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 07:40:13 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 7 -name r2cfII_7 -dft-II -include r2cfII.h */
Chris@10: 
Chris@10: /*
Chris@10:  * This function contains 24 FP additions, 18 FP multiplications,
Chris@10:  * (or, 9 additions, 3 multiplications, 15 fused multiply/add),
Chris@10:  * 25 stack variables, 6 constants, and 14 memory accesses
Chris@10:  */
Chris@10: #include "r2cfII.h"
Chris@10: 
Chris@10: static void r2cfII_7(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(KP900968867, +0.900968867902419126236102319507445051165919162);
Chris@10:      DK(KP692021471, +0.692021471630095869627814897002069140197260599);
Chris@10:      DK(KP801937735, +0.801937735804838252472204639014890102331838324);
Chris@10:      DK(KP974927912, +0.974927912181823607018131682993931217232785801);
Chris@10:      DK(KP554958132, +0.554958132087371191422194871006410481067288862);
Chris@10:      DK(KP356895867, +0.356895867892209443894399510021300583399127187);
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(28, rs), MAKE_VOLATILE_STRIDE(28, csr), MAKE_VOLATILE_STRIDE(28, csi)) {
Chris@10: 	       E Td, Tk;
Chris@10: 	       {
Chris@10: 		    E T4, T3, Te, T5, T9, Tf, T6, Tg, Tj;
Chris@10: 		    Td = R0[0];
Chris@10: 		    {
Chris@10: 			 E T1, T2, T7, T8;
Chris@10: 			 T1 = R0[WS(rs, 1)];
Chris@10: 			 T2 = R1[WS(rs, 2)];
Chris@10: 			 T7 = R1[WS(rs, 1)];
Chris@10: 			 T8 = R0[WS(rs, 2)];
Chris@10: 			 T4 = R1[0];
Chris@10: 			 T3 = T1 + T2;
Chris@10: 			 Te = T1 - T2;
Chris@10: 			 T5 = R0[WS(rs, 3)];
Chris@10: 			 T9 = T7 + T8;
Chris@10: 			 Tf = T8 - T7;
Chris@10: 		    }
Chris@10: 		    T6 = T4 + T5;
Chris@10: 		    Tg = T5 - T4;
Chris@10: 		    Tj = FNMS(KP356895867, Tf, Te);
Chris@10: 		    {
Chris@10: 			 E Ta, Th, Tl, Tb, Ti, Tm, Tc;
Chris@10: 			 Tb = FNMS(KP554958132, T3, T9);
Chris@10: 			 Ta = FMA(KP554958132, T9, T6);
Chris@10: 			 Th = FNMS(KP356895867, Tg, Tf);
Chris@10: 			 Tl = FNMS(KP356895867, Te, Tg);
Chris@10: 			 Ci[WS(csi, 1)] = -(KP974927912 * (FNMS(KP801937735, Tb, T6)));
Chris@10: 			 Ci[WS(csi, 2)] = KP974927912 * (FNMS(KP801937735, Ta, T3));
Chris@10: 			 Ti = FNMS(KP692021471, Th, Te);
Chris@10: 			 Tm = FNMS(KP692021471, Tl, Tf);
Chris@10: 			 Cr[WS(csr, 3)] = Te + Tg + Tf + Td;
Chris@10: 			 Tc = FMA(KP554958132, T6, T3);
Chris@10: 			 Cr[WS(csr, 1)] = FNMS(KP900968867, Ti, Td);
Chris@10: 			 Cr[WS(csr, 2)] = FNMS(KP900968867, Tm, Td);
Chris@10: 			 Tk = FNMS(KP692021471, Tj, Tg);
Chris@10: 			 Ci[0] = -(KP974927912 * (FMA(KP801937735, Tc, T9)));
Chris@10: 		    }
Chris@10: 	       }
Chris@10: 	       Cr[0] = FNMS(KP900968867, Tk, Td);
Chris@10: 	  }
Chris@10:      }
Chris@10: }
Chris@10: 
Chris@10: static const kr2c_desc desc = { 7, "r2cfII_7", {9, 3, 15, 0}, &GENUS };
Chris@10: 
Chris@10: void X(codelet_r2cfII_7) (planner *p) {
Chris@10:      X(kr2c_register) (p, r2cfII_7, &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 7 -name r2cfII_7 -dft-II -include r2cfII.h */
Chris@10: 
Chris@10: /*
Chris@10:  * This function contains 24 FP additions, 18 FP multiplications,
Chris@10:  * (or, 12 additions, 6 multiplications, 12 fused multiply/add),
Chris@10:  * 20 stack variables, 6 constants, and 14 memory accesses
Chris@10:  */
Chris@10: #include "r2cfII.h"
Chris@10: 
Chris@10: static void r2cfII_7(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(KP900968867, +0.900968867902419126236102319507445051165919162);
Chris@10:      DK(KP222520933, +0.222520933956314404288902564496794759466355569);
Chris@10:      DK(KP623489801, +0.623489801858733530525004884004239810632274731);
Chris@10:      DK(KP433883739, +0.433883739117558120475768332848358754609990728);
Chris@10:      DK(KP974927912, +0.974927912181823607018131682993931217232785801);
Chris@10:      DK(KP781831482, +0.781831482468029808708444526674057750232334519);
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(28, rs), MAKE_VOLATILE_STRIDE(28, csr), MAKE_VOLATILE_STRIDE(28, csi)) {
Chris@10: 	       E T1, Ta, Td, T4, Tb, T7, Tc, T8, T9;
Chris@10: 	       T1 = R0[0];
Chris@10: 	       T8 = R1[0];
Chris@10: 	       T9 = R0[WS(rs, 3)];
Chris@10: 	       Ta = T8 - T9;
Chris@10: 	       Td = T8 + T9;
Chris@10: 	       {
Chris@10: 		    E T2, T3, T5, T6;
Chris@10: 		    T2 = R0[WS(rs, 1)];
Chris@10: 		    T3 = R1[WS(rs, 2)];
Chris@10: 		    T4 = T2 - T3;
Chris@10: 		    Tb = T2 + T3;
Chris@10: 		    T5 = R1[WS(rs, 1)];
Chris@10: 		    T6 = R0[WS(rs, 2)];
Chris@10: 		    T7 = T5 - T6;
Chris@10: 		    Tc = T5 + T6;
Chris@10: 	       }
Chris@10: 	       Ci[0] = -(FMA(KP781831482, Tb, KP974927912 * Tc) + (KP433883739 * Td));
Chris@10: 	       Ci[WS(csi, 1)] = FNMS(KP974927912, Td, KP781831482 * Tc) - (KP433883739 * Tb);
Chris@10: 	       Cr[0] = FMA(KP623489801, T4, T1) + FMA(KP222520933, T7, KP900968867 * Ta);
Chris@10: 	       Ci[WS(csi, 2)] = FNMS(KP781831482, Td, KP974927912 * Tb) - (KP433883739 * Tc);
Chris@10: 	       Cr[WS(csr, 2)] = FMA(KP900968867, T7, T1) + FNMA(KP623489801, Ta, KP222520933 * T4);
Chris@10: 	       Cr[WS(csr, 1)] = FMA(KP222520933, Ta, T1) + FNMA(KP623489801, T7, KP900968867 * T4);
Chris@10: 	       Cr[WS(csr, 3)] = T1 + T4 - (T7 + Ta);
Chris@10: 	  }
Chris@10:      }
Chris@10: }
Chris@10: 
Chris@10: static const kr2c_desc desc = { 7, "r2cfII_7", {12, 6, 12, 0}, &GENUS };
Chris@10: 
Chris@10: void X(codelet_r2cfII_7) (planner *p) {
Chris@10:      X(kr2c_register) (p, r2cfII_7, &desc);
Chris@10: }
Chris@10: 
Chris@10: #endif				/* HAVE_FMA */