Chris@42 715 T69 = VSUB(T4T, T4U);
Chris@42 716 }
Chris@42 717 T5y = VFNMS(LDK(KP923879532), T4n, T3S);
Chris@42 718 T4o = VFMA(LDK(KP923879532), T4n, T3S);
Chris@42 719 T8J = VFMA(LDK(KP923879532), T8I, T8H);
Chris@42 720 T8L = VFNMS(LDK(KP923879532), T8I, T8H);
Chris@42 721 T5M = VFNMS(LDK(KP198912367), T5L, T5K);
Chris@42 722 T5Q = VFMA(LDK(KP198912367), T5K, T5L);
Chris@42 723 T5A = VFMA(LDK(KP668178637), T5k, T5v);
Chris@42 724 T5w = VFNMS(LDK(KP668178637), T5v, T5k);
Chris@42 725 T5H = VFMA(LDK(KP707106781), T4V, T4S);
Chris@42 726 T4W = VFNMS(LDK(KP707106781), T4V, T4S);
Chris@42 727 T5O = VFNMS(LDK(KP923879532), T5F, T5C);
Chris@42 728 T5G = VFMA(LDK(KP923879532), T5F, T5C);
Chris@42 729 T8D = VFMA(LDK(KP923879532), T8C, T8B);
Chris@42 730 T8F = VFNMS(LDK(KP923879532), T8C, T8B);
Chris@42 731 }
Chris@42 732 {
Chris@42 733 V T6p, T6q, T6o, T5W, T8W, T63;
Chris@42 734 {
Chris@42 735 V T5J, T5P, T5z, T4X, T5Z, T62;
Chris@42 736 T5J = VFMA(LDK(KP198912367), T5I, T5H);
Chris@42 737 T5P = VFNMS(LDK(KP198912367), T5H, T5I);
Chris@42 738 T5z = VFNMS(LDK(KP668178637), T4L, T4W);
Chris@42 739 T4X = VFMA(LDK(KP668178637), T4W, T4L);
Chris@42 740 T6p = VFNMS(LDK(KP414213562), T5X, T5Y);
Chris@42 741 T5Z = VFMA(LDK(KP414213562), T5Y, T5X);
Chris@42 742 T62 = VFNMS(LDK(KP414213562), T61, T60);
Chris@42 743 T6q = VFMA(LDK(KP414213562), T60, T61);
Chris@42 744 {
Chris@42 745 V T8G, T5N, T5R, T8E;
Chris@42 746 T8G = VSUB(T5M, T5J);
Chris@42 747 T5N = VADD(T5J, T5M);
Chris@42 748 T5R = VSUB(T5P, T5Q);
Chris@42 749 T8E = VADD(T5P, T5Q);
Chris@42 750 {
Chris@42 751 V T5B, T8K, T8M, T5x;
Chris@42 752 T5B = VADD(T5z, T5A);
Chris@42 753 T8K = VSUB(T5A, T5z);
Chris@42 754 T8M = VADD(T4X, T5w);
Chris@42 755 T5x = VSUB(T4X, T5w);
Chris@42 756 T6o = VFNMS(LDK(KP707106781), T5V, T5S);
Chris@42 757 T5W = VFMA(LDK(KP707106781), T5V, T5S);
Chris@42 758 T8W = VADD(T5Z, T62);
Chris@42 759 T63 = VSUB(T5Z, T62);
Chris@42 760 ST(&(ii[WS(rs, 25)]), VFNMS(LDK(KP980785280), T8G, T8F), ms, &(ii[WS(rs, 1)]));
Chris@42 761 ST(&(ii[WS(rs, 9)]), VFMA(LDK(KP980785280), T8G, T8F), ms, &(ii[WS(rs, 1)]));
Chris@42 762 ST(&(ri[WS(rs, 1)]), VFMA(LDK(KP980785280), T5N, T5G), ms, &(ri[WS(rs, 1)]));
Chris@42 763 ST(&(ri[WS(rs, 17)]), VFNMS(LDK(KP980785280), T5N, T5G), ms, &(ri[WS(rs, 1)]));
Chris@42 764 ST(&(ri[WS(rs, 9)]), VFMA(LDK(KP980785280), T5R, T5O), ms, &(ri[WS(rs, 1)]));
Chris@42 765 ST(&(ri[WS(rs, 25)]), VFNMS(LDK(KP980785280), T5R, T5O), ms, &(ri[WS(rs, 1)]));
Chris@42 766 ST(&(ii[WS(rs, 17)]), VFNMS(LDK(KP980785280), T8E, T8D), ms, &(ii[WS(rs, 1)]));
Chris@42 767 ST(&(ii[WS(rs, 1)]), VFMA(LDK(KP980785280), T8E, T8D), ms, &(ii[WS(rs, 1)]));
Chris@42 768 ST(&(ri[WS(rs, 29)]), VFMA(LDK(KP831469612), T5B, T5y), ms, &(ri[WS(rs, 1)]));
Chris@42 769 ST(&(ri[WS(rs, 13)]), VFNMS(LDK(KP831469612), T5B, T5y), ms, &(ri[WS(rs, 1)]));
Chris@42 770 ST(&(ii[WS(rs, 21)]), VFNMS(LDK(KP831469612), T8K, T8J), ms, &(ii[WS(rs, 1)]));
Chris@42 771 ST(&(ii[WS(rs, 5)]), VFMA(LDK(KP831469612), T8K, T8J), ms, &(ii[WS(rs, 1)]));
Chris@42 772 ST(&(ii[WS(rs, 29)]), VFMA(LDK(KP831469612), T8M, T8L), ms, &(ii[WS(rs, 1)]));
Chris@42 773 ST(&(ii[WS(rs, 13)]), VFNMS(LDK(KP831469612), T8M, T8L), ms, &(ii[WS(rs, 1)]));
Chris@42 774 ST(&(ri[WS(rs, 5)]), VFMA(LDK(KP831469612), T5x, T4o), ms, &(ri[WS(rs, 1)]));
Chris@42 775 ST(&(ri[WS(rs, 21)]), VFNMS(LDK(KP831469612), T5x, T4o), ms, &(ri[WS(rs, 1)]));
Chris@42 776 }
Chris@42 777 }
Chris@42 778 }
Chris@42 779 {
Chris@42 780 V T6k, T64, T8V, T6r, T8R, T8T, T6y, T6C, T6m, T6i, T6v, T6B, T6l, T6b, T6A;
Chris@42 781 V T6s, T8X;
Chris@42 782 {
Chris@42 783 V T6x, T6e, T6w, T6h, T6u, T67, T6t, T6a, T8P, T8Q;
Chris@42 784 T6k = VFNMS(LDK(KP923879532), T63, T5W);
Chris@42 785 T64 = VFMA(LDK(KP923879532), T63, T5W);
Chris@42 786 T8V = VFNMS(LDK(KP707106781), T8O, T8N);
Chris@42 787 T8P = VFMA(LDK(KP707106781), T8O, T8N);
Chris@42 788 T8Q = VSUB(T6q, T6p);
Chris@42 789 T6r = VADD(T6p, T6q);
Chris@42 790 T6x = VFMA(LDK(KP707106781), T6d, T6c);
Chris@42 791 T6e = VFNMS(LDK(KP707106781), T6d, T6c);
Chris@42 792 T6w = VFMA(LDK(KP707106781), T6g, T6f);
Chris@42 793 T6h = VFNMS(LDK(KP707106781), T6g, T6f);
Chris@42 794 T6u = VFMA(LDK(KP707106781), T66, T65);
Chris@42 795 T67 = VFNMS(LDK(KP707106781), T66, T65);
Chris@42 796 T6t = VFMA(LDK(KP707106781), T69, T68);
Chris@42 797 T6a = VFNMS(LDK(KP707106781), T69, T68);
Chris@42 798 T8R = VFMA(LDK(KP923879532), T8Q, T8P);
Chris@42 799 T8T = VFNMS(LDK(KP923879532), T8Q, T8P);
Chris@42 800 T6y = VFNMS(LDK(KP198912367), T6x, T6w);
Chris@42 801 T6C = VFMA(LDK(KP198912367), T6w, T6x);
Chris@42 802 T6m = VFMA(LDK(KP668178637), T6e, T6h);
Chris@42 803 T6i = VFNMS(LDK(KP668178637), T6h, T6e);
Chris@42 804 T6v = VFMA(LDK(KP198912367), T6u, T6t);
Chris@42 805 T6B = VFNMS(LDK(KP198912367), T6t, T6u);
Chris@42 806 T6l = VFNMS(LDK(KP668178637), T67, T6a);
Chris@42 807 T6b = VFMA(LDK(KP668178637), T6a, T67);
Chris@42 808 }
Chris@42 809 T6A = VFMA(LDK(KP923879532), T6r, T6o);
Chris@42 810 T6s = VFNMS(LDK(KP923879532), T6r, T6o);
Chris@42 811 T8X = VFNMS(LDK(KP923879532), T8W, T8V);
Chris@42 812 T8Z = VFMA(LDK(KP923879532), T8W, T8V);
Chris@42 813 {
Chris@42 814 V T6z, T6D, T8Y, T6n, T8S, T8U, T6j;
Chris@42 815 T6z = VSUB(T6v, T6y);
Chris@42 816 T90 = VADD(T6v, T6y);
Chris@42 817 T6D = VADD(T6B, T6C);
Chris@42 818 T8Y = VSUB(T6C, T6B);
Chris@42 819 T6n = VSUB(T6l, T6m);
Chris@42 820 T8S = VADD(T6l, T6m);
Chris@42 821 T8U = VSUB(T6i, T6b);
Chris@42 822 T6j = VADD(T6b, T6i);
Chris@42 823 ST(&(ri[WS(rs, 7)]), VFMA(LDK(KP980785280), T6z, T6s), ms, &(ri[WS(rs, 1)]));
Chris@42 824 ST(&(ri[WS(rs, 23)]), VFNMS(LDK(KP980785280), T6z, T6s), ms, &(ri[WS(rs, 1)]));
Chris@42 825 ST(&(ii[WS(rs, 23)]), VFNMS(LDK(KP980785280), T8Y, T8X), ms, &(ii[WS(rs, 1)]));
Chris@42 826 ST(&(ii[WS(rs, 7)]), VFMA(LDK(KP980785280), T8Y, T8X), ms, &(ii[WS(rs, 1)]));
Chris@42 827 ST(&(ri[WS(rs, 11)]), VFMA(LDK(KP831469612), T6n, T6k), ms, &(ri[WS(rs, 1)]));
Chris@42 828 ST(&(ri[WS(rs, 27)]), VFNMS(LDK(KP831469612), T6n, T6k), ms, &(ri[WS(rs, 1)]));
Chris@42 829 ST(&(ii[WS(rs, 19)]), VFNMS(LDK(KP831469612), T8S, T8R), ms, &(ii[WS(rs, 1)]));
Chris@42 830 ST(&(ii[WS(rs, 3)]), VFMA(LDK(KP831469612), T8S, T8R), ms, &(ii[WS(rs, 1)]));
Chris@42 831 ST(&(ii[WS(rs, 27)]), VFNMS(LDK(KP831469612), T8U, T8T), ms, &(ii[WS(rs, 1)]));
Chris@42 832 ST(&(ii[WS(rs, 11)]), VFMA(LDK(KP831469612), T8U, T8T), ms, &(ii[WS(rs, 1)]));
Chris@42 833 ST(&(ri[WS(rs, 3)]), VFMA(LDK(KP831469612), T6j, T64), ms, &(ri[WS(rs, 1)]));
Chris@42 834 ST(&(ri[WS(rs, 19)]), VFNMS(LDK(KP831469612), T6j, T64), ms, &(ri[WS(rs, 1)]));
Chris@42 835 ST(&(ri[WS(rs, 31)]), VFMA(LDK(KP980785280), T6D, T6A), ms, &(ri[WS(rs, 1)]));
Chris@42 836 ST(&(ri[WS(rs, 15)]), VFNMS(LDK(KP980785280), T6D, T6A), ms, &(ri[WS(rs, 1)]));
Chris@42 837 }
Chris@42 838 }
Chris@42 839 }
Chris@42 840 }
Chris@42 841 }
Chris@42 842 }
Chris@42 843 ST(&(ii[WS(rs, 31)]), VFMA(LDK(KP980785280), T90, T8Z), ms, &(ii[WS(rs, 1)]));
Chris@42 844 ST(&(ii[WS(rs, 15)]), VFNMS(LDK(KP980785280), T90, T8Z), ms, &(ii[WS(rs, 1)]));
Chris@42 845 }
Chris@42 846 }
Chris@42 847 VLEAVE();
Chris@42 848 }
Chris@42 849
Chris@42 850 static const tw_instr twinstr[] = {
Chris@42 851 VTW(0, 1),
Chris@42 852 VTW(0, 2),
Chris@42 853 VTW(0, 3),
Chris@42 854 VTW(0, 4),
Chris@42 855 VTW(0, 5),
Chris@42 856 VTW(0, 6),
Chris@42 857 VTW(0, 7),
Chris@42 858 VTW(0, 8),
Chris@42 859 VTW(0, 9),
Chris@42 860 VTW(0, 10),
Chris@42 861 VTW(0, 11),
Chris@42 862 VTW(0, 12),
Chris@42 863 VTW(0, 13),
Chris@42 864 VTW(0, 14),
Chris@42 865 VTW(0, 15),
Chris@42 866 VTW(0, 16),
Chris@42 867 VTW(0, 17),
Chris@42 868 VTW(0, 18),
Chris@42 869 VTW(0, 19),
Chris@42 870 VTW(0, 20),
Chris@42 871 VTW(0, 21),
Chris@42 872 VTW(0, 22),
Chris@42 873 VTW(0, 23),
Chris@42 874 VTW(0, 24),
Chris@42 875 VTW(0, 25),
Chris@42 876 VTW(0, 26),
Chris@42 877 VTW(0, 27),
Chris@42 878 VTW(0, 28),
Chris@42 879 VTW(0, 29),
Chris@42 880 VTW(0, 30),
Chris@42 881 VTW(0, 31),
Chris@42 882 {TW_NEXT, (2 * VL), 0}
Chris@42 883 };
Chris@42 884
Chris@42 885 static const ct_desc desc = { 32, XSIMD_STRING("t1sv_32"), twinstr, &GENUS, {236, 62, 198, 0}, 0, 0, 0 };
Chris@42 886
Chris@42 887 void XSIMD(codelet_t1sv_32) (planner *p) {
Chris@42 888 X(kdft_dit_register) (p, t1sv_32, &desc);
Chris@42 889 }
Chris@42 890 #else /* HAVE_FMA */
Chris@42 891
Chris@42 892 /* Generated by: ../../../genfft/gen_twiddle.native -simd -compact -variables 4 -pipeline-latency 8 -n 32 -name t1sv_32 -include ts.h */
Chris@42 893
Chris@42 894 /*
Chris@42 895 * This function contains 434 FP additions, 208 FP multiplications,
Chris@42 896 * (or, 340 additions, 114 multiplications, 94 fused multiply/add),
Chris@42 897 * 96 stack variables, 7 constants, and 128 memory accesses
Chris@42 898 */
Chris@42 899 #include "ts.h"
Chris@42 900
Chris@42 901 static void t1sv_32(R *ri, R *ii, const R *W, stride rs, INT mb, INT me, INT ms)
Chris@42 902 {
Chris@42 903 DVK(KP195090322, +0.195090322016128267848284868477022240927691618);
Chris@42 904 DVK(KP980785280, +0.980785280403230449126182236134239036973933731);
Chris@42 905 DVK(KP555570233, +0.555570233019602224742830813948532874374937191);
Chris@42 906 DVK(KP831469612, +0.831469612302545237078788377617905756738560812);
Chris@42 907 DVK(KP382683432, +0.382683432365089771728459984030398866761344562);
Chris@42 908 DVK(KP923879532, +0.923879532511286756128183189396788286822416626);
Chris@42 909 DVK(KP707106781, +0.707106781186547524400844362104849039284835938);
Chris@42 910 {
Chris@42 911 INT m;
Chris@42 912 for (m = mb, W = W + (mb * 62); m < me; m = m + (2 * VL), ri = ri + ((2 * VL) * ms), ii = ii + ((2 * VL) * ms), W = W + ((2 * VL) * 62), MAKE_VOLATILE_STRIDE(64, rs)) {
Chris@42 913 V Tj, T5F, T7C, T7Q, T35, T4T, T78, T7m, T1Q, T61, T5Y, T6J, T3K, T59, T41;
Chris@42 914 V T56, T2B, T67, T6e, T6O, T4b, T5d, T4s, T5g, TG, T7l, T5I, T73, T3a, T4U;
Chris@42 915 V T3f, T4V, T14, T5N, T5M, T6E, T3m, T4Y, T3r, T4Z, T1r, T5P, T5S, T6F, T3x;
Chris@42 916 V T51, T3C, T52, T2d, T5Z, T64, T6K, T3V, T57, T44, T5a, T2Y, T6f, T6a, T6P;
Chris@42 917 V T4m, T5h, T4v, T5e;
Chris@42 918 {
Chris@42 919 V T1, T76, T6, T75, Tc, T32, Th, T33;
Chris@42 920 T1 = LD(&(ri[0]), ms, &(ri[0]));
Chris@42 921 T76 = LD(&(ii[0]), ms, &(ii[0]));
Chris@42 922 {
Chris@42 923 V T3, T5, T2, T4;
Chris@42 924 T3 = LD(&(ri[WS(rs, 16)]), ms, &(ri[0]));
Chris@42 925 T5 = LD(&(ii[WS(rs, 16)]), ms, &(ii[0]));
Chris@42 926 T2 = LDW(&(W[TWVL * 30]));
Chris@42 927 T4 = LDW(&(W[TWVL * 31]));
Chris@42 928 T6 = VFMA(T2, T3, VMUL(T4, T5));
Chris@42 929 T75 = VFNMS(T4, T3, VMUL(T2, T5));
Chris@42 930 }
Chris@42 931 {
Chris@42 932 V T9, Tb, T8, Ta;
Chris@42 933 T9 = LD(&(ri[WS(rs, 8)]), ms, &(ri[0]));
Chris@42 934 Tb = LD(&(ii[WS(rs, 8)]), ms, &(ii[0]));
Chris@42 935 T8 = LDW(&(W[TWVL * 14]));
Chris@42 936 Ta = LDW(&(W[TWVL * 15]));
Chris@42 937 Tc = VFMA(T8, T9, VMUL(Ta, Tb));
Chris@42 938 T32 = VFNMS(Ta, T9, VMUL(T8, Tb));
Chris@42 939 }
Chris@42 940 {
Chris@42 941 V Te, Tg, Td, Tf;
Chris@42 942 Te = LD(&(ri[WS(rs, 24)]), ms, &(ri[0]));
Chris@42 943 Tg = LD(&(ii[WS(rs, 24)]), ms, &(ii[0]));
Chris@42 944 Td = LDW(&(W[TWVL * 46]));
Chris@42 945 Tf = LDW(&(W[TWVL * 47]));
Chris@42 946 Th = VFMA(Td, Te, VMUL(Tf, Tg));
Chris@42 947 T33 = VFNMS(Tf, Te, VMUL(Td, Tg));
Chris@42 948 }
Chris@42 949 {
Chris@42 950 V T7, Ti, T7A, T7B;
Chris@42 951 T7 = VADD(T1, T6);
Chris@42 952 Ti = VADD(Tc, Th);
Chris@42 953 Tj = VADD(T7, Ti);
Chris@42 954 T5F = VSUB(T7, Ti);
Chris@42 955 T7A = VSUB(T76, T75);
Chris@42 956 T7B = VSUB(Tc, Th);
Chris@42 957 T7C = VSUB(T7A, T7B);
Chris@42 958 T7Q = VADD(T7B, T7A);
Chris@42 959 }
Chris@42 960 {
Chris@42 961 V T31, T34, T74, T77;
Chris@42 962 T31 = VSUB(T1, T6);
Chris@42 963 T34 = VSUB(T32, T33);
Chris@42 964 T35 = VSUB(T31, T34);
Chris@42 965 T4T = VADD(T31, T34);
Chris@42 966 T74 = VADD(T32, T33);
Chris@42 967 T77 = VADD(T75, T76);
Chris@42 968 T78 = VADD(T74, T77);
Chris@42 969 T7m = VSUB(T77, T74);
Chris@42 970 }
Chris@42 971 }
Chris@42 972 {
Chris@42 973 V T1y, T3G, T1O, T3Z, T1D, T3H, T1J, T3Y;
Chris@42 974 {
Chris@42 975 V T1v, T1x, T1u, T1w;
Chris@42 976 T1v = LD(&(ri[WS(rs, 1)]), ms, &(ri[WS(rs, 1)]));
Chris@42 977 T1x = LD(&(ii[WS(rs, 1)]), ms, &(ii[WS(rs, 1)]));
Chris@42 978 T1u = LDW(&(W[0]));
Chris@42 979 T1w = LDW(&(W[TWVL * 1]));
Chris@42 980 T1y = VFMA(T1u, T1v, VMUL(T1w, T1x));
Chris@42 981 T3G = VFNMS(T1w, T1v, VMUL(T1u, T1x));
Chris@42 982 }
Chris@42 983 {
Chris@42 984 V T1L, T1N, T1K, T1M;
Chris@42 985 T1L = LD(&(ri[WS(rs, 25)]), ms, &(ri[WS(rs, 1)]));
Chris@42 986 T1N = LD(&(ii[WS(rs, 25)]), ms, &(ii[WS(rs, 1)]));
Chris@42 987 T1K = LDW(&(W[TWVL * 48]));
Chris@42 988 T1M = LDW(&(W[TWVL * 49]));
Chris@42 989 T1O = VFMA(T1K, T1L, VMUL(T1M, T1N));
Chris@42 990 T3Z = VFNMS(T1M, T1L, VMUL(T1K, T1N));
Chris@42 991 }
Chris@42 992 {
Chris@42 993 V T1A, T1C, T1z, T1B;
Chris@42 994 T1A = LD(&(ri[WS(rs, 17)]), ms, &(ri[WS(rs, 1)]));
Chris@42 995 T1C = LD(&(ii[WS(rs, 17)]), ms, &(ii[WS(rs, 1)]));
Chris@42 996 T1z = LDW(&(W[TWVL * 32]));
Chris@42 997 T1B = LDW(&(W[TWVL * 33]));
Chris@42 998 T1D = VFMA(T1z, T1A, VMUL(T1B, T1C));
Chris@42 999 T3H = VFNMS(T1B, T1A, VMUL(T1z, T1C));
Chris@42 1000 }
Chris@42 1001 {
Chris@42 1002 V T1G, T1I, T1F, T1H;
Chris@42 1003 T1G = LD(&(ri[WS(rs, 9)]), ms, &(ri[WS(rs, 1)]));
Chris@42 1004 T1I = LD(&(ii[WS(rs, 9)]), ms, &(ii[WS(rs, 1)]));
Chris@42 1005 T1F = LDW(&(W[TWVL * 16]));
Chris@42 1006 T1H = LDW(&(W[TWVL * 17]));
Chris@42 1007 T1J = VFMA(T1F, T1G, VMUL(T1H, T1I));
Chris@42 1008 T3Y = VFNMS(T1H, T1G, VMUL(T1F, T1I));
Chris@42 1009 }
Chris@42 1010 {
Chris@42 1011 V T1E, T1P, T5W, T5X;
Chris@42 1012 T1E = VADD(T1y, T1D);
Chris@42 1013 T1P = VADD(T1J, T1O);
Chris@42 1014 T1Q = VADD(T1E, T1P);
Chris@42 1015 T61 = VSUB(T1E, T1P);
Chris@42 1016 T5W = VADD(T3G, T3H);
Chris@42 1017 T5X = VADD(T3Y, T3Z);
Chris@42 1018 T5Y = VSUB(T5W, T5X);
Chris@42 1019 T6J = VADD(T5W, T5X);
Chris@42 1020 }
Chris@42 1021 {
Chris@42 1022 V T3I, T3J, T3X, T40;
Chris@42 1023 T3I = VSUB(T3G, T3H);
Chris@42 1024 T3J = VSUB(T1J, T1O);
Chris@42 1025 T3K = VADD(T3I, T3J);
Chris@42 1026 T59 = VSUB(T3I, T3J);
Chris@42 1027 T3X = VSUB(T1y, T1D);
Chris@42 1028 T40 = VSUB(T3Y, T3Z);
Chris@42 1029 T41 = VSUB(T3X, T40);
Chris@42 1030 T56 = VADD(T3X, T40);
Chris@42 1031 }
Chris@42 1032 }
Chris@42 1033 {
Chris@42 1034 V T2j, T4o, T2z, T49, T2o, T4p, T2u, T48;
Chris@42 1035 {
Chris@42 1036 V T2g, T2i, T2f, T2h;
Chris@42 1037 T2g = LD(&(ri[WS(rs, 31)]), ms, &(ri[WS(rs, 1)]));
Chris@42 1038 T2i = LD(&(ii[WS(rs, 31)]), ms, &(ii[WS(rs, 1)]));
Chris@42 1039 T2f = LDW(&(W[TWVL * 60]));
Chris@42 1040 T2h = LDW(&(W[TWVL * 61]));
Chris@42 1041 T2j = VFMA(T2f, T2g, VMUL(T2h, T2i));
Chris@42 1042 T4o = VFNMS(T2h, T2g, VMUL(T2f, T2i));
Chris@42 1043 }
Chris@42 1044 {
Chris@42 1045 V T2w, T2y, T2v, T2x;
Chris@42 1046 T2w = LD(&(ri[WS(rs, 23)]), ms, &(ri[WS(rs, 1)]));
Chris@42 1047 T2y = LD(&(ii[WS(rs, 23)]), ms, &(ii[WS(rs, 1)]));
Chris@42 1048 T2v = LDW(&(W[TWVL * 44]));
Chris@42 1049 T2x = LDW(&(W[TWVL * 45]));
Chris@42 1050 T2z = VFMA(T2v, T2w, VMUL(T2x, T2y));
Chris@42 1051 T49 = VFNMS(T2x, T2w, VMUL(T2v, T2y));
Chris@42 1052 }
Chris@42 1053 {
Chris@42 1054 V T2l, T2n, T2k, T2m;
Chris@42 1055 T2l = LD(&(ri[WS(rs, 15)]), ms, &(ri[WS(rs, 1)]));
Chris@42 1056 T2n = LD(&(ii[WS(rs, 15)]), ms, &(ii[WS(rs, 1)]));
Chris@42 1057 T2k = LDW(&(W[TWVL * 28]));
Chris@42 1058 T2m = LDW(&(W[TWVL * 29]));
Chris@42 1059 T2o = VFMA(T2k, T2l, VMUL(T2m, T2n));
Chris@42 1060 T4p = VFNMS(T2m, T2l, VMUL(T2k, T2n));
Chris@42 1061 }
Chris@42 1062 {
Chris@42 1063 V T2r, T2t, T2q, T2s;
Chris@42 1064 T2r = LD(&(ri[WS(rs, 7)]), ms, &(ri[WS(rs, 1)]));
Chris@42 1065 T2t = LD(&(ii[WS(rs, 7)]), ms, &(ii[WS(rs, 1)]));
Chris@42 1066 T2q = LDW(&(W[TWVL * 12]));
Chris@42 1067 T2s = LDW(&(W[TWVL * 13]));
Chris@42 1068 T2u = VFMA(T2q, T2r, VMUL(T2s, T2t));
Chris@42 1069 T48 = VFNMS(T2s, T2r, VMUL(T2q, T2t));
Chris@42 1070 }
Chris@42 1071 {
Chris@42 1072 V T2p, T2A, T6c, T6d;
Chris@42 1073 T2p = VADD(T2j, T2o);
Chris@42 1074 T2A = VADD(T2u, T2z);
Chris@42 1075 T2B = VADD(T2p, T2A);
Chris@42 1076 T67 = VSUB(T2p, T2A);
Chris@42 1077 T6c = VADD(T4o, T4p);
Chris@42 1078 T6d = VADD(T48, T49);
Chris@42 1079 T6e = VSUB(T6c, T6d);
Chris@42 1080 T6O = VADD(T6c, T6d);
Chris@42 1081 }
Chris@42 1082 {
Chris@42 1083 V T47, T4a, T4q, T4r;
Chris@42 1084 T47 = VSUB(T2j, T2o);
Chris@42 1085 T4a = VSUB(T48, T49);
Chris@42 1086 T4b = VSUB(T47, T4a);
Chris@42 1087 T5d = VADD(T47, T4a);
Chris@42 1088 T4q = VSUB(T4o, T4p);
Chris@42 1089 T4r = VSUB(T2u, T2z);
Chris@42 1090 T4s = VADD(T4q, T4r);
Chris@42 1091 T5g = VSUB(T4q, T4r);
Chris@42 1092 }
Chris@42 1093 }
Chris@42 1094 {
Chris@42 1095 V To, T36, TE, T3d, Tt, T37, Tz, T3c;
Chris@42 1096 {
Chris@42 1097 V Tl, Tn, Tk, Tm;
Chris@42 1098 Tl = LD(&(ri[WS(rs, 4)]), ms, &(ri[0]));
Chris@42 1099 Tn = LD(&(ii[WS(rs, 4)]), ms, &(ii[0]));
Chris@42 1100 Tk = LDW(&(W[TWVL * 6]));
Chris@42 1101 Tm = LDW(&(W[TWVL * 7]));
Chris@42 1102 To = VFMA(Tk, Tl, VMUL(Tm, Tn));
Chris@42 1103 T36 = VFNMS(Tm, Tl, VMUL(Tk, Tn));
Chris@42 1104 }
Chris@42 1105 {
Chris@42 1106 V TB, TD, TA, TC;
Chris@42 1107 TB = LD(&(ri[WS(rs, 12)]), ms, &(ri[0]));
Chris@42 1108 TD = LD(&(ii[WS(rs, 12)]), ms, &(ii[0]));
Chris@42 1109 TA = LDW(&(W[TWVL * 22]));
Chris@42 1110 TC = LDW(&(W[TWVL * 23]));
Chris@42 1111 TE = VFMA(TA, TB, VMUL(TC, TD));
Chris@42 1112 T3d = VFNMS(TC, TB, VMUL(TA, TD));
Chris@42 1113 }
Chris@42 1114 {
Chris@42 1115 V Tq, Ts, Tp, Tr;
Chris@42 1116 Tq = LD(&(ri[WS(rs, 20)]), ms, &(ri[0]));
Chris@42 1117 Ts = LD(&(ii[WS(rs, 20)]), ms, &(ii[0]));
Chris@42 1118 Tp = LDW(&(W[TWVL * 38]));
Chris@42 1119 Tr = LDW(&(W[TWVL * 39]));
Chris@42 1120 Tt = VFMA(Tp, Tq, VMUL(Tr, Ts));
Chris@42 1121 T37 = VFNMS(Tr, Tq, VMUL(Tp, Ts));
Chris@42 1122 }
Chris@42 1123 {
Chris@42 1124 V Tw, Ty, Tv, Tx;
Chris@42 1125 Tw = LD(&(ri[WS(rs, 28)]), ms, &(ri[0]));
Chris@42 1126 Ty = LD(&(ii[WS(rs, 28)]), ms, &(ii[0]));
Chris@42 1127 Tv = LDW(&(W[TWVL * 54]));
Chris@42 1128 Tx = LDW(&(W[TWVL * 55]));
Chris@42 1129 Tz = VFMA(Tv, Tw, VMUL(Tx, Ty));
Chris@42 1130 T3c = VFNMS(Tx, Tw, VMUL(Tv, Ty));
Chris@42 1131 }
Chris@42 1132 {
Chris@42 1133 V Tu, TF, T5G, T5H;
Chris@42 1134 Tu = VADD(To, Tt);
Chris@42 1135 TF = VADD(Tz, TE);
Chris@42 1136 TG = VADD(Tu, TF);
Chris@42 1137 T7l = VSUB(TF, Tu);
Chris@42 1138 T5G = VADD(T36, T37);
Chris@42 1139 T5H = VADD(T3c, T3d);
Chris@42 1140 T5I = VSUB(T5G, T5H);
Chris@42 1141 T73 = VADD(T5G, T5H);
Chris@42 1142 }
Chris@42 1143 {
Chris@42 1144 V T38, T39, T3b, T3e;
Chris@42 1145 T38 = VSUB(T36, T37);
Chris@42 1146 T39 = VSUB(To, Tt);
Chris@42 1147 T3a = VSUB(T38, T39);
Chris@42 1148 T4U = VADD(T39, T38);
Chris@42 1149 T3b = VSUB(Tz, TE);
Chris@42 1150 T3e = VSUB(T3c, T3d);
Chris@42 1151 T3f = VADD(T3b, T3e);
Chris@42 1152 T4V = VSUB(T3b, T3e);
Chris@42 1153 }
Chris@42 1154 }
Chris@42 1155 {
Chris@42 1156 V TM, T3i, T12, T3p, TR, T3j, TX, T3o;
Chris@42 1157 {
Chris@42 1158 V TJ, TL, TI, TK;
Chris@42 1159 TJ = LD(&(ri[WS(rs, 2)]), ms, &(ri[0]));
Chris@42 1160 TL = LD(&(ii[WS(rs, 2)]), ms, &(ii[0]));
Chris@42 1161 TI = LDW(&(W[TWVL * 2]));
Chris@42 1162 TK = LDW(&(W[TWVL * 3]));
Chris@42 1163 TM = VFMA(TI, TJ, VMUL(TK, TL));
Chris@42 1164 T3i = VFNMS(TK, TJ, VMUL(TI, TL));
Chris@42 1165 }
Chris@42 1166 {
Chris@42 1167 V TZ, T11, TY, T10;
Chris@42 1168 TZ = LD(&(ri[WS(rs, 26)]), ms, &(ri[0]));
Chris@42 1169 T11 = LD(&(ii[WS(rs, 26)]), ms, &(ii[0]));
Chris@42 1170 TY = LDW(&(W[TWVL * 50]));
Chris@42 1171 T10 = LDW(&(W[TWVL * 51]));
Chris@42 1172 T12 = VFMA(TY, TZ, VMUL(T10, T11));
Chris@42 1173 T3p = VFNMS(T10, TZ, VMUL(TY, T11));
Chris@42 1174 }
Chris@42 1175 {
Chris@42 1176 V TO, TQ, TN, TP;
Chris@42 1177 TO = LD(&(ri[WS(rs, 18)]), ms, &(ri[0]));
Chris@42 1178 TQ = LD(&(ii[WS(rs, 18)]), ms, &(ii[0]));
Chris@42 1179 TN = LDW(&(W[TWVL * 34]));
Chris@42 1180 TP = LDW(&(W[TWVL * 35]));
Chris@42 1181 TR = VFMA(TN, TO, VMUL(TP, TQ));
Chris@42 1182 T3j = VFNMS(TP, TO, VMUL(TN, TQ));
Chris@42 1183 }
Chris@42 1184 {
Chris@42 1185 V TU, TW, TT, TV;
Chris@42 1186 TU = LD(&(ri[WS(rs, 10)]), ms, &(ri[0]));
Chris@42 1187 TW = LD(&(ii[WS(rs, 10)]), ms, &(ii[0]));
Chris@42 1188 TT = LDW(&(W[TWVL * 18]));
Chris@42 1189 TV = LDW(&(W[TWVL * 19]));
Chris@42 1190 TX = VFMA(TT, TU, VMUL(TV, TW));
Chris@42 1191 T3o = VFNMS(TV, TU, VMUL(TT, TW));
Chris@42 1192 }
Chris@42 1193 {
Chris@42 1194 V TS, T13, T5K, T5L;
Chris@42 1195 TS = VADD(TM, TR);
Chris@42 1196 T13 = VADD(TX, T12);
Chris@42 1197 T14 = VADD(TS, T13);
Chris@42 1198 T5N = VSUB(TS, T13);
Chris@42 1199 T5K = VADD(T3i, T3j);
Chris@42 1200 T5L = VADD(T3o, T3p);
Chris@42 1201 T5M = VSUB(T5K, T5L);
Chris@42 1202 T6E = VADD(T5K, T5L);
Chris@42 1203 }
Chris@42 1204 {
Chris@42 1205 V T3k, T3l, T3n, T3q;
Chris@42 1206 T3k = VSUB(T3i, T3j);
Chris@42 1207 T3l = VSUB(TX, T12);
Chris@42 1208 T3m = VADD(T3k, T3l);
Chris@42 1209 T4Y = VSUB(T3k, T3l);
Chris@42 1210 T3n = VSUB(TM, TR);
Chris@42 1211 T3q = VSUB(T3o, T3p);
Chris@42 1212 T3r = VSUB(T3n, T3q);
Chris@42 1213 T4Z = VADD(T3n, T3q);
Chris@42 1214 }
Chris@42 1215 }
Chris@42 1216 {
Chris@42 1217 V T19, T3t, T1p, T3A, T1e, T3u, T1k, T3z;
Chris@42 1218 {
Chris@42 1219 V T16, T18, T15, T17;
Chris@42 1220 T16 = LD(&(ri[WS(rs, 30)]), ms, &(ri[0]));
Chris@42 1221 T18 = LD(&(ii[WS(rs, 30)]), ms, &(ii[0]));
Chris@42 1222 T15 = LDW(&(W[TWVL * 58]));
Chris@42 1223 T17 = LDW(&(W[TWVL * 59]));
Chris@42 1224 T19 = VFMA(T15, T16, VMUL(T17, T18));
Chris@42 1225 T3t = VFNMS(T17, T16, VMUL(T15, T18));
Chris@42 1226 }
Chris@42 1227 {
Chris@42 1228 V T1m, T1o, T1l, T1n;
Chris@42 1229 T1m = LD(&(ri[WS(rs, 22)]), ms, &(ri[0]));
Chris@42 1230 T1o = LD(&(ii[WS(rs, 22)]), ms, &(ii[0]));
Chris@42 1231 T1l = LDW(&(W[TWVL * 42]));
Chris@42 1232 T1n = LDW(&(W[TWVL * 43]));
Chris@42 1233 T1p = VFMA(T1l, T1m, VMUL(T1n, T1o));
Chris@42 1234 T3A = VFNMS(T1n, T1m, VMUL(T1l, T1o));
Chris@42 1235 }
Chris@42 1236 {
Chris@42 1237 V T1b, T1d, T1a, T1c;
Chris@42 1238 T1b = LD(&(ri[WS(rs, 14)]), ms, &(ri[0]));
Chris@42 1239 T1d = LD(&(ii[WS(rs, 14)]), ms, &(ii[0]));
Chris@42 1240 T1a = LDW(&(W[TWVL * 26]));
Chris@42 1241 T1c = LDW(&(W[TWVL * 27]));
Chris@42 1242 T1e = VFMA(T1a, T1b, VMUL(T1c, T1d));
Chris@42 1243 T3u = VFNMS(T1c, T1b, VMUL(T1a, T1d));
Chris@42 1244 }
Chris@42 1245 {
Chris@42 1246 V T1h, T1j, T1g, T1i;
Chris@42 1247 T1h = LD(&(ri[WS(rs, 6)]), ms, &(ri[0]));
Chris@42 1248 T1j = LD(&(ii[WS(rs, 6)]), ms, &(ii[0]));
Chris@42 1249 T1g = LDW(&(W[TWVL * 10]));
Chris@42 1250 T1i = LDW(&(W[TWVL * 11]));
Chris@42 1251 T1k = VFMA(T1g, T1h, VMUL(T1i, T1j));
Chris@42 1252 T3z = VFNMS(T1i, T1h, VMUL(T1g, T1j));
Chris@42 1253 }
Chris@42 1254 {
Chris@42 1255 V T1f, T1q, T5Q, T5R;
Chris@42 1256 T1f = VADD(T19, T1e);
Chris@42 1257 T1q = VADD(T1k, T1p);
Chris@42 1258 T1r = VADD(T1f, T1q);
Chris@42 1259 T5P = VSUB(T1f, T1q);
Chris@42 1260 T5Q = VADD(T3t, T3u);
Chris@42 1261 T5R = VADD(T3z, T3A);
Chris@42 1262 T5S = VSUB(T5Q, T5R);
Chris@42 1263 T6F = VADD(T5Q, T5R);
Chris@42 1264 }
Chris@42 1265 {
Chris@42 1266 V T3v, T3w, T3y, T3B;
Chris@42 1267 T3v = VSUB(T3t, T3u);
Chris@42 1268 T3w = VSUB(T1k, T1p);
Chris@42 1269 T3x = VADD(T3v, T3w);
Chris@42 1270 T51 = VSUB(T3v, T3w);
Chris@42 1271 T3y = VSUB(T19, T1e);
Chris@42 1272 T3B = VSUB(T3z, T3A);
Chris@42 1273 T3C = VSUB(T3y, T3B);
Chris@42 1274 T52 = VADD(T3y, T3B);
Chris@42 1275 }
Chris@42 1276 }
Chris@42 1277 {
Chris@42 1278 V T1V, T3R, T20, T3S, T3Q, T3T, T26, T3M, T2b, T3N, T3L, T3O;
Chris@42 1279 {
Chris@42 1280 V T1S, T1U, T1R, T1T;
Chris@42 1281 T1S = LD(&(ri[WS(rs, 5)]), ms, &(ri[WS(rs, 1)]));
Chris@42 1282 T1U = LD(&(ii[WS(rs, 5)]), ms, &(ii[WS(rs, 1)]));
Chris@42 1283 T1R = LDW(&(W[TWVL * 8]));
Chris@42 1284 T1T = LDW(&(W[TWVL * 9]));
Chris@42 1285 T1V = VFMA(T1R, T1S, VMUL(T1T, T1U));
Chris@42 1286 T3R = VFNMS(T1T, T1S, VMUL(T1R, T1U));
Chris@42 1287 }
Chris@42 1288 {
Chris@42 1289 V T1X, T1Z, T1W, T1Y;
Chris@42 1290 T1X = LD(&(ri[WS(rs, 21)]), ms, &(ri[WS(rs, 1)]));
Chris@42 1291 T1Z = LD(&(ii[WS(rs, 21)]), ms, &(ii[WS(rs, 1)]));
Chris@42 1292 T1W = LDW(&(W[TWVL * 40]));
Chris@42 1293 T1Y = LDW(&(W[TWVL * 41]));
Chris@42 1294 T20 = VFMA(T1W, T1X, VMUL(T1Y, T1Z));
Chris@42 1295 T3S = VFNMS(T1Y, T1X, VMUL(T1W, T1Z));
Chris@42 1296 }
Chris@42 1297 T3Q = VSUB(T1V, T20);
Chris@42 1298 T3T = VSUB(T3R, T3S);
Chris@42 1299 {
Chris@42 1300 V T23, T25, T22, T24;
Chris@42 1301 T23 = LD(&(ri[WS(rs, 29)]), ms, &(ri[WS(rs, 1)]));
Chris@42 1302 T25 = LD(&(ii[WS(rs, 29)]), ms, &(ii[WS(rs, 1)]));
Chris@42 1303 T22 = LDW(&(W[TWVL * 56]));
Chris@42 1304 T24 = LDW(&(W[TWVL * 57]));
Chris@42 1305 T26 = VFMA(T22, T23, VMUL(T24, T25));
Chris@42 1306 T3M = VFNMS(T24, T23, VMUL(T22, T25));
Chris@42 1307 }
Chris@42 1308 {
Chris@42 1309 V T28, T2a, T27, T29;
Chris@42 1310 T28 = LD(&(ri[WS(rs, 13)]), ms, &(ri[WS(rs, 1)]));
Chris@42 1311 T2a = LD(&(ii[WS(rs, 13)]), ms, &(ii[WS(rs, 1)]));
Chris@42 1312 T27 = LDW(&(W[TWVL * 24]));
Chris@42 1313 T29 = LDW(&(W[TWVL * 25]));
Chris@42 1314 T2b = VFMA(T27, T28, VMUL(T29, T2a));
Chris@42 1315 T3N = VFNMS(T29, T28, VMUL(T27, T2a));
Chris@42 1316 }
Chris@42 1317 T3L = VSUB(T26, T2b);
Chris@42 1318 T3O = VSUB(T3M, T3N);
Chris@42 1319 {
Chris@42 1320 V T21, T2c, T62, T63;
Chris@42 1321 T21 = VADD(T1V, T20);
Chris@42 1322 T2c = VADD(T26, T2b);
Chris@42 1323 T2d = VADD(T21, T2c);
Chris@42 1324 T5Z = VSUB(T2c, T21);
Chris@42 1325 T62 = VADD(T3R, T3S);
Chris@42 1326 T63 = VADD(T3M, T3N);
Chris@42 1327 T64 = VSUB(T62, T63);
Chris@42 1328 T6K = VADD(T62, T63);
Chris@42 1329 }
Chris@42 1330 {
Chris@42 1331 V T3P, T3U, T42, T43;
Chris@42 1332 T3P = VSUB(T3L, T3O);
Chris@42 1333 T3U = VADD(T3Q, T3T);
Chris@42 1334 T3V = VMUL(LDK(KP707106781), VSUB(T3P, T3U));
Chris@42 1335 T57 = VMUL(LDK(KP707106781), VADD(T3U, T3P));
Chris@42 1336 T42 = VSUB(T3T, T3Q);
Chris@42 1337 T43 = VADD(T3L, T3O);
Chris@42 1338 T44 = VMUL(LDK(KP707106781), VSUB(T42, T43));
Chris@42 1339 T5a = VMUL(LDK(KP707106781), VADD(T42, T43));
Chris@42 1340 }
Chris@42 1341 }
Chris@42 1342 {
Chris@42 1343 V T2G, T4c, T2L, T4d, T4e, T4f, T2R, T4i, T2W, T4j, T4h, T4k;
Chris@42 1344 {
Chris@42 1345 V T2D, T2F, T2C, T2E;
Chris@42 1346 T2D = LD(&(ri[WS(rs, 3)]), ms, &(ri[WS(rs, 1)]));
Chris@42 1347 T2F = LD(&(ii[WS(rs, 3)]), ms, &(ii[WS(rs, 1)]));
Chris@42 1348 T2C = LDW(&(W[TWVL * 4]));
Chris@42 1349 T2E = LDW(&(W[TWVL * 5]));
Chris@42 1350 T2G = VFMA(T2C, T2D, VMUL(T2E, T2F));
Chris@42 1351 T4c = VFNMS(T2E, T2D, VMUL(T2C, T2F));
Chris@42 1352 }
Chris@42 1353 {
Chris@42 1354 V T2I, T2K, T2H, T2J;
Chris@42 1355 T2I = LD(&(ri[WS(rs, 19)]), ms, &(ri[WS(rs, 1)]));
Chris@42 1356 T2K = LD(&(ii[WS(rs, 19)]), ms, &(ii[WS(rs, 1)]));
Chris@42 1357 T2H = LDW(&(W[TWVL * 36]));
Chris@42 1358 T2J = LDW(&(W[TWVL * 37]));
Chris@42 1359 T2L = VFMA(T2H, T2I, VMUL(T2J, T2K));
Chris@42 1360 T4d = VFNMS(T2J, T2I, VMUL(T2H, T2K));
Chris@42 1361 }
Chris@42 1362 T4e = VSUB(T4c, T4d);
Chris@42 1363 T4f = VSUB(T2G, T2L);
Chris@42 1364 {
Chris@42 1365 V T2O, T2Q, T2N, T2P;
Chris@42 1366 T2O = LD(&(ri[WS(rs, 27)]), ms, &(ri[WS(rs, 1)]));
Chris@42 1367 T2Q = LD(&(ii[WS(rs, 27)]), ms, &(ii[WS(rs, 1)]));
Chris@42 1368 T2N = LDW(&(W[TWVL * 52]));
Chris@42 1369 T2P = LDW(&(W[TWVL * 53]));
Chris@42 1370 T2R = VFMA(T2N, T2O, VMUL(T2P, T2Q));
Chris@42 1371 T4i = VFNMS(T2P, T2O, VMUL(T2N, T2Q));
Chris@42 1372 }
Chris@42 1373 {
Chris@42 1374 V T2T, T2V, T2S, T2U;
Chris@42 1375 T2T = LD(&(ri[WS(rs, 11)]), ms, &(ri[WS(rs, 1)]));
Chris@42 1376 T2V = LD(&(ii[WS(rs, 11)]), ms, &(ii[WS(rs, 1)]));
Chris@42 1377 T2S = LDW(&(W[TWVL * 20]));
Chris@42 1378 T2U = LDW(&(W[TWVL * 21]));
Chris@42 1379 T2W = VFMA(T2S, T2T, VMUL(T2U, T2V));
Chris@42 1380 T4j = VFNMS(T2U, T2T, VMUL(T2S, T2V));
Chris@42 1381 }
Chris@42 1382 T4h = VSUB(T2R, T2W);
Chris@42 1383 T4k = VSUB(T4i, T4j);
Chris@42 1384 {
Chris@42 1385 V T2M, T2X, T68, T69;
Chris@42 1386 T2M = VADD(T2G, T2L);
Chris@42 1387 T2X = VADD(T2R, T2W);
Chris@42 1388 T2Y = VADD(T2M, T2X);
Chris@42 1389 T6f = VSUB(T2X, T2M);
Chris@42 1390 T68 = VADD(T4c, T4d);
Chris@42 1391 T69 = VADD(T4i, T4j);
Chris@42 1392 T6a = VSUB(T68, T69);
Chris@42 1393 T6P = VADD(T68, T69);
Chris@42 1394 }
Chris@42 1395 {
Chris@42 1396 V T4g, T4l, T4t, T4u;
Chris@42 1397 T4g = VSUB(T4e, T4f);
Chris@42 1398 T4l = VADD(T4h, T4k);
Chris@42 1399 T4m = VMUL(LDK(KP707106781), VSUB(T4g, T4l));
Chris@42 1400 T5h = VMUL(LDK(KP707106781), VADD(T4g, T4l));
Chris@42 1401 T4t = VSUB(T4h, T4k);
Chris@42 1402 T4u = VADD(T4f, T4e);
Chris@42 1403 T4v = VMUL(LDK(KP707106781), VSUB(T4t, T4u));
Chris@42 1404 T5e = VMUL(LDK(KP707106781), VADD(T4u, T4t));
Chris@42 1405 }
Chris@42 1406 }
Chris@42 1407 {
Chris@42 1408 V T1t, T6X, T7a, T7c, T30, T7b, T70, T71;
Chris@42 1409 {
Chris@42 1410 V TH, T1s, T72, T79;
Chris@42 1411 TH = VADD(Tj, TG);
Chris@42 1412 T1s = VADD(T14, T1r);
Chris@42 1413 T1t = VADD(TH, T1s);
Chris@42 1414 T6X = VSUB(TH, T1s);
Chris@42 1415 T72 = VADD(T6E, T6F);
Chris@42 1416 T79 = VADD(T73, T78);
Chris@42 1417 T7a = VADD(T72, T79);
Chris@42 1418 T7c = VSUB(T79, T72);
Chris@42 1419 }
Chris@42 1420 {
Chris@42 1421 V T2e, T2Z, T6Y, T6Z;
Chris@42 1422 T2e = VADD(T1Q, T2d);
Chris@42 1423 T2Z = VADD(T2B, T2Y);
Chris@42 1424 T30 = VADD(T2e, T2Z);
Chris@42 1425 T7b = VSUB(T2Z, T2e);
Chris@42 1426 T6Y = VADD(T6J, T6K);
Chris@42 1427 T6Z = VADD(T6O, T6P);
Chris@42 1428 T70 = VSUB(T6Y, T6Z);
Chris@42 1429 T71 = VADD(T6Y, T6Z);
Chris@42 1430 }
Chris@42 1431 ST(&(ri[WS(rs, 16)]), VSUB(T1t, T30), ms, &(ri[0]));
Chris@42 1432 ST(&(ii[WS(rs, 16)]), VSUB(T7a, T71), ms, &(ii[0]));
Chris@42 1433 ST(&(ri[0]), VADD(T1t, T30), ms, &(ri[0]));
Chris@42 1434 ST(&(ii[0]), VADD(T71, T7a), ms, &(ii[0]));
Chris@42 1435 ST(&(ri[WS(rs, 24)]), VSUB(T6X, T70), ms, &(ri[0]));
Chris@42 1436 ST(&(ii[WS(rs, 24)]), VSUB(T7c, T7b), ms, &(ii[0]));
Chris@42 1437 ST(&(ri[WS(rs, 8)]), VADD(T6X, T70), ms, &(ri[0]));
Chris@42 1438 ST(&(ii[WS(rs, 8)]), VADD(T7b, T7c), ms, &(ii[0]));
Chris@42 1439 }
Chris@42 1440 {
Chris@42 1441 V T6H, T6T, T7g, T7i, T6M, T6U, T6R, T6V;
Chris@42 1442 {
Chris@42 1443 V T6D, T6G, T7e, T7f;
Chris@42 1444 T6D = VSUB(Tj, TG);
Chris@42 1445 T6G = VSUB(T6E, T6F);
Chris@42 1446 T6H = VADD(T6D, T6G);
Chris@42 1447 T6T = VSUB(T6D, T6G);
Chris@42 1448 T7e = VSUB(T1r, T14);
Chris@42 1449 T7f = VSUB(T78, T73);
Chris@42 1450 T7g = VADD(T7e, T7f);
Chris@42 1451 T7i = VSUB(T7f, T7e);
Chris@42 1452 }
Chris@42 1453 {
Chris@42 1454 V T6I, T6L, T6N, T6Q;
Chris@42 1455 T6I = VSUB(T1Q, T2d);
Chris@42 1456 T6L = VSUB(T6J, T6K);
Chris@42 1457 T6M = VADD(T6I, T6L);
Chris@42 1458 T6U = VSUB(T6L, T6I);
Chris@42 1459 T6N = VSUB(T2B, T2Y);
Chris@42 1460 T6Q = VSUB(T6O, T6P);
Chris@42 1461 T6R = VSUB(T6N, T6Q);
Chris@42 1462 T6V = VADD(T6N, T6Q);
Chris@42 1463 }
Chris@42 1464 {
Chris@42 1465 V T6S, T7d, T6W, T7h;
Chris@42 1466 T6S = VMUL(LDK(KP707106781), VADD(T6M, T6R));
Chris@42 1467 ST(&(ri[WS(rs, 20)]), VSUB(T6H, T6S), ms, &(ri[0]));
Chris@42 1468 ST(&(ri[WS(rs, 4)]), VADD(T6H, T6S), ms, &(ri[0]));
Chris@42 1469 T7d = VMUL(LDK(KP707106781), VADD(T6U, T6V));
Chris@42 1470 ST(&(ii[WS(rs, 4)]), VADD(T7d, T7g), ms, &(ii[0]));
Chris@42 1471 ST(&(ii[WS(rs, 20)]), VSUB(T7g, T7d), ms, &(ii[0]));
Chris@42 1472 T6W = VMUL(LDK(KP707106781), VSUB(T6U, T6V));
Chris@42 1473 ST(&(ri[WS(rs, 28)]), VSUB(T6T, T6W), ms, &(ri[0]));
Chris@42 1474 ST(&(ri[WS(rs, 12)]), VADD(T6T, T6W), ms, &(ri[0]));
Chris@42 1475 T7h = VMUL(LDK(KP707106781), VSUB(T6R, T6M));
Chris@42 1476 ST(&(ii[WS(rs, 12)]), VADD(T7h, T7i), ms, &(ii[0]));
Chris@42 1477 ST(&(ii[WS(rs, 28)]), VSUB(T7i, T7h), ms, &(ii[0]));
Chris@42 1478 }
Chris@42 1479 }
Chris@42 1480 {
Chris@42 1481 V T5J, T7n, T7t, T6n, T5U, T7k, T6x, T6B, T6q, T7s, T66, T6k, T6u, T6A, T6h;
Chris@42 1482 V T6l;
Chris@42 1483 {
Chris@42 1484 V T5O, T5T, T60, T65;
Chris@42 1485 T5J = VSUB(T5F, T5I);
Chris@42 1486 T7n = VADD(T7l, T7m);
Chris@42 1487 T7t = VSUB(T7m, T7l);
Chris@42 1488 T6n = VADD(T5F, T5I);
Chris@42 1489 T5O = VSUB(T5M, T5N);
Chris@42 1490 T5T = VADD(T5P, T5S);
Chris@42 1491 T5U = VMUL(LDK(KP707106781), VSUB(T5O, T5T));
Chris@42 1492 T7k = VMUL(LDK(KP707106781), VADD(T5O, T5T));
Chris@42 1493 {
Chris@42 1494 V T6v, T6w, T6o, T6p;
Chris@42 1495 T6v = VADD(T67, T6a);
Chris@42 1496 T6w = VADD(T6e, T6f);
Chris@42 1497 T6x = VFNMS(LDK(KP382683432), T6w, VMUL(LDK(KP923879532), T6v));
Chris@42 1498 T6B = VFMA(LDK(KP923879532), T6w, VMUL(LDK(KP382683432), T6v));
Chris@42 1499 T6o = VADD(T5N, T5M);
Chris@42 1500 T6p = VSUB(T5P, T5S);
Chris@42 1501 T6q = VMUL(LDK(KP707106781), VADD(T6o, T6p));
Chris@42 1502 T7s = VMUL(LDK(KP707106781), VSUB(T6p, T6o));
Chris@42 1503 }
Chris@42 1504 T60 = VSUB(T5Y, T5Z);
Chris@42 1505 T65 = VSUB(T61, T64);
Chris@42 1506 T66 = VFMA(LDK(KP923879532), T60, VMUL(LDK(KP382683432), T65));
Chris@42 1507 T6k = VFNMS(LDK(KP923879532), T65, VMUL(LDK(KP382683432), T60));
Chris@42 1508 {
Chris@42 1509 V T6s, T6t, T6b, T6g;
Chris@42 1510 T6s = VADD(T5Y, T5Z);
Chris@42 1511 T6t = VADD(T61, T64);
Chris@42 1512 T6u = VFMA(LDK(KP382683432), T6s, VMUL(LDK(KP923879532), T6t));
Chris@42 1513 T6A = VFNMS(LDK(KP382683432), T6t, VMUL(LDK(KP923879532), T6s));
Chris@42 1514 T6b = VSUB(T67, T6a);
Chris@42 1515 T6g = VSUB(T6e, T6f);
Chris@42 1516 T6h = VFNMS(LDK(KP923879532), T6g, VMUL(LDK(KP382683432), T6b));
Chris@42 1517 T6l = VFMA(LDK(KP382683432), T6g, VMUL(LDK(KP923879532), T6b));
Chris@42 1518 }
Chris@42 1519 }
Chris@42 1520 {
Chris@42 1521 V T5V, T6i, T7r, T7u;
Chris@42 1522 T5V = VADD(T5J, T5U);
Chris@42 1523 T6i = VADD(T66, T6h);
Chris@42 1524 ST(&(ri[WS(rs, 22)]), VSUB(T5V, T6i), ms, &(ri[0]));
Chris@42 1525 ST(&(ri[WS(rs, 6)]), VADD(T5V, T6i), ms, &(ri[0]));
Chris@42 1526 T7r = VADD(T6k, T6l);
Chris@42 1527 T7u = VADD(T7s, T7t);
Chris@42 1528 ST(&(ii[WS(rs, 6)]), VADD(T7r, T7u), ms, &(ii[0]));
Chris@42 1529 ST(&(ii[WS(rs, 22)]), VSUB(T7u, T7r), ms, &(ii[0]));
Chris@42 1530 }
Chris@42 1531 {
Chris@42 1532 V T6j, T6m, T7v, T7w;
Chris@42 1533 T6j = VSUB(T5J, T5U);
Chris@42 1534 T6m = VSUB(T6k, T6l);
Chris@42 1535 ST(&(ri[WS(rs, 30)]), VSUB(T6j, T6m), ms, &(ri[0]));
Chris@42 1536 ST(&(ri[WS(rs, 14)]), VADD(T6j, T6m), ms, &(ri[0]));
Chris@42 1537 T7v = VSUB(T6h, T66);
Chris@42 1538 T7w = VSUB(T7t, T7s);
Chris@42 1539 ST(&(ii[WS(rs, 14)]), VADD(T7v, T7w), ms, &(ii[0]));
Chris@42 1540 ST(&(ii[WS(rs, 30)]), VSUB(T7w, T7v), ms, &(ii[0]));
Chris@42 1541 }
Chris@42 1542 {
Chris@42 1543 V T6r, T6y, T7j, T7o;
Chris@42 1544 T6r = VADD(T6n, T6q);
Chris@42 1545 T6y = VADD(T6u, T6x);
Chris@42 1546 ST(&(ri[WS(rs, 18)]), VSUB(T6r, T6y), ms, &(ri[0]));
Chris@42 1547 ST(&(ri[WS(rs, 2)]), VADD(T6r, T6y), ms, &(ri[0]));
Chris@42 1548 T7j = VADD(T6A, T6B);
Chris@42 1549 T7o = VADD(T7k, T7n);
Chris@42 1550 ST(&(ii[WS(rs, 2)]), VADD(T7j, T7o), ms, &(ii[0]));
Chris@42 1551 ST(&(ii[WS(rs, 18)]), VSUB(T7o, T7j), ms, &(ii[0]));
Chris@42 1552 }
Chris@42 1553 {
Chris@42 1554 V T6z, T6C, T7p, T7q;
Chris@42 1555 T6z = VSUB(T6n, T6q);
Chris@42 1556 T6C = VSUB(T6A, T6B);
Chris@42 1557 ST(&(ri[WS(rs, 26)]), VSUB(T6z, T6C), ms, &(ri[0]));
Chris@42 1558 ST(&(ri[WS(rs, 10)]), VADD(T6z, T6C), ms, &(ri[0]));
Chris@42 1559 T7p = VSUB(T6x, T6u);
Chris@42 1560 T7q = VSUB(T7n, T7k);
Chris@42 1561 ST(&(ii[WS(rs, 10)]), VADD(T7p, T7q), ms, &(ii[0]));
Chris@42 1562 ST(&(ii[WS(rs, 26)]), VSUB(T7q, T7p), ms, &(ii[0]));
Chris@42 1563 }
Chris@42 1564 }
Chris@42 1565 {
Chris@42 1566 V T3h, T4D, T7R, T7X, T3E, T7O, T4N, T4R, T46, T4A, T4G, T7W, T4K, T4Q, T4x;
Chris@42 1567 V T4B, T3g, T7P;
Chris@42 1568 T3g = VMUL(LDK(KP707106781), VSUB(T3a, T3f));
Chris@42 1569 T3h = VSUB(T35, T3g);
Chris@42 1570 T4D = VADD(T35, T3g);
Chris@42 1571 T7P = VMUL(LDK(KP707106781), VSUB(T4V, T4U));
Chris@42 1572 T7R = VADD(T7P, T7Q);
Chris@42 1573 T7X = VSUB(T7Q, T7P);
Chris@42 1574 {
Chris@42 1575 V T3s, T3D, T4L, T4M;
Chris@42 1576 T3s = VFNMS(LDK(KP923879532), T3r, VMUL(LDK(KP382683432), T3m));
Chris@42 1577 T3D = VFMA(LDK(KP382683432), T3x, VMUL(LDK(KP923879532), T3C));
Chris@42 1578 T3E = VSUB(T3s, T3D);
Chris@42 1579 T7O = VADD(T3s, T3D);
Chris@42 1580 T4L = VADD(T4b, T4m);
Chris@42 1581 T4M = VADD(T4s, T4v);
Chris@42 1582 T4N = VFNMS(LDK(KP555570233), T4M, VMUL(LDK(KP831469612), T4L));
Chris@42 1583 T4R = VFMA(LDK(KP831469612), T4M, VMUL(LDK(KP555570233), T4L));
Chris@42 1584 }
Chris@42 1585 {
Chris@42 1586 V T3W, T45, T4E, T4F;
Chris@42 1587 T3W = VSUB(T3K, T3V);
Chris@42 1588 T45 = VSUB(T41, T44);
Chris@42 1589 T46 = VFMA(LDK(KP980785280), T3W, VMUL(LDK(KP195090322), T45));
Chris@42 1590 T4A = VFNMS(LDK(KP980785280), T45, VMUL(LDK(KP195090322), T3W));
Chris@42 1591 T4E = VFMA(LDK(KP923879532), T3m, VMUL(LDK(KP382683432), T3r));
Chris@42 1592 T4F = VFNMS(LDK(KP923879532), T3x, VMUL(LDK(KP382683432), T3C));
Chris@42 1593 T4G = VADD(T4E, T4F);
Chris@42 1594 T7W = VSUB(T4F, T4E);
Chris@42 1595 }
Chris@42 1596 {
Chris@42 1597 V T4I, T4J, T4n, T4w;
Chris@42 1598 T4I = VADD(T3K, T3V);
Chris@42 1599 T4J = VADD(T41, T44);
Chris@42 1600 T4K = VFMA(LDK(KP555570233), T4I, VMUL(LDK(KP831469612), T4J));
Chris@42 1601 T4Q = VFNMS(LDK(KP555570233), T4J, VMUL(LDK(KP831469612), T4I));
Chris@42 1602 T4n = VSUB(T4b, T4m);
Chris@42 1603 T4w = VSUB(T4s, T4v);
Chris@42 1604 T4x = VFNMS(LDK(KP980785280), T4w, VMUL(LDK(KP195090322), T4n));
Chris@42 1605 T4B = VFMA(LDK(KP195090322), T4w, VMUL(LDK(KP980785280), T4n));
Chris@42 1606 }
Chris@42 1607 {
Chris@42 1608 V T3F, T4y, T7V, T7Y;
Chris@42 1609 T3F = VADD(T3h, T3E);
Chris@42 1610 T4y = VADD(T46, T4x);
Chris@42 1611 ST(&(ri[WS(rs, 23)]), VSUB(T3F, T4y), ms, &(ri[WS(rs, 1)]));
Chris@42 1612 ST(&(ri[WS(rs, 7)]), VADD(T3F, T4y), ms, &(ri[WS(rs, 1)]));
Chris@42 1613 T7V = VADD(T4A, T4B);
Chris@42 1614 T7Y = VADD(T7W, T7X);
Chris@42 1615 ST(&(ii[WS(rs, 7)]), VADD(T7V, T7Y), ms, &(ii[WS(rs, 1)]));
Chris@42 1616 ST(&(ii[WS(rs, 23)]), VSUB(T7Y, T7V), ms, &(ii[WS(rs, 1)]));
Chris@42 1617 }
Chris@42 1618 {
Chris@42 1619 V T4z, T4C, T7Z, T80;
Chris@42 1620 T4z = VSUB(T3h, T3E);
Chris@42 1621 T4C = VSUB(T4A, T4B);
Chris@42 1622 ST(&(ri[WS(rs, 31)]), VSUB(T4z, T4C), ms, &(ri[WS(rs, 1)]));
Chris@42 1623 ST(&(ri[WS(rs, 15)]), VADD(T4z, T4C), ms, &(ri[WS(rs, 1)]));
Chris@42 1624 T7Z = VSUB(T4x, T46);
Chris@42 1625 T80 = VSUB(T7X, T7W);
Chris@42 1626 ST(&(ii[WS(rs, 15)]), VADD(T7Z, T80), ms, &(ii[WS(rs, 1)]));
Chris@42 1627 ST(&(ii[WS(rs, 31)]), VSUB(T80, T7Z), ms, &(ii[WS(rs, 1)]));
Chris@42 1628 }
Chris@42 1629 {
Chris@42 1630 V T4H, T4O, T7N, T7S;
Chris@42 1631 T4H = VADD(T4D, T4G);
Chris@42 1632 T4O = VADD(T4K, T4N);
Chris@42 1633 ST(&(ri[WS(rs, 19)]), VSUB(T4H, T4O), ms, &(ri[WS(rs, 1)]));
Chris@42 1634 ST(&(ri[WS(rs, 3)]), VADD(T4H, T4O), ms, &(ri[WS(rs, 1)]));
Chris@42 1635 T7N = VADD(T4Q, T4R);
Chris@42 1636 T7S = VADD(T7O, T7R);
Chris@42 1637 ST(&(ii[WS(rs, 3)]), VADD(T7N, T7S), ms, &(ii[WS(rs, 1)]));
Chris@42 1638 ST(&(ii[WS(rs, 19)]), VSUB(T7S, T7N), ms, &(ii[WS(rs, 1)]));
Chris@42 1639 }
Chris@42 1640 {
Chris@42 1641 V T4P, T4S, T7T, T7U;
Chris@42 1642 T4P = VSUB(T4D, T4G);
Chris@42 1643 T4S = VSUB(T4Q, T4R);
Chris@42 1644 ST(&(ri[WS(rs, 27)]), VSUB(T4P, T4S), ms, &(ri[WS(rs, 1)]));
Chris@42 1645 ST(&(ri[WS(rs, 11)]), VADD(T4P, T4S), ms, &(ri[WS(rs, 1)]));
Chris@42 1646 T7T = VSUB(T4N, T4K);
Chris@42 1647 T7U = VSUB(T7R, T7O);
Chris@42 1648 ST(&(ii[WS(rs, 11)]), VADD(T7T, T7U), ms, &(ii[WS(rs, 1)]));
Chris@42 1649 ST(&(ii[WS(rs, 27)]), VSUB(T7U, T7T), ms, &(ii[WS(rs, 1)]));
Chris@42 1650 }
Chris@42 1651 }
Chris@42 1652 {
Chris@42 1653 V T4X, T5p, T7D, T7J, T54, T7y, T5z, T5D, T5c, T5m, T5s, T7I, T5w, T5C, T5j;
Chris@42 1654 V T5n, T4W, T7z;
Chris@42 1655 T4W = VMUL(LDK(KP707106781), VADD(T4U, T4V));
Chris@42 1656 T4X = VSUB(T4T, T4W);
Chris@42 1657 T5p = VADD(T4T, T4W);
Chris@42 1658 T7z = VMUL(LDK(KP707106781), VADD(T3a, T3f));
Chris@42 1659 T7D = VADD(T7z, T7C);
Chris@42 1660 T7J = VSUB(T7C, T7z);
Chris@42 1661 {
Chris@42 1662 V T50, T53, T5x, T5y;
Chris@42 1663 T50 = VFNMS(LDK(KP382683432), T4Z, VMUL(LDK(KP923879532), T4Y));
Chris@42 1664 T53 = VFMA(LDK(KP923879532), T51, VMUL(LDK(KP382683432), T52));
Chris@42 1665 T54 = VSUB(T50, T53);
Chris@42 1666 T7y = VADD(T50, T53);
Chris@42 1667 T5x = VADD(T5d, T5e);
Chris@42 1668 T5y = VADD(T5g, T5h);
Chris@42 1669 T5z = VFNMS(LDK(KP195090322), T5y, VMUL(LDK(KP980785280), T5x));
Chris@42 1670 T5D = VFMA(LDK(KP195090322), T5x, VMUL(LDK(KP980785280), T5y));
Chris@42 1671 }
Chris@42 1672 {
Chris@42 1673 V T58, T5b, T5q, T5r;
Chris@42 1674 T58 = VSUB(T56, T57);
Chris@42 1675 T5b = VSUB(T59, T5a);
Chris@42 1676 T5c = VFMA(LDK(KP555570233), T58, VMUL(LDK(KP831469612), T5b));
Chris@42 1677 T5m = VFNMS(LDK(KP831469612), T58, VMUL(LDK(KP555570233), T5b));
Chris@42 1678 T5q = VFMA(LDK(KP382683432), T4Y, VMUL(LDK(KP923879532), T4Z));
Chris@42 1679 T5r = VFNMS(LDK(KP382683432), T51, VMUL(LDK(KP923879532), T52));
Chris@42 1680 T5s = VADD(T5q, T5r);
Chris@42 1681 T7I = VSUB(T5r, T5q);
Chris@42 1682 }
Chris@42 1683 {
Chris@42 1684 V T5u, T5v, T5f, T5i;
Chris@42 1685 T5u = VADD(T56, T57);
Chris@42 1686 T5v = VADD(T59, T5a);
Chris@42 1687 T5w = VFMA(LDK(KP980785280), T5u, VMUL(LDK(KP195090322), T5v));
Chris@42 1688 T5C = VFNMS(LDK(KP195090322), T5u, VMUL(LDK(KP980785280), T5v));
Chris@42 1689 T5f = VSUB(T5d, T5e);
Chris@42 1690 T5i = VSUB(T5g, T5h);
Chris@42 1691 T5j = VFNMS(LDK(KP831469612), T5i, VMUL(LDK(KP555570233), T5f));
Chris@42 1692 T5n = VFMA(LDK(KP831469612), T5f, VMUL(LDK(KP555570233), T5i));
Chris@42 1693 }
Chris@42 1694 {
Chris@42 1695 V T55, T5k, T7H, T7K;
Chris@42 1696 T55 = VADD(T4X, T54);
Chris@42 1697 T5k = VADD(T5c, T5j);
Chris@42 1698 ST(&(ri[WS(rs, 21)]), VSUB(T55, T5k), ms, &(ri[WS(rs, 1)]));
Chris@42 1699 ST(&(ri[WS(rs, 5)]), VADD(T55, T5k), ms, &(ri[WS(rs, 1)]));
Chris@42 1700 T7H = VADD(T5m, T5n);
Chris@42 1701 T7K = VADD(T7I, T7J);
Chris@42 1702 ST(&(ii[WS(rs, 5)]), VADD(T7H, T7K), ms, &(ii[WS(rs, 1)]));
Chris@42 1703 ST(&(ii[WS(rs, 21)]), VSUB(T7K, T7H), ms, &(ii[WS(rs, 1)]));
Chris@42 1704 }
Chris@42 1705 {
Chris@42 1706 V T5l, T5o, T7L, T7M;
Chris@42 1707 T5l = VSUB(T4X, T54);
Chris@42 1708 T5o = VSUB(T5m, T5n);
Chris@42 1709 ST(&(ri[WS(rs, 29)]), VSUB(T5l, T5o), ms, &(ri[WS(rs, 1)]));
Chris@42 1710 ST(&(ri[WS(rs, 13)]), VADD(T5l, T5o), ms, &(ri[WS(rs, 1)]));
Chris@42 1711 T7L = VSUB(T5j, T5c);
Chris@42 1712 T7M = VSUB(T7J, T7I);
Chris@42 1713 ST(&(ii[WS(rs, 13)]), VADD(T7L, T7M), ms, &(ii[WS(rs, 1)]));
Chris@42 1714 ST(&(ii[WS(rs, 29)]), VSUB(T7M, T7L), ms, &(ii[WS(rs, 1)]));
Chris@42 1715 }
Chris@42 1716 {
Chris@42 1717 V T5t, T5A, T7x, T7E;
Chris@42 1718 T5t = VADD(T5p, T5s);
Chris@42 1719 T5A = VADD(T5w, T5z);
Chris@42 1720 ST(&(ri[WS(rs, 17)]), VSUB(T5t, T5A), ms, &(ri[WS(rs, 1)]));
Chris@42 1721 ST(&(ri[WS(rs, 1)]), VADD(T5t, T5A), ms, &(ri[WS(rs, 1)]));
Chris@42 1722 T7x = VADD(T5C, T5D);
Chris@42 1723 T7E = VADD(T7y, T7D);
Chris@42 1724 ST(&(ii[WS(rs, 1)]), VADD(T7x, T7E), ms, &(ii[WS(rs, 1)]));
Chris@42 1725 ST(&(ii[WS(rs, 17)]), VSUB(T7E, T7x), ms, &(ii[WS(rs, 1)]));
Chris@42 1726 }
Chris@42 1727 {
Chris@42 1728 V T5B, T5E, T7F, T7G;
Chris@42 1729 T5B = VSUB(T5p, T5s);
Chris@42 1730 T5E = VSUB(T5C, T5D);
Chris@42 1731 ST(&(ri[WS(rs, 25)]), VSUB(T5B, T5E), ms, &(ri[WS(rs, 1)]));
Chris@42 1732 ST(&(ri[WS(rs, 9)]), VADD(T5B, T5E), ms, &(ri[WS(rs, 1)]));
Chris@42 1733 T7F = VSUB(T5z, T5w);
Chris@42 1734 T7G = VSUB(T7D, T7y);
Chris@42 1735 ST(&(ii[WS(rs, 9)]), VADD(T7F, T7G), ms, &(ii[WS(rs, 1)]));
Chris@42 1736 ST(&(ii[WS(rs, 25)]), VSUB(T7G, T7F), ms, &(ii[WS(rs, 1)]));
Chris@42 1737 }
Chris@42 1738 }
Chris@42 1739 }
Chris@42 1740 }
Chris@42 1741 VLEAVE();
Chris@42 1742 }
Chris@42 1743
Chris@42 1744 static const tw_instr twinstr[] = {
Chris@42 1745 VTW(0, 1),
Chris@42 1746 VTW(0, 2),
Chris@42 1747 VTW(0, 3),
Chris@42 1748 VTW(0, 4),
Chris@42 1749 VTW(0, 5),
Chris@42 1750 VTW(0, 6),
Chris@42 1751 VTW(0, 7),
Chris@42 1752 VTW(0, 8),
Chris@42 1753 VTW(0, 9),
Chris@42 1754 VTW(0, 10),
Chris@42 1755 VTW(0, 11),
Chris@42 1756 VTW(0, 12),
Chris@42 1757 VTW(0, 13),
Chris@42 1758 VTW(0, 14),
Chris@42 1759 VTW(0, 15),
Chris@42 1760 VTW(0, 16),
Chris@42 1761 VTW(0, 17),
Chris@42 1762 VTW(0, 18),
Chris@42 1763 VTW(0, 19),
Chris@42 1764 VTW(0, 20),
Chris@42 1765 VTW(0, 21),
Chris@42 1766 VTW(0, 22),
Chris@42 1767 VTW(0, 23),
Chris@42 1768 VTW(0, 24),
Chris@42 1769 VTW(0, 25),
Chris@42 1770 VTW(0, 26),
Chris@42 1771 VTW(0, 27),
Chris@42 1772 VTW(0, 28),
Chris@42 1773 VTW(0, 29),
Chris@42 1774 VTW(0, 30),
Chris@42 1775 VTW(0, 31),
Chris@42 1776 {TW_NEXT, (2 * VL), 0}
Chris@42 1777 };
Chris@42 1778
Chris@42 1779 static const ct_desc desc = { 32, XSIMD_STRING("t1sv_32"), twinstr, &GENUS, {340, 114, 94, 0}, 0, 0, 0 };
Chris@42 1780
Chris@42 1781 void XSIMD(codelet_t1sv_32) (planner *p) {
Chris@42 1782 X(kdft_dit_register) (p, t1sv_32, &desc);
Chris@42 1783 }
Chris@42 1784 #endif /* HAVE_FMA */