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 07:35:44 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_notw.native -fma -reorder-insns -schedule-for-pipeline -compact -variables 4 -pipeline-latency 4 -n 16 -name n1_16 -include n.h */
Chris@10: 
Chris@10: /*
Chris@10:  * This function contains 144 FP additions, 40 FP multiplications,
Chris@10:  * (or, 104 additions, 0 multiplications, 40 fused multiply/add),
Chris@10:  * 82 stack variables, 3 constants, and 64 memory accesses
Chris@10:  */
Chris@10: #include "n.h"
Chris@10: 
Chris@10: static void n1_16(const R *ri, const R *ii, R *ro, R *io, stride is, stride os, INT v, INT ivs, INT ovs)
Chris@10: {
Chris@10:      DK(KP923879532, +0.923879532511286756128183189396788286822416626);
Chris@10:      DK(KP414213562, +0.414213562373095048801688724209698078569671875);
Chris@10:      DK(KP707106781, +0.707106781186547524400844362104849039284835938);
Chris@10:      {
Chris@10: 	  INT i;
Chris@10: 	  for (i = v; i > 0; i = i - 1, ri = ri + ivs, ii = ii + ivs, ro = ro + ovs, io = io + ovs, MAKE_VOLATILE_STRIDE(64, is), MAKE_VOLATILE_STRIDE(64, os)) {
Chris@10: 	       E T1z, T1L, T1M, T1N, T1P, T1J, T1K, T1G, T1O, T1Q;
Chris@10: 	       {
Chris@10: 		    E T1l, T1H, T1R, T7, T1x, TN, TC, T25, T1E, T1b, T1Z, Tt, T2h, T22, T1D;
Chris@10: 		    E T1g, T1n, TQ, Te, T26, TT, T1m, TJ, T1S, Tj, T11, Ti, T1V, TZ, Tk;
Chris@10: 		    E T12, T13;
Chris@10: 		    {
Chris@10: 			 E Tq, T1c, Tp, T20, T1a, Tr, T1d, T1e;
Chris@10: 			 {
Chris@10: 			      E T4, TL, T3, T1k, Ty, T5, Tz, TA;
Chris@10: 			      {
Chris@10: 				   E T1, T2, Tw, Tx;
Chris@10: 				   T1 = ri[0];
Chris@10: 				   T2 = ri[WS(is, 8)];
Chris@10: 				   Tw = ii[0];
Chris@10: 				   Tx = ii[WS(is, 8)];
Chris@10: 				   T4 = ri[WS(is, 4)];
Chris@10: 				   TL = T1 - T2;
Chris@10: 				   T3 = T1 + T2;
Chris@10: 				   T1k = Tw - Tx;
Chris@10: 				   Ty = Tw + Tx;
Chris@10: 				   T5 = ri[WS(is, 12)];
Chris@10: 				   Tz = ii[WS(is, 4)];
Chris@10: 				   TA = ii[WS(is, 12)];
Chris@10: 			      }
Chris@10: 			      {
Chris@10: 				   E Tn, To, T18, T19;
Chris@10: 				   Tn = ri[WS(is, 15)];
Chris@10: 				   {
Chris@10: 					E T1j, T6, TM, TB;
Chris@10: 					T1j = T4 - T5;
Chris@10: 					T6 = T4 + T5;
Chris@10: 					TM = Tz - TA;
Chris@10: 					TB = Tz + TA;
Chris@10: 					T1l = T1j + T1k;
Chris@10: 					T1H = T1k - T1j;
Chris@10: 					T1R = T3 - T6;
Chris@10: 					T7 = T3 + T6;
Chris@10: 					T1x = TL + TM;
Chris@10: 					TN = TL - TM;
Chris@10: 					TC = Ty + TB;
Chris@10: 					T25 = Ty - TB;
Chris@10: 					To = ri[WS(is, 7)];
Chris@10: 				   }
Chris@10: 				   T18 = ii[WS(is, 15)];
Chris@10: 				   T19 = ii[WS(is, 7)];
Chris@10: 				   Tq = ri[WS(is, 3)];
Chris@10: 				   T1c = Tn - To;
Chris@10: 				   Tp = Tn + To;
Chris@10: 				   T20 = T18 + T19;
Chris@10: 				   T1a = T18 - T19;
Chris@10: 				   Tr = ri[WS(is, 11)];
Chris@10: 				   T1d = ii[WS(is, 3)];
Chris@10: 				   T1e = ii[WS(is, 11)];
Chris@10: 			      }
Chris@10: 			 }
Chris@10: 			 {
Chris@10: 			      E Tb, TP, Ta, TO, TF, Tc, TG, TH;
Chris@10: 			      {
Chris@10: 				   E T8, T9, TD, TE;
Chris@10: 				   T8 = ri[WS(is, 2)];
Chris@10: 				   {
Chris@10: 					E T17, Ts, T21, T1f;
Chris@10: 					T17 = Tq - Tr;
Chris@10: 					Ts = Tq + Tr;
Chris@10: 					T21 = T1d + T1e;
Chris@10: 					T1f = T1d - T1e;
Chris@10: 					T1E = T1a - T17;
Chris@10: 					T1b = T17 + T1a;
Chris@10: 					T1Z = Tp - Ts;
Chris@10: 					Tt = Tp + Ts;
Chris@10: 					T2h = T20 + T21;
Chris@10: 					T22 = T20 - T21;
Chris@10: 					T1D = T1c + T1f;
Chris@10: 					T1g = T1c - T1f;
Chris@10: 					T9 = ri[WS(is, 10)];
Chris@10: 				   }
Chris@10: 				   TD = ii[WS(is, 2)];
Chris@10: 				   TE = ii[WS(is, 10)];
Chris@10: 				   Tb = ri[WS(is, 14)];
Chris@10: 				   TP = T8 - T9;
Chris@10: 				   Ta = T8 + T9;
Chris@10: 				   TO = TD - TE;
Chris@10: 				   TF = TD + TE;
Chris@10: 				   Tc = ri[WS(is, 6)];
Chris@10: 				   TG = ii[WS(is, 14)];
Chris@10: 				   TH = ii[WS(is, 6)];
Chris@10: 			      }
Chris@10: 			      {
Chris@10: 				   E TR, Td, TS, TI;
Chris@10: 				   T1n = TP + TO;
Chris@10: 				   TQ = TO - TP;
Chris@10: 				   TR = Tb - Tc;
Chris@10: 				   Td = Tb + Tc;
Chris@10: 				   TS = TG - TH;
Chris@10: 				   TI = TG + TH;
Chris@10: 				   Te = Ta + Td;
Chris@10: 				   T26 = Td - Ta;
Chris@10: 				   TT = TR + TS;
Chris@10: 				   T1m = TR - TS;
Chris@10: 				   TJ = TF + TI;
Chris@10: 				   T1S = TF - TI;
Chris@10: 			      }
Chris@10: 			 }
Chris@10: 			 {
Chris@10: 			      E Tg, Th, TX, TY;
Chris@10: 			      Tg = ri[WS(is, 1)];
Chris@10: 			      Th = ri[WS(is, 9)];
Chris@10: 			      TX = ii[WS(is, 1)];
Chris@10: 			      TY = ii[WS(is, 9)];
Chris@10: 			      Tj = ri[WS(is, 5)];
Chris@10: 			      T11 = Tg - Th;
Chris@10: 			      Ti = Tg + Th;
Chris@10: 			      T1V = TX + TY;
Chris@10: 			      TZ = TX - TY;
Chris@10: 			      Tk = ri[WS(is, 13)];
Chris@10: 			      T12 = ii[WS(is, 5)];
Chris@10: 			      T13 = ii[WS(is, 13)];
Chris@10: 			 }
Chris@10: 		    }
Chris@10: 		    {
Chris@10: 			 E T2f, T1B, T10, T1U, T1X, T1A, T15, Tv, TK, T2i;
Chris@10: 			 {
Chris@10: 			      E Tf, Tu, T2j, T2k, T2g;
Chris@10: 			      T2f = T7 - Te;
Chris@10: 			      Tf = T7 + Te;
Chris@10: 			      {
Chris@10: 				   E TW, Tl, T1W, T14, Tm;
Chris@10: 				   TW = Tj - Tk;
Chris@10: 				   Tl = Tj + Tk;
Chris@10: 				   T1W = T12 + T13;
Chris@10: 				   T14 = T12 - T13;
Chris@10: 				   T1B = TZ - TW;
Chris@10: 				   T10 = TW + TZ;
Chris@10: 				   T1U = Ti - Tl;
Chris@10: 				   Tm = Ti + Tl;
Chris@10: 				   T2g = T1V + T1W;
Chris@10: 				   T1X = T1V - T1W;
Chris@10: 				   T1A = T11 + T14;
Chris@10: 				   T15 = T11 - T14;
Chris@10: 				   Tu = Tm + Tt;
Chris@10: 				   Tv = Tt - Tm;
Chris@10: 			      }
Chris@10: 			      TK = TC - TJ;
Chris@10: 			      T2j = TC + TJ;
Chris@10: 			      T2k = T2g + T2h;
Chris@10: 			      T2i = T2g - T2h;
Chris@10: 			      ro[0] = Tf + Tu;
Chris@10: 			      ro[WS(os, 8)] = Tf - Tu;
Chris@10: 			      io[0] = T2j + T2k;
Chris@10: 			      io[WS(os, 8)] = T2j - T2k;
Chris@10: 			 }
Chris@10: 			 {
Chris@10: 			      E T29, T1T, T27, T2d, T2a, T2b, T28, T24, T1Y, T23;
Chris@10: 			      T29 = T1R - T1S;
Chris@10: 			      T1T = T1R + T1S;
Chris@10: 			      io[WS(os, 12)] = TK - Tv;
Chris@10: 			      io[WS(os, 4)] = Tv + TK;
Chris@10: 			      ro[WS(os, 4)] = T2f + T2i;
Chris@10: 			      ro[WS(os, 12)] = T2f - T2i;
Chris@10: 			      T27 = T25 - T26;
Chris@10: 			      T2d = T26 + T25;
Chris@10: 			      T2a = T1X - T1U;
Chris@10: 			      T1Y = T1U + T1X;
Chris@10: 			      T23 = T1Z - T22;
Chris@10: 			      T2b = T1Z + T22;
Chris@10: 			      T28 = T23 - T1Y;
Chris@10: 			      T24 = T1Y + T23;
Chris@10: 			      {
Chris@10: 				   E T1I, TV, T1v, T1y, T1t, T1s, T1r, T1p, T1q, T1i;
Chris@10: 				   {
Chris@10: 					E T1o, T2e, T2c, TU, T16, T1h;
Chris@10: 					T1I = TQ + TT;
Chris@10: 					TU = TQ - TT;
Chris@10: 					io[WS(os, 14)] = FNMS(KP707106781, T28, T27);
Chris@10: 					io[WS(os, 6)] = FMA(KP707106781, T28, T27);
Chris@10: 					ro[WS(os, 2)] = FMA(KP707106781, T24, T1T);
Chris@10: 					ro[WS(os, 10)] = FNMS(KP707106781, T24, T1T);
Chris@10: 					T2e = T2a + T2b;
Chris@10: 					T2c = T2a - T2b;
Chris@10: 					TV = FMA(KP707106781, TU, TN);
Chris@10: 					T1v = FNMS(KP707106781, TU, TN);
Chris@10: 					io[WS(os, 10)] = FNMS(KP707106781, T2e, T2d);
Chris@10: 					io[WS(os, 2)] = FMA(KP707106781, T2e, T2d);
Chris@10: 					ro[WS(os, 6)] = FMA(KP707106781, T2c, T29);
Chris@10: 					ro[WS(os, 14)] = FNMS(KP707106781, T2c, T29);
Chris@10: 					T1o = T1m - T1n;
Chris@10: 					T1y = T1n + T1m;
Chris@10: 					T1t = FNMS(KP414213562, T10, T15);
Chris@10: 					T16 = FMA(KP414213562, T15, T10);
Chris@10: 					T1h = FNMS(KP414213562, T1g, T1b);
Chris@10: 					T1s = FMA(KP414213562, T1b, T1g);
Chris@10: 					T1r = FMA(KP707106781, T1o, T1l);
Chris@10: 					T1p = FNMS(KP707106781, T1o, T1l);
Chris@10: 					T1q = T16 + T1h;
Chris@10: 					T1i = T16 - T1h;
Chris@10: 				   }
Chris@10: 				   {
Chris@10: 					E T1w, T1u, T1C, T1F;
Chris@10: 					io[WS(os, 15)] = FMA(KP923879532, T1q, T1p);
Chris@10: 					io[WS(os, 7)] = FNMS(KP923879532, T1q, T1p);
Chris@10: 					ro[WS(os, 3)] = FMA(KP923879532, T1i, TV);
Chris@10: 					ro[WS(os, 11)] = FNMS(KP923879532, T1i, TV);
Chris@10: 					T1w = T1t + T1s;
Chris@10: 					T1u = T1s - T1t;
Chris@10: 					T1z = FMA(KP707106781, T1y, T1x);
Chris@10: 					T1L = FNMS(KP707106781, T1y, T1x);
Chris@10: 					ro[WS(os, 15)] = FMA(KP923879532, T1w, T1v);
Chris@10: 					ro[WS(os, 7)] = FNMS(KP923879532, T1w, T1v);
Chris@10: 					io[WS(os, 3)] = FMA(KP923879532, T1u, T1r);
Chris@10: 					io[WS(os, 11)] = FNMS(KP923879532, T1u, T1r);
Chris@10: 					T1M = FNMS(KP414213562, T1A, T1B);
Chris@10: 					T1C = FMA(KP414213562, T1B, T1A);
Chris@10: 					T1F = FNMS(KP414213562, T1E, T1D);
Chris@10: 					T1N = FMA(KP414213562, T1D, T1E);
Chris@10: 					T1P = FMA(KP707106781, T1I, T1H);
Chris@10: 					T1J = FNMS(KP707106781, T1I, T1H);
Chris@10: 					T1K = T1F - T1C;
Chris@10: 					T1G = T1C + T1F;
Chris@10: 				   }
Chris@10: 			      }
Chris@10: 			 }
Chris@10: 		    }
Chris@10: 	       }
Chris@10: 	       io[WS(os, 5)] = FMA(KP923879532, T1K, T1J);
Chris@10: 	       io[WS(os, 13)] = FNMS(KP923879532, T1K, T1J);
Chris@10: 	       ro[WS(os, 1)] = FMA(KP923879532, T1G, T1z);
Chris@10: 	       ro[WS(os, 9)] = FNMS(KP923879532, T1G, T1z);
Chris@10: 	       T1O = T1M - T1N;
Chris@10: 	       T1Q = T1M + T1N;
Chris@10: 	       io[WS(os, 1)] = FMA(KP923879532, T1Q, T1P);
Chris@10: 	       io[WS(os, 9)] = FNMS(KP923879532, T1Q, T1P);
Chris@10: 	       ro[WS(os, 5)] = FMA(KP923879532, T1O, T1L);
Chris@10: 	       ro[WS(os, 13)] = FNMS(KP923879532, T1O, T1L);
Chris@10: 	  }
Chris@10:      }
Chris@10: }
Chris@10: 
Chris@10: static const kdft_desc desc = { 16, "n1_16", {104, 0, 40, 0}, &GENUS, 0, 0, 0, 0 };
Chris@10: 
Chris@10: void X(codelet_n1_16) (planner *p) {
Chris@10:      X(kdft_register) (p, n1_16, &desc);
Chris@10: }
Chris@10: 
Chris@10: #else				/* HAVE_FMA */
Chris@10: 
Chris@10: /* Generated by: ../../../genfft/gen_notw.native -compact -variables 4 -pipeline-latency 4 -n 16 -name n1_16 -include n.h */
Chris@10: 
Chris@10: /*
Chris@10:  * This function contains 144 FP additions, 24 FP multiplications,
Chris@10:  * (or, 136 additions, 16 multiplications, 8 fused multiply/add),
Chris@10:  * 50 stack variables, 3 constants, and 64 memory accesses
Chris@10:  */
Chris@10: #include "n.h"
Chris@10: 
Chris@10: static void n1_16(const R *ri, const R *ii, R *ro, R *io, stride is, stride os, INT v, INT ivs, INT ovs)
Chris@10: {
Chris@10:      DK(KP382683432, +0.382683432365089771728459984030398866761344562);
Chris@10:      DK(KP923879532, +0.923879532511286756128183189396788286822416626);
Chris@10:      DK(KP707106781, +0.707106781186547524400844362104849039284835938);
Chris@10:      {
Chris@10: 	  INT i;
Chris@10: 	  for (i = v; i > 0; i = i - 1, ri = ri + ivs, ii = ii + ivs, ro = ro + ovs, io = io + ovs, MAKE_VOLATILE_STRIDE(64, is), MAKE_VOLATILE_STRIDE(64, os)) {
Chris@10: 	       E T7, T1R, T25, TC, TN, T1x, T1H, T1l, Tt, T22, T2h, T1b, T1g, T1E, T1Z;
Chris@10: 	       E T1D, Te, T1S, T26, TJ, TQ, T1m, T1n, TT, Tm, T1X, T2g, T10, T15, T1B;
Chris@10: 	       E T1U, T1A;
Chris@10: 	       {
Chris@10: 		    E T3, TL, Ty, T1k, T6, T1j, TB, TM;
Chris@10: 		    {
Chris@10: 			 E T1, T2, Tw, Tx;
Chris@10: 			 T1 = ri[0];
Chris@10: 			 T2 = ri[WS(is, 8)];
Chris@10: 			 T3 = T1 + T2;
Chris@10: 			 TL = T1 - T2;
Chris@10: 			 Tw = ii[0];
Chris@10: 			 Tx = ii[WS(is, 8)];
Chris@10: 			 Ty = Tw + Tx;
Chris@10: 			 T1k = Tw - Tx;
Chris@10: 		    }
Chris@10: 		    {
Chris@10: 			 E T4, T5, Tz, TA;
Chris@10: 			 T4 = ri[WS(is, 4)];
Chris@10: 			 T5 = ri[WS(is, 12)];
Chris@10: 			 T6 = T4 + T5;
Chris@10: 			 T1j = T4 - T5;
Chris@10: 			 Tz = ii[WS(is, 4)];
Chris@10: 			 TA = ii[WS(is, 12)];
Chris@10: 			 TB = Tz + TA;
Chris@10: 			 TM = Tz - TA;
Chris@10: 		    }
Chris@10: 		    T7 = T3 + T6;
Chris@10: 		    T1R = T3 - T6;
Chris@10: 		    T25 = Ty - TB;
Chris@10: 		    TC = Ty + TB;
Chris@10: 		    TN = TL - TM;
Chris@10: 		    T1x = TL + TM;
Chris@10: 		    T1H = T1k - T1j;
Chris@10: 		    T1l = T1j + T1k;
Chris@10: 	       }
Chris@10: 	       {
Chris@10: 		    E Tp, T17, T1f, T20, Ts, T1c, T1a, T21;
Chris@10: 		    {
Chris@10: 			 E Tn, To, T1d, T1e;
Chris@10: 			 Tn = ri[WS(is, 15)];
Chris@10: 			 To = ri[WS(is, 7)];
Chris@10: 			 Tp = Tn + To;
Chris@10: 			 T17 = Tn - To;
Chris@10: 			 T1d = ii[WS(is, 15)];
Chris@10: 			 T1e = ii[WS(is, 7)];
Chris@10: 			 T1f = T1d - T1e;
Chris@10: 			 T20 = T1d + T1e;
Chris@10: 		    }
Chris@10: 		    {
Chris@10: 			 E Tq, Tr, T18, T19;
Chris@10: 			 Tq = ri[WS(is, 3)];
Chris@10: 			 Tr = ri[WS(is, 11)];
Chris@10: 			 Ts = Tq + Tr;
Chris@10: 			 T1c = Tq - Tr;
Chris@10: 			 T18 = ii[WS(is, 3)];
Chris@10: 			 T19 = ii[WS(is, 11)];
Chris@10: 			 T1a = T18 - T19;
Chris@10: 			 T21 = T18 + T19;
Chris@10: 		    }
Chris@10: 		    Tt = Tp + Ts;
Chris@10: 		    T22 = T20 - T21;
Chris@10: 		    T2h = T20 + T21;
Chris@10: 		    T1b = T17 - T1a;
Chris@10: 		    T1g = T1c + T1f;
Chris@10: 		    T1E = T1f - T1c;
Chris@10: 		    T1Z = Tp - Ts;
Chris@10: 		    T1D = T17 + T1a;
Chris@10: 	       }
Chris@10: 	       {
Chris@10: 		    E Ta, TP, TF, TO, Td, TR, TI, TS;
Chris@10: 		    {
Chris@10: 			 E T8, T9, TD, TE;
Chris@10: 			 T8 = ri[WS(is, 2)];
Chris@10: 			 T9 = ri[WS(is, 10)];
Chris@10: 			 Ta = T8 + T9;
Chris@10: 			 TP = T8 - T9;
Chris@10: 			 TD = ii[WS(is, 2)];
Chris@10: 			 TE = ii[WS(is, 10)];
Chris@10: 			 TF = TD + TE;
Chris@10: 			 TO = TD - TE;
Chris@10: 		    }
Chris@10: 		    {
Chris@10: 			 E Tb, Tc, TG, TH;
Chris@10: 			 Tb = ri[WS(is, 14)];
Chris@10: 			 Tc = ri[WS(is, 6)];
Chris@10: 			 Td = Tb + Tc;
Chris@10: 			 TR = Tb - Tc;
Chris@10: 			 TG = ii[WS(is, 14)];
Chris@10: 			 TH = ii[WS(is, 6)];
Chris@10: 			 TI = TG + TH;
Chris@10: 			 TS = TG - TH;
Chris@10: 		    }
Chris@10: 		    Te = Ta + Td;
Chris@10: 		    T1S = TF - TI;
Chris@10: 		    T26 = Td - Ta;
Chris@10: 		    TJ = TF + TI;
Chris@10: 		    TQ = TO - TP;
Chris@10: 		    T1m = TR - TS;
Chris@10: 		    T1n = TP + TO;
Chris@10: 		    TT = TR + TS;
Chris@10: 	       }
Chris@10: 	       {
Chris@10: 		    E Ti, T11, TZ, T1V, Tl, TW, T14, T1W;
Chris@10: 		    {
Chris@10: 			 E Tg, Th, TX, TY;
Chris@10: 			 Tg = ri[WS(is, 1)];
Chris@10: 			 Th = ri[WS(is, 9)];
Chris@10: 			 Ti = Tg + Th;
Chris@10: 			 T11 = Tg - Th;
Chris@10: 			 TX = ii[WS(is, 1)];
Chris@10: 			 TY = ii[WS(is, 9)];
Chris@10: 			 TZ = TX - TY;
Chris@10: 			 T1V = TX + TY;
Chris@10: 		    }
Chris@10: 		    {
Chris@10: 			 E Tj, Tk, T12, T13;
Chris@10: 			 Tj = ri[WS(is, 5)];
Chris@10: 			 Tk = ri[WS(is, 13)];
Chris@10: 			 Tl = Tj + Tk;
Chris@10: 			 TW = Tj - Tk;
Chris@10: 			 T12 = ii[WS(is, 5)];
Chris@10: 			 T13 = ii[WS(is, 13)];
Chris@10: 			 T14 = T12 - T13;
Chris@10: 			 T1W = T12 + T13;
Chris@10: 		    }
Chris@10: 		    Tm = Ti + Tl;
Chris@10: 		    T1X = T1V - T1W;
Chris@10: 		    T2g = T1V + T1W;
Chris@10: 		    T10 = TW + TZ;
Chris@10: 		    T15 = T11 - T14;
Chris@10: 		    T1B = T11 + T14;
Chris@10: 		    T1U = Ti - Tl;
Chris@10: 		    T1A = TZ - TW;
Chris@10: 	       }
Chris@10: 	       {
Chris@10: 		    E Tf, Tu, T2j, T2k;
Chris@10: 		    Tf = T7 + Te;
Chris@10: 		    Tu = Tm + Tt;
Chris@10: 		    ro[WS(os, 8)] = Tf - Tu;
Chris@10: 		    ro[0] = Tf + Tu;
Chris@10: 		    T2j = TC + TJ;
Chris@10: 		    T2k = T2g + T2h;
Chris@10: 		    io[WS(os, 8)] = T2j - T2k;
Chris@10: 		    io[0] = T2j + T2k;
Chris@10: 	       }
Chris@10: 	       {
Chris@10: 		    E Tv, TK, T2f, T2i;
Chris@10: 		    Tv = Tt - Tm;
Chris@10: 		    TK = TC - TJ;
Chris@10: 		    io[WS(os, 4)] = Tv + TK;
Chris@10: 		    io[WS(os, 12)] = TK - Tv;
Chris@10: 		    T2f = T7 - Te;
Chris@10: 		    T2i = T2g - T2h;
Chris@10: 		    ro[WS(os, 12)] = T2f - T2i;
Chris@10: 		    ro[WS(os, 4)] = T2f + T2i;
Chris@10: 	       }
Chris@10: 	       {
Chris@10: 		    E T1T, T27, T24, T28, T1Y, T23;
Chris@10: 		    T1T = T1R + T1S;
Chris@10: 		    T27 = T25 - T26;
Chris@10: 		    T1Y = T1U + T1X;
Chris@10: 		    T23 = T1Z - T22;
Chris@10: 		    T24 = KP707106781 * (T1Y + T23);
Chris@10: 		    T28 = KP707106781 * (T23 - T1Y);
Chris@10: 		    ro[WS(os, 10)] = T1T - T24;
Chris@10: 		    io[WS(os, 6)] = T27 + T28;
Chris@10: 		    ro[WS(os, 2)] = T1T + T24;
Chris@10: 		    io[WS(os, 14)] = T27 - T28;
Chris@10: 	       }
Chris@10: 	       {
Chris@10: 		    E T29, T2d, T2c, T2e, T2a, T2b;
Chris@10: 		    T29 = T1R - T1S;
Chris@10: 		    T2d = T26 + T25;
Chris@10: 		    T2a = T1X - T1U;
Chris@10: 		    T2b = T1Z + T22;
Chris@10: 		    T2c = KP707106781 * (T2a - T2b);
Chris@10: 		    T2e = KP707106781 * (T2a + T2b);
Chris@10: 		    ro[WS(os, 14)] = T29 - T2c;
Chris@10: 		    io[WS(os, 2)] = T2d + T2e;
Chris@10: 		    ro[WS(os, 6)] = T29 + T2c;
Chris@10: 		    io[WS(os, 10)] = T2d - T2e;
Chris@10: 	       }
Chris@10: 	       {
Chris@10: 		    E TV, T1r, T1p, T1v, T1i, T1q, T1u, T1w, TU, T1o;
Chris@10: 		    TU = KP707106781 * (TQ - TT);
Chris@10: 		    TV = TN + TU;
Chris@10: 		    T1r = TN - TU;
Chris@10: 		    T1o = KP707106781 * (T1m - T1n);
Chris@10: 		    T1p = T1l - T1o;
Chris@10: 		    T1v = T1l + T1o;
Chris@10: 		    {
Chris@10: 			 E T16, T1h, T1s, T1t;
Chris@10: 			 T16 = FMA(KP923879532, T10, KP382683432 * T15);
Chris@10: 			 T1h = FNMS(KP923879532, T1g, KP382683432 * T1b);
Chris@10: 			 T1i = T16 + T1h;
Chris@10: 			 T1q = T1h - T16;
Chris@10: 			 T1s = FNMS(KP923879532, T15, KP382683432 * T10);
Chris@10: 			 T1t = FMA(KP382683432, T1g, KP923879532 * T1b);
Chris@10: 			 T1u = T1s - T1t;
Chris@10: 			 T1w = T1s + T1t;
Chris@10: 		    }
Chris@10: 		    ro[WS(os, 11)] = TV - T1i;
Chris@10: 		    io[WS(os, 11)] = T1v - T1w;
Chris@10: 		    ro[WS(os, 3)] = TV + T1i;
Chris@10: 		    io[WS(os, 3)] = T1v + T1w;
Chris@10: 		    io[WS(os, 15)] = T1p - T1q;
Chris@10: 		    ro[WS(os, 15)] = T1r - T1u;
Chris@10: 		    io[WS(os, 7)] = T1p + T1q;
Chris@10: 		    ro[WS(os, 7)] = T1r + T1u;
Chris@10: 	       }
Chris@10: 	       {
Chris@10: 		    E T1z, T1L, T1J, T1P, T1G, T1K, T1O, T1Q, T1y, T1I;
Chris@10: 		    T1y = KP707106781 * (T1n + T1m);
Chris@10: 		    T1z = T1x + T1y;
Chris@10: 		    T1L = T1x - T1y;
Chris@10: 		    T1I = KP707106781 * (TQ + TT);
Chris@10: 		    T1J = T1H - T1I;
Chris@10: 		    T1P = T1H + T1I;
Chris@10: 		    {
Chris@10: 			 E T1C, T1F, T1M, T1N;
Chris@10: 			 T1C = FMA(KP382683432, T1A, KP923879532 * T1B);
Chris@10: 			 T1F = FNMS(KP382683432, T1E, KP923879532 * T1D);
Chris@10: 			 T1G = T1C + T1F;
Chris@10: 			 T1K = T1F - T1C;
Chris@10: 			 T1M = FNMS(KP382683432, T1B, KP923879532 * T1A);
Chris@10: 			 T1N = FMA(KP923879532, T1E, KP382683432 * T1D);
Chris@10: 			 T1O = T1M - T1N;
Chris@10: 			 T1Q = T1M + T1N;
Chris@10: 		    }
Chris@10: 		    ro[WS(os, 9)] = T1z - T1G;
Chris@10: 		    io[WS(os, 9)] = T1P - T1Q;
Chris@10: 		    ro[WS(os, 1)] = T1z + T1G;
Chris@10: 		    io[WS(os, 1)] = T1P + T1Q;
Chris@10: 		    io[WS(os, 13)] = T1J - T1K;
Chris@10: 		    ro[WS(os, 13)] = T1L - T1O;
Chris@10: 		    io[WS(os, 5)] = T1J + T1K;
Chris@10: 		    ro[WS(os, 5)] = T1L + T1O;
Chris@10: 	       }
Chris@10: 	  }
Chris@10:      }
Chris@10: }
Chris@10: 
Chris@10: static const kdft_desc desc = { 16, "n1_16", {136, 16, 8, 0}, &GENUS, 0, 0, 0, 0 };
Chris@10: 
Chris@10: void X(codelet_n1_16) (planner *p) {
Chris@10:      X(kdft_register) (p, n1_16, &desc);
Chris@10: }
Chris@10: 
Chris@10: #endif				/* HAVE_FMA */