Chris@10: /*
Chris@10:  * Copyright (c) 2003, 2007-11 Matteo Frigo
Chris@10:  * Copyright (c) 2003, 2007-11 Massachusetts Institute of Technology
Chris@10:  *
Chris@10:  * This program is free software; you can redistribute it and/or modify
Chris@10:  * it under the terms of the GNU General Public License as published by
Chris@10:  * the Free Software Foundation; either version 2 of the License, or
Chris@10:  * (at your option) any later version.
Chris@10:  *
Chris@10:  * This program is distributed in the hope that it will be useful,
Chris@10:  * but WITHOUT ANY WARRANTY; without even the implied warranty of
Chris@10:  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
Chris@10:  * GNU General Public License for more details.
Chris@10:  *
Chris@10:  * You should have received a copy of the GNU General Public License
Chris@10:  * along with this program; if not, write to the Free Software
Chris@10:  * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA  02110-1301  USA
Chris@10:  *
Chris@10:  */
Chris@10: 
Chris@10: /* This file was automatically generated --- DO NOT EDIT */
Chris@10: /* Generated on Sun Nov 25 07:35:49 EST 2012 */
Chris@10: 
Chris@10: #include "codelet-dft.h"
Chris@10: 
Chris@10: #ifdef HAVE_FMA
Chris@10: 
Chris@10: /* Generated by: ../../../genfft/gen_twiddle.native -fma -reorder-insns -schedule-for-pipeline -compact -variables 4 -pipeline-latency 4 -n 10 -name t1_10 -include t.h */
Chris@10: 
Chris@10: /*
Chris@10:  * This function contains 102 FP additions, 72 FP multiplications,
Chris@10:  * (or, 48 additions, 18 multiplications, 54 fused multiply/add),
Chris@10:  * 70 stack variables, 4 constants, and 40 memory accesses
Chris@10:  */
Chris@10: #include "t.h"
Chris@10: 
Chris@10: static void t1_10(R *ri, R *ii, const R *W, stride rs, INT mb, INT me, INT ms)
Chris@10: {
Chris@10:      DK(KP951056516, +0.951056516295153572116439333379382143405698634);
Chris@10:      DK(KP559016994, +0.559016994374947424102293417182819058860154590);
Chris@10:      DK(KP250000000, +0.250000000000000000000000000000000000000000000);
Chris@10:      DK(KP618033988, +0.618033988749894848204586834365638117720309180);
Chris@10:      {
Chris@10: 	  INT m;
Chris@10: 	  for (m = mb, W = W + (mb * 18); m < me; m = m + 1, ri = ri + ms, ii = ii + ms, W = W + 18, MAKE_VOLATILE_STRIDE(20, rs)) {
Chris@10: 	       E T1X, T21, T20, T22;
Chris@10: 	       {
Chris@10: 		    E T23, T1U, T8, T12, T1y, T25, T1P, T1H, T1Y, T18, T10, T2b, T1K, T1O, T15;
Chris@10: 		    E T1Z, T2a, Tz, T24, T1n;
Chris@10: 		    {
Chris@10: 			 E T1, T1T, T3, T6, T2, T5;
Chris@10: 			 T1 = ri[0];
Chris@10: 			 T1T = ii[0];
Chris@10: 			 T3 = ri[WS(rs, 5)];
Chris@10: 			 T6 = ii[WS(rs, 5)];
Chris@10: 			 T2 = W[8];
Chris@10: 			 T5 = W[9];
Chris@10: 			 {
Chris@10: 			      E T1w, TY, T1s, T1F, TM, T16, T1u, TS;
Chris@10: 			      {
Chris@10: 				   E TF, T1p, TO, TR, T1r, TL, TN, TQ, T1t, TP;
Chris@10: 				   {
Chris@10: 					E TU, TX, TT, TW;
Chris@10: 					{
Chris@10: 					     E TB, TE, T1R, T4, TA, TD;
Chris@10: 					     TB = ri[WS(rs, 4)];
Chris@10: 					     TE = ii[WS(rs, 4)];
Chris@10: 					     T1R = T2 * T6;
Chris@10: 					     T4 = T2 * T3;
Chris@10: 					     TA = W[6];
Chris@10: 					     TD = W[7];
Chris@10: 					     {
Chris@10: 						  E T1S, T7, T1o, TC;
Chris@10: 						  T1S = FNMS(T5, T3, T1R);
Chris@10: 						  T7 = FMA(T5, T6, T4);
Chris@10: 						  T1o = TA * TE;
Chris@10: 						  TC = TA * TB;
Chris@10: 						  T23 = T1T - T1S;
Chris@10: 						  T1U = T1S + T1T;
Chris@10: 						  T8 = T1 - T7;
Chris@10: 						  T12 = T1 + T7;
Chris@10: 						  TF = FMA(TD, TE, TC);
Chris@10: 						  T1p = FNMS(TD, TB, T1o);
Chris@10: 					     }
Chris@10: 					}
Chris@10: 					TU = ri[WS(rs, 1)];
Chris@10: 					TX = ii[WS(rs, 1)];
Chris@10: 					TT = W[0];
Chris@10: 					TW = W[1];
Chris@10: 					{
Chris@10: 					     E TH, TK, TJ, T1q, TI, T1v, TV, TG;
Chris@10: 					     TH = ri[WS(rs, 9)];
Chris@10: 					     TK = ii[WS(rs, 9)];
Chris@10: 					     T1v = TT * TX;
Chris@10: 					     TV = TT * TU;
Chris@10: 					     TG = W[16];
Chris@10: 					     TJ = W[17];
Chris@10: 					     T1w = FNMS(TW, TU, T1v);
Chris@10: 					     TY = FMA(TW, TX, TV);
Chris@10: 					     T1q = TG * TK;
Chris@10: 					     TI = TG * TH;
Chris@10: 					     TO = ri[WS(rs, 6)];
Chris@10: 					     TR = ii[WS(rs, 6)];
Chris@10: 					     T1r = FNMS(TJ, TH, T1q);
Chris@10: 					     TL = FMA(TJ, TK, TI);
Chris@10: 					     TN = W[10];
Chris@10: 					     TQ = W[11];
Chris@10: 					}
Chris@10: 				   }
Chris@10: 				   T1s = T1p - T1r;
Chris@10: 				   T1F = T1p + T1r;
Chris@10: 				   TM = TF - TL;
Chris@10: 				   T16 = TF + TL;
Chris@10: 				   T1t = TN * TR;
Chris@10: 				   TP = TN * TO;
Chris@10: 				   T1u = FNMS(TQ, TO, T1t);
Chris@10: 				   TS = FMA(TQ, TR, TP);
Chris@10: 			      }
Chris@10: 			      {
Chris@10: 				   E T1e, Te, T1l, Tx, Tn, Tq, Tp, T1g, Tk, T1i, To;
Chris@10: 				   {
Chris@10: 					E Tt, Tw, Tv, T1k, Tu;
Chris@10: 					{
Chris@10: 					     E Ta, Td, T9, Tc, T1d, Tb, Ts;
Chris@10: 					     Ta = ri[WS(rs, 2)];
Chris@10: 					     Td = ii[WS(rs, 2)];
Chris@10: 					     {
Chris@10: 						  E T1G, T1x, TZ, T17;
Chris@10: 						  T1G = T1u + T1w;
Chris@10: 						  T1x = T1u - T1w;
Chris@10: 						  TZ = TS - TY;
Chris@10: 						  T17 = TS + TY;
Chris@10: 						  T1y = T1s - T1x;
Chris@10: 						  T25 = T1s + T1x;
Chris@10: 						  T1P = T1F + T1G;
Chris@10: 						  T1H = T1F - T1G;
Chris@10: 						  T1Y = T16 - T17;
Chris@10: 						  T18 = T16 + T17;
Chris@10: 						  T10 = TM + TZ;
Chris@10: 						  T2b = TM - TZ;
Chris@10: 						  T9 = W[2];
Chris@10: 					     }
Chris@10: 					     Tc = W[3];
Chris@10: 					     Tt = ri[WS(rs, 3)];
Chris@10: 					     Tw = ii[WS(rs, 3)];
Chris@10: 					     T1d = T9 * Td;
Chris@10: 					     Tb = T9 * Ta;
Chris@10: 					     Ts = W[4];
Chris@10: 					     Tv = W[5];
Chris@10: 					     T1e = FNMS(Tc, Ta, T1d);
Chris@10: 					     Te = FMA(Tc, Td, Tb);
Chris@10: 					     T1k = Ts * Tw;
Chris@10: 					     Tu = Ts * Tt;
Chris@10: 					}
Chris@10: 					{
Chris@10: 					     E Tg, Tj, Tf, Ti, T1f, Th, Tm;
Chris@10: 					     Tg = ri[WS(rs, 7)];
Chris@10: 					     Tj = ii[WS(rs, 7)];
Chris@10: 					     T1l = FNMS(Tv, Tt, T1k);
Chris@10: 					     Tx = FMA(Tv, Tw, Tu);
Chris@10: 					     Tf = W[12];
Chris@10: 					     Ti = W[13];
Chris@10: 					     Tn = ri[WS(rs, 8)];
Chris@10: 					     Tq = ii[WS(rs, 8)];
Chris@10: 					     T1f = Tf * Tj;
Chris@10: 					     Th = Tf * Tg;
Chris@10: 					     Tm = W[14];
Chris@10: 					     Tp = W[15];
Chris@10: 					     T1g = FNMS(Ti, Tg, T1f);
Chris@10: 					     Tk = FMA(Ti, Tj, Th);
Chris@10: 					     T1i = Tm * Tq;
Chris@10: 					     To = Tm * Tn;
Chris@10: 					}
Chris@10: 				   }
Chris@10: 				   {
Chris@10: 					E T1h, T1I, Tl, T13, T1j, Tr;
Chris@10: 					T1h = T1e - T1g;
Chris@10: 					T1I = T1e + T1g;
Chris@10: 					Tl = Te - Tk;
Chris@10: 					T13 = Te + Tk;
Chris@10: 					T1j = FNMS(Tp, Tn, T1i);
Chris@10: 					Tr = FMA(Tp, Tq, To);
Chris@10: 					{
Chris@10: 					     E T1m, T1J, T14, Ty;
Chris@10: 					     T1m = T1j - T1l;
Chris@10: 					     T1J = T1j + T1l;
Chris@10: 					     T14 = Tr + Tx;
Chris@10: 					     Ty = Tr - Tx;
Chris@10: 					     T1K = T1I - T1J;
Chris@10: 					     T1O = T1I + T1J;
Chris@10: 					     T15 = T13 + T14;
Chris@10: 					     T1Z = T13 - T14;
Chris@10: 					     T2a = Tl - Ty;
Chris@10: 					     Tz = Tl + Ty;
Chris@10: 					     T24 = T1h + T1m;
Chris@10: 					     T1n = T1h - T1m;
Chris@10: 					}
Chris@10: 				   }
Chris@10: 			      }
Chris@10: 			 }
Chris@10: 		    }
Chris@10: 		    {
Chris@10: 			 E T2c, T2e, T29, T2d;
Chris@10: 			 {
Chris@10: 			      E T1b, T11, T26, T28, T27;
Chris@10: 			      T1b = Tz - T10;
Chris@10: 			      T11 = Tz + T10;
Chris@10: 			      T26 = T24 + T25;
Chris@10: 			      T28 = T24 - T25;
Chris@10: 			      {
Chris@10: 				   E T1B, T1z, T1a, T1A, T1c;
Chris@10: 				   T1B = FNMS(KP618033988, T1n, T1y);
Chris@10: 				   T1z = FMA(KP618033988, T1y, T1n);
Chris@10: 				   ri[WS(rs, 5)] = T8 + T11;
Chris@10: 				   T1a = FNMS(KP250000000, T11, T8);
Chris@10: 				   T1A = FNMS(KP559016994, T1b, T1a);
Chris@10: 				   T1c = FMA(KP559016994, T1b, T1a);
Chris@10: 				   T27 = FNMS(KP250000000, T26, T23);
Chris@10: 				   T2c = FMA(KP618033988, T2b, T2a);
Chris@10: 				   T2e = FNMS(KP618033988, T2a, T2b);
Chris@10: 				   ri[WS(rs, 1)] = FMA(KP951056516, T1z, T1c);
Chris@10: 				   ri[WS(rs, 9)] = FNMS(KP951056516, T1z, T1c);
Chris@10: 				   ri[WS(rs, 3)] = FMA(KP951056516, T1B, T1A);
Chris@10: 				   ri[WS(rs, 7)] = FNMS(KP951056516, T1B, T1A);
Chris@10: 			      }
Chris@10: 			      ii[WS(rs, 5)] = T26 + T23;
Chris@10: 			      T29 = FMA(KP559016994, T28, T27);
Chris@10: 			      T2d = FNMS(KP559016994, T28, T27);
Chris@10: 			 }
Chris@10: 			 {
Chris@10: 			      E T1E, T1M, T1L, T1N, T19, T1D, T1C, T1Q, T1W, T1V;
Chris@10: 			      T19 = T15 + T18;
Chris@10: 			      T1D = T15 - T18;
Chris@10: 			      ii[WS(rs, 7)] = FMA(KP951056516, T2e, T2d);
Chris@10: 			      ii[WS(rs, 3)] = FNMS(KP951056516, T2e, T2d);
Chris@10: 			      ii[WS(rs, 9)] = FMA(KP951056516, T2c, T29);
Chris@10: 			      ii[WS(rs, 1)] = FNMS(KP951056516, T2c, T29);
Chris@10: 			      T1C = FNMS(KP250000000, T19, T12);
Chris@10: 			      ri[0] = T12 + T19;
Chris@10: 			      T1E = FNMS(KP559016994, T1D, T1C);
Chris@10: 			      T1M = FMA(KP559016994, T1D, T1C);
Chris@10: 			      T1L = FNMS(KP618033988, T1K, T1H);
Chris@10: 			      T1N = FMA(KP618033988, T1H, T1K);
Chris@10: 			      T1Q = T1O + T1P;
Chris@10: 			      T1W = T1O - T1P;
Chris@10: 			      ri[WS(rs, 6)] = FMA(KP951056516, T1N, T1M);
Chris@10: 			      ri[WS(rs, 4)] = FNMS(KP951056516, T1N, T1M);
Chris@10: 			      ri[WS(rs, 8)] = FMA(KP951056516, T1L, T1E);
Chris@10: 			      ri[WS(rs, 2)] = FNMS(KP951056516, T1L, T1E);
Chris@10: 			      T1V = FNMS(KP250000000, T1Q, T1U);
Chris@10: 			      ii[0] = T1Q + T1U;
Chris@10: 			      T1X = FNMS(KP559016994, T1W, T1V);
Chris@10: 			      T21 = FMA(KP559016994, T1W, T1V);
Chris@10: 			      T20 = FNMS(KP618033988, T1Z, T1Y);
Chris@10: 			      T22 = FMA(KP618033988, T1Y, T1Z);
Chris@10: 			 }
Chris@10: 		    }
Chris@10: 	       }
Chris@10: 	       ii[WS(rs, 6)] = FNMS(KP951056516, T22, T21);
Chris@10: 	       ii[WS(rs, 4)] = FMA(KP951056516, T22, T21);
Chris@10: 	       ii[WS(rs, 8)] = FNMS(KP951056516, T20, T1X);
Chris@10: 	       ii[WS(rs, 2)] = FMA(KP951056516, T20, T1X);
Chris@10: 	  }
Chris@10:      }
Chris@10: }
Chris@10: 
Chris@10: static const tw_instr twinstr[] = {
Chris@10:      {TW_FULL, 0, 10},
Chris@10:      {TW_NEXT, 1, 0}
Chris@10: };
Chris@10: 
Chris@10: static const ct_desc desc = { 10, "t1_10", twinstr, &GENUS, {48, 18, 54, 0}, 0, 0, 0 };
Chris@10: 
Chris@10: void X(codelet_t1_10) (planner *p) {
Chris@10:      X(kdft_dit_register) (p, t1_10, &desc);
Chris@10: }
Chris@10: #else				/* HAVE_FMA */
Chris@10: 
Chris@10: /* Generated by: ../../../genfft/gen_twiddle.native -compact -variables 4 -pipeline-latency 4 -n 10 -name t1_10 -include t.h */
Chris@10: 
Chris@10: /*
Chris@10:  * This function contains 102 FP additions, 60 FP multiplications,
Chris@10:  * (or, 72 additions, 30 multiplications, 30 fused multiply/add),
Chris@10:  * 45 stack variables, 4 constants, and 40 memory accesses
Chris@10:  */
Chris@10: #include "t.h"
Chris@10: 
Chris@10: static void t1_10(R *ri, R *ii, const R *W, stride rs, INT mb, INT me, INT ms)
Chris@10: {
Chris@10:      DK(KP587785252, +0.587785252292473129168705954639072768597652438);
Chris@10:      DK(KP951056516, +0.951056516295153572116439333379382143405698634);
Chris@10:      DK(KP250000000, +0.250000000000000000000000000000000000000000000);
Chris@10:      DK(KP559016994, +0.559016994374947424102293417182819058860154590);
Chris@10:      {
Chris@10: 	  INT m;
Chris@10: 	  for (m = mb, W = W + (mb * 18); m < me; m = m + 1, ri = ri + ms, ii = ii + ms, W = W + 18, MAKE_VOLATILE_STRIDE(20, rs)) {
Chris@10: 	       E T7, T1O, TT, T1C, TF, TQ, TR, T1o, T1p, T1y, TX, TY, TZ, T1d, T1g;
Chris@10: 	       E T1M, Ti, Tt, Tu, T1r, T1s, T1x, TU, TV, TW, T16, T19, T1L;
Chris@10: 	       {
Chris@10: 		    E T1, T1B, T6, T1A;
Chris@10: 		    T1 = ri[0];
Chris@10: 		    T1B = ii[0];
Chris@10: 		    {
Chris@10: 			 E T3, T5, T2, T4;
Chris@10: 			 T3 = ri[WS(rs, 5)];
Chris@10: 			 T5 = ii[WS(rs, 5)];
Chris@10: 			 T2 = W[8];
Chris@10: 			 T4 = W[9];
Chris@10: 			 T6 = FMA(T2, T3, T4 * T5);
Chris@10: 			 T1A = FNMS(T4, T3, T2 * T5);
Chris@10: 		    }
Chris@10: 		    T7 = T1 - T6;
Chris@10: 		    T1O = T1B - T1A;
Chris@10: 		    TT = T1 + T6;
Chris@10: 		    T1C = T1A + T1B;
Chris@10: 	       }
Chris@10: 	       {
Chris@10: 		    E Tz, T1b, TP, T1f, TE, T1c, TK, T1e;
Chris@10: 		    {
Chris@10: 			 E Tw, Ty, Tv, Tx;
Chris@10: 			 Tw = ri[WS(rs, 4)];
Chris@10: 			 Ty = ii[WS(rs, 4)];
Chris@10: 			 Tv = W[6];
Chris@10: 			 Tx = W[7];
Chris@10: 			 Tz = FMA(Tv, Tw, Tx * Ty);
Chris@10: 			 T1b = FNMS(Tx, Tw, Tv * Ty);
Chris@10: 		    }
Chris@10: 		    {
Chris@10: 			 E TM, TO, TL, TN;
Chris@10: 			 TM = ri[WS(rs, 1)];
Chris@10: 			 TO = ii[WS(rs, 1)];
Chris@10: 			 TL = W[0];
Chris@10: 			 TN = W[1];
Chris@10: 			 TP = FMA(TL, TM, TN * TO);
Chris@10: 			 T1f = FNMS(TN, TM, TL * TO);
Chris@10: 		    }
Chris@10: 		    {
Chris@10: 			 E TB, TD, TA, TC;
Chris@10: 			 TB = ri[WS(rs, 9)];
Chris@10: 			 TD = ii[WS(rs, 9)];
Chris@10: 			 TA = W[16];
Chris@10: 			 TC = W[17];
Chris@10: 			 TE = FMA(TA, TB, TC * TD);
Chris@10: 			 T1c = FNMS(TC, TB, TA * TD);
Chris@10: 		    }
Chris@10: 		    {
Chris@10: 			 E TH, TJ, TG, TI;
Chris@10: 			 TH = ri[WS(rs, 6)];
Chris@10: 			 TJ = ii[WS(rs, 6)];
Chris@10: 			 TG = W[10];
Chris@10: 			 TI = W[11];
Chris@10: 			 TK = FMA(TG, TH, TI * TJ);
Chris@10: 			 T1e = FNMS(TI, TH, TG * TJ);
Chris@10: 		    }
Chris@10: 		    TF = Tz - TE;
Chris@10: 		    TQ = TK - TP;
Chris@10: 		    TR = TF + TQ;
Chris@10: 		    T1o = T1b + T1c;
Chris@10: 		    T1p = T1e + T1f;
Chris@10: 		    T1y = T1o + T1p;
Chris@10: 		    TX = Tz + TE;
Chris@10: 		    TY = TK + TP;
Chris@10: 		    TZ = TX + TY;
Chris@10: 		    T1d = T1b - T1c;
Chris@10: 		    T1g = T1e - T1f;
Chris@10: 		    T1M = T1d + T1g;
Chris@10: 	       }
Chris@10: 	       {
Chris@10: 		    E Tc, T14, Ts, T18, Th, T15, Tn, T17;
Chris@10: 		    {
Chris@10: 			 E T9, Tb, T8, Ta;
Chris@10: 			 T9 = ri[WS(rs, 2)];
Chris@10: 			 Tb = ii[WS(rs, 2)];
Chris@10: 			 T8 = W[2];
Chris@10: 			 Ta = W[3];
Chris@10: 			 Tc = FMA(T8, T9, Ta * Tb);
Chris@10: 			 T14 = FNMS(Ta, T9, T8 * Tb);
Chris@10: 		    }
Chris@10: 		    {
Chris@10: 			 E Tp, Tr, To, Tq;
Chris@10: 			 Tp = ri[WS(rs, 3)];
Chris@10: 			 Tr = ii[WS(rs, 3)];
Chris@10: 			 To = W[4];
Chris@10: 			 Tq = W[5];
Chris@10: 			 Ts = FMA(To, Tp, Tq * Tr);
Chris@10: 			 T18 = FNMS(Tq, Tp, To * Tr);
Chris@10: 		    }
Chris@10: 		    {
Chris@10: 			 E Te, Tg, Td, Tf;
Chris@10: 			 Te = ri[WS(rs, 7)];
Chris@10: 			 Tg = ii[WS(rs, 7)];
Chris@10: 			 Td = W[12];
Chris@10: 			 Tf = W[13];
Chris@10: 			 Th = FMA(Td, Te, Tf * Tg);
Chris@10: 			 T15 = FNMS(Tf, Te, Td * Tg);
Chris@10: 		    }
Chris@10: 		    {
Chris@10: 			 E Tk, Tm, Tj, Tl;
Chris@10: 			 Tk = ri[WS(rs, 8)];
Chris@10: 			 Tm = ii[WS(rs, 8)];
Chris@10: 			 Tj = W[14];
Chris@10: 			 Tl = W[15];
Chris@10: 			 Tn = FMA(Tj, Tk, Tl * Tm);
Chris@10: 			 T17 = FNMS(Tl, Tk, Tj * Tm);
Chris@10: 		    }
Chris@10: 		    Ti = Tc - Th;
Chris@10: 		    Tt = Tn - Ts;
Chris@10: 		    Tu = Ti + Tt;
Chris@10: 		    T1r = T14 + T15;
Chris@10: 		    T1s = T17 + T18;
Chris@10: 		    T1x = T1r + T1s;
Chris@10: 		    TU = Tc + Th;
Chris@10: 		    TV = Tn + Ts;
Chris@10: 		    TW = TU + TV;
Chris@10: 		    T16 = T14 - T15;
Chris@10: 		    T19 = T17 - T18;
Chris@10: 		    T1L = T16 + T19;
Chris@10: 	       }
Chris@10: 	       {
Chris@10: 		    E T11, TS, T12, T1i, T1k, T1a, T1h, T1j, T13;
Chris@10: 		    T11 = KP559016994 * (Tu - TR);
Chris@10: 		    TS = Tu + TR;
Chris@10: 		    T12 = FNMS(KP250000000, TS, T7);
Chris@10: 		    T1a = T16 - T19;
Chris@10: 		    T1h = T1d - T1g;
Chris@10: 		    T1i = FMA(KP951056516, T1a, KP587785252 * T1h);
Chris@10: 		    T1k = FNMS(KP587785252, T1a, KP951056516 * T1h);
Chris@10: 		    ri[WS(rs, 5)] = T7 + TS;
Chris@10: 		    T1j = T12 - T11;
Chris@10: 		    ri[WS(rs, 7)] = T1j - T1k;
Chris@10: 		    ri[WS(rs, 3)] = T1j + T1k;
Chris@10: 		    T13 = T11 + T12;
Chris@10: 		    ri[WS(rs, 9)] = T13 - T1i;
Chris@10: 		    ri[WS(rs, 1)] = T13 + T1i;
Chris@10: 	       }
Chris@10: 	       {
Chris@10: 		    E T1N, T1P, T1Q, T1U, T1W, T1S, T1T, T1V, T1R;
Chris@10: 		    T1N = KP559016994 * (T1L - T1M);
Chris@10: 		    T1P = T1L + T1M;
Chris@10: 		    T1Q = FNMS(KP250000000, T1P, T1O);
Chris@10: 		    T1S = Ti - Tt;
Chris@10: 		    T1T = TF - TQ;
Chris@10: 		    T1U = FMA(KP951056516, T1S, KP587785252 * T1T);
Chris@10: 		    T1W = FNMS(KP587785252, T1S, KP951056516 * T1T);
Chris@10: 		    ii[WS(rs, 5)] = T1P + T1O;
Chris@10: 		    T1V = T1Q - T1N;
Chris@10: 		    ii[WS(rs, 3)] = T1V - T1W;
Chris@10: 		    ii[WS(rs, 7)] = T1W + T1V;
Chris@10: 		    T1R = T1N + T1Q;
Chris@10: 		    ii[WS(rs, 1)] = T1R - T1U;
Chris@10: 		    ii[WS(rs, 9)] = T1U + T1R;
Chris@10: 	       }
Chris@10: 	       {
Chris@10: 		    E T1m, T10, T1l, T1u, T1w, T1q, T1t, T1v, T1n;
Chris@10: 		    T1m = KP559016994 * (TW - TZ);
Chris@10: 		    T10 = TW + TZ;
Chris@10: 		    T1l = FNMS(KP250000000, T10, TT);
Chris@10: 		    T1q = T1o - T1p;
Chris@10: 		    T1t = T1r - T1s;
Chris@10: 		    T1u = FNMS(KP587785252, T1t, KP951056516 * T1q);
Chris@10: 		    T1w = FMA(KP951056516, T1t, KP587785252 * T1q);
Chris@10: 		    ri[0] = TT + T10;
Chris@10: 		    T1v = T1m + T1l;
Chris@10: 		    ri[WS(rs, 4)] = T1v - T1w;
Chris@10: 		    ri[WS(rs, 6)] = T1v + T1w;
Chris@10: 		    T1n = T1l - T1m;
Chris@10: 		    ri[WS(rs, 2)] = T1n - T1u;
Chris@10: 		    ri[WS(rs, 8)] = T1n + T1u;
Chris@10: 	       }
Chris@10: 	       {
Chris@10: 		    E T1H, T1z, T1G, T1F, T1J, T1D, T1E, T1K, T1I;
Chris@10: 		    T1H = KP559016994 * (T1x - T1y);
Chris@10: 		    T1z = T1x + T1y;
Chris@10: 		    T1G = FNMS(KP250000000, T1z, T1C);
Chris@10: 		    T1D = TX - TY;
Chris@10: 		    T1E = TU - TV;
Chris@10: 		    T1F = FNMS(KP587785252, T1E, KP951056516 * T1D);
Chris@10: 		    T1J = FMA(KP951056516, T1E, KP587785252 * T1D);
Chris@10: 		    ii[0] = T1z + T1C;
Chris@10: 		    T1K = T1H + T1G;
Chris@10: 		    ii[WS(rs, 4)] = T1J + T1K;
Chris@10: 		    ii[WS(rs, 6)] = T1K - T1J;
Chris@10: 		    T1I = T1G - T1H;
Chris@10: 		    ii[WS(rs, 2)] = T1F + T1I;
Chris@10: 		    ii[WS(rs, 8)] = T1I - T1F;
Chris@10: 	       }
Chris@10: 	  }
Chris@10:      }
Chris@10: }
Chris@10: 
Chris@10: static const tw_instr twinstr[] = {
Chris@10:      {TW_FULL, 0, 10},
Chris@10:      {TW_NEXT, 1, 0}
Chris@10: };
Chris@10: 
Chris@10: static const ct_desc desc = { 10, "t1_10", twinstr, &GENUS, {72, 30, 30, 0}, 0, 0, 0 };
Chris@10: 
Chris@10: void X(codelet_t1_10) (planner *p) {
Chris@10:      X(kdft_dit_register) (p, t1_10, &desc);
Chris@10: }
Chris@10: #endif				/* HAVE_FMA */