Mercurial > hg > batch-feature-extraction-tool

annotate Lib/fftw-3.2.1/cell/spu/spu_t1fv_15.spuc @ 2:c649e493c30a

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
Removed a redundant cout<<
author Geogaddi\David <d.m.ronan@qmul.ac.uk>
date Thu, 09 Jul 2015 21:45:55 +0100
parents 25bf17994ef1
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_twiddle_c -standalone -fma -reorder-insns -simd -compact -variables 100000 -include fftw-spu.h -trivial-stores -n 15 -name X(spu_t1fv_15) */
d@0 21
d@0 22 /*
d@0 23 * This function contains 92 FP additions, 77 FP multiplications,
d@0 24 * (or, 50 additions, 35 multiplications, 42 fused multiply/add),
d@0 25 * 117 stack variables, 8 constants, and 30 memory accesses
d@0 26 */
d@0 27 #include "fftw-spu.h"
d@0 28
d@0 29 void X(spu_t1fv_15) (R *ri, R *ii, const R *W, stride rs, INT mb, INT me, INT ms) {
d@0 30 DVK(KP823639103, +0.823639103546331925877420039278190003029660514);
d@0 31 DVK(KP910592997, +0.910592997310029334643087372129977886038870291);
d@0 32 DVK(KP559016994, +0.559016994374947424102293417182819058860154590);
d@0 33 DVK(KP618033988, +0.618033988749894848204586834365638117720309180);
d@0 34 DVK(KP951056516, +0.951056516295153572116439333379382143405698634);
d@0 35 DVK(KP866025403, +0.866025403784438646763723170752936183471402627);
d@0 36 DVK(KP250000000, +0.250000000000000000000000000000000000000000000);
d@0 37 DVK(KP500000000, +0.500000000000000000000000000000000000000000000);
d@0 38 INT m;
d@0 39 R *x;
d@0 40 x = ri;
d@0 41 for (m = mb, W = W + (mb * ((TWVL / VL) * 28)); m < me; m = m + VL, x = x + (VL * ms), W = W + (TWVL * 28), MAKE_VOLATILE_STRIDE(rs)) {
d@0 42 V T1g, T7, TU, T17, T1a, To, TL, TK, TF, T1j, T1l, T1d, T1e, T11, T13;
d@0 43 V T1, T5, T3, T4, T2, T6, T9, Tq, Ty, Th, Te, T15, Tv, T18, TD;
d@0 44 V T19, Tm, T16, T8, Tp, Tx, Tg, Tb, Td, Ta, Tc, Ts, Tu, Tr, Tt;
d@0 45 V TA, TC, Tz, TB, Tj, Tl, Ti, Tk, T1h, T1i, TV, TW, Tf, Tn, TY;
d@0 46 V TZ, Tw, TE, TX, T10, T12, T1k, T1J, T1I, T1G, T1H, TQ, TM, TT, TJ;
d@0 47 V TP, TI, TH, TG, TR, TS, TO, TN, T1r, T1n, T1D, T1z, T1q, T1c, T1C;
d@0 48 V T1w, T1f, T1x, T1y, T1m, T1v, T1b, T1u, T14, T1p, T1F, T1o, T1E, T1t, T1B;
d@0 49 V T1s, T1A;
d@0 50 T1 = LD(&(x[0]), ms, &(x[0]));
d@0 51 T4 = LD(&(x[WS(rs, 10)]), ms, &(x[0]));
d@0 52 T5 = BYTWJ(&(W[TWVL * 18]), T4);
d@0 53 T2 = LD(&(x[WS(rs, 5)]), ms, &(x[WS(rs, 1)]));
d@0 54 T3 = BYTWJ(&(W[TWVL * 8]), T2);
d@0 55 T1g = VSUB(T5, T3);
d@0 56 T6 = VADD(T3, T5);
d@0 57 T7 = VADD(T1, T6);
d@0 58 TU = VFNMS(LDK(KP500000000), T6, T1);
d@0 59 T8 = LD(&(x[WS(rs, 3)]), ms, &(x[WS(rs, 1)]));
d@0 60 T9 = BYTWJ(&(W[TWVL * 4]), T8);
d@0 61 Tp = LD(&(x[WS(rs, 6)]), ms, &(x[0]));
d@0 62 Tq = BYTWJ(&(W[TWVL * 10]), Tp);
d@0 63 Tx = LD(&(x[WS(rs, 9)]), ms, &(x[WS(rs, 1)]));
d@0 64 Ty = BYTWJ(&(W[TWVL * 16]), Tx);
d@0 65 Tg = LD(&(x[WS(rs, 12)]), ms, &(x[0]));
d@0 66 Th = BYTWJ(&(W[TWVL * 22]), Tg);
d@0 67 Ta = LD(&(x[WS(rs, 8)]), ms, &(x[0]));
d@0 68 Tb = BYTWJ(&(W[TWVL * 14]), Ta);
d@0 69 Tc = LD(&(x[WS(rs, 13)]), ms, &(x[WS(rs, 1)]));
d@0 70 Td = BYTWJ(&(W[TWVL * 24]), Tc);
d@0 71 Te = VADD(Tb, Td);
d@0 72 T15 = VSUB(Td, Tb);
d@0 73 Tr = LD(&(x[WS(rs, 11)]), ms, &(x[WS(rs, 1)]));
d@0 74 Ts = BYTWJ(&(W[TWVL * 20]), Tr);
d@0 75 Tt = LD(&(x[WS(rs, 1)]), ms, &(x[WS(rs, 1)]));
d@0 76 Tu = BYTWJ(&(W[0]), Tt);
d@0 77 Tv = VADD(Ts, Tu);
d@0 78 T18 = VSUB(Tu, Ts);
d@0 79 Tz = LD(&(x[WS(rs, 14)]), ms, &(x[0]));
d@0 80 TA = BYTWJ(&(W[TWVL * 26]), Tz);
d@0 81 TB = LD(&(x[WS(rs, 4)]), ms, &(x[0]));
d@0 82 TC = BYTWJ(&(W[TWVL * 6]), TB);
d@0 83 TD = VADD(TA, TC);
d@0 84 T19 = VSUB(TC, TA);
d@0 85 Ti = LD(&(x[WS(rs, 2)]), ms, &(x[0]));
d@0 86 Tj = BYTWJ(&(W[TWVL * 2]), Ti);
d@0 87 Tk = LD(&(x[WS(rs, 7)]), ms, &(x[WS(rs, 1)]));
d@0 88 Tl = BYTWJ(&(W[TWVL * 12]), Tk);
d@0 89 Tm = VADD(Tj, Tl);
d@0 90 T16 = VSUB(Tl, Tj);
d@0 91 T17 = VSUB(T15, T16);
d@0 92 T1h = VADD(T15, T16);
d@0 93 T1i = VADD(T18, T19);
d@0 94 T1a = VSUB(T18, T19);
d@0 95 Tf = VADD(T9, Te);
d@0 96 TV = VFNMS(LDK(KP500000000), Te, T9);
d@0 97 TW = VFNMS(LDK(KP500000000), Tm, Th);
d@0 98 Tn = VADD(Th, Tm);
d@0 99 To = VADD(Tf, Tn);
d@0 100 TL = VSUB(Tf, Tn);
d@0 101 TY = VFNMS(LDK(KP500000000), Tv, Tq);
d@0 102 Tw = VADD(Tq, Tv);
d@0 103 TE = VADD(Ty, TD);
d@0 104 TZ = VFNMS(LDK(KP500000000), TD, Ty);
d@0 105 TK = VSUB(Tw, TE);
d@0 106 TF = VADD(Tw, TE);
d@0 107 T1j = VADD(T1h, T1i);
d@0 108 T1l = VSUB(T1h, T1i);
d@0 109 TX = VADD(TV, TW);
d@0 110 T1d = VSUB(TV, TW);
d@0 111 T1e = VSUB(TY, TZ);
d@0 112 T10 = VADD(TY, TZ);
d@0 113 T11 = VADD(TX, T10);
d@0 114 T13 = VSUB(TX, T10);
d@0 115 T12 = VFNMS(LDK(KP250000000), T11, TU);
d@0 116 T1G = VADD(TU, T11);
d@0 117 T1H = VMUL(LDK(KP866025403), VADD(T1g, T1j));
d@0 118 T1k = VFNMS(LDK(KP250000000), T1j, T1g);
d@0 119 T1J = VFMAI(T1H, T1G);
d@0 120 T1I = VFNMSI(T1H, T1G);
d@0 121 ST(&(x[WS(rs, 5)]), T1I, ms, &(x[WS(rs, 1)]));
d@0 122 ST(&(x[WS(rs, 10)]), T1J, ms, &(x[0]));
d@0 123 TQ = VMUL(LDK(KP951056516), VFMA(LDK(KP618033988), TK, TL));
d@0 124 TM = VMUL(LDK(KP951056516), VFNMS(LDK(KP618033988), TL, TK));
d@0 125 TG = VADD(To, TF);
d@0 126 TI = VSUB(To, TF);
d@0 127 TT = VADD(T7, TG);
d@0 128 TH = VFNMS(LDK(KP250000000), TG, T7);
d@0 129 TJ = VFNMS(LDK(KP559016994), TI, TH);
d@0 130 TP = VFMA(LDK(KP559016994), TI, TH);
d@0 131 ST(&(x[0]), TT, ms, &(x[0]));
d@0 132 TS = VFMAI(TQ, TP);
d@0 133 TR = VFNMSI(TQ, TP);
d@0 134 ST(&(x[WS(rs, 9)]), TS, ms, &(x[WS(rs, 1)]));
d@0 135 TN = VFNMSI(TM, TJ);
d@0 136 TO = VFMAI(TM, TJ);
d@0 137 ST(&(x[WS(rs, 3)]), TN, ms, &(x[WS(rs, 1)]));
d@0 138 ST(&(x[WS(rs, 12)]), TO, ms, &(x[0]));
d@0 139 ST(&(x[WS(rs, 6)]), TR, ms, &(x[0]));
d@0 140 T1f = VFMA(LDK(KP618033988), T1e, T1d);
d@0 141 T1x = VFNMS(LDK(KP618033988), T1d, T1e);
d@0 142 T1y = VFNMS(LDK(KP559016994), T1l, T1k);
d@0 143 T1m = VFMA(LDK(KP559016994), T1l, T1k);
d@0 144 T1r = VMUL(LDK(KP951056516), VFMA(LDK(KP910592997), T1m, T1f));
d@0 145 T1n = VMUL(LDK(KP951056516), VFNMS(LDK(KP910592997), T1m, T1f));
d@0 146 T1D = VMUL(LDK(KP951056516), VFMA(LDK(KP910592997), T1y, T1x));
d@0 147 T1z = VMUL(LDK(KP951056516), VFNMS(LDK(KP910592997), T1y, T1x));
d@0 148 T1v = VFNMS(LDK(KP618033988), T17, T1a);
d@0 149 T1b = VFMA(LDK(KP618033988), T1a, T17);
d@0 150 T1u = VFNMS(LDK(KP559016994), T13, T12);
d@0 151 T14 = VFMA(LDK(KP559016994), T13, T12);
d@0 152 T1q = VFNMS(LDK(KP823639103), T1b, T14);
d@0 153 T1c = VFMA(LDK(KP823639103), T1b, T14);
d@0 154 T1C = VFNMS(LDK(KP823639103), T1v, T1u);
d@0 155 T1w = VFMA(LDK(KP823639103), T1v, T1u);
d@0 156 T1p = VFMAI(T1n, T1c);
d@0 157 T1o = VFNMSI(T1n, T1c);
d@0 158 ST(&(x[WS(rs, 1)]), T1o, ms, &(x[WS(rs, 1)]));
d@0 159 T1F = VFMAI(T1D, T1C);
d@0 160 T1E = VFNMSI(T1D, T1C);
d@0 161 ST(&(x[WS(rs, 8)]), T1E, ms, &(x[0]));
d@0 162 ST(&(x[WS(rs, 7)]), T1F, ms, &(x[WS(rs, 1)]));
d@0 163 ST(&(x[WS(rs, 14)]), T1p, ms, &(x[0]));
d@0 164 T1t = VFMAI(T1r, T1q);
d@0 165 T1s = VFNMSI(T1r, T1q);
d@0 166 ST(&(x[WS(rs, 11)]), T1s, ms, &(x[WS(rs, 1)]));
d@0 167 T1B = VFMAI(T1z, T1w);
d@0 168 T1A = VFNMSI(T1z, T1w);
d@0 169 ST(&(x[WS(rs, 13)]), T1A, ms, &(x[WS(rs, 1)]));
d@0 170 ST(&(x[WS(rs, 2)]), T1B, ms, &(x[0]));
d@0 171 ST(&(x[WS(rs, 4)]), T1t, ms, &(x[0]));
d@0 172 }
d@0 173 }