/* * Copyright (c) 2003 Matteo Frigo * Copyright (c) 2003 Massachusetts Institute of Technology * * This program is free software; you can redistribute it and/or modify * it under the terms of the GNU General Public License as published by * the Free Software Foundation; either version 2 of the License, or * (at your option) any later version. * * This program is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU General Public License for more details. * * You should have received a copy of the GNU General Public License * along with this program; if not, write to the Free Software * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA * */ /* This file was automatically generated --- DO NOT EDIT */ /* Generated on Sat Jul 5 21:57:02 EDT 2003 */ #include "codelet-rdft.h" /* Generated by: /homee/stevenj/cvs/fftw3.0.1/genfft/gen_hc2hc -compact -variables 4 -n 10 -dit -name hf_10 -include hf.h */ /* * This function contains 102 FP additions, 60 FP multiplications, * (or, 72 additions, 30 multiplications, 30 fused multiply/add), * 45 stack variables, and 40 memory accesses */ /* * Generator Id's : * $Id: hf_10.c,v 1.1 2008/10/17 06:12:34 scuri Exp $ * $Id: hf_10.c,v 1.1 2008/10/17 06:12:34 scuri Exp $ * $Id: hf_10.c,v 1.1 2008/10/17 06:12:34 scuri Exp $ */ #include "hf.h" static const R *hf_10(R *rio, R *iio, const R *W, stride ios, int m, int dist) { DK(KP587785252, +0.587785252292473129168705954639072768597652438); DK(KP951056516, +0.951056516295153572116439333379382143405698634); DK(KP250000000, +0.250000000000000000000000000000000000000000000); DK(KP559016994, +0.559016994374947424102293417182819058860154590); int i; for (i = m - 2; i > 0; i = i - 2, rio = rio + dist, iio = iio - dist, W = W + 18) { E T7, T1O, TT, T1C, TF, TQ, TR, T1r, T1s, T1L, TX, TY, TZ, T16, T19; E T1y, Ti, Tt, Tu, T1o, T1p, T1M, TU, TV, TW, T1d, T1g, T1x; { E T1, T1B, T6, T1A; T1 = rio[0]; T1B = iio[-WS(ios, 9)]; { E T3, T5, T2, T4; T3 = rio[WS(ios, 5)]; T5 = iio[-WS(ios, 4)]; T2 = W[8]; T4 = W[9]; T6 = FMA(T2, T3, T4 * T5); T1A = FNMS(T4, T3, T2 * T5); } T7 = T1 - T6; T1O = T1B - T1A; TT = T1 + T6; T1C = T1A + T1B; } { E Tz, T14, TP, T18, TE, T15, TK, T17; { E Tw, Ty, Tv, Tx; Tw = rio[WS(ios, 4)]; Ty = iio[-WS(ios, 5)]; Tv = W[6]; Tx = W[7]; Tz = FMA(Tv, Tw, Tx * Ty); T14 = FNMS(Tx, Tw, Tv * Ty); } { E TM, TO, TL, TN; TM = rio[WS(ios, 1)]; TO = iio[-WS(ios, 8)]; TL = W[0]; TN = W[1]; TP = FMA(TL, TM, TN * TO); T18 = FNMS(TN, TM, TL * TO); } { E TB, TD, TA, TC; TB = rio[WS(ios, 9)]; TD = iio[0]; TA = W[16]; TC = W[17]; TE = FMA(TA, TB, TC * TD); T15 = FNMS(TC, TB, TA * TD); } { E TH, TJ, TG, TI; TH = rio[WS(ios, 6)]; TJ = iio[-WS(ios, 3)]; TG = W[10]; TI = W[11]; TK = FMA(TG, TH, TI * TJ); T17 = FNMS(TI, TH, TG * TJ); } TF = Tz - TE; TQ = TK - TP; TR = TF + TQ; T1r = T14 - T15; T1s = T18 - T17; T1L = T1s - T1r; TX = Tz + TE; TY = TK + TP; TZ = TX + TY; T16 = T14 + T15; T19 = T17 + T18; T1y = T16 + T19; } { E Tc, T1b, Th, T1c, Tn, T1e, Ts, T1f; { E T9, Tb, T8, Ta; T9 = rio[WS(ios, 2)]; Tb = iio[-WS(ios, 7)]; T8 = W[2]; Ta = W[3]; Tc = FMA(T8, T9, Ta * Tb); T1b = FNMS(Ta, T9, T8 * Tb); } { E Te, Tg, Td, Tf; Te = rio[WS(ios, 7)]; Tg = iio[-WS(ios, 2)]; Td = W[12]; Tf = W[13]; Th = FMA(Td, Te, Tf * Tg); T1c = FNMS(Tf, Te, Td * Tg); } { E Tk, Tm, Tj, Tl; Tk = rio[WS(ios, 8)]; Tm = iio[-WS(ios, 1)]; Tj = W[14]; Tl = W[15]; Tn = FMA(Tj, Tk, Tl * Tm); T1e = FNMS(Tl, Tk, Tj * Tm); } { E Tp, Tr, To, Tq; Tp = rio[WS(ios, 3)]; Tr = iio[-WS(ios, 6)]; To = W[4]; Tq = W[5]; Ts = FMA(To, Tp, Tq * Tr); T1f = FNMS(Tq, Tp, To * Tr); } Ti = Tc - Th; Tt = Tn - Ts; Tu = Ti + Tt; T1o = T1b - T1c; T1p = T1e - T1f; T1M = T1o + T1p; TU = Tc + Th; TV = Tn + Ts; TW = TU + TV; T1d = T1b + T1c; T1g = T1e + T1f; T1x = T1d + T1g; } { E T1l, TS, T1m, T1u, T1w, T1q, T1t, T1v, T1n; T1l = KP559016994 * (Tu - TR); TS = Tu + TR; T1m = FNMS(KP250000000, TS, T7); T1q = T1o - T1p; T1t = T1r + T1s; T1u = FMA(KP951056516, T1q, KP587785252 * T1t); T1w = FNMS(KP587785252, T1q, KP951056516 * T1t); iio[-WS(ios, 5)] = T7 + TS; T1v = T1m - T1l; iio[-WS(ios, 7)] = T1v - T1w; rio[WS(ios, 3)] = T1v + T1w; T1n = T1l + T1m; iio[-WS(ios, 9)] = T1n - T1u; rio[WS(ios, 1)] = T1n + T1u; } { E T1S, T1N, T1T, T1R, T1V, T1P, T1Q, T1W, T1U; T1S = KP559016994 * (T1M + T1L); T1N = T1L - T1M; T1T = FMA(KP250000000, T1N, T1O); T1P = TQ - TF; T1Q = Ti - Tt; T1R = FNMS(KP951056516, T1Q, KP587785252 * T1P); T1V = FMA(KP587785252, T1Q, KP951056516 * T1P); rio[WS(ios, 5)] = T1N - T1O; T1W = T1T - T1S; rio[WS(ios, 7)] = T1V - T1W; iio[-WS(ios, 3)] = T1V + T1W; T1U = T1S + T1T; rio[WS(ios, 9)] = T1R - T1U; iio[-WS(ios, 1)] = T1R + T1U; } { E T12, T10, T11, T1i, T1k, T1a, T1h, T1j, T13; T12 = KP559016994 * (TW - TZ); T10 = TW + TZ; T11 = FNMS(KP250000000, T10, TT); T1a = T16 - T19; T1h = T1d - T1g; T1i = FNMS(KP587785252, T1h, KP951056516 * T1a); T1k = FMA(KP951056516, T1h, KP587785252 * T1a); rio[0] = TT + T10; T1j = T12 + T11; rio[WS(ios, 4)] = T1j - T1k; iio[-WS(ios, 6)] = T1j + T1k; T13 = T11 - T12; rio[WS(ios, 2)] = T13 - T1i; iio[-WS(ios, 8)] = T13 + T1i; } { E T1H, T1z, T1G, T1F, T1J, T1D, T1E, T1K, T1I; T1H = KP559016994 * (T1x - T1y); T1z = T1x + T1y; T1G = FNMS(KP250000000, T1z, T1C); T1D = TX - TY; T1E = TU - TV; T1F = FNMS(KP587785252, T1E, KP951056516 * T1D); T1J = FMA(KP951056516, T1E, KP587785252 * T1D); iio[0] = T1z + T1C; T1K = T1H + T1G; rio[WS(ios, 6)] = T1J - T1K; iio[-WS(ios, 4)] = T1J + T1K; T1I = T1G - T1H; rio[WS(ios, 8)] = T1F - T1I; iio[-WS(ios, 2)] = T1F + T1I; } } return W; } static const tw_instr twinstr[] = { {TW_FULL, 0, 10}, {TW_NEXT, 1, 0} }; static const hc2hc_desc desc = { 10, "hf_10", twinstr, {72, 30, 30, 0}, &GENUS, 0, 0, 0 }; void X(codelet_hf_10) (planner *p) { X(khc2hc_dit_register) (p, hf_10, &desc); }