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 5 -name r2cfII_5 -dft-II -include r2cfII.h */
Chris@10: 
Chris@10: /*
Chris@10:  * This function contains 12 FP additions, 7 FP multiplications,
Chris@10:  * (or, 7 additions, 2 multiplications, 5 fused multiply/add),
Chris@10:  * 17 stack variables, 4 constants, and 10 memory accesses
Chris@10:  */
Chris@10: #include "r2cfII.h"
Chris@10: 
Chris@10: static void r2cfII_5(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(KP559016994, +0.559016994374947424102293417182819058860154590);
Chris@10:      DK(KP250000000, +0.250000000000000000000000000000000000000000000);
Chris@10:      DK(KP618033988, +0.618033988749894848204586834365638117720309180);
Chris@10:      DK(KP951056516, +0.951056516295153572116439333379382143405698634);
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(20, rs), MAKE_VOLATILE_STRIDE(20, csr), MAKE_VOLATILE_STRIDE(20, csi)) {
Chris@10: 	       E T1, T2, T3, T5, T6;
Chris@10: 	       T1 = R0[0];
Chris@10: 	       T2 = R0[WS(rs, 1)];
Chris@10: 	       T3 = R1[WS(rs, 1)];
Chris@10: 	       T5 = R0[WS(rs, 2)];
Chris@10: 	       T6 = R1[0];
Chris@10: 	       {
Chris@10: 		    E Tb, T4, Tc, T7, Ta, T8, T9;
Chris@10: 		    Tb = T2 + T3;
Chris@10: 		    T4 = T2 - T3;
Chris@10: 		    Tc = T5 + T6;
Chris@10: 		    T7 = T5 - T6;
Chris@10: 		    Ci[0] = -(KP951056516 * (FMA(KP618033988, Tc, Tb)));
Chris@10: 		    Ci[WS(csi, 1)] = -(KP951056516 * (FNMS(KP618033988, Tb, Tc)));
Chris@10: 		    Ta = T4 - T7;
Chris@10: 		    T8 = T4 + T7;
Chris@10: 		    T9 = FNMS(KP250000000, T8, T1);
Chris@10: 		    Cr[WS(csr, 2)] = T1 + T8;
Chris@10: 		    Cr[WS(csr, 1)] = FNMS(KP559016994, Ta, T9);
Chris@10: 		    Cr[0] = FMA(KP559016994, Ta, T9);
Chris@10: 	       }
Chris@10: 	  }
Chris@10:      }
Chris@10: }
Chris@10: 
Chris@10: static const kr2c_desc desc = { 5, "r2cfII_5", {7, 2, 5, 0}, &GENUS };
Chris@10: 
Chris@10: void X(codelet_r2cfII_5) (planner *p) {
Chris@10:      X(kr2c_register) (p, r2cfII_5, &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 5 -name r2cfII_5 -dft-II -include r2cfII.h */
Chris@10: 
Chris@10: /*
Chris@10:  * This function contains 12 FP additions, 6 FP multiplications,
Chris@10:  * (or, 9 additions, 3 multiplications, 3 fused multiply/add),
Chris@10:  * 17 stack variables, 4 constants, and 10 memory accesses
Chris@10:  */
Chris@10: #include "r2cfII.h"
Chris@10: 
Chris@10: static void r2cfII_5(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(KP250000000, +0.250000000000000000000000000000000000000000000);
Chris@10:      DK(KP559016994, +0.559016994374947424102293417182819058860154590);
Chris@10:      DK(KP951056516, +0.951056516295153572116439333379382143405698634);
Chris@10:      DK(KP587785252, +0.587785252292473129168705954639072768597652438);
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(20, rs), MAKE_VOLATILE_STRIDE(20, csr), MAKE_VOLATILE_STRIDE(20, csi)) {
Chris@10: 	       E T8, T3, T6, T9, Tc, Tb, T7, Ta;
Chris@10: 	       T8 = R0[0];
Chris@10: 	       {
Chris@10: 		    E T1, T2, T4, T5;
Chris@10: 		    T1 = R0[WS(rs, 1)];
Chris@10: 		    T2 = R1[WS(rs, 1)];
Chris@10: 		    T3 = T1 - T2;
Chris@10: 		    T4 = R0[WS(rs, 2)];
Chris@10: 		    T5 = R1[0];
Chris@10: 		    T6 = T4 - T5;
Chris@10: 		    T9 = T3 + T6;
Chris@10: 		    Tc = T4 + T5;
Chris@10: 		    Tb = T1 + T2;
Chris@10: 	       }
Chris@10: 	       Cr[WS(csr, 2)] = T8 + T9;
Chris@10: 	       Ci[WS(csi, 1)] = FNMS(KP951056516, Tc, KP587785252 * Tb);
Chris@10: 	       Ci[0] = -(FMA(KP951056516, Tb, KP587785252 * Tc));
Chris@10: 	       T7 = KP559016994 * (T3 - T6);
Chris@10: 	       Ta = FNMS(KP250000000, T9, T8);
Chris@10: 	       Cr[0] = T7 + Ta;
Chris@10: 	       Cr[WS(csr, 1)] = Ta - T7;
Chris@10: 	  }
Chris@10:      }
Chris@10: }
Chris@10: 
Chris@10: static const kr2c_desc desc = { 5, "r2cfII_5", {9, 3, 3, 0}, &GENUS };
Chris@10: 
Chris@10: void X(codelet_r2cfII_5) (planner *p) {
Chris@10:      X(kr2c_register) (p, r2cfII_5, &desc);
Chris@10: }
Chris@10: 
Chris@10: #endif				/* HAVE_FMA */