d@0: /*
d@0:  * Copyright (c) 2003, 2007-8 Matteo Frigo
d@0:  * Copyright (c) 2003, 2007-8 Massachusetts Institute of Technology
d@0:  *
d@0:  * This program is free software; you can redistribute it and/or modify
d@0:  * it under the terms of the GNU General Public License as published by
d@0:  * the Free Software Foundation; either version 2 of the License, or
d@0:  * (at your option) any later version.
d@0:  *
d@0:  * This program is distributed in the hope that it will be useful,
d@0:  * but WITHOUT ANY WARRANTY; without even the implied warranty of
d@0:  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
d@0:  * GNU General Public License for more details.
d@0:  *
d@0:  * You should have received a copy of the GNU General Public License
d@0:  * along with this program; if not, write to the Free Software
d@0:  * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
d@0:  *
d@0:  */
d@0: /* Generated by: ../../genfft/gen_notw_c -standalone -fma -reorder-insns -simd -compact -variables 100000 -with-ostride 2 -include fftw-spu.h -store-multiple 2 -n 6 -name X(spu_n2fv_6) */
d@0: 
d@0: /*
d@0:  * This function contains 18 FP additions, 8 FP multiplications,
d@0:  * (or, 12 additions, 2 multiplications, 6 fused multiply/add),
d@0:  * 29 stack variables, 2 constants, and 15 memory accesses
d@0:  */
d@0: #include "fftw-spu.h"
d@0: 
d@0: void X(spu_n2fv_6) (const R *ri, const R *ii, R *ro, R *io, stride is, stride os, INT v, INT ivs, INT ovs) {
d@0:      DVK(KP500000000, +0.500000000000000000000000000000000000000000000);
d@0:      DVK(KP866025403, +0.866025403784438646763723170752936183471402627);
d@0:      INT i;
d@0:      const R *xi;
d@0:      R *xo;
d@0:      xi = ri;
d@0:      xo = ro;
d@0:      for (i = v; i > 0; i = i - VL, xi = xi + (VL * ivs), xo = xo + (VL * ovs), MAKE_VOLATILE_STRIDE(is), MAKE_VOLATILE_STRIDE(os)) {
d@0: 	  V Td, T3, Tc, Ta, Tg, Ti, T1, T2, Te, T6, Tf, T9, T4, T5, T7;
d@0: 	  V T8, Tj, Tk, Tl, Tb, Tm, Th, Tn, To;
d@0: 	  T1 = LD(&(xi[0]), ivs, &(xi[0]));
d@0: 	  T2 = LD(&(xi[WS(is, 3)]), ivs, &(xi[WS(is, 1)]));
d@0: 	  Td = VADD(T1, T2);
d@0: 	  T3 = VSUB(T1, T2);
d@0: 	  T4 = LD(&(xi[WS(is, 2)]), ivs, &(xi[0]));
d@0: 	  T5 = LD(&(xi[WS(is, 5)]), ivs, &(xi[WS(is, 1)]));
d@0: 	  Te = VADD(T4, T5);
d@0: 	  T6 = VSUB(T4, T5);
d@0: 	  T7 = LD(&(xi[WS(is, 4)]), ivs, &(xi[0]));
d@0: 	  T8 = LD(&(xi[WS(is, 1)]), ivs, &(xi[WS(is, 1)]));
d@0: 	  Tf = VADD(T7, T8);
d@0: 	  T9 = VSUB(T7, T8);
d@0: 	  Tc = VMUL(LDK(KP866025403), VSUB(T9, T6));
d@0: 	  Ta = VADD(T6, T9);
d@0: 	  Tg = VADD(Te, Tf);
d@0: 	  Ti = VMUL(LDK(KP866025403), VSUB(Tf, Te));
d@0: 	  Tj = VADD(T3, Ta);
d@0: 	  STM2(&(xo[6]), Tj, ovs, &(xo[2]));
d@0: 	  Tk = VADD(Td, Tg);
d@0: 	  STM2(&(xo[0]), Tk, ovs, &(xo[0]));
d@0: 	  Tb = VFNMS(LDK(KP500000000), Ta, T3);
d@0: 	  Tl = VFNMSI(Tc, Tb);
d@0: 	  STM2(&(xo[10]), Tl, ovs, &(xo[2]));
d@0: 	  Tm = VFMAI(Tc, Tb);
d@0: 	  STM2(&(xo[2]), Tm, ovs, &(xo[2]));
d@0: 	  STN2(&(xo[0]), Tk, Tm, ovs);
d@0: 	  Th = VFNMS(LDK(KP500000000), Tg, Td);
d@0: 	  Tn = VFNMSI(Ti, Th);
d@0: 	  STM2(&(xo[4]), Tn, ovs, &(xo[0]));
d@0: 	  STN2(&(xo[4]), Tn, Tj, ovs);
d@0: 	  To = VFMAI(Ti, Th);
d@0: 	  STM2(&(xo[8]), To, ovs, &(xo[0]));
d@0: 	  STN2(&(xo[8]), To, Tl, ovs);
d@0:      }
d@0: }