Mercurial > hg > batch-feature-extraction-tool

annotate Lib/fftw-3.2.1/cell/spu/spu_n2fv_14.spuc @ 0:25bf17994ef1

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
First commit. VS2013, Codeblocks and Mac OSX configuration
author Geogaddi\David <d.m.ronan@qmul.ac.uk>
date Thu, 09 Jul 2015 01:12:16 +0100
parents
children
rev   line source
d@0 1 /*
d@0 2 * Copyright (c) 2003, 2007-8 Matteo Frigo
d@0 3 * Copyright (c) 2003, 2007-8 Massachusetts Institute of Technology
d@0 4 *
d@0 5 * This program is free software; you can redistribute it and/or modify
d@0 6 * it under the terms of the GNU General Public License as published by
d@0 7 * the Free Software Foundation; either version 2 of the License, or
d@0 8 * (at your option) any later version.
d@0 9 *
d@0 10 * This program is distributed in the hope that it will be useful,
d@0 11 * but WITHOUT ANY WARRANTY; without even the implied warranty of
d@0 12 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
d@0 13 * GNU General Public License for more details.
d@0 14 *
d@0 15 * You should have received a copy of the GNU General Public License
d@0 16 * along with this program; if not, write to the Free Software
d@0 17 * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
d@0 18 *
d@0 19 */
d@0 20 /* Generated by: ../../genfft/gen_notw_c -standalone -fma -reorder-insns -simd -compact -variables 100000 -with-ostride 2 -include fftw-spu.h -store-multiple 2 -n 14 -name X(spu_n2fv_14) */
d@0 21
d@0 22 /*
d@0 23 * This function contains 74 FP additions, 48 FP multiplications,
d@0 24 * (or, 32 additions, 6 multiplications, 42 fused multiply/add),
d@0 25 * 93 stack variables, 6 constants, and 35 memory accesses
d@0 26 */
d@0 27 #include "fftw-spu.h"
d@0 28
d@0 29 void X(spu_n2fv_14) (const R *ri, const R *ii, R *ro, R *io, stride is, stride os, INT v, INT ivs, INT ovs) {
d@0 30 DVK(KP801937735, +0.801937735804838252472204639014890102331838324);
d@0 31 DVK(KP974927912, +0.974927912181823607018131682993931217232785801);
d@0 32 DVK(KP554958132, +0.554958132087371191422194871006410481067288862);
d@0 33 DVK(KP900968867, +0.900968867902419126236102319507445051165919162);
d@0 34 DVK(KP692021471, +0.692021471630095869627814897002069140197260599);
d@0 35 DVK(KP356895867, +0.356895867892209443894399510021300583399127187);
d@0 36 INT i;
d@0 37 const R *xi;
d@0 38 R *xo;
d@0 39 xi = ri;
d@0 40 xo = ro;
d@0 41 for (i = v; i > 0; i = i - VL, xi = xi + (VL * ivs), xo = xo + (VL * ovs), MAKE_VOLATILE_STRIDE(is), MAKE_VOLATILE_STRIDE(os)) {
d@0 42 V TH, T3, Ta, Ts, TV, TN, TK, TW, Tt, To, Th, Tu, TU, TQ, Tp;
d@0 43 V TC, Tx, TR, T14, TZ, T1, T2, TI, T6, TJ, T9, TP, Tn, TO, Tk;
d@0 44 V TM, Tg, TL, Td, T4, T5, T7, T8, Tl, Tm, Ti, Tj, Te, Tf, Tb;
d@0 45 V Tc, T1a, T1b, T19, T1c, Tr, Tw, Tq, Tv, T16, T18, T15, T17, T1d, T1e;
d@0 46 V T1f, T1i, Tz, TB, Ty, TA, T1g, TT, TY, TS, TX, T1h, T1j, T11, T13;
d@0 47 V T10, T12, T1k, TE, TG, TD, TF, T1l, T1m;
d@0 48 T1 = LD(&(xi[0]), ivs, &(xi[0]));
d@0 49 T2 = LD(&(xi[WS(is, 7)]), ivs, &(xi[WS(is, 1)]));
d@0 50 TH = VADD(T1, T2);
d@0 51 T3 = VSUB(T1, T2);
d@0 52 T4 = LD(&(xi[WS(is, 2)]), ivs, &(xi[0]));
d@0 53 T5 = LD(&(xi[WS(is, 9)]), ivs, &(xi[WS(is, 1)]));
d@0 54 TI = VADD(T4, T5);
d@0 55 T6 = VSUB(T4, T5);
d@0 56 T7 = LD(&(xi[WS(is, 12)]), ivs, &(xi[0]));
d@0 57 T8 = LD(&(xi[WS(is, 5)]), ivs, &(xi[WS(is, 1)]));
d@0 58 TJ = VADD(T7, T8);
d@0 59 T9 = VSUB(T7, T8);
d@0 60 Tl = LD(&(xi[WS(is, 8)]), ivs, &(xi[0]));
d@0 61 Tm = LD(&(xi[WS(is, 1)]), ivs, &(xi[WS(is, 1)]));
d@0 62 TP = VADD(Tl, Tm);
d@0 63 Tn = VSUB(Tl, Tm);
d@0 64 Ti = LD(&(xi[WS(is, 6)]), ivs, &(xi[0]));
d@0 65 Tj = LD(&(xi[WS(is, 13)]), ivs, &(xi[WS(is, 1)]));
d@0 66 TO = VADD(Ti, Tj);
d@0 67 Tk = VSUB(Ti, Tj);
d@0 68 Te = LD(&(xi[WS(is, 10)]), ivs, &(xi[0]));
d@0 69 Tf = LD(&(xi[WS(is, 3)]), ivs, &(xi[WS(is, 1)]));
d@0 70 TM = VADD(Te, Tf);
d@0 71 Tg = VSUB(Te, Tf);
d@0 72 Tb = LD(&(xi[WS(is, 4)]), ivs, &(xi[0]));
d@0 73 Tc = LD(&(xi[WS(is, 11)]), ivs, &(xi[WS(is, 1)]));
d@0 74 TL = VADD(Tb, Tc);
d@0 75 Td = VSUB(Tb, Tc);
d@0 76 Ta = VADD(T6, T9);
d@0 77 Ts = VSUB(T9, T6);
d@0 78 TV = VSUB(TL, TM);
d@0 79 TN = VADD(TL, TM);
d@0 80 TK = VADD(TI, TJ);
d@0 81 TW = VSUB(TJ, TI);
d@0 82 Tt = VSUB(Tn, Tk);
d@0 83 To = VADD(Tk, Tn);
d@0 84 Th = VADD(Td, Tg);
d@0 85 Tu = VSUB(Tg, Td);
d@0 86 TU = VSUB(TO, TP);
d@0 87 TQ = VADD(TO, TP);
d@0 88 Tp = VFNMS(LDK(KP356895867), Ta, To);
d@0 89 TC = VFNMS(LDK(KP356895867), To, Th);
d@0 90 Tx = VFNMS(LDK(KP356895867), Th, Ta);
d@0 91 TR = VFNMS(LDK(KP356895867), TQ, TN);
d@0 92 T14 = VFNMS(LDK(KP356895867), TN, TK);
d@0 93 TZ = VFNMS(LDK(KP356895867), TK, TQ);
d@0 94 T19 = VADD(T3, VADD(Ta, VADD(Th, To)));
d@0 95 STM2(&(xo[14]), T19, ovs, &(xo[2]));
d@0 96 T1a = VADD(TH, VADD(TK, VADD(TN, TQ)));
d@0 97 STM2(&(xo[0]), T1a, ovs, &(xo[0]));
d@0 98 Tq = VFNMS(LDK(KP692021471), Tp, Th);
d@0 99 Tr = VFNMS(LDK(KP900968867), Tq, T3);
d@0 100 Tv = VFMA(LDK(KP554958132), Tu, Tt);
d@0 101 Tw = VMUL(LDK(KP974927912), VFNMS(LDK(KP801937735), Tv, Ts));
d@0 102 T1b = VFNMSI(Tw, Tr);
d@0 103 STM2(&(xo[10]), T1b, ovs, &(xo[2]));
d@0 104 T1c = VFMAI(Tw, Tr);
d@0 105 STM2(&(xo[18]), T1c, ovs, &(xo[2]));
d@0 106 T15 = VFNMS(LDK(KP692021471), T14, TQ);
d@0 107 T16 = VFNMS(LDK(KP900968867), T15, TH);
d@0 108 T17 = VFNMS(LDK(KP554958132), TU, TW);
d@0 109 T18 = VMUL(LDK(KP974927912), VFNMS(LDK(KP801937735), T17, TV));
d@0 110 T1d = VFMAI(T18, T16);
d@0 111 STM2(&(xo[12]), T1d, ovs, &(xo[0]));
d@0 112 STN2(&(xo[12]), T1d, T19, ovs);
d@0 113 T1e = VFNMSI(T18, T16);
d@0 114 STM2(&(xo[16]), T1e, ovs, &(xo[0]));
d@0 115 STN2(&(xo[16]), T1e, T1c, ovs);
d@0 116 Ty = VFNMS(LDK(KP692021471), Tx, To);
d@0 117 Tz = VFNMS(LDK(KP900968867), Ty, T3);
d@0 118 TA = VFMA(LDK(KP554958132), Tt, Ts);
d@0 119 TB = VMUL(LDK(KP974927912), VFMA(LDK(KP801937735), TA, Tu));
d@0 120 T1f = VFNMSI(TB, Tz);
d@0 121 STM2(&(xo[26]), T1f, ovs, &(xo[2]));
d@0 122 T1g = VFMAI(TB, Tz);
d@0 123 STM2(&(xo[2]), T1g, ovs, &(xo[2]));
d@0 124 STN2(&(xo[0]), T1a, T1g, ovs);
d@0 125 TS = VFNMS(LDK(KP692021471), TR, TK);
d@0 126 TT = VFNMS(LDK(KP900968867), TS, TH);
d@0 127 TX = VFMA(LDK(KP554958132), TW, TV);
d@0 128 TY = VMUL(LDK(KP974927912), VFNMS(LDK(KP801937735), TX, TU));
d@0 129 T1h = VFMAI(TY, TT);
d@0 130 STM2(&(xo[8]), T1h, ovs, &(xo[0]));
d@0 131 STN2(&(xo[8]), T1h, T1b, ovs);
d@0 132 T1i = VFNMSI(TY, TT);
d@0 133 STM2(&(xo[20]), T1i, ovs, &(xo[0]));
d@0 134 T10 = VFNMS(LDK(KP692021471), TZ, TN);
d@0 135 T11 = VFNMS(LDK(KP900968867), T10, TH);
d@0 136 T12 = VFMA(LDK(KP554958132), TV, TU);
d@0 137 T13 = VMUL(LDK(KP974927912), VFMA(LDK(KP801937735), T12, TW));
d@0 138 T1j = VFMAI(T13, T11);
d@0 139 STM2(&(xo[4]), T1j, ovs, &(xo[0]));
d@0 140 T1k = VFNMSI(T13, T11);
d@0 141 STM2(&(xo[24]), T1k, ovs, &(xo[0]));
d@0 142 STN2(&(xo[24]), T1k, T1f, ovs);
d@0 143 TD = VFNMS(LDK(KP692021471), TC, Ta);
d@0 144 TE = VFNMS(LDK(KP900968867), TD, T3);
d@0 145 TF = VFNMS(LDK(KP554958132), Ts, Tu);
d@0 146 TG = VMUL(LDK(KP974927912), VFNMS(LDK(KP801937735), TF, Tt));
d@0 147 T1l = VFNMSI(TG, TE);
d@0 148 STM2(&(xo[22]), T1l, ovs, &(xo[2]));
d@0 149 STN2(&(xo[20]), T1i, T1l, ovs);
d@0 150 T1m = VFMAI(TG, TE);
d@0 151 STM2(&(xo[6]), T1m, ovs, &(xo[2]));
d@0 152 STN2(&(xo[4]), T1j, T1m, ovs);
d@0 153 }
d@0 154 }