Mercurial > hg > batch-feature-extraction-tool

annotate Lib/fftw-3.2.1/cell/spu/spu_t1fv_16.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 16 -name X(spu_t1fv_16) */
d@0 21
d@0 22 /*
d@0 23 * This function contains 87 FP additions, 64 FP multiplications,
d@0 24 * (or, 53 additions, 30 multiplications, 34 fused multiply/add),
d@0 25 * 108 stack variables, 3 constants, and 32 memory accesses
d@0 26 */
d@0 27 #include "fftw-spu.h"
d@0 28
d@0 29 void X(spu_t1fv_16) (R *ri, R *ii, const R *W, stride rs, INT mb, INT me, INT ms) {
d@0 30 DVK(KP923879532, +0.923879532511286756128183189396788286822416626);
d@0 31 DVK(KP707106781, +0.707106781186547524400844362104849039284835938);
d@0 32 DVK(KP414213562, +0.414213562373095048801688724209698078569671875);
d@0 33 INT m;
d@0 34 R *x;
d@0 35 x = ri;
d@0 36 for (m = mb, W = W + (mb * ((TWVL / VL) * 30)); m < me; m = m + VL, x = x + (VL * ms), W = W + (TWVL * 30), MAKE_VOLATILE_STRIDE(rs)) {
d@0 37 V T14, T1h, Ta, TS, TJ, TT, T17, T1i, Tl, TV, T1b, T1k, Tw, TW, T1e;
d@0 38 V T1l, T1, T8, T3, T6, T7, T2, T5, T4, T9, TA, TH, TC, TF, Tz;
d@0 39 V TG, TB, TE, T16, T15, TD, TI, Tc, Tj, Te, Th, Tb, Ti, Td, Tg;
d@0 40 V T19, T1a, Tf, Tk, Tn, Tu, Tp, Ts, Tm, Tt, To, Tr, T1c, T1d, Tq;
d@0 41 V Tv, T10, T11, TY, TZ, TU, TX, T12, T13, TO, Ty, TL, TP, TK, Tx;
d@0 42 V TN, TQ, TM, TR, T1u, T1y, T1x, T1v, T1g, T1q, T1n, T1r, T18, T1f, T1j;
d@0 43 V T1m, T1p, T1s, T1o, T1t, T1w, T1C, T1z, T1D, T1B, T1E, T1A, T1F;
d@0 44 T1 = LD(&(x[0]), ms, &(x[0]));
d@0 45 T7 = LD(&(x[WS(rs, 12)]), ms, &(x[0]));
d@0 46 T8 = BYTWJ(&(W[TWVL * 22]), T7);
d@0 47 T2 = LD(&(x[WS(rs, 8)]), ms, &(x[0]));
d@0 48 T3 = BYTWJ(&(W[TWVL * 14]), T2);
d@0 49 T5 = LD(&(x[WS(rs, 4)]), ms, &(x[0]));
d@0 50 T6 = BYTWJ(&(W[TWVL * 6]), T5);
d@0 51 T14 = VSUB(T1, T3);
d@0 52 T4 = VADD(T1, T3);
d@0 53 T9 = VADD(T6, T8);
d@0 54 T1h = VSUB(T6, T8);
d@0 55 Ta = VSUB(T4, T9);
d@0 56 TS = VADD(T4, T9);
d@0 57 Tz = LD(&(x[WS(rs, 14)]), ms, &(x[0]));
d@0 58 TA = BYTWJ(&(W[TWVL * 26]), Tz);
d@0 59 TG = LD(&(x[WS(rs, 10)]), ms, &(x[0]));
d@0 60 TH = BYTWJ(&(W[TWVL * 18]), TG);
d@0 61 TB = LD(&(x[WS(rs, 6)]), ms, &(x[0]));
d@0 62 TC = BYTWJ(&(W[TWVL * 10]), TB);
d@0 63 TE = LD(&(x[WS(rs, 2)]), ms, &(x[0]));
d@0 64 TF = BYTWJ(&(W[TWVL * 2]), TE);
d@0 65 TD = VADD(TA, TC);
d@0 66 T16 = VSUB(TA, TC);
d@0 67 T15 = VSUB(TF, TH);
d@0 68 TI = VADD(TF, TH);
d@0 69 TJ = VSUB(TD, TI);
d@0 70 TT = VADD(TI, TD);
d@0 71 T17 = VADD(T15, T16);
d@0 72 T1i = VSUB(T16, T15);
d@0 73 Tb = LD(&(x[WS(rs, 1)]), ms, &(x[WS(rs, 1)]));
d@0 74 Tc = BYTWJ(&(W[0]), Tb);
d@0 75 Ti = LD(&(x[WS(rs, 13)]), ms, &(x[WS(rs, 1)]));
d@0 76 Tj = BYTWJ(&(W[TWVL * 24]), Ti);
d@0 77 Td = LD(&(x[WS(rs, 9)]), ms, &(x[WS(rs, 1)]));
d@0 78 Te = BYTWJ(&(W[TWVL * 16]), Td);
d@0 79 Tg = LD(&(x[WS(rs, 5)]), ms, &(x[WS(rs, 1)]));
d@0 80 Th = BYTWJ(&(W[TWVL * 8]), Tg);
d@0 81 Tf = VADD(Tc, Te);
d@0 82 T19 = VSUB(Tc, Te);
d@0 83 T1a = VSUB(Th, Tj);
d@0 84 Tk = VADD(Th, Tj);
d@0 85 Tl = VSUB(Tf, Tk);
d@0 86 TV = VADD(Tf, Tk);
d@0 87 T1b = VFNMS(LDK(KP414213562), T1a, T19);
d@0 88 T1k = VFMA(LDK(KP414213562), T19, T1a);
d@0 89 Tm = LD(&(x[WS(rs, 15)]), ms, &(x[WS(rs, 1)]));
d@0 90 Tn = BYTWJ(&(W[TWVL * 28]), Tm);
d@0 91 Tt = LD(&(x[WS(rs, 11)]), ms, &(x[WS(rs, 1)]));
d@0 92 Tu = BYTWJ(&(W[TWVL * 20]), Tt);
d@0 93 To = LD(&(x[WS(rs, 7)]), ms, &(x[WS(rs, 1)]));
d@0 94 Tp = BYTWJ(&(W[TWVL * 12]), To);
d@0 95 Tr = LD(&(x[WS(rs, 3)]), ms, &(x[WS(rs, 1)]));
d@0 96 Ts = BYTWJ(&(W[TWVL * 4]), Tr);
d@0 97 Tq = VADD(Tn, Tp);
d@0 98 T1c = VSUB(Tn, Tp);
d@0 99 T1d = VSUB(Tu, Ts);
d@0 100 Tv = VADD(Ts, Tu);
d@0 101 Tw = VSUB(Tq, Tv);
d@0 102 TW = VADD(Tq, Tv);
d@0 103 T1e = VFNMS(LDK(KP414213562), T1d, T1c);
d@0 104 T1l = VFMA(LDK(KP414213562), T1c, T1d);
d@0 105 TU = VADD(TS, TT);
d@0 106 T10 = VSUB(TS, TT);
d@0 107 T11 = VSUB(TW, TV);
d@0 108 TX = VADD(TV, TW);
d@0 109 TY = VSUB(TU, TX);
d@0 110 TZ = VADD(TU, TX);
d@0 111 ST(&(x[WS(rs, 8)]), TY, ms, &(x[0]));
d@0 112 ST(&(x[0]), TZ, ms, &(x[0]));
d@0 113 T12 = VFNMSI(T11, T10);
d@0 114 T13 = VFMAI(T11, T10);
d@0 115 ST(&(x[WS(rs, 12)]), T12, ms, &(x[0]));
d@0 116 ST(&(x[WS(rs, 4)]), T13, ms, &(x[0]));
d@0 117 Tx = VADD(Tl, Tw);
d@0 118 TK = VSUB(Tw, Tl);
d@0 119 TO = VFMA(LDK(KP707106781), Tx, Ta);
d@0 120 Ty = VFNMS(LDK(KP707106781), Tx, Ta);
d@0 121 TL = VFNMS(LDK(KP707106781), TK, TJ);
d@0 122 TP = VFMA(LDK(KP707106781), TK, TJ);
d@0 123 TM = VFNMSI(TL, Ty);
d@0 124 TN = VFMAI(TL, Ty);
d@0 125 ST(&(x[WS(rs, 6)]), TM, ms, &(x[0]));
d@0 126 TR = VFMAI(TP, TO);
d@0 127 TQ = VFNMSI(TP, TO);
d@0 128 ST(&(x[WS(rs, 2)]), TR, ms, &(x[0]));
d@0 129 ST(&(x[WS(rs, 10)]), TN, ms, &(x[0]));
d@0 130 ST(&(x[WS(rs, 14)]), TQ, ms, &(x[0]));
d@0 131 T18 = VFMA(LDK(KP707106781), T17, T14);
d@0 132 T1u = VFNMS(LDK(KP707106781), T17, T14);
d@0 133 T1y = VSUB(T1e, T1b);
d@0 134 T1f = VADD(T1b, T1e);
d@0 135 T1g = VFNMS(LDK(KP923879532), T1f, T18);
d@0 136 T1q = VFMA(LDK(KP923879532), T1f, T18);
d@0 137 T1j = VFNMS(LDK(KP707106781), T1i, T1h);
d@0 138 T1x = VFMA(LDK(KP707106781), T1i, T1h);
d@0 139 T1v = VADD(T1k, T1l);
d@0 140 T1m = VSUB(T1k, T1l);
d@0 141 T1n = VFNMS(LDK(KP923879532), T1m, T1j);
d@0 142 T1r = VFMA(LDK(KP923879532), T1m, T1j);
d@0 143 T1o = VFNMSI(T1n, T1g);
d@0 144 T1p = VFMAI(T1n, T1g);
d@0 145 ST(&(x[WS(rs, 9)]), T1o, ms, &(x[WS(rs, 1)]));
d@0 146 T1t = VFMAI(T1r, T1q);
d@0 147 T1s = VFNMSI(T1r, T1q);
d@0 148 ST(&(x[WS(rs, 15)]), T1t, ms, &(x[WS(rs, 1)]));
d@0 149 ST(&(x[WS(rs, 7)]), T1p, ms, &(x[WS(rs, 1)]));
d@0 150 ST(&(x[WS(rs, 1)]), T1s, ms, &(x[WS(rs, 1)]));
d@0 151 T1w = VFNMS(LDK(KP923879532), T1v, T1u);
d@0 152 T1C = VFMA(LDK(KP923879532), T1v, T1u);
d@0 153 T1z = VFNMS(LDK(KP923879532), T1y, T1x);
d@0 154 T1D = VFMA(LDK(KP923879532), T1y, T1x);
d@0 155 T1A = VFNMSI(T1z, T1w);
d@0 156 T1B = VFMAI(T1z, T1w);
d@0 157 ST(&(x[WS(rs, 5)]), T1A, ms, &(x[WS(rs, 1)]));
d@0 158 T1F = VFNMSI(T1D, T1C);
d@0 159 T1E = VFMAI(T1D, T1C);
d@0 160 ST(&(x[WS(rs, 13)]), T1F, ms, &(x[WS(rs, 1)]));
d@0 161 ST(&(x[WS(rs, 11)]), T1B, ms, &(x[WS(rs, 1)]));
d@0 162 ST(&(x[WS(rs, 3)]), T1E, ms, &(x[WS(rs, 1)]));
d@0 163 }
d@0 164 }