First commit - moving from LuaForge to SourceForge

1 files changed, 168 insertions, 0 deletions

diff --git a/src/fftw3/dft/codelets/inplace/q1_3.c b/src/fftw3/dft/codelets/inplace/q1_3.c
new file mode 100644
index 0000000..a05996f
--- /dev/null
+++ b/src/fftw3/dft/codelets/inplace/q1_3.c
@@ -0,0 +1,168 @@
+/*
+ * Copyright (c) 2003 Matteo Frigo
+ * Copyright (c) 2003 Massachusetts Institute of Technology
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
+ *
+ */
+
+/* This file was automatically generated --- DO NOT EDIT */
+/* Generated on Sat Jul  5 21:39:14 EDT 2003 */
+
+#include "codelet-dft.h"
+
+/* Generated by: /homee/stevenj/cvs/fftw3.0.1/genfft/gen_twidsq -compact -variables 4 -reload-twiddle -dif -n 3 -name q1_3 -include q.h */
+
+/*
+ * This function contains 48 FP additions, 36 FP multiplications,
+ * (or, 30 additions, 18 multiplications, 18 fused multiply/add),
+ * 35 stack variables, and 36 memory accesses
+ */
+/*
+ * Generator Id's : 
+ * $Id: q1_3.c,v 1.1 2008/10/17 06:11:08 scuri Exp $
+ * $Id: q1_3.c,v 1.1 2008/10/17 06:11:08 scuri Exp $
+ * $Id: q1_3.c,v 1.1 2008/10/17 06:11:08 scuri Exp $
+ */
+
+#include "q.h"
+
+static const R *q1_3(R *rio, R *iio, const R *W, stride is, stride vs, int m, int dist)
+{
+     DK(KP866025403, +0.866025403784438646763723170752936183471402627);
+     DK(KP500000000, +0.500000000000000000000000000000000000000000000);
+     int i;
+     for (i = m; i > 0; i = i - 1, rio = rio + dist, iio = iio + dist, W = W + 4) {
+	  E T1, T4, T6, Tc, Td, Te, T9, Tf, Tl, To, Tq, Tw, Tx, Ty, Tt;
+	  E Tz, TR, TS, TN, TT, TF, TI, TK, TQ;
+	  {
+	       E T2, T3, Tr, Ts;
+	       T1 = rio[0];
+	       T2 = rio[WS(is, 1)];
+	       T3 = rio[WS(is, 2)];
+	       T4 = T2 + T3;
+	       T6 = FNMS(KP500000000, T4, T1);
+	       Tc = KP866025403 * (T3 - T2);
+	       {
+		    E T7, T8, Tm, Tn;
+		    Td = iio[0];
+		    T7 = iio[WS(is, 1)];
+		    T8 = iio[WS(is, 2)];
+		    Te = T7 + T8;
+		    T9 = KP866025403 * (T7 - T8);
+		    Tf = FNMS(KP500000000, Te, Td);
+		    Tl = rio[WS(vs, 1)];
+		    Tm = rio[WS(vs, 1) + WS(is, 1)];
+		    Tn = rio[WS(vs, 1) + WS(is, 2)];
+		    To = Tm + Tn;
+		    Tq = FNMS(KP500000000, To, Tl);
+		    Tw = KP866025403 * (Tn - Tm);
+	       }
+	       Tx = iio[WS(vs, 1)];
+	       Tr = iio[WS(vs, 1) + WS(is, 1)];
+	       Ts = iio[WS(vs, 1) + WS(is, 2)];
+	       Ty = Tr + Ts;
+	       Tt = KP866025403 * (Tr - Ts);
+	       Tz = FNMS(KP500000000, Ty, Tx);
+	       {
+		    E TL, TM, TG, TH;
+		    TR = iio[WS(vs, 2)];
+		    TL = iio[WS(vs, 2) + WS(is, 1)];
+		    TM = iio[WS(vs, 2) + WS(is, 2)];
+		    TS = TL + TM;
+		    TN = KP866025403 * (TL - TM);
+		    TT = FNMS(KP500000000, TS, TR);
+		    TF = rio[WS(vs, 2)];
+		    TG = rio[WS(vs, 2) + WS(is, 1)];
+		    TH = rio[WS(vs, 2) + WS(is, 2)];
+		    TI = TG + TH;
+		    TK = FNMS(KP500000000, TI, TF);
+		    TQ = KP866025403 * (TH - TG);
+	       }
+	  }
+	  rio[0] = T1 + T4;
+	  iio[0] = Td + Te;
+	  rio[WS(is, 1)] = Tl + To;
+	  iio[WS(is, 1)] = Tx + Ty;
+	  iio[WS(is, 2)] = TR + TS;
+	  rio[WS(is, 2)] = TF + TI;
+	  {
+	       E Ta, Tg, T5, Tb;
+	       Ta = T6 + T9;
+	       Tg = Tc + Tf;
+	       T5 = W[0];
+	       Tb = W[1];
+	       rio[WS(vs, 1)] = FMA(T5, Ta, Tb * Tg);
+	       iio[WS(vs, 1)] = FNMS(Tb, Ta, T5 * Tg);
+	  }
+	  {
+	       E TW, TY, TV, TX;
+	       TW = TK - TN;
+	       TY = TT - TQ;
+	       TV = W[2];
+	       TX = W[3];
+	       rio[WS(vs, 2) + WS(is, 2)] = FMA(TV, TW, TX * TY);
+	       iio[WS(vs, 2) + WS(is, 2)] = FNMS(TX, TW, TV * TY);
+	  }
+	  {
+	       E TC, TE, TB, TD;
+	       TC = Tq - Tt;
+	       TE = Tz - Tw;
+	       TB = W[2];
+	       TD = W[3];
+	       rio[WS(vs, 2) + WS(is, 1)] = FMA(TB, TC, TD * TE);
+	       iio[WS(vs, 2) + WS(is, 1)] = FNMS(TD, TC, TB * TE);
+	  }
+	  {
+	       E Tu, TA, Tp, Tv;
+	       Tu = Tq + Tt;
+	       TA = Tw + Tz;
+	       Tp = W[0];
+	       Tv = W[1];
+	       rio[WS(vs, 1) + WS(is, 1)] = FMA(Tp, Tu, Tv * TA);
+	       iio[WS(vs, 1) + WS(is, 1)] = FNMS(Tv, Tu, Tp * TA);
+	  }
+	  {
+	       E TO, TU, TJ, TP;
+	       TO = TK + TN;
+	       TU = TQ + TT;
+	       TJ = W[0];
+	       TP = W[1];
+	       rio[WS(vs, 1) + WS(is, 2)] = FMA(TJ, TO, TP * TU);
+	       iio[WS(vs, 1) + WS(is, 2)] = FNMS(TP, TO, TJ * TU);
+	  }
+	  {
+	       E Ti, Tk, Th, Tj;
+	       Ti = T6 - T9;
+	       Tk = Tf - Tc;
+	       Th = W[2];
+	       Tj = W[3];
+	       rio[WS(vs, 2)] = FMA(Th, Ti, Tj * Tk);
+	       iio[WS(vs, 2)] = FNMS(Tj, Ti, Th * Tk);
+	  }
+     }
+     return W;
+}
+
+static const tw_instr twinstr[] = {
+     {TW_FULL, 0, 3},
+     {TW_NEXT, 1, 0}
+};
+
+static const ct_desc desc = { 3, "q1_3", twinstr, {30, 18, 18, 0}, &GENUS, 0, 0, 0 };
+
+void X(codelet_q1_3) (planner *p) {
+     X(kdft_difsq_register) (p, q1_3, &desc);
+}