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 10 -name X(spu_n2fv_10) */
d@0: 
d@0: /*
d@0:  * This function contains 42 FP additions, 22 FP multiplications,
d@0:  * (or, 24 additions, 4 multiplications, 18 fused multiply/add),
d@0:  * 59 stack variables, 4 constants, and 25 memory accesses
d@0:  */
d@0: #include "fftw-spu.h"
d@0: 
d@0: void X(spu_n2fv_10) (const R *ri, const R *ii, R *ro, R *io, stride is, stride os, INT v, INT ivs, INT ovs) {
d@0:      DVK(KP559016994, +0.559016994374947424102293417182819058860154590);
d@0:      DVK(KP250000000, +0.250000000000000000000000000000000000000000000);
d@0:      DVK(KP618033988, +0.618033988749894848204586834365638117720309180);
d@0:      DVK(KP951056516, +0.951056516295153572116439333379382143405698634);
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 Tr, T3, Tm, Tn, TD, TC, TA, Ty, Ti, Tk, T1, T2, Ts, T6, Tw;
d@0: 	  V Tg, Tt, T9, Tv, Td, T4, T5, Te, Tf, T7, T8, Tb, Tc, Ta, Th;
d@0: 	  V Tu, Tx, TH, TI, TK, TL, TM, Tq, To, Tp, Tl, Tj, TJ, TG, TE;
d@0: 	  V TF, TB, Tz, TN, TO, TP, TQ;
d@0: 	  T1 = LD(&(xi[0]), ivs, &(xi[0]));
d@0: 	  T2 = LD(&(xi[WS(is, 5)]), ivs, &(xi[WS(is, 1)]));
d@0: 	  Tr = 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, 7)]), ivs, &(xi[WS(is, 1)]));
d@0: 	  Ts = VADD(T4, T5);
d@0: 	  T6 = VSUB(T4, T5);
d@0: 	  Te = LD(&(xi[WS(is, 6)]), ivs, &(xi[0]));
d@0: 	  Tf = LD(&(xi[WS(is, 1)]), ivs, &(xi[WS(is, 1)]));
d@0: 	  Tw = VADD(Te, Tf);
d@0: 	  Tg = VSUB(Te, Tf);
d@0: 	  T7 = LD(&(xi[WS(is, 8)]), ivs, &(xi[0]));
d@0: 	  T8 = LD(&(xi[WS(is, 3)]), ivs, &(xi[WS(is, 1)]));
d@0: 	  Tt = VADD(T7, T8);
d@0: 	  T9 = VSUB(T7, T8);
d@0: 	  Tb = LD(&(xi[WS(is, 4)]), ivs, &(xi[0]));
d@0: 	  Tc = LD(&(xi[WS(is, 9)]), ivs, &(xi[WS(is, 1)]));
d@0: 	  Tv = VADD(Tb, Tc);
d@0: 	  Td = VSUB(Tb, Tc);
d@0: 	  Ta = VADD(T6, T9);
d@0: 	  Tm = VSUB(T6, T9);
d@0: 	  Tn = VSUB(Td, Tg);
d@0: 	  Th = VADD(Td, Tg);
d@0: 	  Tu = VADD(Ts, Tt);
d@0: 	  TD = VSUB(Ts, Tt);
d@0: 	  TC = VSUB(Tv, Tw);
d@0: 	  Tx = VADD(Tv, Tw);
d@0: 	  TA = VSUB(Tu, Tx);
d@0: 	  Ty = VADD(Tu, Tx);
d@0: 	  Ti = VADD(Ta, Th);
d@0: 	  Tk = VSUB(Ta, Th);
d@0: 	  TH = VADD(T3, Ti);
d@0: 	  STM2(&(xo[10]), TH, ovs, &(xo[2]));
d@0: 	  TI = VADD(Tr, Ty);
d@0: 	  STM2(&(xo[0]), TI, ovs, &(xo[0]));
d@0: 	  Tq = VMUL(LDK(KP951056516), VFNMS(LDK(KP618033988), Tm, Tn));
d@0: 	  To = VMUL(LDK(KP951056516), VFMA(LDK(KP618033988), Tn, Tm));
d@0: 	  Tj = VFNMS(LDK(KP250000000), Ti, T3);
d@0: 	  Tp = VFNMS(LDK(KP559016994), Tk, Tj);
d@0: 	  Tl = VFMA(LDK(KP559016994), Tk, Tj);
d@0: 	  TJ = VFNMSI(To, Tl);
d@0: 	  STM2(&(xo[2]), TJ, ovs, &(xo[2]));
d@0: 	  STN2(&(xo[0]), TI, TJ, ovs);
d@0: 	  TK = VFMAI(Tq, Tp);
d@0: 	  STM2(&(xo[14]), TK, ovs, &(xo[2]));
d@0: 	  TL = VFMAI(To, Tl);
d@0: 	  STM2(&(xo[18]), TL, ovs, &(xo[2]));
d@0: 	  TM = VFNMSI(Tq, Tp);
d@0: 	  STM2(&(xo[6]), TM, ovs, &(xo[2]));
d@0: 	  TG = VMUL(LDK(KP951056516), VFMA(LDK(KP618033988), TC, TD));
d@0: 	  TE = VMUL(LDK(KP951056516), VFNMS(LDK(KP618033988), TD, TC));
d@0: 	  Tz = VFNMS(LDK(KP250000000), Ty, Tr);
d@0: 	  TF = VFMA(LDK(KP559016994), TA, Tz);
d@0: 	  TB = VFNMS(LDK(KP559016994), TA, Tz);
d@0: 	  TN = VFMAI(TE, TB);
d@0: 	  STM2(&(xo[4]), TN, ovs, &(xo[0]));
d@0: 	  STN2(&(xo[4]), TN, TM, ovs);
d@0: 	  TO = VFNMSI(TG, TF);
d@0: 	  STM2(&(xo[12]), TO, ovs, &(xo[0]));
d@0: 	  STN2(&(xo[12]), TO, TK, ovs);
d@0: 	  TP = VFNMSI(TE, TB);
d@0: 	  STM2(&(xo[16]), TP, ovs, &(xo[0]));
d@0: 	  STN2(&(xo[16]), TP, TL, ovs);
d@0: 	  TQ = VFMAI(TG, TF);
d@0: 	  STM2(&(xo[8]), TQ, ovs, &(xo[0]));
d@0: 	  STN2(&(xo[8]), TQ, TH, ovs);
d@0:      }
d@0: }