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 07:39:45 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 11 -name r2cf_11 -include r2cf.h */
Chris@10: 
Chris@10: /*
Chris@10:  * This function contains 60 FP additions, 50 FP multiplications,
Chris@10:  * (or, 15 additions, 5 multiplications, 45 fused multiply/add),
Chris@10:  * 51 stack variables, 10 constants, and 22 memory accesses
Chris@10:  */
Chris@10: #include "r2cf.h"
Chris@10: 
Chris@10: static void r2cf_11(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(KP959492973, +0.959492973614497389890368057066327699062454848);
Chris@10:      DK(KP876768831, +0.876768831002589333891339807079336796764054852);
Chris@10:      DK(KP918985947, +0.918985947228994779780736114132655398124909697);
Chris@10:      DK(KP989821441, +0.989821441880932732376092037776718787376519372);
Chris@10:      DK(KP778434453, +0.778434453334651800608337670740821884709317477);
Chris@10:      DK(KP830830026, +0.830830026003772851058548298459246407048009821);
Chris@10:      DK(KP715370323, +0.715370323453429719112414662767260662417897278);
Chris@10:      DK(KP634356270, +0.634356270682424498893150776899916060542806975);
Chris@10:      DK(KP342584725, +0.342584725681637509502641509861112333758894680);
Chris@10:      DK(KP521108558, +0.521108558113202722944698153526659300680427422);
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(44, rs), MAKE_VOLATILE_STRIDE(44, csr), MAKE_VOLATILE_STRIDE(44, csi)) {
Chris@10: 	       E T1, Tg, TF, TB, TI, TL, Tz, TA;
Chris@10: 	       {
Chris@10: 		    E T4, TC, TE, T7, TD, Ta, TS, TG, TJ, Td, TP, TM, Ty, Tq, Th;
Chris@10: 		    E Tt, Tl;
Chris@10: 		    T1 = R0[0];
Chris@10: 		    {
Chris@10: 			 E Tb, Tc, Tx, Tp;
Chris@10: 			 {
Chris@10: 			      E T2, T3, Te, Tf;
Chris@10: 			      T2 = R1[0];
Chris@10: 			      T3 = R0[WS(rs, 5)];
Chris@10: 			      Te = R1[WS(rs, 2)];
Chris@10: 			      Tf = R0[WS(rs, 3)];
Chris@10: 			      {
Chris@10: 				   E T5, T6, T8, T9;
Chris@10: 				   T5 = R0[WS(rs, 1)];
Chris@10: 				   T4 = T2 + T3;
Chris@10: 				   TC = T3 - T2;
Chris@10: 				   Tg = Te + Tf;
Chris@10: 				   TE = Tf - Te;
Chris@10: 				   T6 = R1[WS(rs, 4)];
Chris@10: 				   T8 = R1[WS(rs, 1)];
Chris@10: 				   T9 = R0[WS(rs, 4)];
Chris@10: 				   Tb = R0[WS(rs, 2)];
Chris@10: 				   T7 = T5 + T6;
Chris@10: 				   TD = T5 - T6;
Chris@10: 				   Ta = T8 + T9;
Chris@10: 				   TF = T9 - T8;
Chris@10: 				   Tc = R1[WS(rs, 3)];
Chris@10: 			      }
Chris@10: 			 }
Chris@10: 			 TS = FMA(KP521108558, TC, TD);
Chris@10: 			 TG = FMA(KP521108558, TF, TE);
Chris@10: 			 TJ = FMA(KP521108558, TE, TC);
Chris@10: 			 Td = Tb + Tc;
Chris@10: 			 TB = Tb - Tc;
Chris@10: 			 Tx = FNMS(KP342584725, Ta, T7);
Chris@10: 			 Tp = FNMS(KP342584725, T4, Ta);
Chris@10: 			 TP = FNMS(KP521108558, TB, TF);
Chris@10: 			 TM = FNMS(KP521108558, TD, TB);
Chris@10: 			 Ty = FNMS(KP634356270, Tx, Td);
Chris@10: 			 Tq = FNMS(KP634356270, Tp, Tg);
Chris@10: 			 Th = FNMS(KP342584725, Tg, Td);
Chris@10: 			 Tt = FNMS(KP342584725, Td, T4);
Chris@10: 			 Tl = FNMS(KP342584725, T7, Tg);
Chris@10: 		    }
Chris@10: 		    {
Chris@10: 			 E Tu, Ts, TN, Tv;
Chris@10: 			 {
Chris@10: 			      E Tm, TU, Tj, Ti, TT;
Chris@10: 			      TT = FMA(KP715370323, TS, TF);
Chris@10: 			      Ti = FNMS(KP634356270, Th, Ta);
Chris@10: 			      Tu = FNMS(KP634356270, Tt, T7);
Chris@10: 			      Tm = FNMS(KP634356270, Tl, T4);
Chris@10: 			      TU = FMA(KP830830026, TT, TB);
Chris@10: 			      Tj = FNMS(KP778434453, Ti, T7);
Chris@10: 			      {
Chris@10: 				   E Tk, TR, To, Tn, TQ, Tr;
Chris@10: 				   TQ = FMA(KP715370323, TP, TC);
Chris@10: 				   Tn = FNMS(KP778434453, Tm, Ta);
Chris@10: 				   Ci[WS(csi, 5)] = KP989821441 * (FMA(KP918985947, TU, TE));
Chris@10: 				   Tk = FNMS(KP876768831, Tj, T4);
Chris@10: 				   TR = FNMS(KP830830026, TQ, TE);
Chris@10: 				   To = FNMS(KP876768831, Tn, Td);
Chris@10: 				   Tr = FNMS(KP778434453, Tq, Td);
Chris@10: 				   Cr[WS(csr, 5)] = FNMS(KP959492973, Tk, T1);
Chris@10: 				   Ci[WS(csi, 4)] = KP989821441 * (FNMS(KP918985947, TR, TD));
Chris@10: 				   Cr[WS(csr, 4)] = FNMS(KP959492973, To, T1);
Chris@10: 				   Ts = FNMS(KP876768831, Tr, T7);
Chris@10: 			      }
Chris@10: 			 }
Chris@10: 			 TN = FNMS(KP715370323, TM, TE);
Chris@10: 			 Tv = FNMS(KP778434453, Tu, Tg);
Chris@10: 			 Cr[0] = T1 + T4 + T7 + Ta + Td + Tg;
Chris@10: 			 Cr[WS(csr, 3)] = FNMS(KP959492973, Ts, T1);
Chris@10: 			 {
Chris@10: 			      E TO, Tw, TH, TK;
Chris@10: 			      TO = FNMS(KP830830026, TN, TF);
Chris@10: 			      Tw = FNMS(KP876768831, Tv, Ta);
Chris@10: 			      TH = FMA(KP715370323, TG, TD);
Chris@10: 			      TK = FNMS(KP715370323, TJ, TB);
Chris@10: 			      Ci[WS(csi, 3)] = KP989821441 * (FNMS(KP918985947, TO, TC));
Chris@10: 			      Cr[WS(csr, 2)] = FNMS(KP959492973, Tw, T1);
Chris@10: 			      TI = FNMS(KP830830026, TH, TC);
Chris@10: 			      TL = FMA(KP830830026, TK, TD);
Chris@10: 			      Tz = FNMS(KP778434453, Ty, T4);
Chris@10: 			 }
Chris@10: 		    }
Chris@10: 	       }
Chris@10: 	       Ci[WS(csi, 2)] = KP989821441 * (FMA(KP918985947, TI, TB));
Chris@10: 	       Ci[WS(csi, 1)] = KP989821441 * (FNMS(KP918985947, TL, TF));
Chris@10: 	       TA = FNMS(KP876768831, Tz, Tg);
Chris@10: 	       Cr[WS(csr, 1)] = FNMS(KP959492973, TA, T1);
Chris@10: 	  }
Chris@10:      }
Chris@10: }
Chris@10: 
Chris@10: static const kr2c_desc desc = { 11, "r2cf_11", {15, 5, 45, 0}, &GENUS };
Chris@10: 
Chris@10: void X(codelet_r2cf_11) (planner *p) {
Chris@10:      X(kr2c_register) (p, r2cf_11, &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 11 -name r2cf_11 -include r2cf.h */
Chris@10: 
Chris@10: /*
Chris@10:  * This function contains 60 FP additions, 50 FP multiplications,
Chris@10:  * (or, 20 additions, 10 multiplications, 40 fused multiply/add),
Chris@10:  * 28 stack variables, 10 constants, and 22 memory accesses
Chris@10:  */
Chris@10: #include "r2cf.h"
Chris@10: 
Chris@10: static void r2cf_11(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(KP654860733, +0.654860733945285064056925072466293553183791199);
Chris@10:      DK(KP142314838, +0.142314838273285140443792668616369668791051361);
Chris@10:      DK(KP959492973, +0.959492973614497389890368057066327699062454848);
Chris@10:      DK(KP415415013, +0.415415013001886425529274149229623203524004910);
Chris@10:      DK(KP841253532, +0.841253532831181168861811648919367717513292498);
Chris@10:      DK(KP989821441, +0.989821441880932732376092037776718787376519372);
Chris@10:      DK(KP909631995, +0.909631995354518371411715383079028460060241051);
Chris@10:      DK(KP281732556, +0.281732556841429697711417915346616899035777899);
Chris@10:      DK(KP540640817, +0.540640817455597582107635954318691695431770608);
Chris@10:      DK(KP755749574, +0.755749574354258283774035843972344420179717445);
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(44, rs), MAKE_VOLATILE_STRIDE(44, csr), MAKE_VOLATILE_STRIDE(44, csi)) {
Chris@10: 	       E T1, T4, Tl, Tg, Th, Td, Ti, Ta, Tk, T7, Tj, Tb, Tc;
Chris@10: 	       T1 = R0[0];
Chris@10: 	       {
Chris@10: 		    E T2, T3, Te, Tf;
Chris@10: 		    T2 = R0[WS(rs, 1)];
Chris@10: 		    T3 = R1[WS(rs, 4)];
Chris@10: 		    T4 = T2 + T3;
Chris@10: 		    Tl = T3 - T2;
Chris@10: 		    Te = R1[0];
Chris@10: 		    Tf = R0[WS(rs, 5)];
Chris@10: 		    Tg = Te + Tf;
Chris@10: 		    Th = Tf - Te;
Chris@10: 	       }
Chris@10: 	       Tb = R1[WS(rs, 1)];
Chris@10: 	       Tc = R0[WS(rs, 4)];
Chris@10: 	       Td = Tb + Tc;
Chris@10: 	       Ti = Tc - Tb;
Chris@10: 	       {
Chris@10: 		    E T8, T9, T5, T6;
Chris@10: 		    T8 = R1[WS(rs, 2)];
Chris@10: 		    T9 = R0[WS(rs, 3)];
Chris@10: 		    Ta = T8 + T9;
Chris@10: 		    Tk = T9 - T8;
Chris@10: 		    T5 = R0[WS(rs, 2)];
Chris@10: 		    T6 = R1[WS(rs, 3)];
Chris@10: 		    T7 = T5 + T6;
Chris@10: 		    Tj = T6 - T5;
Chris@10: 	       }
Chris@10: 	       Ci[WS(csi, 4)] = FMA(KP755749574, Th, KP540640817 * Ti) + FNMS(KP909631995, Tk, KP281732556 * Tj) - (KP989821441 * Tl);
Chris@10: 	       Cr[WS(csr, 4)] = FMA(KP841253532, Td, T1) + FNMS(KP959492973, T7, KP415415013 * Ta) + FNMA(KP142314838, T4, KP654860733 * Tg);
Chris@10: 	       Ci[WS(csi, 2)] = FMA(KP909631995, Th, KP755749574 * Tl) + FNMA(KP540640817, Tk, KP989821441 * Tj) - (KP281732556 * Ti);
Chris@10: 	       Ci[WS(csi, 5)] = FMA(KP281732556, Th, KP755749574 * Ti) + FNMS(KP909631995, Tj, KP989821441 * Tk) - (KP540640817 * Tl);
Chris@10: 	       Ci[WS(csi, 1)] = FMA(KP540640817, Th, KP909631995 * Tl) + FMA(KP989821441, Ti, KP755749574 * Tj) + (KP281732556 * Tk);
Chris@10: 	       Ci[WS(csi, 3)] = FMA(KP989821441, Th, KP540640817 * Tj) + FNMS(KP909631995, Ti, KP755749574 * Tk) - (KP281732556 * Tl);
Chris@10: 	       Cr[WS(csr, 3)] = FMA(KP415415013, Td, T1) + FNMS(KP654860733, Ta, KP841253532 * T7) + FNMA(KP959492973, T4, KP142314838 * Tg);
Chris@10: 	       Cr[WS(csr, 1)] = FMA(KP841253532, Tg, T1) + FNMS(KP959492973, Ta, KP415415013 * T4) + FNMA(KP654860733, T7, KP142314838 * Td);
Chris@10: 	       Cr[0] = T1 + Tg + T4 + Td + T7 + Ta;
Chris@10: 	       Cr[WS(csr, 2)] = FMA(KP415415013, Tg, T1) + FNMS(KP142314838, T7, KP841253532 * Ta) + FNMA(KP959492973, Td, KP654860733 * T4);
Chris@10: 	       Cr[WS(csr, 5)] = FMA(KP841253532, T4, T1) + FNMS(KP142314838, Ta, KP415415013 * T7) + FNMA(KP654860733, Td, KP959492973 * Tg);
Chris@10: 	  }
Chris@10:      }
Chris@10: }
Chris@10: 
Chris@10: static const kr2c_desc desc = { 11, "r2cf_11", {20, 10, 40, 0}, &GENUS };
Chris@10: 
Chris@10: void X(codelet_r2cf_11) (planner *p) {
Chris@10:      X(kr2c_register) (p, r2cf_11, &desc);
Chris@10: }
Chris@10: 
Chris@10: #endif				/* HAVE_FMA */