Mercurial > hg > sv-dependency-builds

annotate src/fftw-3.3.5/rdft/scalar/r2cb/r2cbIII_16.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 16:50:42 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_r2cb.native -fma -reorder-insns -schedule-for-pipeline -compact -variables 4 -pipeline-latency 4 -sign 1 -n 16 -name r2cbIII_16 -dft-III -include r2cbIII.h */
Chris@42 29
Chris@42 30 /*
Chris@42 31 * This function contains 66 FP additions, 36 FP multiplications,
Chris@42 32 * (or, 46 additions, 16 multiplications, 20 fused multiply/add),
Chris@42 33 * 55 stack variables, 9 constants, and 32 memory accesses
Chris@42 34 */
Chris@42 35 #include "r2cbIII.h"
Chris@42 36
Chris@42 37 static void r2cbIII_16(R *R0, R *R1, R *Cr, R *Ci, stride rs, stride csr, stride csi, INT v, INT ivs, INT ovs)
Chris@42 38 {
Chris@42 39 DK(KP668178637, +0.668178637919298919997757686523080761552472251);
Chris@42 40 DK(KP1_662939224, +1.662939224605090474157576755235811513477121624);
Chris@42 41 DK(KP198912367, +0.198912367379658006911597622644676228597850501);
Chris@42 42 DK(KP1_961570560, +1.961570560806460898252364472268478073947867462);
Chris@42 43 DK(KP707106781, +0.707106781186547524400844362104849039284835938);
Chris@42 44 DK(KP1_414213562, +1.414213562373095048801688724209698078569671875);
Chris@42 45 DK(KP414213562, +0.414213562373095048801688724209698078569671875);
Chris@42 46 DK(KP1_847759065, +1.847759065022573512256366378793576573644833252);
Chris@42 47 DK(KP2_000000000, +2.000000000000000000000000000000000000000000000);
Chris@42 48 {
Chris@42 49 INT i;
Chris@42 50 for (i = v; i > 0; i = i - 1, R0 = R0 + ovs, R1 = R1 + ovs, Cr = Cr + ivs, Ci = Ci + ivs, MAKE_VOLATILE_STRIDE(64, rs), MAKE_VOLATILE_STRIDE(64, csr), MAKE_VOLATILE_STRIDE(64, csi)) {
Chris@42 51 E TA, TD, Tv, TG, TE, TF;
Chris@42 52 {
Chris@42 53 E TK, TP, T7, T13, TW, TH, Tj, TC, To, Te, TX, TS, T12, Tt, TB;
Chris@42 54 {
Chris@42 55 E T4, Tf, T3, TU, Tz, T5, Tg, Th;
Chris@42 56 {
Chris@42 57 E T1, T2, Tx, Ty;
Chris@42 58 T1 = Cr[0];
Chris@42 59 T2 = Cr[WS(csr, 7)];
Chris@42 60 Tx = Ci[0];
Chris@42 61 Ty = Ci[WS(csi, 7)];
Chris@42 62 T4 = Cr[WS(csr, 4)];
Chris@42 63 Tf = T1 - T2;
Chris@42 64 T3 = T1 + T2;
Chris@42 65 TU = Ty - Tx;
Chris@42 66 Tz = Tx + Ty;
Chris@42 67 T5 = Cr[WS(csr, 3)];
Chris@42 68 Tg = Ci[WS(csi, 4)];
Chris@42 69 Th = Ci[WS(csi, 3)];
Chris@42 70 }
Chris@42 71 {
Chris@42 72 E Tb, Tk, Ta, TR, Tn, Tc, Tq, Tr;
Chris@42 73 {
Chris@42 74 E T8, T9, Tl, Tm;
Chris@42 75 T8 = Cr[WS(csr, 2)];
Chris@42 76 {
Chris@42 77 E Tw, T6, TV, Ti;
Chris@42 78 Tw = T4 - T5;
Chris@42 79 T6 = T4 + T5;
Chris@42 80 TV = Th - Tg;
Chris@42 81 Ti = Tg + Th;
Chris@42 82 TK = Tw - Tz;
Chris@42 83 TA = Tw + Tz;
Chris@42 84 TP = T3 - T6;
Chris@42 85 T7 = T3 + T6;
Chris@42 86 T13 = TV + TU;
Chris@42 87 TW = TU - TV;
Chris@42 88 TH = Tf + Ti;
Chris@42 89 Tj = Tf - Ti;
Chris@42 90 T9 = Cr[WS(csr, 5)];
Chris@42 91 }
Chris@42 92 Tl = Ci[WS(csi, 2)];
Chris@42 93 Tm = Ci[WS(csi, 5)];
Chris@42 94 Tb = Cr[WS(csr, 1)];
Chris@42 95 Tk = T8 - T9;
Chris@42 96 Ta = T8 + T9;
Chris@42 97 TR = Tl - Tm;
Chris@42 98 Tn = Tl + Tm;
Chris@42 99 Tc = Cr[WS(csr, 6)];
Chris@42 100 Tq = Ci[WS(csi, 1)];
Chris@42 101 Tr = Ci[WS(csi, 6)];
Chris@42 102 }
Chris@42 103 TC = Tk + Tn;
Chris@42 104 To = Tk - Tn;
Chris@42 105 {
Chris@42 106 E Tp, Td, TQ, Ts;
Chris@42 107 Tp = Tb - Tc;
Chris@42 108 Td = Tb + Tc;
Chris@42 109 TQ = Tr - Tq;
Chris@42 110 Ts = Tq + Tr;
Chris@42 111 Te = Ta + Td;
Chris@42 112 TX = Ta - Td;
Chris@42 113 TS = TQ - TR;
Chris@42 114 T12 = TR + TQ;
Chris@42 115 Tt = Tp - Ts;
Chris@42 116 TB = Tp + Ts;
Chris@42 117 }
Chris@42 118 }
Chris@42 119 }
Chris@42 120 {
Chris@42 121 E T10, TT, TY, TZ;
Chris@42 122 R0[0] = KP2_000000000 * (T7 + Te);
Chris@42 123 R0[WS(rs, 4)] = KP2_000000000 * (T13 - T12);
Chris@42 124 T10 = TP - TS;
Chris@42 125 TT = TP + TS;
Chris@42 126 TY = TW - TX;
Chris@42 127 TZ = TX + TW;
Chris@42 128 {
Chris@42 129 E T11, T14, TI, TL, Tu;
Chris@42 130 T11 = T7 - Te;
Chris@42 131 T14 = T12 + T13;
Chris@42 132 R0[WS(rs, 5)] = KP1_847759065 * (FNMS(KP414213562, TT, TY));
Chris@42 133 R0[WS(rs, 1)] = KP1_847759065 * (FMA(KP414213562, TY, TT));
Chris@42 134 R0[WS(rs, 6)] = KP1_414213562 * (T14 - T11);
Chris@42 135 R0[WS(rs, 2)] = KP1_414213562 * (T11 + T14);
Chris@42 136 TD = TB - TC;
Chris@42 137 TI = TC + TB;
Chris@42 138 TL = To - Tt;
Chris@42 139 Tu = To + Tt;
Chris@42 140 {
Chris@42 141 E TO, TJ, TN, TM;
Chris@42 142 R0[WS(rs, 7)] = -(KP1_847759065 * (FNMS(KP414213562, TZ, T10)));
Chris@42 143 R0[WS(rs, 3)] = KP1_847759065 * (FMA(KP414213562, T10, TZ));
Chris@42 144 TO = FMA(KP707106781, TI, TH);
Chris@42 145 TJ = FNMS(KP707106781, TI, TH);
Chris@42 146 TN = FMA(KP707106781, TL, TK);
Chris@42 147 TM = FNMS(KP707106781, TL, TK);
Chris@42 148 Tv = FMA(KP707106781, Tu, Tj);
Chris@42 149 TG = FNMS(KP707106781, Tu, Tj);
Chris@42 150 R1[WS(rs, 3)] = KP1_961570560 * (FMA(KP198912367, TO, TN));
Chris@42 151 R1[WS(rs, 7)] = -(KP1_961570560 * (FNMS(KP198912367, TN, TO)));
Chris@42 152 R1[WS(rs, 5)] = KP1_662939224 * (FNMS(KP668178637, TJ, TM));
Chris@42 153 R1[WS(rs, 1)] = KP1_662939224 * (FMA(KP668178637, TM, TJ));
Chris@42 154 }
Chris@42 155 }
Chris@42 156 }
Chris@42 157 }
Chris@42 158 TE = FNMS(KP707106781, TD, TA);
Chris@42 159 TF = FMA(KP707106781, TD, TA);
Chris@42 160 R1[WS(rs, 2)] = -(KP1_662939224 * (FNMS(KP668178637, TG, TF)));
Chris@42 161 R1[WS(rs, 6)] = -(KP1_662939224 * (FMA(KP668178637, TF, TG)));
Chris@42 162 R1[WS(rs, 4)] = -(KP1_961570560 * (FMA(KP198912367, Tv, TE)));
Chris@42 163 R1[0] = KP1_961570560 * (FNMS(KP198912367, TE, Tv));
Chris@42 164 }
Chris@42 165 }
Chris@42 166 }
Chris@42 167
Chris@42 168 static const kr2c_desc desc = { 16, "r2cbIII_16", {46, 16, 20, 0}, &GENUS };
Chris@42 169
Chris@42 170 void X(codelet_r2cbIII_16) (planner *p) {
Chris@42 171 X(kr2c_register) (p, r2cbIII_16, &desc);
Chris@42 172 }
Chris@42 173
Chris@42 174 #else /* HAVE_FMA */
Chris@42 175
Chris@42 176 /* Generated by: ../../../genfft/gen_r2cb.native -compact -variables 4 -pipeline-latency 4 -sign 1 -n 16 -name r2cbIII_16 -dft-III -include r2cbIII.h */
Chris@42 177
Chris@42 178 /*
Chris@42 179 * This function contains 66 FP additions, 32 FP multiplications,
Chris@42 180 * (or, 54 additions, 20 multiplications, 12 fused multiply/add),
Chris@42 181 * 40 stack variables, 9 constants, and 32 memory accesses
Chris@42 182 */
Chris@42 183 #include "r2cbIII.h"
Chris@42 184
Chris@42 185 static void r2cbIII_16(R *R0, R *R1, R *Cr, R *Ci, stride rs, stride csr, stride csi, INT v, INT ivs, INT ovs)
Chris@42 186 {
Chris@42 187 DK(KP1_961570560, +1.961570560806460898252364472268478073947867462);
Chris@42 188 DK(KP390180644, +0.390180644032256535696569736954044481855383236);
Chris@42 189 DK(KP1_111140466, +1.111140466039204449485661627897065748749874382);
Chris@42 190 DK(KP1_662939224, +1.662939224605090474157576755235811513477121624);
Chris@42 191 DK(KP707106781, +0.707106781186547524400844362104849039284835938);
Chris@42 192 DK(KP1_414213562, +1.414213562373095048801688724209698078569671875);
Chris@42 193 DK(KP765366864, +0.765366864730179543456919968060797733522689125);
Chris@42 194 DK(KP1_847759065, +1.847759065022573512256366378793576573644833252);
Chris@42 195 DK(KP2_000000000, +2.000000000000000000000000000000000000000000000);
Chris@42 196 {
Chris@42 197 INT i;
Chris@42 198 for (i = v; i > 0; i = i - 1, R0 = R0 + ovs, R1 = R1 + ovs, Cr = Cr + ivs, Ci = Ci + ivs, MAKE_VOLATILE_STRIDE(64, rs), MAKE_VOLATILE_STRIDE(64, csr), MAKE_VOLATILE_STRIDE(64, csi)) {
Chris@42 199 E T7, TW, T13, Tj, TD, TK, TP, TH, Te, TX, T12, To, Tt, Tx, TS;
Chris@42 200 E Tw, TT, TY;
Chris@42 201 {
Chris@42 202 E T3, Tf, TC, TV, T6, Tz, Ti, TU;
Chris@42 203 {
Chris@42 204 E T1, T2, TA, TB;
Chris@42 205 T1 = Cr[0];
Chris@42 206 T2 = Cr[WS(csr, 7)];
Chris@42 207 T3 = T1 + T2;
Chris@42 208 Tf = T1 - T2;
Chris@42 209 TA = Ci[0];
Chris@42 210 TB = Ci[WS(csi, 7)];
Chris@42 211 TC = TA + TB;
Chris@42 212 TV = TB - TA;
Chris@42 213 }
Chris@42 214 {
Chris@42 215 E T4, T5, Tg, Th;
Chris@42 216 T4 = Cr[WS(csr, 4)];
Chris@42 217 T5 = Cr[WS(csr, 3)];
Chris@42 218 T6 = T4 + T5;
Chris@42 219 Tz = T4 - T5;
Chris@42 220 Tg = Ci[WS(csi, 4)];
Chris@42 221 Th = Ci[WS(csi, 3)];
Chris@42 222 Ti = Tg + Th;
Chris@42 223 TU = Tg - Th;
Chris@42 224 }
Chris@42 225 T7 = T3 + T6;
Chris@42 226 TW = TU + TV;
Chris@42 227 T13 = TV - TU;
Chris@42 228 Tj = Tf - Ti;
Chris@42 229 TD = Tz + TC;
Chris@42 230 TK = Tz - TC;
Chris@42 231 TP = T3 - T6;
Chris@42 232 TH = Tf + Ti;
Chris@42 233 }
Chris@42 234 {
Chris@42 235 E Ta, Tk, Tn, TR, Td, Tp, Ts, TQ;
Chris@42 236 {
Chris@42 237 E T8, T9, Tl, Tm;
Chris@42 238 T8 = Cr[WS(csr, 2)];
Chris@42 239 T9 = Cr[WS(csr, 5)];
Chris@42 240 Ta = T8 + T9;
Chris@42 241 Tk = T8 - T9;
Chris@42 242 Tl = Ci[WS(csi, 2)];
Chris@42 243 Tm = Ci[WS(csi, 5)];
Chris@42 244 Tn = Tl + Tm;
Chris@42 245 TR = Tl - Tm;
Chris@42 246 }
Chris@42 247 {
Chris@42 248 E Tb, Tc, Tq, Tr;
Chris@42 249 Tb = Cr[WS(csr, 1)];
Chris@42 250 Tc = Cr[WS(csr, 6)];
Chris@42 251 Td = Tb + Tc;
Chris@42 252 Tp = Tb - Tc;
Chris@42 253 Tq = Ci[WS(csi, 1)];
Chris@42 254 Tr = Ci[WS(csi, 6)];
Chris@42 255 Ts = Tq + Tr;
Chris@42 256 TQ = Tr - Tq;
Chris@42 257 }
Chris@42 258 Te = Ta + Td;
Chris@42 259 TX = Ta - Td;
Chris@42 260 T12 = TR + TQ;
Chris@42 261 To = Tk - Tn;
Chris@42 262 Tt = Tp - Ts;
Chris@42 263 Tx = Tp + Ts;
Chris@42 264 TS = TQ - TR;
Chris@42 265 Tw = Tk + Tn;
Chris@42 266 }
Chris@42 267 R0[0] = KP2_000000000 * (T7 + Te);
Chris@42 268 R0[WS(rs, 4)] = KP2_000000000 * (T13 - T12);
Chris@42 269 TT = TP + TS;
Chris@42 270 TY = TW - TX;
Chris@42 271 R0[WS(rs, 1)] = FMA(KP1_847759065, TT, KP765366864 * TY);
Chris@42 272 R0[WS(rs, 5)] = FNMS(KP765366864, TT, KP1_847759065 * TY);
Chris@42 273 {
Chris@42 274 E T11, T14, TZ, T10;
Chris@42 275 T11 = T7 - Te;
Chris@42 276 T14 = T12 + T13;
Chris@42 277 R0[WS(rs, 2)] = KP1_414213562 * (T11 + T14);
Chris@42 278 R0[WS(rs, 6)] = KP1_414213562 * (T14 - T11);
Chris@42 279 TZ = TP - TS;
Chris@42 280 T10 = TX + TW;
Chris@42 281 R0[WS(rs, 3)] = FMA(KP765366864, TZ, KP1_847759065 * T10);
Chris@42 282 R0[WS(rs, 7)] = FNMS(KP1_847759065, TZ, KP765366864 * T10);
Chris@42 283 }
Chris@42 284 {
Chris@42 285 E TJ, TN, TM, TO, TI, TL;
Chris@42 286 TI = KP707106781 * (Tw + Tx);
Chris@42 287 TJ = TH - TI;
Chris@42 288 TN = TH + TI;
Chris@42 289 TL = KP707106781 * (To - Tt);
Chris@42 290 TM = TK - TL;
Chris@42 291 TO = TL + TK;
Chris@42 292 R1[WS(rs, 1)] = FMA(KP1_662939224, TJ, KP1_111140466 * TM);
Chris@42 293 R1[WS(rs, 7)] = FNMS(KP1_961570560, TN, KP390180644 * TO);
Chris@42 294 R1[WS(rs, 5)] = FNMS(KP1_111140466, TJ, KP1_662939224 * TM);
Chris@42 295 R1[WS(rs, 3)] = FMA(KP390180644, TN, KP1_961570560 * TO);
Chris@42 296 }
Chris@42 297 {
Chris@42 298 E Tv, TF, TE, TG, Tu, Ty;
Chris@42 299 Tu = KP707106781 * (To + Tt);
Chris@42 300 Tv = Tj + Tu;
Chris@42 301 TF = Tj - Tu;
Chris@42 302 Ty = KP707106781 * (Tw - Tx);
Chris@42 303 TE = Ty + TD;
Chris@42 304 TG = Ty - TD;
Chris@42 305 R1[0] = FNMS(KP390180644, TE, KP1_961570560 * Tv);
Chris@42 306 R1[WS(rs, 6)] = FNMS(KP1_662939224, TF, KP1_111140466 * TG);
Chris@42 307 R1[WS(rs, 4)] = -(FMA(KP390180644, Tv, KP1_961570560 * TE));
Chris@42 308 R1[WS(rs, 2)] = FMA(KP1_111140466, TF, KP1_662939224 * TG);
Chris@42 309 }
Chris@42 310 }
Chris@42 311 }
Chris@42 312 }
Chris@42 313
Chris@42 314 static const kr2c_desc desc = { 16, "r2cbIII_16", {54, 20, 12, 0}, &GENUS };
Chris@42 315
Chris@42 316 void X(codelet_r2cbIII_16) (planner *p) {
Chris@42 317 X(kr2c_register) (p, r2cbIII_16, &desc);
Chris@42 318 }
Chris@42 319
Chris@42 320 #endif /* HAVE_FMA */