t1bv_7.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 20:50:54 EST 2008 */

#include "codelet-dft.h"

#ifdef HAVE_FMA

/* Generated by: ../../../genfft/gen_twiddle_c -fma -reorder-insns -schedule-for-pipeline -simd -compact -variables 4 -pipeline-latency 8 -n 7 -name t1bv_7 -include t1b.h -sign 1 */

/*
 * This function contains 36 FP additions, 36 FP multiplications,
 * (or, 15 additions, 15 multiplications, 21 fused multiply/add),
 * 42 stack variables, 6 constants, and 14 memory accesses
 */
#include "t1b.h"

static void t1bv_7(R *ri, R *ii, const R *W, stride rs, INT mb, INT me, INT ms)
{
     DVK(KP900968867, +0.900968867902419126236102319507445051165919162);
     DVK(KP801937735, +0.801937735804838252472204639014890102331838324);
     DVK(KP974927912, +0.974927912181823607018131682993931217232785801);
     DVK(KP692021471, +0.692021471630095869627814897002069140197260599);
     DVK(KP554958132, +0.554958132087371191422194871006410481067288862);
     DVK(KP356895867, +0.356895867892209443894399510021300583399127187);
     INT m;
     R *x;
     x = ii;
     for (m = mb, W = W + (mb * ((TWVL / VL) * 12)); m < me; m = m + VL, x = x + (VL * ms), W = W + (TWVL * 12), MAKE_VOLATILE_STRIDE(rs)) {
	  V T1, T2, T4, Te, Tc, T9, T7;
	  T1 = LD(&(x[0]), ms, &(x[0]));
	  T2 = LD(&(x[WS(rs, 1)]), ms, &(x[WS(rs, 1)]));
	  T4 = LD(&(x[WS(rs, 6)]), ms, &(x[0]));
	  Te = LD(&(x[WS(rs, 4)]), ms, &(x[0]));
	  Tc = LD(&(x[WS(rs, 3)]), ms, &(x[WS(rs, 1)]));
	  T9 = LD(&(x[WS(rs, 5)]), ms, &(x[WS(rs, 1)]));
	  T7 = LD(&(x[WS(rs, 2)]), ms, &(x[0]));
	  {
	       V T3, T5, Tf, Td, Ta, T8;
	       T3 = BYTW(&(W[0]), T2);
	       T5 = BYTW(&(W[TWVL * 10]), T4);
	       Tf = BYTW(&(W[TWVL * 6]), Te);
	       Td = BYTW(&(W[TWVL * 4]), Tc);
	       Ta = BYTW(&(W[TWVL * 8]), T9);
	       T8 = BYTW(&(W[TWVL * 2]), T7);
	       {
		    V T6, Tm, Tg, Tk, Tb, Tl;
		    T6 = VADD(T3, T5);
		    Tm = VSUB(T3, T5);
		    Tg = VADD(Td, Tf);
		    Tk = VSUB(Td, Tf);
		    Tb = VADD(T8, Ta);
		    Tl = VSUB(T8, Ta);
		    {
			 V Tp, Tx, Tu, Th, Ts, Tn, Tq, Ty;
			 Tp = VFNMS(LDK(KP356895867), T6, Tg);
			 Tx = VFMA(LDK(KP554958132), Tk, Tm);
			 ST(&(x[0]), VADD(T1, VADD(T6, VADD(Tb, Tg))), ms, &(x[0]));
			 Tu = VFNMS(LDK(KP356895867), Tb, T6);
			 Th = VFNMS(LDK(KP356895867), Tg, Tb);
			 Ts = VFMA(LDK(KP554958132), Tl, Tk);
			 Tn = VFNMS(LDK(KP554958132), Tm, Tl);
			 Tq = VFNMS(LDK(KP692021471), Tp, Tb);
			 Ty = VMUL(LDK(KP974927912), VFMA(LDK(KP801937735), Tx, Tl));
			 {
			      V Tv, Ti, Tt, To, Tr, Tw, Tj;
			      Tv = VFNMS(LDK(KP692021471), Tu, Tg);
			      Ti = VFNMS(LDK(KP692021471), Th, T6);
			      Tt = VMUL(LDK(KP974927912), VFNMS(LDK(KP801937735), Ts, Tm));
			      To = VMUL(LDK(KP974927912), VFNMS(LDK(KP801937735), Tn, Tk));
			      Tr = VFNMS(LDK(KP900968867), Tq, T1);
			      Tw = VFNMS(LDK(KP900968867), Tv, T1);
			      Tj = VFNMS(LDK(KP900968867), Ti, T1);
			      ST(&(x[WS(rs, 5)]), VFNMSI(Tt, Tr), ms, &(x[WS(rs, 1)]));
			      ST(&(x[WS(rs, 2)]), VFMAI(Tt, Tr), ms, &(x[0]));
			      ST(&(x[WS(rs, 6)]), VFNMSI(Ty, Tw), ms, &(x[0]));
			      ST(&(x[WS(rs, 1)]), VFMAI(Ty, Tw), ms, &(x[WS(rs, 1)]));
			      ST(&(x[WS(rs, 4)]), VFNMSI(To, Tj), ms, &(x[0]));
			      ST(&(x[WS(rs, 3)]), VFMAI(To, Tj), ms, &(x[WS(rs, 1)]));
			 }
		    }
	       }
	  }
     }
}

static const tw_instr twinstr[] = {
     VTW(0, 1),
     VTW(0, 2),
     VTW(0, 3),
     VTW(0, 4),
     VTW(0, 5),
     VTW(0, 6),
     {TW_NEXT, VL, 0}
};

static const ct_desc desc = { 7, "t1bv_7", twinstr, &GENUS, {15, 15, 21, 0}, 0, 0, 0 };

void X(codelet_t1bv_7) (planner *p) {
     X(kdft_dit_register) (p, t1bv_7, &desc);
}
#else				/* HAVE_FMA */

/* Generated by: ../../../genfft/gen_twiddle_c -simd -compact -variables 4 -pipeline-latency 8 -n 7 -name t1bv_7 -include t1b.h -sign 1 */

/*
 * This function contains 36 FP additions, 30 FP multiplications,
 * (or, 24 additions, 18 multiplications, 12 fused multiply/add),
 * 21 stack variables, 6 constants, and 14 memory accesses
 */
#include "t1b.h"

static void t1bv_7(R *ri, R *ii, const R *W, stride rs, INT mb, INT me, INT ms)
{
     DVK(KP222520933, +0.222520933956314404288902564496794759466355569);
     DVK(KP900968867, +0.900968867902419126236102319507445051165919162);
     DVK(KP623489801, +0.623489801858733530525004884004239810632274731);
     DVK(KP433883739, +0.433883739117558120475768332848358754609990728);
     DVK(KP781831482, +0.781831482468029808708444526674057750232334519);
     DVK(KP974927912, +0.974927912181823607018131682993931217232785801);
     INT m;
     R *x;
     x = ii;
     for (m = mb, W = W + (mb * ((TWVL / VL) * 12)); m < me; m = m + VL, x = x + (VL * ms), W = W + (TWVL * 12), MAKE_VOLATILE_STRIDE(rs)) {
	  V Th, Tf, Ti, T5, Tk, Ta, Tj, To, Tp;
	  Th = LD(&(x[0]), ms, &(x[0]));
	  {
	       V Tc, Te, Tb, Td;
	       Tb = LD(&(x[WS(rs, 2)]), ms, &(x[0]));
	       Tc = BYTW(&(W[TWVL * 2]), Tb);
	       Td = LD(&(x[WS(rs, 5)]), ms, &(x[WS(rs, 1)]));
	       Te = BYTW(&(W[TWVL * 8]), Td);
	       Tf = VSUB(Tc, Te);
	       Ti = VADD(Tc, Te);
	  }
	  {
	       V T2, T4, T1, T3;
	       T1 = LD(&(x[WS(rs, 1)]), ms, &(x[WS(rs, 1)]));
	       T2 = BYTW(&(W[0]), T1);
	       T3 = LD(&(x[WS(rs, 6)]), ms, &(x[0]));
	       T4 = BYTW(&(W[TWVL * 10]), T3);
	       T5 = VSUB(T2, T4);
	       Tk = VADD(T2, T4);
	  }
	  {
	       V T7, T9, T6, T8;
	       T6 = LD(&(x[WS(rs, 3)]), ms, &(x[WS(rs, 1)]));
	       T7 = BYTW(&(W[TWVL * 4]), T6);
	       T8 = LD(&(x[WS(rs, 4)]), ms, &(x[0]));
	       T9 = BYTW(&(W[TWVL * 6]), T8);
	       Ta = VSUB(T7, T9);
	       Tj = VADD(T7, T9);
	  }
	  ST(&(x[0]), VADD(Th, VADD(Tk, VADD(Ti, Tj))), ms, &(x[0]));
	  To = VBYI(VFNMS(LDK(KP781831482), Ta, VFNMS(LDK(KP433883739), Tf, VMUL(LDK(KP974927912), T5))));
	  Tp = VFMA(LDK(KP623489801), Tj, VFNMS(LDK(KP900968867), Ti, VFNMS(LDK(KP222520933), Tk, Th)));
	  ST(&(x[WS(rs, 2)]), VADD(To, Tp), ms, &(x[0]));
	  ST(&(x[WS(rs, 5)]), VSUB(Tp, To), ms, &(x[WS(rs, 1)]));
	  {
	       V Tg, Tl, Tm, Tn;
	       Tg = VBYI(VFMA(LDK(KP433883739), T5, VFNMS(LDK(KP781831482), Tf, VMUL(LDK(KP974927912), Ta))));
	       Tl = VFMA(LDK(KP623489801), Ti, VFNMS(LDK(KP222520933), Tj, VFNMS(LDK(KP900968867), Tk, Th)));
	       ST(&(x[WS(rs, 3)]), VADD(Tg, Tl), ms, &(x[WS(rs, 1)]));
	       ST(&(x[WS(rs, 4)]), VSUB(Tl, Tg), ms, &(x[0]));
	       Tm = VBYI(VFMA(LDK(KP781831482), T5, VFMA(LDK(KP974927912), Tf, VMUL(LDK(KP433883739), Ta))));
	       Tn = VFMA(LDK(KP623489801), Tk, VFNMS(LDK(KP900968867), Tj, VFNMS(LDK(KP222520933), Ti, Th)));
	       ST(&(x[WS(rs, 1)]), VADD(Tm, Tn), ms, &(x[WS(rs, 1)]));
	       ST(&(x[WS(rs, 6)]), VSUB(Tn, Tm), ms, &(x[0]));
	  }
     }
}

static const tw_instr twinstr[] = {
     VTW(0, 1),
     VTW(0, 2),
     VTW(0, 3),
     VTW(0, 4),
     VTW(0, 5),
     VTW(0, 6),
     {TW_NEXT, VL, 0}
};

static const ct_desc desc = { 7, "t1bv_7", twinstr, &GENUS, {24, 18, 12, 0}, 0, 0, 0 };

void X(codelet_t1bv_7) (planner *p) {
     X(kdft_dit_register) (p, t1bv_7, &desc);
}
#endif				/* HAVE_FMA */