e01_8.c

快速fft变换

/*
 * Copyright (c) 2003, 2007-8 Matteo Frigo
 * Copyright (c) 2003, 2007-8 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 Nov 15 21:14:08 EST 2008 */

#include "codelet-rdft.h"

#ifdef HAVE_FMA

/* Generated by: ../../../genfft/gen_r2r -fma -reorder-insns -schedule-for-pipeline -compact -variables 4 -pipeline-latency 4 -redft01 -n 8 -name e01_8 -include r2r.h */

/*
 * This function contains 26 FP additions, 24 FP multiplications,
 * (or, 2 additions, 0 multiplications, 24 fused multiply/add),
 * 27 stack variables, 8 constants, and 16 memory accesses
 */
#include "r2r.h"

static void e01_8(const R *I, R *O, stride is, stride os, INT v, INT ivs, INT ovs)
{
     DK(KP1_961570560, +1.961570560806460898252364472268478073947867462);
     DK(KP1_662939224, +1.662939224605090474157576755235811513477121624);
     DK(KP198912367, +0.198912367379658006911597622644676228597850501);
     DK(KP668178637, +0.668178637919298919997757686523080761552472251);
     DK(KP707106781, +0.707106781186547524400844362104849039284835938);
     DK(KP1_847759065, +1.847759065022573512256366378793576573644833252);
     DK(KP414213562, +0.414213562373095048801688724209698078569671875);
     DK(KP1_414213562, +1.414213562373095048801688724209698078569671875);
     INT i;
     for (i = v; i > 0; i = i - 1, I = I + ivs, O = O + ovs, MAKE_VOLATILE_STRIDE(is), MAKE_VOLATILE_STRIDE(os)) {
	  E T8, Td, Th, T7, Tp, Tl, Te, Tb;
	  {
	       E Tj, T3, Tk, T6, T9, Ta;
	       {
		    E T1, T2, T4, T5;
		    T1 = I[0];
		    T2 = I[WS(is, 4)];
		    T4 = I[WS(is, 2)];
		    T5 = I[WS(is, 6)];
		    T8 = I[WS(is, 1)];
		    Tj = FNMS(KP1_414213562, T2, T1);
		    T3 = FMA(KP1_414213562, T2, T1);
		    Tk = FMS(KP414213562, T4, T5);
		    T6 = FMA(KP414213562, T5, T4);
		    Td = I[WS(is, 7)];
		    T9 = I[WS(is, 5)];
		    Ta = I[WS(is, 3)];
	       }
	       Th = FNMS(KP1_847759065, T6, T3);
	       T7 = FMA(KP1_847759065, T6, T3);
	       Tp = FNMS(KP1_847759065, Tk, Tj);
	       Tl = FMA(KP1_847759065, Tk, Tj);
	       Te = Ta - T9;
	       Tb = T9 + Ta;
	  }
	  {
	       E Tn, Tf, Tc, Tm;
	       Tn = FNMS(KP707106781, Te, Td);
	       Tf = FMA(KP707106781, Te, Td);
	       Tc = FMA(KP707106781, Tb, T8);
	       Tm = FNMS(KP707106781, Tb, T8);
	       {
		    E Tq, To, Tg, Ti;
		    Tq = FMA(KP668178637, Tm, Tn);
		    To = FNMS(KP668178637, Tn, Tm);
		    Tg = FMA(KP198912367, Tf, Tc);
		    Ti = FNMS(KP198912367, Tc, Tf);
		    O[WS(os, 1)] = FMA(KP1_662939224, To, Tl);
		    O[WS(os, 6)] = FNMS(KP1_662939224, To, Tl);
		    O[WS(os, 2)] = FMA(KP1_662939224, Tq, Tp);
		    O[WS(os, 5)] = FNMS(KP1_662939224, Tq, Tp);
		    O[WS(os, 4)] = FMA(KP1_961570560, Ti, Th);
		    O[WS(os, 3)] = FNMS(KP1_961570560, Ti, Th);
		    O[0] = FMA(KP1_961570560, Tg, T7);
		    O[WS(os, 7)] = FNMS(KP1_961570560, Tg, T7);
	       }
	  }
     }
}

static const kr2r_desc desc = { 8, "e01_8", {2, 0, 24, 0}, &GENUS, REDFT01 };

void X(codelet_e01_8) (planner *p) {
     X(kr2r_register) (p, e01_8, &desc);
}

#else				/* HAVE_FMA */

/* Generated by: ../../../genfft/gen_r2r -compact -variables 4 -pipeline-latency 4 -redft01 -n 8 -name e01_8 -include r2r.h */

/*
 * This function contains 26 FP additions, 15 FP multiplications,
 * (or, 20 additions, 9 multiplications, 6 fused multiply/add),
 * 28 stack variables, 8 constants, and 16 memory accesses
 */
#include "r2r.h"

static void e01_8(const R *I, R *O, stride is, stride os, INT v, INT ivs, INT ovs)
{
     DK(KP1_662939224, +1.662939224605090474157576755235811513477121624);
     DK(KP1_111140466, +1.111140466039204449485661627897065748749874382);
     DK(KP390180644, +0.390180644032256535696569736954044481855383236);
     DK(KP1_961570560, +1.961570560806460898252364472268478073947867462);
     DK(KP707106781, +0.707106781186547524400844362104849039284835938);
     DK(KP1_414213562, +1.414213562373095048801688724209698078569671875);
     DK(KP765366864, +0.765366864730179543456919968060797733522689125);
     DK(KP1_847759065, +1.847759065022573512256366378793576573644833252);
     INT i;
     for (i = v; i > 0; i = i - 1, I = I + ivs, O = O + ovs, MAKE_VOLATILE_STRIDE(is), MAKE_VOLATILE_STRIDE(os)) {
	  E T7, Tl, T4, Tk, Td, To, Tg, Tn;
	  {
	       E T5, T6, T1, T3, T2;
	       T5 = I[WS(is, 2)];
	       T6 = I[WS(is, 6)];
	       T7 = FMA(KP1_847759065, T5, KP765366864 * T6);
	       Tl = FNMS(KP1_847759065, T6, KP765366864 * T5);
	       T1 = I[0];
	       T2 = I[WS(is, 4)];
	       T3 = KP1_414213562 * T2;
	       T4 = T1 + T3;
	       Tk = T1 - T3;
	       {
		    E T9, Tf, Tc, Te, Ta, Tb;
		    T9 = I[WS(is, 1)];
		    Tf = I[WS(is, 7)];
		    Ta = I[WS(is, 5)];
		    Tb = I[WS(is, 3)];
		    Tc = KP707106781 * (Ta + Tb);
		    Te = KP707106781 * (Ta - Tb);
		    Td = T9 + Tc;
		    To = Te + Tf;
		    Tg = Te - Tf;
		    Tn = T9 - Tc;
	       }
	  }
	  {
	       E T8, Th, Tq, Tr;
	       T8 = T4 + T7;
	       Th = FNMS(KP390180644, Tg, KP1_961570560 * Td);
	       O[WS(os, 7)] = T8 - Th;
	       O[0] = T8 + Th;
	       Tq = Tk - Tl;
	       Tr = FMA(KP1_111140466, Tn, KP1_662939224 * To);
	       O[WS(os, 5)] = Tq - Tr;
	       O[WS(os, 2)] = Tq + Tr;
	  }
	  {
	       E Ti, Tj, Tm, Tp;
	       Ti = T4 - T7;
	       Tj = FMA(KP390180644, Td, KP1_961570560 * Tg);
	       O[WS(os, 4)] = Ti - Tj;
	       O[WS(os, 3)] = Ti + Tj;
	       Tm = Tk + Tl;
	       Tp = FNMS(KP1_111140466, To, KP1_662939224 * Tn);
	       O[WS(os, 6)] = Tm - Tp;
	       O[WS(os, 1)] = Tm + Tp;
	  }
     }
}

static const kr2r_desc desc = { 8, "e01_8", {20, 9, 6, 0}, &GENUS, REDFT01 };

void X(codelet_e01_8) (planner *p) {
     X(kr2r_register) (p, e01_8, &desc);
}

#endif				/* HAVE_FMA */