comparison src/fftw-3.3.5/rdft/scalar/r2cf/hf_2.c @ 42:2cd0e3b3e1fd

Current fftw source

+/*
+ * Copyright (c) 2003, 2007-14 Matteo Frigo
+ * Copyright (c) 2003, 2007-14 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., 51 Franklin Street, Fifth Floor, Boston, MA  02110-1301  USA
+ *
+ */
+
+/* This file was automatically generated --- DO NOT EDIT */
+/* Generated on Sat Jul 30 16:46:14 EDT 2016 */
+
+#include "codelet-rdft.h"
+
+#ifdef HAVE_FMA
+
+/* Generated by: ../../../genfft/gen_hc2hc.native -fma -reorder-insns -schedule-for-pipeline -compact -variables 4 -pipeline-latency 4 -n 2 -dit -name hf_2 -include hf.h */
+
+/*
+ * This function contains 6 FP additions, 4 FP multiplications,
+ * (or, 4 additions, 2 multiplications, 2 fused multiply/add),
+ * 11 stack variables, 0 constants, and 8 memory accesses
+ */
+#include "hf.h"
+
+static void hf_2(R *cr, R *ci, const R *W, stride rs, INT mb, INT me, INT ms)
+{
+     {
+	  INT m;
+	  for (m = mb, W = W + ((mb - 1) * 2); m < me; m = m + 1, cr = cr + ms, ci = ci - ms, W = W + 2, MAKE_VOLATILE_STRIDE(4, rs)) {
+	       E T1, Ta, T3, T6, T2, T5;
+	       T1 = cr[0];
+	       Ta = ci[0];
+	       T3 = cr[WS(rs, 1)];
+	       T6 = ci[WS(rs, 1)];
+	       T2 = W[0];
+	       T5 = W[1];
+	       {
+		    E T8, T4, T9, T7;
+		    T8 = T2 * T6;
+		    T4 = T2 * T3;
+		    T9 = FNMS(T5, T3, T8);
+		    T7 = FMA(T5, T6, T4);
+		    ci[WS(rs, 1)] = T9 + Ta;
+		    cr[WS(rs, 1)] = T9 - Ta;
+		    cr[0] = T1 + T7;
+		    ci[0] = T1 - T7;
+	       }
+	  }
+     }
+}
+
+static const tw_instr twinstr[] = {
+     {TW_FULL, 1, 2},
+     {TW_NEXT, 1, 0}
+};
+
+static const hc2hc_desc desc = { 2, "hf_2", twinstr, &GENUS, {4, 2, 2, 0} };
+
+void X(codelet_hf_2) (planner *p) {
+     X(khc2hc_register) (p, hf_2, &desc);
+}
+#else				/* HAVE_FMA */
+
+/* Generated by: ../../../genfft/gen_hc2hc.native -compact -variables 4 -pipeline-latency 4 -n 2 -dit -name hf_2 -include hf.h */
+
+/*
+ * This function contains 6 FP additions, 4 FP multiplications,
+ * (or, 4 additions, 2 multiplications, 2 fused multiply/add),
+ * 9 stack variables, 0 constants, and 8 memory accesses
+ */
+#include "hf.h"
+
+static void hf_2(R *cr, R *ci, const R *W, stride rs, INT mb, INT me, INT ms)
+{
+     {
+	  INT m;
+	  for (m = mb, W = W + ((mb - 1) * 2); m < me; m = m + 1, cr = cr + ms, ci = ci - ms, W = W + 2, MAKE_VOLATILE_STRIDE(4, rs)) {
+	       E T1, T8, T6, T7;
+	       T1 = cr[0];
+	       T8 = ci[0];
+	       {
+		    E T3, T5, T2, T4;
+		    T3 = cr[WS(rs, 1)];
+		    T5 = ci[WS(rs, 1)];
+		    T2 = W[0];
+		    T4 = W[1];
+		    T6 = FMA(T2, T3, T4 * T5);
+		    T7 = FNMS(T4, T3, T2 * T5);
+	       }
+	       ci[0] = T1 - T6;
+	       cr[0] = T1 + T6;
+	       cr[WS(rs, 1)] = T7 - T8;
+	       ci[WS(rs, 1)] = T7 + T8;
+	  }
+     }
+}
+
+static const tw_instr twinstr[] = {
+     {TW_FULL, 1, 2},
+     {TW_NEXT, 1, 0}
+};
+
+static const hc2hc_desc desc = { 2, "hf_2", twinstr, &GENUS, {4, 2, 2, 0} };
+
+void X(codelet_hf_2) (planner *p) {
+     X(khc2hc_register) (p, hf_2, &desc);
+}
+#endif				/* HAVE_FMA */