Chris@42: /*
Chris@42:  * Copyright (c) 2003, 2007-14 Matteo Frigo
Chris@42:  * Copyright (c) 2003, 2007-14 Massachusetts Institute of Technology
Chris@42:  *
Chris@42:  * This program is free software; you can redistribute it and/or modify
Chris@42:  * it under the terms of the GNU General Public License as published by
Chris@42:  * the Free Software Foundation; either version 2 of the License, or
Chris@42:  * (at your option) any later version.
Chris@42:  *
Chris@42:  * This program is distributed in the hope that it will be useful,
Chris@42:  * but WITHOUT ANY WARRANTY; without even the implied warranty of
Chris@42:  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
Chris@42:  * GNU General Public License for more details.
Chris@42:  *
Chris@42:  * You should have received a copy of the GNU General Public License
Chris@42:  * along with this program; if not, write to the Free Software
Chris@42:  * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA  02110-1301  USA
Chris@42:  *
Chris@42:  */
Chris@42: 
Chris@42: /* This file was automatically generated --- DO NOT EDIT */
Chris@42: /* Generated on Sat Jul 30 16:40:10 EDT 2016 */
Chris@42: 
Chris@42: #include "codelet-dft.h"
Chris@42: 
Chris@42: #ifdef HAVE_FMA
Chris@42: 
Chris@42: /* Generated by: ../../../genfft/gen_notw_c.native -fma -reorder-insns -schedule-for-pipeline -simd -compact -variables 4 -pipeline-latency 8 -n 14 -name n2fv_14 -with-ostride 2 -include n2f.h -store-multiple 2 */
Chris@42: 
Chris@42: /*
Chris@42:  * This function contains 74 FP additions, 48 FP multiplications,
Chris@42:  * (or, 32 additions, 6 multiplications, 42 fused multiply/add),
Chris@42:  * 65 stack variables, 6 constants, and 35 memory accesses
Chris@42:  */
Chris@42: #include "n2f.h"
Chris@42: 
Chris@42: static void n2fv_14(const R *ri, const R *ii, R *ro, R *io, stride is, stride os, INT v, INT ivs, INT ovs)
Chris@42: {
Chris@42:      DVK(KP900968867, +0.900968867902419126236102319507445051165919162);
Chris@42:      DVK(KP801937735, +0.801937735804838252472204639014890102331838324);
Chris@42:      DVK(KP974927912, +0.974927912181823607018131682993931217232785801);
Chris@42:      DVK(KP692021471, +0.692021471630095869627814897002069140197260599);
Chris@42:      DVK(KP554958132, +0.554958132087371191422194871006410481067288862);
Chris@42:      DVK(KP356895867, +0.356895867892209443894399510021300583399127187);
Chris@42:      {
Chris@42: 	  INT i;
Chris@42: 	  const R *xi;
Chris@42: 	  R *xo;
Chris@42: 	  xi = ri;
Chris@42: 	  xo = ro;
Chris@42: 	  for (i = v; i > 0; i = i - VL, xi = xi + (VL * ivs), xo = xo + (VL * ovs), MAKE_VOLATILE_STRIDE(28, is), MAKE_VOLATILE_STRIDE(28, os)) {
Chris@42: 	       V TH, T3, TP, Tn, Ta, Ts, TW, TK, TO, Tk, TM, Tg, TL, Td, T1;
Chris@42: 	       V T2;
Chris@42: 	       T1 = LD(&(xi[0]), ivs, &(xi[0]));
Chris@42: 	       T2 = LD(&(xi[WS(is, 7)]), ivs, &(xi[WS(is, 1)]));
Chris@42: 	       {
Chris@42: 		    V Ti, TI, T6, TJ, T9, Tj, Te, Tf, Tb, Tc;
Chris@42: 		    {
Chris@42: 			 V T4, T5, T7, T8, Tl, Tm;
Chris@42: 			 T4 = LD(&(xi[WS(is, 2)]), ivs, &(xi[0]));
Chris@42: 			 T5 = LD(&(xi[WS(is, 9)]), ivs, &(xi[WS(is, 1)]));
Chris@42: 			 T7 = LD(&(xi[WS(is, 12)]), ivs, &(xi[0]));
Chris@42: 			 T8 = LD(&(xi[WS(is, 5)]), ivs, &(xi[WS(is, 1)]));
Chris@42: 			 Tl = LD(&(xi[WS(is, 8)]), ivs, &(xi[0]));
Chris@42: 			 Tm = LD(&(xi[WS(is, 1)]), ivs, &(xi[WS(is, 1)]));
Chris@42: 			 Ti = LD(&(xi[WS(is, 6)]), ivs, &(xi[0]));
Chris@42: 			 TH = VADD(T1, T2);
Chris@42: 			 T3 = VSUB(T1, T2);
Chris@42: 			 TI = VADD(T4, T5);
Chris@42: 			 T6 = VSUB(T4, T5);
Chris@42: 			 TJ = VADD(T7, T8);
Chris@42: 			 T9 = VSUB(T7, T8);
Chris@42: 			 TP = VADD(Tl, Tm);
Chris@42: 			 Tn = VSUB(Tl, Tm);
Chris@42: 			 Tj = LD(&(xi[WS(is, 13)]), ivs, &(xi[WS(is, 1)]));
Chris@42: 			 Te = LD(&(xi[WS(is, 10)]), ivs, &(xi[0]));
Chris@42: 			 Tf = LD(&(xi[WS(is, 3)]), ivs, &(xi[WS(is, 1)]));
Chris@42: 			 Tb = LD(&(xi[WS(is, 4)]), ivs, &(xi[0]));
Chris@42: 			 Tc = LD(&(xi[WS(is, 11)]), ivs, &(xi[WS(is, 1)]));
Chris@42: 		    }
Chris@42: 		    Ta = VADD(T6, T9);
Chris@42: 		    Ts = VSUB(T9, T6);
Chris@42: 		    TW = VSUB(TJ, TI);
Chris@42: 		    TK = VADD(TI, TJ);
Chris@42: 		    TO = VADD(Ti, Tj);
Chris@42: 		    Tk = VSUB(Ti, Tj);
Chris@42: 		    TM = VADD(Te, Tf);
Chris@42: 		    Tg = VSUB(Te, Tf);
Chris@42: 		    TL = VADD(Tb, Tc);
Chris@42: 		    Td = VSUB(Tb, Tc);
Chris@42: 	       }
Chris@42: 	       {
Chris@42: 		    V T19, T1a, T18, TB, T13, TY, TG, Tw, T11, Tr, T16, TT, Tz, TE, TU;
Chris@42: 		    V TQ;
Chris@42: 		    TU = VSUB(TO, TP);
Chris@42: 		    TQ = VADD(TO, TP);
Chris@42: 		    {
Chris@42: 			 V Tt, To, TV, TN;
Chris@42: 			 Tt = VSUB(Tn, Tk);
Chris@42: 			 To = VADD(Tk, Tn);
Chris@42: 			 TV = VSUB(TL, TM);
Chris@42: 			 TN = VADD(TL, TM);
Chris@42: 			 {
Chris@42: 			      V Tu, Th, TZ, T17;
Chris@42: 			      Tu = VSUB(Tg, Td);
Chris@42: 			      Th = VADD(Td, Tg);
Chris@42: 			      TZ = VFNMS(LDK(KP356895867), TK, TQ);
Chris@42: 			      T17 = VFNMS(LDK(KP554958132), TU, TW);
Chris@42: 			      {
Chris@42: 				   V Tp, TA, T14, TR;
Chris@42: 				   Tp = VFNMS(LDK(KP356895867), Ta, To);
Chris@42: 				   TA = VFMA(LDK(KP554958132), Tt, Ts);
Chris@42: 				   T19 = VADD(TH, VADD(TK, VADD(TN, TQ)));
Chris@42: 				   STM2(&(xo[0]), T19, ovs, &(xo[0]));
Chris@42: 				   T14 = VFNMS(LDK(KP356895867), TN, TK);
Chris@42: 				   TR = VFNMS(LDK(KP356895867), TQ, TN);
Chris@42: 				   {
Chris@42: 					V T12, TX, Tx, TC;
Chris@42: 					T12 = VFMA(LDK(KP554958132), TV, TU);
Chris@42: 					TX = VFMA(LDK(KP554958132), TW, TV);
Chris@42: 					T1a = VADD(T3, VADD(Ta, VADD(Th, To)));
Chris@42: 					STM2(&(xo[14]), T1a, ovs, &(xo[2]));
Chris@42: 					Tx = VFNMS(LDK(KP356895867), Th, Ta);
Chris@42: 					TC = VFNMS(LDK(KP356895867), To, Th);
Chris@42: 					{
Chris@42: 					     V TF, Tv, T10, Tq;
Chris@42: 					     TF = VFNMS(LDK(KP554958132), Ts, Tu);
Chris@42: 					     Tv = VFMA(LDK(KP554958132), Tu, Tt);
Chris@42: 					     T10 = VFNMS(LDK(KP692021471), TZ, TN);
Chris@42: 					     T18 = VMUL(LDK(KP974927912), VFNMS(LDK(KP801937735), T17, TV));
Chris@42: 					     Tq = VFNMS(LDK(KP692021471), Tp, Th);
Chris@42: 					     TB = VMUL(LDK(KP974927912), VFMA(LDK(KP801937735), TA, Tu));
Chris@42: 					     {
Chris@42: 						  V T15, TS, Ty, TD;
Chris@42: 						  T15 = VFNMS(LDK(KP692021471), T14, TQ);
Chris@42: 						  TS = VFNMS(LDK(KP692021471), TR, TK);
Chris@42: 						  T13 = VMUL(LDK(KP974927912), VFMA(LDK(KP801937735), T12, TW));
Chris@42: 						  TY = VMUL(LDK(KP974927912), VFNMS(LDK(KP801937735), TX, TU));
Chris@42: 						  Ty = VFNMS(LDK(KP692021471), Tx, To);
Chris@42: 						  TD = VFNMS(LDK(KP692021471), TC, Ta);
Chris@42: 						  TG = VMUL(LDK(KP974927912), VFNMS(LDK(KP801937735), TF, Tt));
Chris@42: 						  Tw = VMUL(LDK(KP974927912), VFNMS(LDK(KP801937735), Tv, Ts));
Chris@42: 						  T11 = VFNMS(LDK(KP900968867), T10, TH);
Chris@42: 						  Tr = VFNMS(LDK(KP900968867), Tq, T3);
Chris@42: 						  T16 = VFNMS(LDK(KP900968867), T15, TH);
Chris@42: 						  TT = VFNMS(LDK(KP900968867), TS, TH);
Chris@42: 						  Tz = VFNMS(LDK(KP900968867), Ty, T3);
Chris@42: 						  TE = VFNMS(LDK(KP900968867), TD, T3);
Chris@42: 					     }
Chris@42: 					}
Chris@42: 				   }
Chris@42: 			      }
Chris@42: 			 }
Chris@42: 		    }
Chris@42: 		    {
Chris@42: 			 V T1b, T1c, T1d, T1e;
Chris@42: 			 T1b = VFNMSI(T13, T11);
Chris@42: 			 STM2(&(xo[24]), T1b, ovs, &(xo[0]));
Chris@42: 			 T1c = VFMAI(T13, T11);
Chris@42: 			 STM2(&(xo[4]), T1c, ovs, &(xo[0]));
Chris@42: 			 T1d = VFMAI(Tw, Tr);
Chris@42: 			 STM2(&(xo[18]), T1d, ovs, &(xo[2]));
Chris@42: 			 T1e = VFNMSI(Tw, Tr);
Chris@42: 			 STM2(&(xo[10]), T1e, ovs, &(xo[2]));
Chris@42: 			 {
Chris@42: 			      V T1f, T1g, T1h, T1i;
Chris@42: 			      T1f = VFNMSI(T18, T16);
Chris@42: 			      STM2(&(xo[16]), T1f, ovs, &(xo[0]));
Chris@42: 			      STN2(&(xo[16]), T1f, T1d, ovs);
Chris@42: 			      T1g = VFMAI(T18, T16);
Chris@42: 			      STM2(&(xo[12]), T1g, ovs, &(xo[0]));
Chris@42: 			      STN2(&(xo[12]), T1g, T1a, ovs);
Chris@42: 			      T1h = VFNMSI(TY, TT);
Chris@42: 			      STM2(&(xo[20]), T1h, ovs, &(xo[0]));
Chris@42: 			      T1i = VFMAI(TY, TT);
Chris@42: 			      STM2(&(xo[8]), T1i, ovs, &(xo[0]));
Chris@42: 			      STN2(&(xo[8]), T1i, T1e, ovs);
Chris@42: 			      {
Chris@42: 				   V T1j, T1k, T1l, T1m;
Chris@42: 				   T1j = VFMAI(TB, Tz);
Chris@42: 				   STM2(&(xo[2]), T1j, ovs, &(xo[2]));
Chris@42: 				   STN2(&(xo[0]), T19, T1j, ovs);
Chris@42: 				   T1k = VFNMSI(TB, Tz);
Chris@42: 				   STM2(&(xo[26]), T1k, ovs, &(xo[2]));
Chris@42: 				   STN2(&(xo[24]), T1b, T1k, ovs);
Chris@42: 				   T1l = VFMAI(TG, TE);
Chris@42: 				   STM2(&(xo[6]), T1l, ovs, &(xo[2]));
Chris@42: 				   STN2(&(xo[4]), T1c, T1l, ovs);
Chris@42: 				   T1m = VFNMSI(TG, TE);
Chris@42: 				   STM2(&(xo[22]), T1m, ovs, &(xo[2]));
Chris@42: 				   STN2(&(xo[20]), T1h, T1m, ovs);
Chris@42: 			      }
Chris@42: 			 }
Chris@42: 		    }
Chris@42: 	       }
Chris@42: 	  }
Chris@42:      }
Chris@42:      VLEAVE();
Chris@42: }
Chris@42: 
Chris@42: static const kdft_desc desc = { 14, XSIMD_STRING("n2fv_14"), {32, 6, 42, 0}, &GENUS, 0, 2, 0, 0 };
Chris@42: 
Chris@42: void XSIMD(codelet_n2fv_14) (planner *p) {
Chris@42:      X(kdft_register) (p, n2fv_14, &desc);
Chris@42: }
Chris@42: 
Chris@42: #else				/* HAVE_FMA */
Chris@42: 
Chris@42: /* Generated by: ../../../genfft/gen_notw_c.native -simd -compact -variables 4 -pipeline-latency 8 -n 14 -name n2fv_14 -with-ostride 2 -include n2f.h -store-multiple 2 */
Chris@42: 
Chris@42: /*
Chris@42:  * This function contains 74 FP additions, 36 FP multiplications,
Chris@42:  * (or, 50 additions, 12 multiplications, 24 fused multiply/add),
Chris@42:  * 39 stack variables, 6 constants, and 35 memory accesses
Chris@42:  */
Chris@42: #include "n2f.h"
Chris@42: 
Chris@42: static void n2fv_14(const R *ri, const R *ii, R *ro, R *io, stride is, stride os, INT v, INT ivs, INT ovs)
Chris@42: {
Chris@42:      DVK(KP222520933, +0.222520933956314404288902564496794759466355569);
Chris@42:      DVK(KP900968867, +0.900968867902419126236102319507445051165919162);
Chris@42:      DVK(KP623489801, +0.623489801858733530525004884004239810632274731);
Chris@42:      DVK(KP433883739, +0.433883739117558120475768332848358754609990728);
Chris@42:      DVK(KP781831482, +0.781831482468029808708444526674057750232334519);
Chris@42:      DVK(KP974927912, +0.974927912181823607018131682993931217232785801);
Chris@42:      {
Chris@42: 	  INT i;
Chris@42: 	  const R *xi;
Chris@42: 	  R *xo;
Chris@42: 	  xi = ri;
Chris@42: 	  xo = ro;
Chris@42: 	  for (i = v; i > 0; i = i - VL, xi = xi + (VL * ivs), xo = xo + (VL * ovs), MAKE_VOLATILE_STRIDE(28, is), MAKE_VOLATILE_STRIDE(28, os)) {
Chris@42: 	       V T3, Ty, To, TK, Tr, TE, Ta, TJ, Tq, TB, Th, TL, Ts, TH, T1;
Chris@42: 	       V T2;
Chris@42: 	       T1 = LD(&(xi[0]), ivs, &(xi[0]));
Chris@42: 	       T2 = LD(&(xi[WS(is, 7)]), ivs, &(xi[WS(is, 1)]));
Chris@42: 	       T3 = VSUB(T1, T2);
Chris@42: 	       Ty = VADD(T1, T2);
Chris@42: 	       {
Chris@42: 		    V Tk, TC, Tn, TD;
Chris@42: 		    {
Chris@42: 			 V Ti, Tj, Tl, Tm;
Chris@42: 			 Ti = LD(&(xi[WS(is, 6)]), ivs, &(xi[0]));
Chris@42: 			 Tj = LD(&(xi[WS(is, 13)]), ivs, &(xi[WS(is, 1)]));
Chris@42: 			 Tk = VSUB(Ti, Tj);
Chris@42: 			 TC = VADD(Ti, Tj);
Chris@42: 			 Tl = LD(&(xi[WS(is, 8)]), ivs, &(xi[0]));
Chris@42: 			 Tm = LD(&(xi[WS(is, 1)]), ivs, &(xi[WS(is, 1)]));
Chris@42: 			 Tn = VSUB(Tl, Tm);
Chris@42: 			 TD = VADD(Tl, Tm);
Chris@42: 		    }
Chris@42: 		    To = VADD(Tk, Tn);
Chris@42: 		    TK = VSUB(TC, TD);
Chris@42: 		    Tr = VSUB(Tn, Tk);
Chris@42: 		    TE = VADD(TC, TD);
Chris@42: 	       }
Chris@42: 	       {
Chris@42: 		    V T6, Tz, T9, TA;
Chris@42: 		    {
Chris@42: 			 V T4, T5, T7, T8;
Chris@42: 			 T4 = LD(&(xi[WS(is, 2)]), ivs, &(xi[0]));
Chris@42: 			 T5 = LD(&(xi[WS(is, 9)]), ivs, &(xi[WS(is, 1)]));
Chris@42: 			 T6 = VSUB(T4, T5);
Chris@42: 			 Tz = VADD(T4, T5);
Chris@42: 			 T7 = LD(&(xi[WS(is, 12)]), ivs, &(xi[0]));
Chris@42: 			 T8 = LD(&(xi[WS(is, 5)]), ivs, &(xi[WS(is, 1)]));
Chris@42: 			 T9 = VSUB(T7, T8);
Chris@42: 			 TA = VADD(T7, T8);
Chris@42: 		    }
Chris@42: 		    Ta = VADD(T6, T9);
Chris@42: 		    TJ = VSUB(TA, Tz);
Chris@42: 		    Tq = VSUB(T9, T6);
Chris@42: 		    TB = VADD(Tz, TA);
Chris@42: 	       }
Chris@42: 	       {
Chris@42: 		    V Td, TF, Tg, TG;
Chris@42: 		    {
Chris@42: 			 V Tb, Tc, Te, Tf;
Chris@42: 			 Tb = LD(&(xi[WS(is, 4)]), ivs, &(xi[0]));
Chris@42: 			 Tc = LD(&(xi[WS(is, 11)]), ivs, &(xi[WS(is, 1)]));
Chris@42: 			 Td = VSUB(Tb, Tc);
Chris@42: 			 TF = VADD(Tb, Tc);
Chris@42: 			 Te = LD(&(xi[WS(is, 10)]), ivs, &(xi[0]));
Chris@42: 			 Tf = LD(&(xi[WS(is, 3)]), ivs, &(xi[WS(is, 1)]));
Chris@42: 			 Tg = VSUB(Te, Tf);
Chris@42: 			 TG = VADD(Te, Tf);
Chris@42: 		    }
Chris@42: 		    Th = VADD(Td, Tg);
Chris@42: 		    TL = VSUB(TF, TG);
Chris@42: 		    Ts = VSUB(Tg, Td);
Chris@42: 		    TH = VADD(TF, TG);
Chris@42: 	       }
Chris@42: 	       {
Chris@42: 		    V TR, TS, TT, TU, TV, TW;
Chris@42: 		    TR = VADD(T3, VADD(Ta, VADD(Th, To)));
Chris@42: 		    STM2(&(xo[14]), TR, ovs, &(xo[2]));
Chris@42: 		    TS = VADD(Ty, VADD(TB, VADD(TH, TE)));
Chris@42: 		    STM2(&(xo[0]), TS, ovs, &(xo[0]));
Chris@42: 		    {
Chris@42: 			 V Tt, Tp, TP, TQ;
Chris@42: 			 Tt = VBYI(VFNMS(LDK(KP781831482), Tr, VFNMS(LDK(KP433883739), Ts, VMUL(LDK(KP974927912), Tq))));
Chris@42: 			 Tp = VFMA(LDK(KP623489801), To, VFNMS(LDK(KP900968867), Th, VFNMS(LDK(KP222520933), Ta, T3)));
Chris@42: 			 TT = VSUB(Tp, Tt);
Chris@42: 			 STM2(&(xo[10]), TT, ovs, &(xo[2]));
Chris@42: 			 TU = VADD(Tp, Tt);
Chris@42: 			 STM2(&(xo[18]), TU, ovs, &(xo[2]));
Chris@42: 			 TP = VBYI(VFMA(LDK(KP974927912), TJ, VFMA(LDK(KP433883739), TL, VMUL(LDK(KP781831482), TK))));
Chris@42: 			 TQ = VFMA(LDK(KP623489801), TE, VFNMS(LDK(KP900968867), TH, VFNMS(LDK(KP222520933), TB, Ty)));
Chris@42: 			 TV = VADD(TP, TQ);
Chris@42: 			 STM2(&(xo[4]), TV, ovs, &(xo[0]));
Chris@42: 			 TW = VSUB(TQ, TP);
Chris@42: 			 STM2(&(xo[24]), TW, ovs, &(xo[0]));
Chris@42: 		    }
Chris@42: 		    {
Chris@42: 			 V Tv, Tu, TX, TY;
Chris@42: 			 Tv = VBYI(VFMA(LDK(KP781831482), Tq, VFMA(LDK(KP974927912), Ts, VMUL(LDK(KP433883739), Tr))));
Chris@42: 			 Tu = VFMA(LDK(KP623489801), Ta, VFNMS(LDK(KP900968867), To, VFNMS(LDK(KP222520933), Th, T3)));
Chris@42: 			 TX = VSUB(Tu, Tv);
Chris@42: 			 STM2(&(xo[26]), TX, ovs, &(xo[2]));
Chris@42: 			 STN2(&(xo[24]), TW, TX, ovs);
Chris@42: 			 TY = VADD(Tu, Tv);
Chris@42: 			 STM2(&(xo[2]), TY, ovs, &(xo[2]));
Chris@42: 			 STN2(&(xo[0]), TS, TY, ovs);
Chris@42: 		    }
Chris@42: 		    {
Chris@42: 			 V TM, TI, TZ, T10;
Chris@42: 			 TM = VBYI(VFNMS(LDK(KP433883739), TK, VFNMS(LDK(KP974927912), TL, VMUL(LDK(KP781831482), TJ))));
Chris@42: 			 TI = VFMA(LDK(KP623489801), TB, VFNMS(LDK(KP900968867), TE, VFNMS(LDK(KP222520933), TH, Ty)));
Chris@42: 			 TZ = VSUB(TI, TM);
Chris@42: 			 STM2(&(xo[12]), TZ, ovs, &(xo[0]));
Chris@42: 			 STN2(&(xo[12]), TZ, TR, ovs);
Chris@42: 			 T10 = VADD(TM, TI);
Chris@42: 			 STM2(&(xo[16]), T10, ovs, &(xo[0]));
Chris@42: 			 STN2(&(xo[16]), T10, TU, ovs);
Chris@42: 		    }
Chris@42: 		    {
Chris@42: 			 V T12, TO, TN, T11;
Chris@42: 			 TO = VBYI(VFMA(LDK(KP433883739), TJ, VFNMS(LDK(KP974927912), TK, VMUL(LDK(KP781831482), TL))));
Chris@42: 			 TN = VFMA(LDK(KP623489801), TH, VFNMS(LDK(KP222520933), TE, VFNMS(LDK(KP900968867), TB, Ty)));
Chris@42: 			 T11 = VSUB(TN, TO);
Chris@42: 			 STM2(&(xo[8]), T11, ovs, &(xo[0]));
Chris@42: 			 STN2(&(xo[8]), T11, TT, ovs);
Chris@42: 			 T12 = VADD(TO, TN);
Chris@42: 			 STM2(&(xo[20]), T12, ovs, &(xo[0]));
Chris@42: 			 {
Chris@42: 			      V Tx, Tw, T13, T14;
Chris@42: 			      Tx = VBYI(VFMA(LDK(KP433883739), Tq, VFNMS(LDK(KP781831482), Ts, VMUL(LDK(KP974927912), Tr))));
Chris@42: 			      Tw = VFMA(LDK(KP623489801), Th, VFNMS(LDK(KP222520933), To, VFNMS(LDK(KP900968867), Ta, T3)));
Chris@42: 			      T13 = VSUB(Tw, Tx);
Chris@42: 			      STM2(&(xo[22]), T13, ovs, &(xo[2]));
Chris@42: 			      STN2(&(xo[20]), T12, T13, ovs);
Chris@42: 			      T14 = VADD(Tw, Tx);
Chris@42: 			      STM2(&(xo[6]), T14, ovs, &(xo[2]));
Chris@42: 			      STN2(&(xo[4]), TV, T14, ovs);
Chris@42: 			 }
Chris@42: 		    }
Chris@42: 	       }
Chris@42: 	  }
Chris@42:      }
Chris@42:      VLEAVE();
Chris@42: }
Chris@42: 
Chris@42: static const kdft_desc desc = { 14, XSIMD_STRING("n2fv_14"), {50, 12, 24, 0}, &GENUS, 0, 2, 0, 0 };
Chris@42: 
Chris@42: void XSIMD(codelet_n2fv_14) (planner *p) {
Chris@42:      X(kdft_register) (p, n2fv_14, &desc);
Chris@42: }
Chris@42: 
Chris@42: #endif				/* HAVE_FMA */