annotate fft/fftw/fftw-3.3.4/rdft/scalar/r2cb/r2cb_6.c @ 40:223f770b5341 kissfft-double tip

Try a double-precision kissfft

/*
 * Copyright (c) 2003, 2007-14 Matteo Frigo
 * Copyright (c) 2003, 2007-14 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., 51 Franklin Street, Fifth Floor, Boston, MA  02110-1301  USA
 *
 */

/* This file was automatically generated --- DO NOT EDIT */
/* Generated on Tue Mar  4 13:50:24 EST 2014 */

#include "codelet-rdft.h"

#ifdef HAVE_FMA

/* Generated by: ../../../genfft/gen_r2cb.native -fma -reorder-insns -schedule-for-pipeline -compact -variables 4 -pipeline-latency 4 -sign 1 -n 6 -name r2cb_6 -include r2cb.h */

/*
 * This function contains 14 FP additions, 6 FP multiplications,
 * (or, 8 additions, 0 multiplications, 6 fused multiply/add),
 * 13 stack variables, 2 constants, and 12 memory accesses
 */
#include "r2cb.h"

static void r2cb_6(R *R0, R *R1, R *Cr, R *Ci, stride rs, stride csr, stride csi, INT v, INT ivs, INT ovs)
{
     DK(KP1_732050807, +1.732050807568877293527446341505872366942805254);
     DK(KP2_000000000, +2.000000000000000000000000000000000000000000000);
     {
	  INT i;
	  for (i = v; i > 0; i = i - 1, R0 = R0 + ovs, R1 = R1 + ovs, Cr = Cr + ivs, Ci = Ci + ivs, MAKE_VOLATILE_STRIDE(24, rs), MAKE_VOLATILE_STRIDE(24, csr), MAKE_VOLATILE_STRIDE(24, csi)) {
	       E T4, T7, T3, Te, Tc, T5;
	       {
		    E T1, T2, Ta, Tb;
		    T1 = Cr[0];
		    T2 = Cr[WS(csr, 3)];
		    Ta = Ci[WS(csi, 2)];
		    Tb = Ci[WS(csi, 1)];
		    T4 = Cr[WS(csr, 2)];
		    T7 = T1 - T2;
		    T3 = T1 + T2;
		    Te = Ta + Tb;
		    Tc = Ta - Tb;
		    T5 = Cr[WS(csr, 1)];
	       }
	       {
		    E T6, T8, Td, T9;
		    T6 = T4 + T5;
		    T8 = T5 - T4;
		    Td = T7 + T8;
		    R1[WS(rs, 1)] = FNMS(KP2_000000000, T8, T7);
		    T9 = T3 - T6;
		    R0[0] = FMA(KP2_000000000, T6, T3);
		    R1[WS(rs, 2)] = FMA(KP1_732050807, Te, Td);
		    R1[0] = FNMS(KP1_732050807, Te, Td);
		    R0[WS(rs, 1)] = FMA(KP1_732050807, Tc, T9);
		    R0[WS(rs, 2)] = FNMS(KP1_732050807, Tc, T9);
	       }
	  }
     }
}

static const kr2c_desc desc = { 6, "r2cb_6", {8, 0, 6, 0}, &GENUS };

void X(codelet_r2cb_6) (planner *p) {
     X(kr2c_register) (p, r2cb_6, &desc);
}

#else				/* HAVE_FMA */

/* Generated by: ../../../genfft/gen_r2cb.native -compact -variables 4 -pipeline-latency 4 -sign 1 -n 6 -name r2cb_6 -include r2cb.h */

/*
 * This function contains 14 FP additions, 4 FP multiplications,
 * (or, 12 additions, 2 multiplications, 2 fused multiply/add),
 * 17 stack variables, 2 constants, and 12 memory accesses
 */
#include "r2cb.h"

static void r2cb_6(R *R0, R *R1, R *Cr, R *Ci, stride rs, stride csr, stride csi, INT v, INT ivs, INT ovs)
{
     DK(KP2_000000000, +2.000000000000000000000000000000000000000000000);
     DK(KP1_732050807, +1.732050807568877293527446341505872366942805254);
     {
	  INT i;
	  for (i = v; i > 0; i = i - 1, R0 = R0 + ovs, R1 = R1 + ovs, Cr = Cr + ivs, Ci = Ci + ivs, MAKE_VOLATILE_STRIDE(24, rs), MAKE_VOLATILE_STRIDE(24, csr), MAKE_VOLATILE_STRIDE(24, csi)) {
	       E T3, T7, Tc, Te, T6, T8, T1, T2, T9, Td;
	       T1 = Cr[0];
	       T2 = Cr[WS(csr, 3)];
	       T3 = T1 - T2;
	       T7 = T1 + T2;
	       {
		    E Ta, Tb, T4, T5;
		    Ta = Ci[WS(csi, 2)];
		    Tb = Ci[WS(csi, 1)];
		    Tc = KP1_732050807 * (Ta - Tb);
		    Te = KP1_732050807 * (Ta + Tb);
		    T4 = Cr[WS(csr, 2)];
		    T5 = Cr[WS(csr, 1)];
		    T6 = T4 - T5;
		    T8 = T4 + T5;
	       }
	       R1[WS(rs, 1)] = FMA(KP2_000000000, T6, T3);
	       R0[0] = FMA(KP2_000000000, T8, T7);
	       T9 = T7 - T8;
	       R0[WS(rs, 2)] = T9 - Tc;
	       R0[WS(rs, 1)] = T9 + Tc;
	       Td = T3 - T6;
	       R1[0] = Td - Te;
	       R1[WS(rs, 2)] = Td + Te;
	  }
     }
}

static const kr2c_desc desc = { 6, "r2cb_6", {12, 2, 2, 0}, &GENUS };

void X(codelet_r2cb_6) (planner *p) {
     X(kr2c_register) (p, r2cb_6, &desc);
}

#endif				/* HAVE_FMA */