Mercurial > hg > sv-dependency-builds

annotate src/fftw-3.3.5/rdft/scalar/r2cf/r2cfII_15.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 16:47:26 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 15 -name r2cfII_15 -dft-II -include r2cfII.h */
Chris@42 29
Chris@42 30 /*
Chris@42 31 * This function contains 72 FP additions, 41 FP multiplications,
Chris@42 32 * (or, 38 additions, 7 multiplications, 34 fused multiply/add),
Chris@42 33 * 57 stack variables, 12 constants, and 30 memory accesses
Chris@42 34 */
Chris@42 35 #include "r2cfII.h"
Chris@42 36
Chris@42 37 static void r2cfII_15(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(KP823639103, +0.823639103546331925877420039278190003029660514);
Chris@42 40 DK(KP910592997, +0.910592997310029334643087372129977886038870291);
Chris@42 41 DK(KP951056516, +0.951056516295153572116439333379382143405698634);
Chris@42 42 DK(KP866025403, +0.866025403784438646763723170752936183471402627);
Chris@42 43 DK(KP500000000, +0.500000000000000000000000000000000000000000000);
Chris@42 44 DK(KP559016994, +0.559016994374947424102293417182819058860154590);
Chris@42 45 DK(KP690983005, +0.690983005625052575897706582817180941139845410);
Chris@42 46 DK(KP552786404, +0.552786404500042060718165266253744752911876328);
Chris@42 47 DK(KP447213595, +0.447213595499957939281834733746255247088123672);
Chris@42 48 DK(KP809016994, +0.809016994374947424102293417182819058860154590);
Chris@42 49 DK(KP618033988, +0.618033988749894848204586834365638117720309180);
Chris@42 50 DK(KP250000000, +0.250000000000000000000000000000000000000000000);
Chris@42 51 {
Chris@42 52 INT i;
Chris@42 53 for (i = v; i > 0; i = i - 1, R0 = R0 + ivs, R1 = R1 + ivs, Cr = Cr + ovs, Ci = Ci + ovs, MAKE_VOLATILE_STRIDE(60, rs), MAKE_VOLATILE_STRIDE(60, csr), MAKE_VOLATILE_STRIDE(60, csi)) {
Chris@42 54 E T9, TQ, TV, TW, Tw, TJ;
Chris@42 55 {
Chris@42 56 E Ta, Tl, Tg, T8, T7, TF, TX, TT, Tm, Th, TM, TZ, Tr, Tn, Tj;
Chris@42 57 E Tz, To, TN, TH, Tp, TO;
Chris@42 58 Ta = R0[WS(rs, 5)];
Chris@42 59 Tl = R1[WS(rs, 2)];
Chris@42 60 {
Chris@42 61 E T1, T2, T5, T3, T4;
Chris@42 62 T1 = R0[0];
Chris@42 63 T2 = R0[WS(rs, 3)];
Chris@42 64 T5 = R1[WS(rs, 4)];
Chris@42 65 T3 = R0[WS(rs, 6)];
Chris@42 66 T4 = R1[WS(rs, 1)];
Chris@42 67 {
Chris@42 68 E Tb, TL, Te, TK, TR, Tf, Ti, Ty;
Chris@42 69 Tb = R1[0];
Chris@42 70 TR = T2 + T5;
Chris@42 71 Tg = R0[WS(rs, 2)];
Chris@42 72 {
Chris@42 73 E T6, TS, Tc, Td;
Chris@42 74 T6 = T2 + T3 - T4 - T5;
Chris@42 75 T8 = (T3 + T5 - T2) - T4;
Chris@42 76 TS = T3 + T4;
Chris@42 77 Tc = R1[WS(rs, 3)];
Chris@42 78 Td = R1[WS(rs, 6)];
Chris@42 79 T7 = FNMS(KP250000000, T6, T1);
Chris@42 80 TF = T1 + T6;
Chris@42 81 TX = FNMS(KP618033988, TR, TS);
Chris@42 82 TT = FMA(KP618033988, TS, TR);
Chris@42 83 TL = Tc - Td;
Chris@42 84 Te = Tc + Td;
Chris@42 85 }
Chris@42 86 TK = Tg + Tb;
Chris@42 87 Tm = R0[WS(rs, 7)];
Chris@42 88 Tf = Tb - Te;
Chris@42 89 Th = Tb + Te;
Chris@42 90 TM = FMA(KP618033988, TL, TK);
Chris@42 91 TZ = FNMS(KP618033988, TK, TL);
Chris@42 92 Ti = FMA(KP809016994, Th, Tg);
Chris@42 93 Ty = FMA(KP447213595, Th, Tf);
Chris@42 94 Tr = R1[WS(rs, 5)];
Chris@42 95 Tn = R0[WS(rs, 1)];
Chris@42 96 Tj = FNMS(KP552786404, Ti, Tf);
Chris@42 97 Tz = FNMS(KP690983005, Ty, Tg);
Chris@42 98 To = R0[WS(rs, 4)];
Chris@42 99 TN = Tr + Tm;
Chris@42 100 }
Chris@42 101 }
Chris@42 102 TH = Ta + Tg - Th;
Chris@42 103 Tp = Tn + To;
Chris@42 104 TO = To - Tn;
Chris@42 105 {
Chris@42 106 E Tx, TA, TP, T14, T11, Tu, TD;
Chris@42 107 {
Chris@42 108 E T10, TI, TC, TY;
Chris@42 109 T9 = FNMS(KP559016994, T8, T7);
Chris@42 110 Tx = FMA(KP559016994, T8, T7);
Chris@42 111 TA = FNMS(KP809016994, Tz, Ta);
Chris@42 112 TP = FMA(KP618033988, TO, TN);
Chris@42 113 TY = FNMS(KP618033988, TN, TO);
Chris@42 114 {
Chris@42 115 E Tq, Ts, TG, Tt, TB;
Chris@42 116 Tq = Tm - Tp;
Chris@42 117 Ts = Tm + Tp;
Chris@42 118 T14 = TZ - TY;
Chris@42 119 T10 = TY + TZ;
Chris@42 120 TG = Ts - Tr - Tl;
Chris@42 121 Tt = FMA(KP809016994, Ts, Tr);
Chris@42 122 TB = FMA(KP447213595, Ts, Tq);
Chris@42 123 T11 = FMA(KP500000000, T10, TX);
Chris@42 124 Ci[WS(csi, 2)] = KP866025403 * (TH - TG);
Chris@42 125 TI = TG + TH;
Chris@42 126 Tu = FNMS(KP552786404, Tt, Tq);
Chris@42 127 TC = FNMS(KP690983005, TB, Tr);
Chris@42 128 }
Chris@42 129 Ci[WS(csi, 1)] = KP951056516 * (T10 - TX);
Chris@42 130 Cr[WS(csr, 7)] = TF + TI;
Chris@42 131 Cr[WS(csr, 2)] = FNMS(KP500000000, TI, TF);
Chris@42 132 TD = FNMS(KP809016994, TC, Tl);
Chris@42 133 }
Chris@42 134 {
Chris@42 135 E TU, Tk, T13, Tv, T12, TE;
Chris@42 136 TQ = TM - TP;
Chris@42 137 TU = TP + TM;
Chris@42 138 T12 = TD + TA;
Chris@42 139 TE = TA - TD;
Chris@42 140 Tk = FNMS(KP559016994, Tj, Ta);
Chris@42 141 TV = FMA(KP500000000, TU, TT);
Chris@42 142 Ci[WS(csi, 6)] = -(KP951056516 * (FMA(KP910592997, T12, T11)));
Chris@42 143 Ci[WS(csi, 3)] = KP951056516 * (FNMS(KP910592997, T12, T11));
Chris@42 144 T13 = FNMS(KP500000000, TE, Tx);
Chris@42 145 Cr[WS(csr, 1)] = Tx + TE;
Chris@42 146 Tv = FNMS(KP559016994, Tu, Tl);
Chris@42 147 Ci[WS(csi, 4)] = KP951056516 * (TT - TU);
Chris@42 148 Cr[WS(csr, 6)] = FMA(KP823639103, T14, T13);
Chris@42 149 Cr[WS(csr, 3)] = FNMS(KP823639103, T14, T13);
Chris@42 150 TW = Tv + Tk;
Chris@42 151 Tw = Tk - Tv;
Chris@42 152 }
Chris@42 153 }
Chris@42 154 }
Chris@42 155 Ci[WS(csi, 5)] = -(KP951056516 * (FNMS(KP910592997, TW, TV)));
Chris@42 156 Ci[0] = -(KP951056516 * (FMA(KP910592997, TW, TV)));
Chris@42 157 TJ = FNMS(KP500000000, Tw, T9);
Chris@42 158 Cr[WS(csr, 4)] = T9 + Tw;
Chris@42 159 Cr[0] = FMA(KP823639103, TQ, TJ);
Chris@42 160 Cr[WS(csr, 5)] = FNMS(KP823639103, TQ, TJ);
Chris@42 161 }
Chris@42 162 }
Chris@42 163 }
Chris@42 164
Chris@42 165 static const kr2c_desc desc = { 15, "r2cfII_15", {38, 7, 34, 0}, &GENUS };
Chris@42 166
Chris@42 167 void X(codelet_r2cfII_15) (planner *p) {
Chris@42 168 X(kr2c_register) (p, r2cfII_15, &desc);
Chris@42 169 }
Chris@42 170
Chris@42 171 #else /* HAVE_FMA */
Chris@42 172
Chris@42 173 /* Generated by: ../../../genfft/gen_r2cf.native -compact -variables 4 -pipeline-latency 4 -n 15 -name r2cfII_15 -dft-II -include r2cfII.h */
Chris@42 174
Chris@42 175 /*
Chris@42 176 * This function contains 72 FP additions, 33 FP multiplications,
Chris@42 177 * (or, 54 additions, 15 multiplications, 18 fused multiply/add),
Chris@42 178 * 37 stack variables, 8 constants, and 30 memory accesses
Chris@42 179 */
Chris@42 180 #include "r2cfII.h"
Chris@42 181
Chris@42 182 static void r2cfII_15(R *R0, R *R1, R *Cr, R *Ci, stride rs, stride csr, stride csi, INT v, INT ivs, INT ovs)
Chris@42 183 {
Chris@42 184 DK(KP500000000, +0.500000000000000000000000000000000000000000000);
Chris@42 185 DK(KP866025403, +0.866025403784438646763723170752936183471402627);
Chris@42 186 DK(KP809016994, +0.809016994374947424102293417182819058860154590);
Chris@42 187 DK(KP309016994, +0.309016994374947424102293417182819058860154590);
Chris@42 188 DK(KP250000000, +0.250000000000000000000000000000000000000000000);
Chris@42 189 DK(KP559016994, +0.559016994374947424102293417182819058860154590);
Chris@42 190 DK(KP587785252, +0.587785252292473129168705954639072768597652438);
Chris@42 191 DK(KP951056516, +0.951056516295153572116439333379382143405698634);
Chris@42 192 {
Chris@42 193 INT i;
Chris@42 194 for (i = v; i > 0; i = i - 1, R0 = R0 + ivs, R1 = R1 + ivs, Cr = Cr + ovs, Ci = Ci + ovs, MAKE_VOLATILE_STRIDE(60, rs), MAKE_VOLATILE_STRIDE(60, csr), MAKE_VOLATILE_STRIDE(60, csi)) {
Chris@42 195 E T1, T2, Tx, TR, TE, T7, TD, Th, Tm, Tr, TQ, TA, TB, Tf, Te;
Chris@42 196 E Tu, TS, Td, TH, TO;
Chris@42 197 T1 = R0[WS(rs, 5)];
Chris@42 198 {
Chris@42 199 E T3, Tv, T6, Tw, T4, T5;
Chris@42 200 T2 = R0[WS(rs, 2)];
Chris@42 201 T3 = R1[0];
Chris@42 202 Tv = T2 + T3;
Chris@42 203 T4 = R1[WS(rs, 3)];
Chris@42 204 T5 = R1[WS(rs, 6)];
Chris@42 205 T6 = T4 + T5;
Chris@42 206 Tw = T4 - T5;
Chris@42 207 Tx = FMA(KP951056516, Tv, KP587785252 * Tw);
Chris@42 208 TR = FNMS(KP587785252, Tv, KP951056516 * Tw);
Chris@42 209 TE = KP559016994 * (T3 - T6);
Chris@42 210 T7 = T3 + T6;
Chris@42 211 TD = KP250000000 * T7;
Chris@42 212 }
Chris@42 213 {
Chris@42 214 E Ti, Tl, Tj, Tk, Tp, Tq;
Chris@42 215 Th = R0[0];
Chris@42 216 Ti = R1[WS(rs, 4)];
Chris@42 217 Tl = R0[WS(rs, 6)];
Chris@42 218 Tj = R1[WS(rs, 1)];
Chris@42 219 Tk = R0[WS(rs, 3)];
Chris@42 220 Tp = Tk + Ti;
Chris@42 221 Tq = Tl + Tj;
Chris@42 222 Tm = Ti + Tj - (Tk + Tl);
Chris@42 223 Tr = FMA(KP951056516, Tp, KP587785252 * Tq);
Chris@42 224 TQ = FNMS(KP951056516, Tq, KP587785252 * Tp);
Chris@42 225 TA = FMA(KP250000000, Tm, Th);
Chris@42 226 TB = KP559016994 * (Tl + Ti - (Tk + Tj));
Chris@42 227 }
Chris@42 228 {
Chris@42 229 E T9, Tt, Tc, Ts, Ta, Tb, TG;
Chris@42 230 Tf = R1[WS(rs, 2)];
Chris@42 231 T9 = R0[WS(rs, 7)];
Chris@42 232 Te = R1[WS(rs, 5)];
Chris@42 233 Tt = T9 + Te;
Chris@42 234 Ta = R0[WS(rs, 1)];
Chris@42 235 Tb = R0[WS(rs, 4)];
Chris@42 236 Tc = Ta + Tb;
Chris@42 237 Ts = Ta - Tb;
Chris@42 238 Tu = FNMS(KP951056516, Tt, KP587785252 * Ts);
Chris@42 239 TS = FMA(KP951056516, Ts, KP587785252 * Tt);
Chris@42 240 Td = T9 + Tc;
Chris@42 241 TG = KP559016994 * (T9 - Tc);
Chris@42 242 TH = FNMS(KP309016994, Te, TG) + FNMA(KP250000000, Td, Tf);
Chris@42 243 TO = FMS(KP809016994, Te, Tf) + FNMA(KP250000000, Td, TG);
Chris@42 244 }
Chris@42 245 {
Chris@42 246 E Tn, T8, Tg, To;
Chris@42 247 Tn = Th - Tm;
Chris@42 248 T8 = T1 + T2 - T7;
Chris@42 249 Tg = Td - Te - Tf;
Chris@42 250 To = T8 + Tg;
Chris@42 251 Ci[WS(csi, 2)] = KP866025403 * (T8 - Tg);
Chris@42 252 Cr[WS(csr, 2)] = FNMS(KP500000000, To, Tn);
Chris@42 253 Cr[WS(csr, 7)] = Tn + To;
Chris@42 254 }
Chris@42 255 {
Chris@42 256 E TM, TX, TT, TV, TP, TU, TN, TW;
Chris@42 257 TM = TB + TA;
Chris@42 258 TX = KP866025403 * (TR + TS);
Chris@42 259 TT = TR - TS;
Chris@42 260 TV = FMS(KP500000000, TT, TQ);
Chris@42 261 TN = T1 + TE + FNMS(KP809016994, T2, TD);
Chris@42 262 TP = TN + TO;
Chris@42 263 TU = KP866025403 * (TO - TN);
Chris@42 264 Cr[WS(csr, 1)] = TM + TP;
Chris@42 265 Ci[WS(csi, 1)] = TQ + TT;
Chris@42 266 Ci[WS(csi, 6)] = TU - TV;
Chris@42 267 Ci[WS(csi, 3)] = TU + TV;
Chris@42 268 TW = FNMS(KP500000000, TP, TM);
Chris@42 269 Cr[WS(csr, 3)] = TW - TX;
Chris@42 270 Cr[WS(csr, 6)] = TW + TX;
Chris@42 271 }
Chris@42 272 {
Chris@42 273 E Tz, TC, Ty, TK, TI, TL, TF, TJ;
Chris@42 274 Tz = KP866025403 * (Tx + Tu);
Chris@42 275 TC = TA - TB;
Chris@42 276 Ty = Tu - Tx;
Chris@42 277 TK = FMS(KP500000000, Ty, Tr);
Chris@42 278 TF = FMA(KP309016994, T2, T1) + TD - TE;
Chris@42 279 TI = TF + TH;
Chris@42 280 TL = KP866025403 * (TH - TF);
Chris@42 281 Ci[WS(csi, 4)] = Tr + Ty;
Chris@42 282 Cr[WS(csr, 4)] = TC + TI;
Chris@42 283 Ci[WS(csi, 5)] = TK - TL;
Chris@42 284 Ci[0] = TK + TL;
Chris@42 285 TJ = FNMS(KP500000000, TI, TC);
Chris@42 286 Cr[0] = Tz + TJ;
Chris@42 287 Cr[WS(csr, 5)] = TJ - Tz;
Chris@42 288 }
Chris@42 289 }
Chris@42 290 }
Chris@42 291 }
Chris@42 292
Chris@42 293 static const kr2c_desc desc = { 15, "r2cfII_15", {54, 15, 18, 0}, &GENUS };
Chris@42 294
Chris@42 295 void X(codelet_r2cfII_15) (planner *p) {
Chris@42 296 X(kr2c_register) (p, r2cfII_15, &desc);
Chris@42 297 }
Chris@42 298
Chris@42 299 #endif /* HAVE_FMA */