Mercurial > hg > sv-dependency-builds

annotate src/fftw-3.3.5/rdft/scalar/r2cf/r2cfII_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 16:47:27 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_r2cf.native -fma -reorder-insns -schedule-for-pipeline -compact -variables 4 -pipeline-latency 4 -n 16 -name r2cfII_16 -dft-II -include r2cfII.h */
Chris@42 29
Chris@42 30 /*
Chris@42 31 * This function contains 66 FP additions, 48 FP multiplications,
Chris@42 32 * (or, 18 additions, 0 multiplications, 48 fused multiply/add),
Chris@42 33 * 54 stack variables, 7 constants, and 32 memory accesses
Chris@42 34 */
Chris@42 35 #include "r2cfII.h"
Chris@42 36
Chris@42 37 static void r2cfII_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(KP980785280, +0.980785280403230449126182236134239036973933731);
Chris@42 40 DK(KP198912367, +0.198912367379658006911597622644676228597850501);
Chris@42 41 DK(KP831469612, +0.831469612302545237078788377617905756738560812);
Chris@42 42 DK(KP923879532, +0.923879532511286756128183189396788286822416626);
Chris@42 43 DK(KP668178637, +0.668178637919298919997757686523080761552472251);
Chris@42 44 DK(KP414213562, +0.414213562373095048801688724209698078569671875);
Chris@42 45 DK(KP707106781, +0.707106781186547524400844362104849039284835938);
Chris@42 46 {
Chris@42 47 INT i;
Chris@42 48 for (i = v; i > 0; i = i - 1, R0 = R0 + ivs, R1 = R1 + ivs, Cr = Cr + ovs, Ci = Ci + ovs, MAKE_VOLATILE_STRIDE(64, rs), MAKE_VOLATILE_STRIDE(64, csr), MAKE_VOLATILE_STRIDE(64, csi)) {
Chris@42 49 E TN, TF, TX, TV, TO, TP, TY, TM, TQ, TW;
Chris@42 50 {
Chris@42 51 E TT, TZ, TB, T5, Tu, TK, TJ, Tr, T9, TC, T8, Tl, TH, TG, Ti;
Chris@42 52 E Ta;
Chris@42 53 {
Chris@42 54 E T1, TR, Tn, Ts, To, TS, T4, Tp, T2, T3;
Chris@42 55 T1 = R0[0];
Chris@42 56 TR = R0[WS(rs, 4)];
Chris@42 57 T2 = R0[WS(rs, 2)];
Chris@42 58 T3 = R0[WS(rs, 6)];
Chris@42 59 Tn = R1[WS(rs, 7)];
Chris@42 60 Ts = R1[WS(rs, 3)];
Chris@42 61 To = R1[WS(rs, 1)];
Chris@42 62 TS = T2 + T3;
Chris@42 63 T4 = T2 - T3;
Chris@42 64 Tp = R1[WS(rs, 5)];
Chris@42 65 {
Chris@42 66 E Te, Tj, Tf, Tg, Tt, Tq;
Chris@42 67 Te = R1[0];
Chris@42 68 TT = FMA(KP707106781, TS, TR);
Chris@42 69 TZ = FNMS(KP707106781, TS, TR);
Chris@42 70 TB = FMA(KP707106781, T4, T1);
Chris@42 71 T5 = FNMS(KP707106781, T4, T1);
Chris@42 72 Tt = To + Tp;
Chris@42 73 Tq = To - Tp;
Chris@42 74 Tj = R1[WS(rs, 4)];
Chris@42 75 Tf = R1[WS(rs, 2)];
Chris@42 76 Tu = FNMS(KP707106781, Tt, Ts);
Chris@42 77 TK = FMA(KP707106781, Tt, Ts);
Chris@42 78 TJ = FMS(KP707106781, Tq, Tn);
Chris@42 79 Tr = FMA(KP707106781, Tq, Tn);
Chris@42 80 Tg = R1[WS(rs, 6)];
Chris@42 81 {
Chris@42 82 E T6, T7, Tk, Th;
Chris@42 83 T6 = R0[WS(rs, 5)];
Chris@42 84 T7 = R0[WS(rs, 1)];
Chris@42 85 T9 = R0[WS(rs, 3)];
Chris@42 86 Tk = Tf + Tg;
Chris@42 87 Th = Tf - Tg;
Chris@42 88 TC = FNMS(KP414213562, T6, T7);
Chris@42 89 T8 = FMA(KP414213562, T7, T6);
Chris@42 90 Tl = FNMS(KP707106781, Tk, Tj);
Chris@42 91 TH = FMA(KP707106781, Tk, Tj);
Chris@42 92 TG = FMA(KP707106781, Th, Te);
Chris@42 93 Ti = FNMS(KP707106781, Th, Te);
Chris@42 94 Ta = R0[WS(rs, 7)];
Chris@42 95 }
Chris@42 96 }
Chris@42 97 }
Chris@42 98 {
Chris@42 99 E TE, TU, Ty, Tv, TI, TL;
Chris@42 100 Ty = FNMS(KP668178637, Tr, Tu);
Chris@42 101 Tv = FMA(KP668178637, Tu, Tr);
Chris@42 102 {
Chris@42 103 E Tw, T14, T12, TA, T11, T13, Tx, Td;
Chris@42 104 {
Chris@42 105 E Tz, Tm, TD, Tb, T10, Tc;
Chris@42 106 Tz = FNMS(KP668178637, Ti, Tl);
Chris@42 107 Tm = FMA(KP668178637, Tl, Ti);
Chris@42 108 TD = FMS(KP414213562, T9, Ta);
Chris@42 109 Tb = FMA(KP414213562, Ta, T9);
Chris@42 110 Tw = Tm - Tv;
Chris@42 111 T14 = Tm + Tv;
Chris@42 112 T10 = TD - TC;
Chris@42 113 TE = TC + TD;
Chris@42 114 Tc = T8 - Tb;
Chris@42 115 TU = T8 + Tb;
Chris@42 116 T12 = Tz + Ty;
Chris@42 117 TA = Ty - Tz;
Chris@42 118 T11 = FMA(KP923879532, T10, TZ);
Chris@42 119 T13 = FNMS(KP923879532, T10, TZ);
Chris@42 120 Tx = FNMS(KP923879532, Tc, T5);
Chris@42 121 Td = FMA(KP923879532, Tc, T5);
Chris@42 122 }
Chris@42 123 Ci[WS(csi, 2)] = -(FMA(KP831469612, T14, T13));
Chris@42 124 Ci[WS(csi, 5)] = FNMS(KP831469612, T14, T13);
Chris@42 125 Cr[WS(csr, 1)] = FMA(KP831469612, Tw, Td);
Chris@42 126 Cr[WS(csr, 6)] = FNMS(KP831469612, Tw, Td);
Chris@42 127 Cr[WS(csr, 5)] = FNMS(KP831469612, TA, Tx);
Chris@42 128 Ci[WS(csi, 1)] = FMA(KP831469612, T12, T11);
Chris@42 129 Cr[WS(csr, 2)] = FMA(KP831469612, TA, Tx);
Chris@42 130 Ci[WS(csi, 6)] = FMS(KP831469612, T12, T11);
Chris@42 131 }
Chris@42 132 TN = FNMS(KP923879532, TE, TB);
Chris@42 133 TF = FMA(KP923879532, TE, TB);
Chris@42 134 TX = FNMS(KP923879532, TU, TT);
Chris@42 135 TV = FMA(KP923879532, TU, TT);
Chris@42 136 TO = FMA(KP198912367, TG, TH);
Chris@42 137 TI = FNMS(KP198912367, TH, TG);
Chris@42 138 TL = FMA(KP198912367, TK, TJ);
Chris@42 139 TP = FNMS(KP198912367, TJ, TK);
Chris@42 140 TY = TL - TI;
Chris@42 141 TM = TI + TL;
Chris@42 142 }
Chris@42 143 }
Chris@42 144 Ci[WS(csi, 4)] = FMS(KP980785280, TY, TX);
Chris@42 145 Ci[WS(csi, 3)] = FMA(KP980785280, TY, TX);
Chris@42 146 Cr[0] = FMA(KP980785280, TM, TF);
Chris@42 147 Cr[WS(csr, 7)] = FNMS(KP980785280, TM, TF);
Chris@42 148 TQ = TO - TP;
Chris@42 149 TW = TO + TP;
Chris@42 150 Ci[0] = -(FMA(KP980785280, TW, TV));
Chris@42 151 Ci[WS(csi, 7)] = FNMS(KP980785280, TW, TV);
Chris@42 152 Cr[WS(csr, 3)] = FMA(KP980785280, TQ, TN);
Chris@42 153 Cr[WS(csr, 4)] = FNMS(KP980785280, TQ, TN);
Chris@42 154 }
Chris@42 155 }
Chris@42 156 }
Chris@42 157
Chris@42 158 static const kr2c_desc desc = { 16, "r2cfII_16", {18, 0, 48, 0}, &GENUS };
Chris@42 159
Chris@42 160 void X(codelet_r2cfII_16) (planner *p) {
Chris@42 161 X(kr2c_register) (p, r2cfII_16, &desc);
Chris@42 162 }
Chris@42 163
Chris@42 164 #else /* HAVE_FMA */
Chris@42 165
Chris@42 166 /* Generated by: ../../../genfft/gen_r2cf.native -compact -variables 4 -pipeline-latency 4 -n 16 -name r2cfII_16 -dft-II -include r2cfII.h */
Chris@42 167
Chris@42 168 /*
Chris@42 169 * This function contains 66 FP additions, 30 FP multiplications,
Chris@42 170 * (or, 54 additions, 18 multiplications, 12 fused multiply/add),
Chris@42 171 * 32 stack variables, 7 constants, and 32 memory accesses
Chris@42 172 */
Chris@42 173 #include "r2cfII.h"
Chris@42 174
Chris@42 175 static void r2cfII_16(R *R0, R *R1, R *Cr, R *Ci, stride rs, stride csr, stride csi, INT v, INT ivs, INT ovs)
Chris@42 176 {
Chris@42 177 DK(KP555570233, +0.555570233019602224742830813948532874374937191);
Chris@42 178 DK(KP831469612, +0.831469612302545237078788377617905756738560812);
Chris@42 179 DK(KP980785280, +0.980785280403230449126182236134239036973933731);
Chris@42 180 DK(KP195090322, +0.195090322016128267848284868477022240927691618);
Chris@42 181 DK(KP382683432, +0.382683432365089771728459984030398866761344562);
Chris@42 182 DK(KP923879532, +0.923879532511286756128183189396788286822416626);
Chris@42 183 DK(KP707106781, +0.707106781186547524400844362104849039284835938);
Chris@42 184 {
Chris@42 185 INT i;
Chris@42 186 for (i = v; i > 0; i = i - 1, R0 = R0 + ivs, R1 = R1 + ivs, Cr = Cr + ovs, Ci = Ci + ovs, MAKE_VOLATILE_STRIDE(64, rs), MAKE_VOLATILE_STRIDE(64, csr), MAKE_VOLATILE_STRIDE(64, csi)) {
Chris@42 187 E T5, T11, TB, TV, Tr, TK, Tu, TJ, Ti, TH, Tl, TG, Tc, T10, TE;
Chris@42 188 E TS;
Chris@42 189 {
Chris@42 190 E T1, TU, T4, TT, T2, T3;
Chris@42 191 T1 = R0[0];
Chris@42 192 TU = R0[WS(rs, 4)];
Chris@42 193 T2 = R0[WS(rs, 2)];
Chris@42 194 T3 = R0[WS(rs, 6)];
Chris@42 195 T4 = KP707106781 * (T2 - T3);
Chris@42 196 TT = KP707106781 * (T2 + T3);
Chris@42 197 T5 = T1 + T4;
Chris@42 198 T11 = TU - TT;
Chris@42 199 TB = T1 - T4;
Chris@42 200 TV = TT + TU;
Chris@42 201 }
Chris@42 202 {
Chris@42 203 E Tq, Tt, Tp, Ts, Tn, To;
Chris@42 204 Tq = R1[WS(rs, 7)];
Chris@42 205 Tt = R1[WS(rs, 3)];
Chris@42 206 Tn = R1[WS(rs, 1)];
Chris@42 207 To = R1[WS(rs, 5)];
Chris@42 208 Tp = KP707106781 * (Tn - To);
Chris@42 209 Ts = KP707106781 * (Tn + To);
Chris@42 210 Tr = Tp - Tq;
Chris@42 211 TK = Tt - Ts;
Chris@42 212 Tu = Ts + Tt;
Chris@42 213 TJ = Tp + Tq;
Chris@42 214 }
Chris@42 215 {
Chris@42 216 E Te, Tk, Th, Tj, Tf, Tg;
Chris@42 217 Te = R1[0];
Chris@42 218 Tk = R1[WS(rs, 4)];
Chris@42 219 Tf = R1[WS(rs, 2)];
Chris@42 220 Tg = R1[WS(rs, 6)];
Chris@42 221 Th = KP707106781 * (Tf - Tg);
Chris@42 222 Tj = KP707106781 * (Tf + Tg);
Chris@42 223 Ti = Te + Th;
Chris@42 224 TH = Tk - Tj;
Chris@42 225 Tl = Tj + Tk;
Chris@42 226 TG = Te - Th;
Chris@42 227 }
Chris@42 228 {
Chris@42 229 E T8, TC, Tb, TD;
Chris@42 230 {
Chris@42 231 E T6, T7, T9, Ta;
Chris@42 232 T6 = R0[WS(rs, 1)];
Chris@42 233 T7 = R0[WS(rs, 5)];
Chris@42 234 T8 = FNMS(KP382683432, T7, KP923879532 * T6);
Chris@42 235 TC = FMA(KP382683432, T6, KP923879532 * T7);
Chris@42 236 T9 = R0[WS(rs, 3)];
Chris@42 237 Ta = R0[WS(rs, 7)];
Chris@42 238 Tb = FNMS(KP923879532, Ta, KP382683432 * T9);
Chris@42 239 TD = FMA(KP923879532, T9, KP382683432 * Ta);
Chris@42 240 }
Chris@42 241 Tc = T8 + Tb;
Chris@42 242 T10 = Tb - T8;
Chris@42 243 TE = TC - TD;
Chris@42 244 TS = TC + TD;
Chris@42 245 }
Chris@42 246 {
Chris@42 247 E Td, TW, Tw, TR, Tm, Tv;
Chris@42 248 Td = T5 - Tc;
Chris@42 249 TW = TS + TV;
Chris@42 250 Tm = FMA(KP195090322, Ti, KP980785280 * Tl);
Chris@42 251 Tv = FNMS(KP980785280, Tu, KP195090322 * Tr);
Chris@42 252 Tw = Tm + Tv;
Chris@42 253 TR = Tv - Tm;
Chris@42 254 Cr[WS(csr, 4)] = Td - Tw;
Chris@42 255 Ci[WS(csi, 7)] = TR + TW;
Chris@42 256 Cr[WS(csr, 3)] = Td + Tw;
Chris@42 257 Ci[0] = TR - TW;
Chris@42 258 }
Chris@42 259 {
Chris@42 260 E Tx, TY, TA, TX, Ty, Tz;
Chris@42 261 Tx = T5 + Tc;
Chris@42 262 TY = TV - TS;
Chris@42 263 Ty = FNMS(KP195090322, Tl, KP980785280 * Ti);
Chris@42 264 Tz = FMA(KP980785280, Tr, KP195090322 * Tu);
Chris@42 265 TA = Ty + Tz;
Chris@42 266 TX = Tz - Ty;
Chris@42 267 Cr[WS(csr, 7)] = Tx - TA;
Chris@42 268 Ci[WS(csi, 3)] = TX + TY;
Chris@42 269 Cr[0] = Tx + TA;
Chris@42 270 Ci[WS(csi, 4)] = TX - TY;
Chris@42 271 }
Chris@42 272 {
Chris@42 273 E TF, T12, TM, TZ, TI, TL;
Chris@42 274 TF = TB + TE;
Chris@42 275 T12 = T10 - T11;
Chris@42 276 TI = FMA(KP831469612, TG, KP555570233 * TH);
Chris@42 277 TL = FMA(KP831469612, TJ, KP555570233 * TK);
Chris@42 278 TM = TI - TL;
Chris@42 279 TZ = TI + TL;
Chris@42 280 Cr[WS(csr, 6)] = TF - TM;
Chris@42 281 Ci[WS(csi, 2)] = T12 - TZ;
Chris@42 282 Cr[WS(csr, 1)] = TF + TM;
Chris@42 283 Ci[WS(csi, 5)] = -(TZ + T12);
Chris@42 284 }
Chris@42 285 {
Chris@42 286 E TN, T14, TQ, T13, TO, TP;
Chris@42 287 TN = TB - TE;
Chris@42 288 T14 = T10 + T11;
Chris@42 289 TO = FNMS(KP555570233, TJ, KP831469612 * TK);
Chris@42 290 TP = FNMS(KP555570233, TG, KP831469612 * TH);
Chris@42 291 TQ = TO - TP;
Chris@42 292 T13 = TP + TO;
Chris@42 293 Cr[WS(csr, 5)] = TN - TQ;
Chris@42 294 Ci[WS(csi, 1)] = T13 + T14;
Chris@42 295 Cr[WS(csr, 2)] = TN + TQ;
Chris@42 296 Ci[WS(csi, 6)] = T13 - T14;
Chris@42 297 }
Chris@42 298 }
Chris@42 299 }
Chris@42 300 }
Chris@42 301
Chris@42 302 static const kr2c_desc desc = { 16, "r2cfII_16", {54, 18, 12, 0}, &GENUS };
Chris@42 303
Chris@42 304 void X(codelet_r2cfII_16) (planner *p) {
Chris@42 305 X(kr2c_register) (p, r2cfII_16, &desc);
Chris@42 306 }
Chris@42 307
Chris@42 308 #endif /* HAVE_FMA */