Mercurial > hg > sv-dependency-builds

annotate src/fftw-3.3.8/rdft/scalar/r2cf/r2cfII_16.c @ 82:d0c2a83c1364

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
Add FFTW 3.3.8 source, and a Linux build
author Chris Cannam
date Tue, 19 Nov 2019 14:52:55 +0000
parents
children
rev   line source
Chris@82 1 /*
Chris@82 2 * Copyright (c) 2003, 2007-14 Matteo Frigo
Chris@82 3 * Copyright (c) 2003, 2007-14 Massachusetts Institute of Technology
Chris@82 4 *
Chris@82 5 * This program is free software; you can redistribute it and/or modify
Chris@82 6 * it under the terms of the GNU General Public License as published by
Chris@82 7 * the Free Software Foundation; either version 2 of the License, or
Chris@82 8 * (at your option) any later version.
Chris@82 9 *
Chris@82 10 * This program is distributed in the hope that it will be useful,
Chris@82 11 * but WITHOUT ANY WARRANTY; without even the implied warranty of
Chris@82 12 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
Chris@82 13 * GNU General Public License for more details.
Chris@82 14 *
Chris@82 15 * You should have received a copy of the GNU General Public License
Chris@82 16 * along with this program; if not, write to the Free Software
Chris@82 17 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
Chris@82 18 *
Chris@82 19 */
Chris@82 20
Chris@82 21 /* This file was automatically generated --- DO NOT EDIT */
Chris@82 22 /* Generated on Thu May 24 08:06:43 EDT 2018 */
Chris@82 23
Chris@82 24 #include "rdft/codelet-rdft.h"
Chris@82 25
Chris@82 26 #if defined(ARCH_PREFERS_FMA) || defined(ISA_EXTENSION_PREFERS_FMA)
Chris@82 27
Chris@82 28 /* Generated by: ../../../genfft/gen_r2cf.native -fma -compact -variables 4 -pipeline-latency 4 -n 16 -name r2cfII_16 -dft-II -include rdft/scalar/r2cfII.h */
Chris@82 29
Chris@82 30 /*
Chris@82 31 * This function contains 66 FP additions, 48 FP multiplications,
Chris@82 32 * (or, 18 additions, 0 multiplications, 48 fused multiply/add),
Chris@82 33 * 32 stack variables, 7 constants, and 32 memory accesses
Chris@82 34 */
Chris@82 35 #include "rdft/scalar/r2cfII.h"
Chris@82 36
Chris@82 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@82 38 {
Chris@82 39 DK(KP980785280, +0.980785280403230449126182236134239036973933731);
Chris@82 40 DK(KP198912367, +0.198912367379658006911597622644676228597850501);
Chris@82 41 DK(KP831469612, +0.831469612302545237078788377617905756738560812);
Chris@82 42 DK(KP668178637, +0.668178637919298919997757686523080761552472251);
Chris@82 43 DK(KP923879532, +0.923879532511286756128183189396788286822416626);
Chris@82 44 DK(KP414213562, +0.414213562373095048801688724209698078569671875);
Chris@82 45 DK(KP707106781, +0.707106781186547524400844362104849039284835938);
Chris@82 46 {
Chris@82 47 INT i;
Chris@82 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@82 49 E T5, TZ, TB, TT, Tr, TK, Tu, TJ, Ti, TH, Tl, TG, Tc, T10, TE;
Chris@82 50 E TU;
Chris@82 51 {
Chris@82 52 E T1, TR, T4, TS, T2, T3;
Chris@82 53 T1 = R0[0];
Chris@82 54 TR = R0[WS(rs, 4)];
Chris@82 55 T2 = R0[WS(rs, 2)];
Chris@82 56 T3 = R0[WS(rs, 6)];
Chris@82 57 T4 = T2 - T3;
Chris@82 58 TS = T2 + T3;
Chris@82 59 T5 = FNMS(KP707106781, T4, T1);
Chris@82 60 TZ = FNMS(KP707106781, TS, TR);
Chris@82 61 TB = FMA(KP707106781, T4, T1);
Chris@82 62 TT = FMA(KP707106781, TS, TR);
Chris@82 63 }
Chris@82 64 {
Chris@82 65 E Tn, Ts, Tq, Tt, To, Tp;
Chris@82 66 Tn = R1[WS(rs, 7)];
Chris@82 67 Ts = R1[WS(rs, 3)];
Chris@82 68 To = R1[WS(rs, 1)];
Chris@82 69 Tp = R1[WS(rs, 5)];
Chris@82 70 Tq = To - Tp;
Chris@82 71 Tt = To + Tp;
Chris@82 72 Tr = FMA(KP707106781, Tq, Tn);
Chris@82 73 TK = FMA(KP707106781, Tt, Ts);
Chris@82 74 Tu = FNMS(KP707106781, Tt, Ts);
Chris@82 75 TJ = FMS(KP707106781, Tq, Tn);
Chris@82 76 }
Chris@82 77 {
Chris@82 78 E Te, Tj, Th, Tk, Tf, Tg;
Chris@82 79 Te = R1[0];
Chris@82 80 Tj = R1[WS(rs, 4)];
Chris@82 81 Tf = R1[WS(rs, 2)];
Chris@82 82 Tg = R1[WS(rs, 6)];
Chris@82 83 Th = Tf - Tg;
Chris@82 84 Tk = Tf + Tg;
Chris@82 85 Ti = FNMS(KP707106781, Th, Te);
Chris@82 86 TH = FMA(KP707106781, Tk, Tj);
Chris@82 87 Tl = FNMS(KP707106781, Tk, Tj);
Chris@82 88 TG = FMA(KP707106781, Th, Te);
Chris@82 89 }
Chris@82 90 {
Chris@82 91 E T8, TC, Tb, TD;
Chris@82 92 {
Chris@82 93 E T6, T7, T9, Ta;
Chris@82 94 T6 = R0[WS(rs, 5)];
Chris@82 95 T7 = R0[WS(rs, 1)];
Chris@82 96 T8 = FMA(KP414213562, T7, T6);
Chris@82 97 TC = FNMS(KP414213562, T6, T7);
Chris@82 98 T9 = R0[WS(rs, 3)];
Chris@82 99 Ta = R0[WS(rs, 7)];
Chris@82 100 Tb = FMA(KP414213562, Ta, T9);
Chris@82 101 TD = FMS(KP414213562, T9, Ta);
Chris@82 102 }
Chris@82 103 Tc = T8 - Tb;
Chris@82 104 T10 = TD - TC;
Chris@82 105 TE = TC + TD;
Chris@82 106 TU = T8 + Tb;
Chris@82 107 }
Chris@82 108 {
Chris@82 109 E Td, T13, Tw, T14, Tm, Tv;
Chris@82 110 Td = FMA(KP923879532, Tc, T5);
Chris@82 111 T13 = FNMS(KP923879532, T10, TZ);
Chris@82 112 Tm = FMA(KP668178637, Tl, Ti);
Chris@82 113 Tv = FMA(KP668178637, Tu, Tr);
Chris@82 114 Tw = Tm - Tv;
Chris@82 115 T14 = Tm + Tv;
Chris@82 116 Cr[WS(csr, 6)] = FNMS(KP831469612, Tw, Td);
Chris@82 117 Ci[WS(csi, 5)] = FNMS(KP831469612, T14, T13);
Chris@82 118 Cr[WS(csr, 1)] = FMA(KP831469612, Tw, Td);
Chris@82 119 Ci[WS(csi, 2)] = -(FMA(KP831469612, T14, T13));
Chris@82 120 }
Chris@82 121 {
Chris@82 122 E Tx, T11, TA, T12, Ty, Tz;
Chris@82 123 Tx = FNMS(KP923879532, Tc, T5);
Chris@82 124 T11 = FMA(KP923879532, T10, TZ);
Chris@82 125 Ty = FNMS(KP668178637, Tr, Tu);
Chris@82 126 Tz = FNMS(KP668178637, Ti, Tl);
Chris@82 127 TA = Ty - Tz;
Chris@82 128 T12 = Tz + Ty;
Chris@82 129 Cr[WS(csr, 5)] = FNMS(KP831469612, TA, Tx);
Chris@82 130 Ci[WS(csi, 1)] = FMA(KP831469612, T12, T11);
Chris@82 131 Cr[WS(csr, 2)] = FMA(KP831469612, TA, Tx);
Chris@82 132 Ci[WS(csi, 6)] = FMS(KP831469612, T12, T11);
Chris@82 133 }
Chris@82 134 {
Chris@82 135 E TF, TX, TM, TY, TI, TL;
Chris@82 136 TF = FMA(KP923879532, TE, TB);
Chris@82 137 TX = FNMS(KP923879532, TU, TT);
Chris@82 138 TI = FNMS(KP198912367, TH, TG);
Chris@82 139 TL = FMA(KP198912367, TK, TJ);
Chris@82 140 TM = TI + TL;
Chris@82 141 TY = TL - TI;
Chris@82 142 Cr[WS(csr, 7)] = FNMS(KP980785280, TM, TF);
Chris@82 143 Ci[WS(csi, 3)] = FMA(KP980785280, TY, TX);
Chris@82 144 Cr[0] = FMA(KP980785280, TM, TF);
Chris@82 145 Ci[WS(csi, 4)] = FMS(KP980785280, TY, TX);
Chris@82 146 }
Chris@82 147 {
Chris@82 148 E TN, TV, TQ, TW, TO, TP;
Chris@82 149 TN = FNMS(KP923879532, TE, TB);
Chris@82 150 TV = FMA(KP923879532, TU, TT);
Chris@82 151 TO = FMA(KP198912367, TG, TH);
Chris@82 152 TP = FNMS(KP198912367, TJ, TK);
Chris@82 153 TQ = TO - TP;
Chris@82 154 TW = TO + TP;
Chris@82 155 Cr[WS(csr, 4)] = FNMS(KP980785280, TQ, TN);
Chris@82 156 Ci[WS(csi, 7)] = FNMS(KP980785280, TW, TV);
Chris@82 157 Cr[WS(csr, 3)] = FMA(KP980785280, TQ, TN);
Chris@82 158 Ci[0] = -(FMA(KP980785280, TW, TV));
Chris@82 159 }
Chris@82 160 }
Chris@82 161 }
Chris@82 162 }
Chris@82 163
Chris@82 164 static const kr2c_desc desc = { 16, "r2cfII_16", {18, 0, 48, 0}, &GENUS };
Chris@82 165
Chris@82 166 void X(codelet_r2cfII_16) (planner *p) {
Chris@82 167 X(kr2c_register) (p, r2cfII_16, &desc);
Chris@82 168 }
Chris@82 169
Chris@82 170 #else
Chris@82 171
Chris@82 172 /* Generated by: ../../../genfft/gen_r2cf.native -compact -variables 4 -pipeline-latency 4 -n 16 -name r2cfII_16 -dft-II -include rdft/scalar/r2cfII.h */
Chris@82 173
Chris@82 174 /*
Chris@82 175 * This function contains 66 FP additions, 30 FP multiplications,
Chris@82 176 * (or, 54 additions, 18 multiplications, 12 fused multiply/add),
Chris@82 177 * 32 stack variables, 7 constants, and 32 memory accesses
Chris@82 178 */
Chris@82 179 #include "rdft/scalar/r2cfII.h"
Chris@82 180
Chris@82 181 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@82 182 {
Chris@82 183 DK(KP555570233, +0.555570233019602224742830813948532874374937191);
Chris@82 184 DK(KP831469612, +0.831469612302545237078788377617905756738560812);
Chris@82 185 DK(KP980785280, +0.980785280403230449126182236134239036973933731);
Chris@82 186 DK(KP195090322, +0.195090322016128267848284868477022240927691618);
Chris@82 187 DK(KP382683432, +0.382683432365089771728459984030398866761344562);
Chris@82 188 DK(KP923879532, +0.923879532511286756128183189396788286822416626);
Chris@82 189 DK(KP707106781, +0.707106781186547524400844362104849039284835938);
Chris@82 190 {
Chris@82 191 INT i;
Chris@82 192 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@82 193 E T5, T11, TB, TV, Tr, TK, Tu, TJ, Ti, TH, Tl, TG, Tc, T10, TE;
Chris@82 194 E TS;
Chris@82 195 {
Chris@82 196 E T1, TU, T4, TT, T2, T3;
Chris@82 197 T1 = R0[0];
Chris@82 198 TU = R0[WS(rs, 4)];
Chris@82 199 T2 = R0[WS(rs, 2)];
Chris@82 200 T3 = R0[WS(rs, 6)];
Chris@82 201 T4 = KP707106781 * (T2 - T3);
Chris@82 202 TT = KP707106781 * (T2 + T3);
Chris@82 203 T5 = T1 + T4;
Chris@82 204 T11 = TU - TT;
Chris@82 205 TB = T1 - T4;
Chris@82 206 TV = TT + TU;
Chris@82 207 }
Chris@82 208 {
Chris@82 209 E Tq, Tt, Tp, Ts, Tn, To;
Chris@82 210 Tq = R1[WS(rs, 7)];
Chris@82 211 Tt = R1[WS(rs, 3)];
Chris@82 212 Tn = R1[WS(rs, 1)];
Chris@82 213 To = R1[WS(rs, 5)];
Chris@82 214 Tp = KP707106781 * (Tn - To);
Chris@82 215 Ts = KP707106781 * (Tn + To);
Chris@82 216 Tr = Tp - Tq;
Chris@82 217 TK = Tt - Ts;
Chris@82 218 Tu = Ts + Tt;
Chris@82 219 TJ = Tp + Tq;
Chris@82 220 }
Chris@82 221 {
Chris@82 222 E Te, Tk, Th, Tj, Tf, Tg;
Chris@82 223 Te = R1[0];
Chris@82 224 Tk = R1[WS(rs, 4)];
Chris@82 225 Tf = R1[WS(rs, 2)];
Chris@82 226 Tg = R1[WS(rs, 6)];
Chris@82 227 Th = KP707106781 * (Tf - Tg);
Chris@82 228 Tj = KP707106781 * (Tf + Tg);
Chris@82 229 Ti = Te + Th;
Chris@82 230 TH = Tk - Tj;
Chris@82 231 Tl = Tj + Tk;
Chris@82 232 TG = Te - Th;
Chris@82 233 }
Chris@82 234 {
Chris@82 235 E T8, TC, Tb, TD;
Chris@82 236 {
Chris@82 237 E T6, T7, T9, Ta;
Chris@82 238 T6 = R0[WS(rs, 1)];
Chris@82 239 T7 = R0[WS(rs, 5)];
Chris@82 240 T8 = FNMS(KP382683432, T7, KP923879532 * T6);
Chris@82 241 TC = FMA(KP382683432, T6, KP923879532 * T7);
Chris@82 242 T9 = R0[WS(rs, 3)];
Chris@82 243 Ta = R0[WS(rs, 7)];
Chris@82 244 Tb = FNMS(KP923879532, Ta, KP382683432 * T9);
Chris@82 245 TD = FMA(KP923879532, T9, KP382683432 * Ta);
Chris@82 246 }
Chris@82 247 Tc = T8 + Tb;
Chris@82 248 T10 = Tb - T8;
Chris@82 249 TE = TC - TD;
Chris@82 250 TS = TC + TD;
Chris@82 251 }
Chris@82 252 {
Chris@82 253 E Td, TW, Tw, TR, Tm, Tv;
Chris@82 254 Td = T5 - Tc;
Chris@82 255 TW = TS + TV;
Chris@82 256 Tm = FMA(KP195090322, Ti, KP980785280 * Tl);
Chris@82 257 Tv = FNMS(KP980785280, Tu, KP195090322 * Tr);
Chris@82 258 Tw = Tm + Tv;
Chris@82 259 TR = Tv - Tm;
Chris@82 260 Cr[WS(csr, 4)] = Td - Tw;
Chris@82 261 Ci[WS(csi, 7)] = TR + TW;
Chris@82 262 Cr[WS(csr, 3)] = Td + Tw;
Chris@82 263 Ci[0] = TR - TW;
Chris@82 264 }
Chris@82 265 {
Chris@82 266 E Tx, TY, TA, TX, Ty, Tz;
Chris@82 267 Tx = T5 + Tc;
Chris@82 268 TY = TV - TS;
Chris@82 269 Ty = FNMS(KP195090322, Tl, KP980785280 * Ti);
Chris@82 270 Tz = FMA(KP980785280, Tr, KP195090322 * Tu);
Chris@82 271 TA = Ty + Tz;
Chris@82 272 TX = Tz - Ty;
Chris@82 273 Cr[WS(csr, 7)] = Tx - TA;
Chris@82 274 Ci[WS(csi, 3)] = TX + TY;
Chris@82 275 Cr[0] = Tx + TA;
Chris@82 276 Ci[WS(csi, 4)] = TX - TY;
Chris@82 277 }
Chris@82 278 {
Chris@82 279 E TF, T12, TM, TZ, TI, TL;
Chris@82 280 TF = TB + TE;
Chris@82 281 T12 = T10 - T11;
Chris@82 282 TI = FMA(KP831469612, TG, KP555570233 * TH);
Chris@82 283 TL = FMA(KP831469612, TJ, KP555570233 * TK);
Chris@82 284 TM = TI - TL;
Chris@82 285 TZ = TI + TL;
Chris@82 286 Cr[WS(csr, 6)] = TF - TM;
Chris@82 287 Ci[WS(csi, 2)] = T12 - TZ;
Chris@82 288 Cr[WS(csr, 1)] = TF + TM;
Chris@82 289 Ci[WS(csi, 5)] = -(TZ + T12);
Chris@82 290 }
Chris@82 291 {
Chris@82 292 E TN, T14, TQ, T13, TO, TP;
Chris@82 293 TN = TB - TE;
Chris@82 294 T14 = T10 + T11;
Chris@82 295 TO = FNMS(KP555570233, TJ, KP831469612 * TK);
Chris@82 296 TP = FNMS(KP555570233, TG, KP831469612 * TH);
Chris@82 297 TQ = TO - TP;
Chris@82 298 T13 = TP + TO;
Chris@82 299 Cr[WS(csr, 5)] = TN - TQ;
Chris@82 300 Ci[WS(csi, 1)] = T13 + T14;
Chris@82 301 Cr[WS(csr, 2)] = TN + TQ;
Chris@82 302 Ci[WS(csi, 6)] = T13 - T14;
Chris@82 303 }
Chris@82 304 }
Chris@82 305 }
Chris@82 306 }
Chris@82 307
Chris@82 308 static const kr2c_desc desc = { 16, "r2cfII_16", {54, 18, 12, 0}, &GENUS };
Chris@82 309
Chris@82 310 void X(codelet_r2cfII_16) (planner *p) {
Chris@82 311 X(kr2c_register) (p, r2cfII_16, &desc);
Chris@82 312 }
Chris@82 313
Chris@82 314 #endif