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_twiddle_c -standalone -fma -reorder-insns -simd -compact -variables 100000 -include fftw-spu.h -trivial-stores -n 7 -name X(spu_t1fv_7) */
d@0: 
d@0: /*
d@0:  * This function contains 36 FP additions, 36 FP multiplications,
d@0:  * (or, 15 additions, 15 multiplications, 21 fused multiply/add),
d@0:  * 49 stack variables, 6 constants, and 14 memory accesses
d@0:  */
d@0: #include "fftw-spu.h"
d@0: 
d@0: void X(spu_t1fv_7) (R *ri, R *ii, const R *W, stride rs, INT mb, INT me, INT ms) {
d@0:      DVK(KP900968867, +0.900968867902419126236102319507445051165919162);
d@0:      DVK(KP692021471, +0.692021471630095869627814897002069140197260599);
d@0:      DVK(KP801937735, +0.801937735804838252472204639014890102331838324);
d@0:      DVK(KP974927912, +0.974927912181823607018131682993931217232785801);
d@0:      DVK(KP554958132, +0.554958132087371191422194871006410481067288862);
d@0:      DVK(KP356895867, +0.356895867892209443894399510021300583399127187);
d@0:      INT m;
d@0:      R *x;
d@0:      x = ri;
d@0:      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)) {
d@0: 	  V T1, T6, Tl, Tm, Tg, Tb, Tn, Ti, Tz, Ts, To, TC, Tv, T3, T5;
d@0: 	  V Tf, Td, Ta, T8, T2, T4, Te, Tc, T9, T7, Th, TD, TB, TA, TF;
d@0: 	  V TE, Tp, Tk, Tj, Tr, Tq, Tw, Tu, Tt, Ty, Tx;
d@0: 	  T1 = LD(&(x[0]), ms, &(x[0]));
d@0: 	  T2 = LD(&(x[WS(rs, 1)]), ms, &(x[WS(rs, 1)]));
d@0: 	  T3 = BYTWJ(&(W[0]), T2);
d@0: 	  T4 = LD(&(x[WS(rs, 6)]), ms, &(x[0]));
d@0: 	  T5 = BYTWJ(&(W[TWVL * 10]), T4);
d@0: 	  Te = LD(&(x[WS(rs, 4)]), ms, &(x[0]));
d@0: 	  Tf = BYTWJ(&(W[TWVL * 6]), Te);
d@0: 	  Tc = LD(&(x[WS(rs, 3)]), ms, &(x[WS(rs, 1)]));
d@0: 	  Td = BYTWJ(&(W[TWVL * 4]), Tc);
d@0: 	  T9 = LD(&(x[WS(rs, 5)]), ms, &(x[WS(rs, 1)]));
d@0: 	  Ta = BYTWJ(&(W[TWVL * 8]), T9);
d@0: 	  T7 = LD(&(x[WS(rs, 2)]), ms, &(x[0]));
d@0: 	  T8 = BYTWJ(&(W[TWVL * 2]), T7);
d@0: 	  T6 = VADD(T3, T5);
d@0: 	  Tl = VSUB(T5, T3);
d@0: 	  Tm = VSUB(Tf, Td);
d@0: 	  Tg = VADD(Td, Tf);
d@0: 	  Tb = VADD(T8, Ta);
d@0: 	  Tn = VSUB(Ta, T8);
d@0: 	  Ti = VFNMS(LDK(KP356895867), T6, Tg);
d@0: 	  Tz = VFNMS(LDK(KP356895867), Tg, Tb);
d@0: 	  Ts = VFNMS(LDK(KP356895867), Tb, T6);
d@0: 	  To = VFMA(LDK(KP554958132), Tn, Tm);
d@0: 	  TC = VFNMS(LDK(KP554958132), Tl, Tn);
d@0: 	  Tv = VFMA(LDK(KP554958132), Tm, Tl);
d@0: 	  Th = VADD(T1, VADD(T6, VADD(Tb, Tg)));
d@0: 	  ST(&(x[0]), Th, ms, &(x[0]));
d@0: 	  TD = VMUL(LDK(KP974927912), VFNMS(LDK(KP801937735), TC, Tm));
d@0: 	  TA = VFNMS(LDK(KP692021471), Tz, T6);
d@0: 	  TB = VFNMS(LDK(KP900968867), TA, T1);
d@0: 	  TF = VFMAI(TD, TB);
d@0: 	  TE = VFNMSI(TD, TB);
d@0: 	  ST(&(x[WS(rs, 4)]), TE, ms, &(x[0]));
d@0: 	  ST(&(x[WS(rs, 3)]), TF, ms, &(x[WS(rs, 1)]));
d@0: 	  Tp = VMUL(LDK(KP974927912), VFNMS(LDK(KP801937735), To, Tl));
d@0: 	  Tj = VFNMS(LDK(KP692021471), Ti, Tb);
d@0: 	  Tk = VFNMS(LDK(KP900968867), Tj, T1);
d@0: 	  Tq = VFNMSI(Tp, Tk);
d@0: 	  Tr = VFMAI(Tp, Tk);
d@0: 	  ST(&(x[WS(rs, 5)]), Tq, ms, &(x[WS(rs, 1)]));
d@0: 	  ST(&(x[WS(rs, 2)]), Tr, ms, &(x[0]));
d@0: 	  Tw = VMUL(LDK(KP974927912), VFMA(LDK(KP801937735), Tv, Tn));
d@0: 	  Tt = VFNMS(LDK(KP692021471), Ts, Tg);
d@0: 	  Tu = VFNMS(LDK(KP900968867), Tt, T1);
d@0: 	  Ty = VFMAI(Tw, Tu);
d@0: 	  Tx = VFNMSI(Tw, Tu);
d@0: 	  ST(&(x[WS(rs, 6)]), Tx, ms, &(x[0]));
d@0: 	  ST(&(x[WS(rs, 1)]), Ty, ms, &(x[WS(rs, 1)]));
d@0:      }
d@0: }