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 07:40:23 EST 2012 */
Chris@10: 
Chris@10: #include "codelet-rdft.h"
Chris@10: 
Chris@10: #ifdef HAVE_FMA
Chris@10: 
Chris@10: /* Generated by: ../../../genfft/gen_r2cf.native -fma -reorder-insns -schedule-for-pipeline -compact -variables 4 -pipeline-latency 4 -n 20 -name r2cfII_20 -dft-II -include r2cfII.h */
Chris@10: 
Chris@10: /*
Chris@10:  * This function contains 102 FP additions, 63 FP multiplications,
Chris@10:  * (or, 39 additions, 0 multiplications, 63 fused multiply/add),
Chris@10:  * 67 stack variables, 10 constants, and 40 memory accesses
Chris@10:  */
Chris@10: #include "r2cfII.h"
Chris@10: 
Chris@10: static void r2cfII_20(R *R0, R *R1, R *Cr, R *Ci, stride rs, stride csr, stride csi, INT v, INT ivs, INT ovs)
Chris@10: {
Chris@10:      DK(KP707106781, +0.707106781186547524400844362104849039284835938);
Chris@10:      DK(KP951056516, +0.951056516295153572116439333379382143405698634);
Chris@10:      DK(KP559016994, +0.559016994374947424102293417182819058860154590);
Chris@10:      DK(KP690983005, +0.690983005625052575897706582817180941139845410);
Chris@10:      DK(KP552786404, +0.552786404500042060718165266253744752911876328);
Chris@10:      DK(KP447213595, +0.447213595499957939281834733746255247088123672);
Chris@10:      DK(KP809016994, +0.809016994374947424102293417182819058860154590);
Chris@10:      DK(KP250000000, +0.250000000000000000000000000000000000000000000);
Chris@10:      DK(KP618033988, +0.618033988749894848204586834365638117720309180);
Chris@10:      DK(KP381966011, +0.381966011250105151795413165634361882279690820);
Chris@10:      {
Chris@10: 	  INT i;
Chris@10: 	  for (i = v; i > 0; i = i - 1, R0 = R0 + ivs, R1 = R1 + ivs, Cr = Cr + ovs, Ci = Ci + ovs, MAKE_VOLATILE_STRIDE(80, rs), MAKE_VOLATILE_STRIDE(80, csr), MAKE_VOLATILE_STRIDE(80, csi)) {
Chris@10: 	       E Tv, TK, TN, Th, T1l, T1n, Ts, TH;
Chris@10: 	       {
Chris@10: 		    E Ti, T1d, T1f, T1e, T1g, T1p, TS, Tg, To, T8, T7, T19, T1r, T1k, Tx;
Chris@10: 		    E Tp, TX, Ty, TF, Tr, TV, Tz, TA, TI;
Chris@10: 		    {
Chris@10: 			 E Ta, Tb, Td, Te;
Chris@10: 			 Ti = R1[WS(rs, 2)];
Chris@10: 			 T1d = R0[WS(rs, 5)];
Chris@10: 			 Ta = R0[WS(rs, 9)];
Chris@10: 			 Tb = R0[WS(rs, 1)];
Chris@10: 			 Td = R0[WS(rs, 3)];
Chris@10: 			 Te = R0[WS(rs, 7)];
Chris@10: 			 {
Chris@10: 			      E T1, T2, T5, T3, T4, T1i, Tc, Tf;
Chris@10: 			      T1 = R0[0];
Chris@10: 			      T1f = Ta + Tb;
Chris@10: 			      Tc = Ta - Tb;
Chris@10: 			      T1e = Td + Te;
Chris@10: 			      Tf = Td - Te;
Chris@10: 			      T2 = R0[WS(rs, 4)];
Chris@10: 			      T5 = R0[WS(rs, 6)];
Chris@10: 			      T1g = FMA(KP381966011, T1f, T1e);
Chris@10: 			      T1p = FMA(KP381966011, T1e, T1f);
Chris@10: 			      TS = FMA(KP618033988, Tc, Tf);
Chris@10: 			      Tg = FNMS(KP618033988, Tf, Tc);
Chris@10: 			      T3 = R0[WS(rs, 8)];
Chris@10: 			      T4 = R0[WS(rs, 2)];
Chris@10: 			      T1i = T2 + T5;
Chris@10: 			      {
Chris@10: 				   E Tj, Tu, Tm, Tt, Tn, Tq, TU;
Chris@10: 				   Tj = R1[WS(rs, 8)];
Chris@10: 				   To = R1[WS(rs, 6)];
Chris@10: 				   {
Chris@10: 					E T6, T1j, Tk, Tl;
Chris@10: 					T6 = T2 + T3 - T4 - T5;
Chris@10: 					T8 = (T3 + T5 - T2) - T4;
Chris@10: 					T1j = T3 + T4;
Chris@10: 					Tk = R1[0];
Chris@10: 					Tl = R1[WS(rs, 4)];
Chris@10: 					T7 = FNMS(KP250000000, T6, T1);
Chris@10: 					T19 = T1 + T6;
Chris@10: 					T1r = FNMS(KP618033988, T1i, T1j);
Chris@10: 					T1k = FMA(KP618033988, T1j, T1i);
Chris@10: 					Tu = Tk - Tl;
Chris@10: 					Tm = Tk + Tl;
Chris@10: 				   }
Chris@10: 				   Tt = To + Tj;
Chris@10: 				   Tx = R1[WS(rs, 7)];
Chris@10: 				   Tn = Tj - Tm;
Chris@10: 				   Tp = Tj + Tm;
Chris@10: 				   Tv = FNMS(KP618033988, Tu, Tt);
Chris@10: 				   TX = FMA(KP618033988, Tt, Tu);
Chris@10: 				   Tq = FMA(KP809016994, Tp, To);
Chris@10: 				   TU = FMA(KP447213595, Tp, Tn);
Chris@10: 				   Ty = R1[WS(rs, 1)];
Chris@10: 				   TF = R1[WS(rs, 3)];
Chris@10: 				   Tr = FNMS(KP552786404, Tq, Tn);
Chris@10: 				   TV = FNMS(KP690983005, TU, To);
Chris@10: 				   Tz = R1[WS(rs, 5)];
Chris@10: 				   TA = R1[WS(rs, 9)];
Chris@10: 				   TI = TF + Ty;
Chris@10: 			      }
Chris@10: 			 }
Chris@10: 		    }
Chris@10: 		    {
Chris@10: 			 E T1w, TJ, TB, T1a;
Chris@10: 			 T1w = T1f + T1d - T1e;
Chris@10: 			 TJ = Tz - TA;
Chris@10: 			 TB = Tz + TA;
Chris@10: 			 T1a = Ti + To - Tp;
Chris@10: 			 {
Chris@10: 			      E T9, T12, TT, T15, TG, TD, T1s, T1u, TW, T11, T10, T1h;
Chris@10: 			      {
Chris@10: 				   E TE, TC, TR, T1b;
Chris@10: 				   T9 = FNMS(KP559016994, T8, T7);
Chris@10: 				   TR = FMA(KP559016994, T8, T7);
Chris@10: 				   TK = FMA(KP618033988, TJ, TI);
Chris@10: 				   T12 = FNMS(KP618033988, TI, TJ);
Chris@10: 				   TE = Ty - TB;
Chris@10: 				   TC = Ty + TB;
Chris@10: 				   TT = FMA(KP951056516, TS, TR);
Chris@10: 				   T15 = FNMS(KP951056516, TS, TR);
Chris@10: 				   TG = FNMS(KP552786404, TF, TE);
Chris@10: 				   T1b = TC - TF - Tx;
Chris@10: 				   {
Chris@10: 					E TZ, T1q, T1c, T1x;
Chris@10: 					TZ = FMA(KP447213595, TC, TE);
Chris@10: 					TD = FMA(KP250000000, TC, Tx);
Chris@10: 					T1q = FNMS(KP809016994, T1p, T1d);
Chris@10: 					T1c = T1a + T1b;
Chris@10: 					T1x = T1a - T1b;
Chris@10: 					T10 = FNMS(KP690983005, TZ, TF);
Chris@10: 					T1s = FNMS(KP951056516, T1r, T1q);
Chris@10: 					T1u = FMA(KP951056516, T1r, T1q);
Chris@10: 					Ci[WS(csi, 7)] = FMA(KP707106781, T1x, T1w);
Chris@10: 					Ci[WS(csi, 2)] = FMS(KP707106781, T1x, T1w);
Chris@10: 					Cr[WS(csr, 7)] = FMA(KP707106781, T1c, T19);
Chris@10: 					Cr[WS(csr, 2)] = FNMS(KP707106781, T1c, T19);
Chris@10: 				   }
Chris@10: 			      }
Chris@10: 			      TW = FNMS(KP809016994, TV, Ti);
Chris@10: 			      T11 = FNMS(KP809016994, T10, Tx);
Chris@10: 			      T1h = FMA(KP809016994, T1g, T1d);
Chris@10: 			      {
Chris@10: 				   E T17, TY, T16, T13;
Chris@10: 				   T17 = FNMS(KP951056516, TX, TW);
Chris@10: 				   TY = FMA(KP951056516, TX, TW);
Chris@10: 				   T16 = FMA(KP951056516, T12, T11);
Chris@10: 				   T13 = FNMS(KP951056516, T12, T11);
Chris@10: 				   TN = FMA(KP951056516, Tg, T9);
Chris@10: 				   Th = FNMS(KP951056516, Tg, T9);
Chris@10: 				   {
Chris@10: 					E T18, T1v, T1t, T14;
Chris@10: 					T18 = T16 - T17;
Chris@10: 					T1v = T17 + T16;
Chris@10: 					T1t = TY + T13;
Chris@10: 					T14 = TY - T13;
Chris@10: 					Cr[WS(csr, 1)] = FMA(KP707106781, T18, T15);
Chris@10: 					Cr[WS(csr, 8)] = FNMS(KP707106781, T18, T15);
Chris@10: 					Ci[WS(csi, 3)] = FMA(KP707106781, T1v, T1u);
Chris@10: 					Ci[WS(csi, 6)] = FMS(KP707106781, T1v, T1u);
Chris@10: 					Ci[WS(csi, 1)] = FNMS(KP707106781, T1t, T1s);
Chris@10: 					Ci[WS(csi, 8)] = -(FMA(KP707106781, T1t, T1s));
Chris@10: 					Cr[WS(csr, 3)] = FMA(KP707106781, T14, TT);
Chris@10: 					Cr[WS(csr, 6)] = FNMS(KP707106781, T14, TT);
Chris@10: 					T1l = FMA(KP951056516, T1k, T1h);
Chris@10: 					T1n = FNMS(KP951056516, T1k, T1h);
Chris@10: 				   }
Chris@10: 			      }
Chris@10: 			      Ts = FNMS(KP559016994, Tr, Ti);
Chris@10: 			      TH = FNMS(KP559016994, TG, TD);
Chris@10: 			 }
Chris@10: 		    }
Chris@10: 	       }
Chris@10: 	       {
Chris@10: 		    E TO, Tw, TP, TL;
Chris@10: 		    TO = FMA(KP951056516, Tv, Ts);
Chris@10: 		    Tw = FNMS(KP951056516, Tv, Ts);
Chris@10: 		    TP = FMA(KP951056516, TK, TH);
Chris@10: 		    TL = FNMS(KP951056516, TK, TH);
Chris@10: 		    {
Chris@10: 			 E TQ, T1m, T1o, TM;
Chris@10: 			 TQ = TO - TP;
Chris@10: 			 T1m = TO + TP;
Chris@10: 			 T1o = Tw + TL;
Chris@10: 			 TM = Tw - TL;
Chris@10: 			 Cr[WS(csr, 4)] = FMA(KP707106781, TQ, TN);
Chris@10: 			 Cr[WS(csr, 5)] = FNMS(KP707106781, TQ, TN);
Chris@10: 			 Ci[WS(csi, 9)] = FNMS(KP707106781, T1m, T1l);
Chris@10: 			 Ci[0] = -(FMA(KP707106781, T1m, T1l));
Chris@10: 			 Ci[WS(csi, 5)] = FNMS(KP707106781, T1o, T1n);
Chris@10: 			 Ci[WS(csi, 4)] = -(FMA(KP707106781, T1o, T1n));
Chris@10: 			 Cr[0] = FMA(KP707106781, TM, Th);
Chris@10: 			 Cr[WS(csr, 9)] = FNMS(KP707106781, TM, Th);
Chris@10: 		    }
Chris@10: 	       }
Chris@10: 	  }
Chris@10:      }
Chris@10: }
Chris@10: 
Chris@10: static const kr2c_desc desc = { 20, "r2cfII_20", {39, 0, 63, 0}, &GENUS };
Chris@10: 
Chris@10: void X(codelet_r2cfII_20) (planner *p) {
Chris@10:      X(kr2c_register) (p, r2cfII_20, &desc);
Chris@10: }
Chris@10: 
Chris@10: #else				/* HAVE_FMA */
Chris@10: 
Chris@10: /* Generated by: ../../../genfft/gen_r2cf.native -compact -variables 4 -pipeline-latency 4 -n 20 -name r2cfII_20 -dft-II -include r2cfII.h */
Chris@10: 
Chris@10: /*
Chris@10:  * This function contains 102 FP additions, 34 FP multiplications,
Chris@10:  * (or, 86 additions, 18 multiplications, 16 fused multiply/add),
Chris@10:  * 60 stack variables, 13 constants, and 40 memory accesses
Chris@10:  */
Chris@10: #include "r2cfII.h"
Chris@10: 
Chris@10: static void r2cfII_20(R *R0, R *R1, R *Cr, R *Ci, stride rs, stride csr, stride csi, INT v, INT ivs, INT ovs)
Chris@10: {
Chris@10:      DK(KP572061402, +0.572061402817684297600072783580302076536153377);
Chris@10:      DK(KP218508012, +0.218508012224410535399650602527877556893735408);
Chris@10:      DK(KP309016994, +0.309016994374947424102293417182819058860154590);
Chris@10:      DK(KP809016994, +0.809016994374947424102293417182819058860154590);
Chris@10:      DK(KP559016994, +0.559016994374947424102293417182819058860154590);
Chris@10:      DK(KP951056516, +0.951056516295153572116439333379382143405698634);
Chris@10:      DK(KP587785252, +0.587785252292473129168705954639072768597652438);
Chris@10:      DK(KP250000000, +0.250000000000000000000000000000000000000000000);
Chris@10:      DK(KP176776695, +0.176776695296636881100211090526212259821208984);
Chris@10:      DK(KP395284707, +0.395284707521047416499861693054089816714944392);
Chris@10:      DK(KP672498511, +0.672498511963957326960058968885748755876783111);
Chris@10:      DK(KP415626937, +0.415626937777453428589967464113135184222253485);
Chris@10:      DK(KP707106781, +0.707106781186547524400844362104849039284835938);
Chris@10:      {
Chris@10: 	  INT i;
Chris@10: 	  for (i = v; i > 0; i = i - 1, R0 = R0 + ivs, R1 = R1 + ivs, Cr = Cr + ovs, Ci = Ci + ovs, MAKE_VOLATILE_STRIDE(80, rs), MAKE_VOLATILE_STRIDE(80, csr), MAKE_VOLATILE_STRIDE(80, csi)) {
Chris@10: 	       E T8, TD, Tm, TN, T9, TC, TY, TE, Te, TF, Tl, TK, T12, TL, Tk;
Chris@10: 	       E TM, T1, T6, Tq, T1l, T1c, Tp, T1f, T1e, T1d, Ty, TW, T1g, T1m, Tx;
Chris@10: 	       E Tu;
Chris@10: 	       T8 = R1[WS(rs, 2)];
Chris@10: 	       TD = KP707106781 * T8;
Chris@10: 	       Tm = R1[WS(rs, 7)];
Chris@10: 	       TN = KP707106781 * Tm;
Chris@10: 	       {
Chris@10: 		    E Ta, TA, Td, TB, Tb, Tc;
Chris@10: 		    T9 = R1[WS(rs, 6)];
Chris@10: 		    Ta = R1[WS(rs, 8)];
Chris@10: 		    TA = T9 + Ta;
Chris@10: 		    Tb = R1[0];
Chris@10: 		    Tc = R1[WS(rs, 4)];
Chris@10: 		    Td = Tb + Tc;
Chris@10: 		    TB = Tb - Tc;
Chris@10: 		    TC = FMA(KP415626937, TA, KP672498511 * TB);
Chris@10: 		    TY = FNMS(KP415626937, TB, KP672498511 * TA);
Chris@10: 		    TE = KP395284707 * (Ta - Td);
Chris@10: 		    Te = Ta + Td;
Chris@10: 		    TF = KP176776695 * Te;
Chris@10: 	       }
Chris@10: 	       {
Chris@10: 		    E Tg, TJ, Tj, TI, Th, Ti;
Chris@10: 		    Tg = R1[WS(rs, 1)];
Chris@10: 		    Tl = R1[WS(rs, 3)];
Chris@10: 		    TJ = Tg + Tl;
Chris@10: 		    Th = R1[WS(rs, 5)];
Chris@10: 		    Ti = R1[WS(rs, 9)];
Chris@10: 		    Tj = Th + Ti;
Chris@10: 		    TI = Th - Ti;
Chris@10: 		    TK = FNMS(KP415626937, TJ, KP672498511 * TI);
Chris@10: 		    T12 = FMA(KP415626937, TI, KP672498511 * TJ);
Chris@10: 		    TL = KP395284707 * (Tg - Tj);
Chris@10: 		    Tk = Tg + Tj;
Chris@10: 		    TM = KP176776695 * Tk;
Chris@10: 	       }
Chris@10: 	       {
Chris@10: 		    E T2, T5, T3, T4, T1a, T1b;
Chris@10: 		    T1 = R0[0];
Chris@10: 		    T2 = R0[WS(rs, 6)];
Chris@10: 		    T5 = R0[WS(rs, 8)];
Chris@10: 		    T3 = R0[WS(rs, 2)];
Chris@10: 		    T4 = R0[WS(rs, 4)];
Chris@10: 		    T1a = T4 + T2;
Chris@10: 		    T1b = T5 + T3;
Chris@10: 		    T6 = T2 + T3 - (T4 + T5);
Chris@10: 		    Tq = FMA(KP250000000, T6, T1);
Chris@10: 		    T1l = FNMS(KP951056516, T1b, KP587785252 * T1a);
Chris@10: 		    T1c = FMA(KP951056516, T1a, KP587785252 * T1b);
Chris@10: 		    Tp = KP559016994 * (T5 + T2 - (T4 + T3));
Chris@10: 	       }
Chris@10: 	       T1f = R0[WS(rs, 5)];
Chris@10: 	       {
Chris@10: 		    E Tv, Tw, Ts, Tt;
Chris@10: 		    Tv = R0[WS(rs, 9)];
Chris@10: 		    Tw = R0[WS(rs, 1)];
Chris@10: 		    Tx = Tv - Tw;
Chris@10: 		    T1e = Tv + Tw;
Chris@10: 		    Ts = R0[WS(rs, 3)];
Chris@10: 		    Tt = R0[WS(rs, 7)];
Chris@10: 		    Tu = Ts - Tt;
Chris@10: 		    T1d = Ts + Tt;
Chris@10: 	       }
Chris@10: 	       Ty = FMA(KP951056516, Tu, KP587785252 * Tx);
Chris@10: 	       TW = FNMS(KP951056516, Tx, KP587785252 * Tu);
Chris@10: 	       T1g = FMA(KP809016994, T1d, KP309016994 * T1e) + T1f;
Chris@10: 	       T1m = FNMS(KP809016994, T1e, T1f) - (KP309016994 * T1d);
Chris@10: 	       {
Chris@10: 		    E T7, T1r, To, T1q, Tf, Tn;
Chris@10: 		    T7 = T1 - T6;
Chris@10: 		    T1r = T1e + T1f - T1d;
Chris@10: 		    Tf = T8 + (T9 - Te);
Chris@10: 		    Tn = (Tk - Tl) - Tm;
Chris@10: 		    To = KP707106781 * (Tf + Tn);
Chris@10: 		    T1q = KP707106781 * (Tf - Tn);
Chris@10: 		    Cr[WS(csr, 2)] = T7 - To;
Chris@10: 		    Ci[WS(csi, 2)] = T1q - T1r;
Chris@10: 		    Cr[WS(csr, 7)] = T7 + To;
Chris@10: 		    Ci[WS(csi, 7)] = T1q + T1r;
Chris@10: 	       }
Chris@10: 	       {
Chris@10: 		    E T1h, T1j, TX, T15, T10, T16, T13, T17, TV, TZ, T11;
Chris@10: 		    T1h = T1c - T1g;
Chris@10: 		    T1j = T1c + T1g;
Chris@10: 		    TV = Tq - Tp;
Chris@10: 		    TX = TV - TW;
Chris@10: 		    T15 = TV + TW;
Chris@10: 		    TZ = FMA(KP218508012, T9, TD) + TF - TE;
Chris@10: 		    T10 = TY + TZ;
Chris@10: 		    T16 = TZ - TY;
Chris@10: 		    T11 = FNMS(KP218508012, Tl, TL) - (TM + TN);
Chris@10: 		    T13 = T11 - T12;
Chris@10: 		    T17 = T11 + T12;
Chris@10: 		    {
Chris@10: 			 E T14, T19, T18, T1i;
Chris@10: 			 T14 = T10 + T13;
Chris@10: 			 Cr[WS(csr, 5)] = TX - T14;
Chris@10: 			 Cr[WS(csr, 4)] = TX + T14;
Chris@10: 			 T19 = T17 - T16;
Chris@10: 			 Ci[WS(csi, 5)] = T19 - T1h;
Chris@10: 			 Ci[WS(csi, 4)] = T19 + T1h;
Chris@10: 			 T18 = T16 + T17;
Chris@10: 			 Cr[WS(csr, 9)] = T15 - T18;
Chris@10: 			 Cr[0] = T15 + T18;
Chris@10: 			 T1i = T13 - T10;
Chris@10: 			 Ci[0] = T1i - T1j;
Chris@10: 			 Ci[WS(csi, 9)] = T1i + T1j;
Chris@10: 		    }
Chris@10: 	       }
Chris@10: 	       {
Chris@10: 		    E T1n, T1p, Tz, TR, TH, TS, TP, TT, Tr, TG, TO;
Chris@10: 		    T1n = T1l + T1m;
Chris@10: 		    T1p = T1m - T1l;
Chris@10: 		    Tr = Tp + Tq;
Chris@10: 		    Tz = Tr + Ty;
Chris@10: 		    TR = Tr - Ty;
Chris@10: 		    TG = TD + TE + FNMS(KP572061402, T9, TF);
Chris@10: 		    TH = TC + TG;
Chris@10: 		    TS = TC - TG;
Chris@10: 		    TO = TL + TM + FNMS(KP572061402, Tl, TN);
Chris@10: 		    TP = TK - TO;
Chris@10: 		    TT = TK + TO;
Chris@10: 		    {
Chris@10: 			 E TQ, T1o, TU, T1k;
Chris@10: 			 TQ = TH + TP;
Chris@10: 			 Cr[WS(csr, 6)] = Tz - TQ;
Chris@10: 			 Cr[WS(csr, 3)] = Tz + TQ;
Chris@10: 			 T1o = TT - TS;
Chris@10: 			 Ci[WS(csi, 6)] = T1o - T1p;
Chris@10: 			 Ci[WS(csi, 3)] = T1o + T1p;
Chris@10: 			 TU = TS + TT;
Chris@10: 			 Cr[WS(csr, 8)] = TR - TU;
Chris@10: 			 Cr[WS(csr, 1)] = TR + TU;
Chris@10: 			 T1k = TP - TH;
Chris@10: 			 Ci[WS(csi, 8)] = T1k - T1n;
Chris@10: 			 Ci[WS(csi, 1)] = T1k + T1n;
Chris@10: 		    }
Chris@10: 	       }
Chris@10: 	  }
Chris@10:      }
Chris@10: }
Chris@10: 
Chris@10: static const kr2c_desc desc = { 20, "r2cfII_20", {86, 18, 16, 0}, &GENUS };
Chris@10: 
Chris@10: void X(codelet_r2cfII_20) (planner *p) {
Chris@10:      X(kr2c_register) (p, r2cfII_20, &desc);
Chris@10: }
Chris@10: 
Chris@10: #endif				/* HAVE_FMA */