Chris@19: /*
Chris@19:  * Copyright (c) 2003, 2007-14 Matteo Frigo
Chris@19:  * Copyright (c) 2003, 2007-14 Massachusetts Institute of Technology
Chris@19:  *
Chris@19:  * This program is free software; you can redistribute it and/or modify
Chris@19:  * it under the terms of the GNU General Public License as published by
Chris@19:  * the Free Software Foundation; either version 2 of the License, or
Chris@19:  * (at your option) any later version.
Chris@19:  *
Chris@19:  * This program is distributed in the hope that it will be useful,
Chris@19:  * but WITHOUT ANY WARRANTY; without even the implied warranty of
Chris@19:  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
Chris@19:  * GNU General Public License for more details.
Chris@19:  *
Chris@19:  * You should have received a copy of the GNU General Public License
Chris@19:  * along with this program; if not, write to the Free Software
Chris@19:  * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA  02110-1301  USA
Chris@19:  *
Chris@19:  */
Chris@19: 
Chris@19: /* This file was automatically generated --- DO NOT EDIT */
Chris@19: /* Generated on Tue Mar  4 13:45:51 EST 2014 */
Chris@19: 
Chris@19: #include "codelet-dft.h"
Chris@19: 
Chris@19: #ifdef HAVE_FMA
Chris@19: 
Chris@19: /* Generated by: ../../../genfft/gen_twiddle.native -fma -reorder-insns -schedule-for-pipeline -compact -variables 4 -pipeline-latency 4 -n 7 -name t1_7 -include t.h */
Chris@19: 
Chris@19: /*
Chris@19:  * This function contains 72 FP additions, 66 FP multiplications,
Chris@19:  * (or, 18 additions, 12 multiplications, 54 fused multiply/add),
Chris@19:  * 66 stack variables, 6 constants, and 28 memory accesses
Chris@19:  */
Chris@19: #include "t.h"
Chris@19: 
Chris@19: static void t1_7(R *ri, R *ii, const R *W, stride rs, INT mb, INT me, INT ms)
Chris@19: {
Chris@19:      DK(KP974927912, +0.974927912181823607018131682993931217232785801);
Chris@19:      DK(KP801937735, +0.801937735804838252472204639014890102331838324);
Chris@19:      DK(KP900968867, +0.900968867902419126236102319507445051165919162);
Chris@19:      DK(KP692021471, +0.692021471630095869627814897002069140197260599);
Chris@19:      DK(KP554958132, +0.554958132087371191422194871006410481067288862);
Chris@19:      DK(KP356895867, +0.356895867892209443894399510021300583399127187);
Chris@19:      {
Chris@19: 	  INT m;
Chris@19: 	  for (m = mb, W = W + (mb * 12); m < me; m = m + 1, ri = ri + ms, ii = ii + ms, W = W + 12, MAKE_VOLATILE_STRIDE(14, rs)) {
Chris@19: 	       E T1c, T19, T1i, T18, T16, T1q, T1t, T1r, T1u, T1s;
Chris@19: 	       {
Chris@19: 		    E T1, TR, T1h, Te, Tt, Tw, T1a, TM, T1g, Tr, Tu, TS, Tz, TC, Ty;
Chris@19: 		    E Tv, TB;
Chris@19: 		    T1 = ri[0];
Chris@19: 		    T1c = ii[0];
Chris@19: 		    {
Chris@19: 			 E T9, Tc, TP, Ta, Tb, TO, T7;
Chris@19: 			 {
Chris@19: 			      E T3, T6, T8, TN, T4, T2, T5;
Chris@19: 			      T3 = ri[WS(rs, 1)];
Chris@19: 			      T6 = ii[WS(rs, 1)];
Chris@19: 			      T2 = W[0];
Chris@19: 			      T9 = ri[WS(rs, 6)];
Chris@19: 			      Tc = ii[WS(rs, 6)];
Chris@19: 			      T8 = W[10];
Chris@19: 			      TN = T2 * T6;
Chris@19: 			      T4 = T2 * T3;
Chris@19: 			      T5 = W[1];
Chris@19: 			      TP = T8 * Tc;
Chris@19: 			      Ta = T8 * T9;
Chris@19: 			      Tb = W[11];
Chris@19: 			      TO = FNMS(T5, T3, TN);
Chris@19: 			      T7 = FMA(T5, T6, T4);
Chris@19: 			 }
Chris@19: 			 {
Chris@19: 			      E Tg, Tj, Th, TI, Tm, Tp, Tl, Ti, To, TQ, Td, Tf;
Chris@19: 			      Tg = ri[WS(rs, 2)];
Chris@19: 			      TQ = FNMS(Tb, T9, TP);
Chris@19: 			      Td = FMA(Tb, Tc, Ta);
Chris@19: 			      Tj = ii[WS(rs, 2)];
Chris@19: 			      Tf = W[2];
Chris@19: 			      T19 = TO + TQ;
Chris@19: 			      TR = TO - TQ;
Chris@19: 			      T1h = Td - T7;
Chris@19: 			      Te = T7 + Td;
Chris@19: 			      Th = Tf * Tg;
Chris@19: 			      TI = Tf * Tj;
Chris@19: 			      Tm = ri[WS(rs, 5)];
Chris@19: 			      Tp = ii[WS(rs, 5)];
Chris@19: 			      Tl = W[8];
Chris@19: 			      Ti = W[3];
Chris@19: 			      To = W[9];
Chris@19: 			      {
Chris@19: 				   E TJ, Tk, TL, Tq, TK, Tn, Ts;
Chris@19: 				   Tt = ri[WS(rs, 3)];
Chris@19: 				   TK = Tl * Tp;
Chris@19: 				   Tn = Tl * Tm;
Chris@19: 				   TJ = FNMS(Ti, Tg, TI);
Chris@19: 				   Tk = FMA(Ti, Tj, Th);
Chris@19: 				   TL = FNMS(To, Tm, TK);
Chris@19: 				   Tq = FMA(To, Tp, Tn);
Chris@19: 				   Tw = ii[WS(rs, 3)];
Chris@19: 				   Ts = W[4];
Chris@19: 				   T1a = TJ + TL;
Chris@19: 				   TM = TJ - TL;
Chris@19: 				   T1g = Tq - Tk;
Chris@19: 				   Tr = Tk + Tq;
Chris@19: 				   Tu = Ts * Tt;
Chris@19: 				   TS = Ts * Tw;
Chris@19: 			      }
Chris@19: 			      Tz = ri[WS(rs, 4)];
Chris@19: 			      TC = ii[WS(rs, 4)];
Chris@19: 			      Ty = W[6];
Chris@19: 			      Tv = W[5];
Chris@19: 			      TB = W[7];
Chris@19: 			 }
Chris@19: 		    }
Chris@19: 		    {
Chris@19: 			 E TF, TT, Tx, TV, TD, T1d, TU, TA;
Chris@19: 			 TF = FNMS(KP356895867, Tr, Te);
Chris@19: 			 TU = Ty * TC;
Chris@19: 			 TA = Ty * Tz;
Chris@19: 			 TT = FNMS(Tv, Tt, TS);
Chris@19: 			 Tx = FMA(Tv, Tw, Tu);
Chris@19: 			 TV = FNMS(TB, Tz, TU);
Chris@19: 			 TD = FMA(TB, TC, TA);
Chris@19: 			 T1d = FNMS(KP356895867, T1a, T19);
Chris@19: 			 {
Chris@19: 			      E T1b, T15, T17, TW;
Chris@19: 			      T17 = FNMS(KP554958132, TR, TM);
Chris@19: 			      T1b = TT + TV;
Chris@19: 			      TW = TT - TV;
Chris@19: 			      {
Chris@19: 				   E TE, T1l, T1e, T12;
Chris@19: 				   T1i = TD - Tx;
Chris@19: 				   TE = Tx + TD;
Chris@19: 				   T1l = FNMS(KP356895867, T19, T1b);
Chris@19: 				   T1e = FNMS(KP692021471, T1d, T1b);
Chris@19: 				   ii[0] = T19 + T1a + T1b + T1c;
Chris@19: 				   T12 = FMA(KP554958132, TM, TW);
Chris@19: 				   {
Chris@19: 					E TX, T1o, T1j, T14;
Chris@19: 					TX = FMA(KP554958132, TW, TR);
Chris@19: 					T1o = FMA(KP554958132, T1g, T1i);
Chris@19: 					T1j = FMA(KP554958132, T1i, T1h);
Chris@19: 					T14 = FNMS(KP356895867, TE, Tr);
Chris@19: 					{
Chris@19: 					     E TZ, TG, T1m, T1f;
Chris@19: 					     TZ = FNMS(KP356895867, Te, TE);
Chris@19: 					     TG = FNMS(KP692021471, TF, TE);
Chris@19: 					     ri[0] = T1 + Te + Tr + TE;
Chris@19: 					     T1m = FNMS(KP692021471, T1l, T1a);
Chris@19: 					     T1f = FNMS(KP900968867, T1e, T1c);
Chris@19: 					     {
Chris@19: 						  E T13, TY, T1p, T1k;
Chris@19: 						  T13 = FNMS(KP801937735, T12, TR);
Chris@19: 						  TY = FMA(KP801937735, TX, TM);
Chris@19: 						  T1p = FNMS(KP801937735, T1o, T1h);
Chris@19: 						  T1k = FMA(KP801937735, T1j, T1g);
Chris@19: 						  T15 = FNMS(KP692021471, T14, Te);
Chris@19: 						  {
Chris@19: 						       E T10, TH, T1n, T11;
Chris@19: 						       T10 = FNMS(KP692021471, TZ, Tr);
Chris@19: 						       TH = FNMS(KP900968867, TG, T1);
Chris@19: 						       T1n = FNMS(KP900968867, T1m, T1c);
Chris@19: 						       ii[WS(rs, 6)] = FNMS(KP974927912, T1k, T1f);
Chris@19: 						       ii[WS(rs, 1)] = FMA(KP974927912, T1k, T1f);
Chris@19: 						       T11 = FNMS(KP900968867, T10, T1);
Chris@19: 						       ri[WS(rs, 1)] = FMA(KP974927912, TY, TH);
Chris@19: 						       ri[WS(rs, 6)] = FNMS(KP974927912, TY, TH);
Chris@19: 						       ii[WS(rs, 5)] = FNMS(KP974927912, T1p, T1n);
Chris@19: 						       ii[WS(rs, 2)] = FMA(KP974927912, T1p, T1n);
Chris@19: 						       ri[WS(rs, 2)] = FMA(KP974927912, T13, T11);
Chris@19: 						       ri[WS(rs, 5)] = FNMS(KP974927912, T13, T11);
Chris@19: 						       T18 = FNMS(KP801937735, T17, TW);
Chris@19: 						  }
Chris@19: 					     }
Chris@19: 					}
Chris@19: 				   }
Chris@19: 			      }
Chris@19: 			      T16 = FNMS(KP900968867, T15, T1);
Chris@19: 			      T1q = FNMS(KP356895867, T1b, T1a);
Chris@19: 			      T1t = FNMS(KP554958132, T1h, T1g);
Chris@19: 			 }
Chris@19: 		    }
Chris@19: 	       }
Chris@19: 	       ri[WS(rs, 3)] = FMA(KP974927912, T18, T16);
Chris@19: 	       ri[WS(rs, 4)] = FNMS(KP974927912, T18, T16);
Chris@19: 	       T1r = FNMS(KP692021471, T1q, T19);
Chris@19: 	       T1u = FNMS(KP801937735, T1t, T1i);
Chris@19: 	       T1s = FNMS(KP900968867, T1r, T1c);
Chris@19: 	       ii[WS(rs, 4)] = FNMS(KP974927912, T1u, T1s);
Chris@19: 	       ii[WS(rs, 3)] = FMA(KP974927912, T1u, T1s);
Chris@19: 	  }
Chris@19:      }
Chris@19: }
Chris@19: 
Chris@19: static const tw_instr twinstr[] = {
Chris@19:      {TW_FULL, 0, 7},
Chris@19:      {TW_NEXT, 1, 0}
Chris@19: };
Chris@19: 
Chris@19: static const ct_desc desc = { 7, "t1_7", twinstr, &GENUS, {18, 12, 54, 0}, 0, 0, 0 };
Chris@19: 
Chris@19: void X(codelet_t1_7) (planner *p) {
Chris@19:      X(kdft_dit_register) (p, t1_7, &desc);
Chris@19: }
Chris@19: #else				/* HAVE_FMA */
Chris@19: 
Chris@19: /* Generated by: ../../../genfft/gen_twiddle.native -compact -variables 4 -pipeline-latency 4 -n 7 -name t1_7 -include t.h */
Chris@19: 
Chris@19: /*
Chris@19:  * This function contains 72 FP additions, 60 FP multiplications,
Chris@19:  * (or, 36 additions, 24 multiplications, 36 fused multiply/add),
Chris@19:  * 29 stack variables, 6 constants, and 28 memory accesses
Chris@19:  */
Chris@19: #include "t.h"
Chris@19: 
Chris@19: static void t1_7(R *ri, R *ii, const R *W, stride rs, INT mb, INT me, INT ms)
Chris@19: {
Chris@19:      DK(KP222520933, +0.222520933956314404288902564496794759466355569);
Chris@19:      DK(KP900968867, +0.900968867902419126236102319507445051165919162);
Chris@19:      DK(KP623489801, +0.623489801858733530525004884004239810632274731);
Chris@19:      DK(KP433883739, +0.433883739117558120475768332848358754609990728);
Chris@19:      DK(KP781831482, +0.781831482468029808708444526674057750232334519);
Chris@19:      DK(KP974927912, +0.974927912181823607018131682993931217232785801);
Chris@19:      {
Chris@19: 	  INT m;
Chris@19: 	  for (m = mb, W = W + (mb * 12); m < me; m = m + 1, ri = ri + ms, ii = ii + ms, W = W + 12, MAKE_VOLATILE_STRIDE(14, rs)) {
Chris@19: 	       E T1, TR, Tc, TS, TC, TO, Tn, TT, TI, TP, Ty, TU, TF, TQ;
Chris@19: 	       T1 = ri[0];
Chris@19: 	       TR = ii[0];
Chris@19: 	       {
Chris@19: 		    E T6, TA, Tb, TB;
Chris@19: 		    {
Chris@19: 			 E T3, T5, T2, T4;
Chris@19: 			 T3 = ri[WS(rs, 1)];
Chris@19: 			 T5 = ii[WS(rs, 1)];
Chris@19: 			 T2 = W[0];
Chris@19: 			 T4 = W[1];
Chris@19: 			 T6 = FMA(T2, T3, T4 * T5);
Chris@19: 			 TA = FNMS(T4, T3, T2 * T5);
Chris@19: 		    }
Chris@19: 		    {
Chris@19: 			 E T8, Ta, T7, T9;
Chris@19: 			 T8 = ri[WS(rs, 6)];
Chris@19: 			 Ta = ii[WS(rs, 6)];
Chris@19: 			 T7 = W[10];
Chris@19: 			 T9 = W[11];
Chris@19: 			 Tb = FMA(T7, T8, T9 * Ta);
Chris@19: 			 TB = FNMS(T9, T8, T7 * Ta);
Chris@19: 		    }
Chris@19: 		    Tc = T6 + Tb;
Chris@19: 		    TS = Tb - T6;
Chris@19: 		    TC = TA - TB;
Chris@19: 		    TO = TA + TB;
Chris@19: 	       }
Chris@19: 	       {
Chris@19: 		    E Th, TG, Tm, TH;
Chris@19: 		    {
Chris@19: 			 E Te, Tg, Td, Tf;
Chris@19: 			 Te = ri[WS(rs, 2)];
Chris@19: 			 Tg = ii[WS(rs, 2)];
Chris@19: 			 Td = W[2];
Chris@19: 			 Tf = W[3];
Chris@19: 			 Th = FMA(Td, Te, Tf * Tg);
Chris@19: 			 TG = FNMS(Tf, Te, Td * Tg);
Chris@19: 		    }
Chris@19: 		    {
Chris@19: 			 E Tj, Tl, Ti, Tk;
Chris@19: 			 Tj = ri[WS(rs, 5)];
Chris@19: 			 Tl = ii[WS(rs, 5)];
Chris@19: 			 Ti = W[8];
Chris@19: 			 Tk = W[9];
Chris@19: 			 Tm = FMA(Ti, Tj, Tk * Tl);
Chris@19: 			 TH = FNMS(Tk, Tj, Ti * Tl);
Chris@19: 		    }
Chris@19: 		    Tn = Th + Tm;
Chris@19: 		    TT = Tm - Th;
Chris@19: 		    TI = TG - TH;
Chris@19: 		    TP = TG + TH;
Chris@19: 	       }
Chris@19: 	       {
Chris@19: 		    E Ts, TD, Tx, TE;
Chris@19: 		    {
Chris@19: 			 E Tp, Tr, To, Tq;
Chris@19: 			 Tp = ri[WS(rs, 3)];
Chris@19: 			 Tr = ii[WS(rs, 3)];
Chris@19: 			 To = W[4];
Chris@19: 			 Tq = W[5];
Chris@19: 			 Ts = FMA(To, Tp, Tq * Tr);
Chris@19: 			 TD = FNMS(Tq, Tp, To * Tr);
Chris@19: 		    }
Chris@19: 		    {
Chris@19: 			 E Tu, Tw, Tt, Tv;
Chris@19: 			 Tu = ri[WS(rs, 4)];
Chris@19: 			 Tw = ii[WS(rs, 4)];
Chris@19: 			 Tt = W[6];
Chris@19: 			 Tv = W[7];
Chris@19: 			 Tx = FMA(Tt, Tu, Tv * Tw);
Chris@19: 			 TE = FNMS(Tv, Tu, Tt * Tw);
Chris@19: 		    }
Chris@19: 		    Ty = Ts + Tx;
Chris@19: 		    TU = Tx - Ts;
Chris@19: 		    TF = TD - TE;
Chris@19: 		    TQ = TD + TE;
Chris@19: 	       }
Chris@19: 	       ri[0] = T1 + Tc + Tn + Ty;
Chris@19: 	       ii[0] = TO + TP + TQ + TR;
Chris@19: 	       {
Chris@19: 		    E TJ, Tz, TX, TY;
Chris@19: 		    TJ = FNMS(KP781831482, TF, KP974927912 * TC) - (KP433883739 * TI);
Chris@19: 		    Tz = FMA(KP623489801, Ty, T1) + FNMA(KP900968867, Tn, KP222520933 * Tc);
Chris@19: 		    ri[WS(rs, 5)] = Tz - TJ;
Chris@19: 		    ri[WS(rs, 2)] = Tz + TJ;
Chris@19: 		    TX = FNMS(KP781831482, TU, KP974927912 * TS) - (KP433883739 * TT);
Chris@19: 		    TY = FMA(KP623489801, TQ, TR) + FNMA(KP900968867, TP, KP222520933 * TO);
Chris@19: 		    ii[WS(rs, 2)] = TX + TY;
Chris@19: 		    ii[WS(rs, 5)] = TY - TX;
Chris@19: 	       }
Chris@19: 	       {
Chris@19: 		    E TL, TK, TV, TW;
Chris@19: 		    TL = FMA(KP781831482, TC, KP974927912 * TI) + (KP433883739 * TF);
Chris@19: 		    TK = FMA(KP623489801, Tc, T1) + FNMA(KP900968867, Ty, KP222520933 * Tn);
Chris@19: 		    ri[WS(rs, 6)] = TK - TL;
Chris@19: 		    ri[WS(rs, 1)] = TK + TL;
Chris@19: 		    TV = FMA(KP781831482, TS, KP974927912 * TT) + (KP433883739 * TU);
Chris@19: 		    TW = FMA(KP623489801, TO, TR) + FNMA(KP900968867, TQ, KP222520933 * TP);
Chris@19: 		    ii[WS(rs, 1)] = TV + TW;
Chris@19: 		    ii[WS(rs, 6)] = TW - TV;
Chris@19: 	       }
Chris@19: 	       {
Chris@19: 		    E TN, TM, TZ, T10;
Chris@19: 		    TN = FMA(KP433883739, TC, KP974927912 * TF) - (KP781831482 * TI);
Chris@19: 		    TM = FMA(KP623489801, Tn, T1) + FNMA(KP222520933, Ty, KP900968867 * Tc);
Chris@19: 		    ri[WS(rs, 4)] = TM - TN;
Chris@19: 		    ri[WS(rs, 3)] = TM + TN;
Chris@19: 		    TZ = FMA(KP433883739, TS, KP974927912 * TU) - (KP781831482 * TT);
Chris@19: 		    T10 = FMA(KP623489801, TP, TR) + FNMA(KP222520933, TQ, KP900968867 * TO);
Chris@19: 		    ii[WS(rs, 3)] = TZ + T10;
Chris@19: 		    ii[WS(rs, 4)] = T10 - TZ;
Chris@19: 	       }
Chris@19: 	  }
Chris@19:      }
Chris@19: }
Chris@19: 
Chris@19: static const tw_instr twinstr[] = {
Chris@19:      {TW_FULL, 0, 7},
Chris@19:      {TW_NEXT, 1, 0}
Chris@19: };
Chris@19: 
Chris@19: static const ct_desc desc = { 7, "t1_7", twinstr, &GENUS, {36, 24, 36, 0}, 0, 0, 0 };
Chris@19: 
Chris@19: void X(codelet_t1_7) (planner *p) {
Chris@19:      X(kdft_dit_register) (p, t1_7, &desc);
Chris@19: }
Chris@19: #endif				/* HAVE_FMA */