Mercurial > hg > sv-dependency-builds

diff src/fftw-3.3.3/dft/simd/common/t1bv_10.c @ 10:37bf6b4a2645

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
Add FFTW3
author Chris Cannam
date Wed, 20 Mar 2013 15:35:50 +0000
parents
children
line wrap: on
line diff
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/src/fftw-3.3.3/dft/simd/common/t1bv_10.c	Wed Mar 20 15:35:50 2013 +0000
@@ -0,0 +1,280 @@
+/*
+ * Copyright (c) 2003, 2007-11 Matteo Frigo
+ * Copyright (c) 2003, 2007-11 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 Sun Nov 25 07:39:04 EST 2012 */
+
+#include "codelet-dft.h"
+
+#ifdef HAVE_FMA
+
+/* Generated by: ../../../genfft/gen_twiddle_c.native -fma -reorder-insns -schedule-for-pipeline -simd -compact -variables 4 -pipeline-latency 8 -n 10 -name t1bv_10 -include t1b.h -sign 1 */
+
+/*
+ * This function contains 51 FP additions, 40 FP multiplications,
+ * (or, 33 additions, 22 multiplications, 18 fused multiply/add),
+ * 43 stack variables, 4 constants, and 20 memory accesses
+ */
+#include "t1b.h"
+
+static void t1bv_10(R *ri, R *ii, const R *W, stride rs, INT mb, INT me, INT ms)
+{
+     DVK(KP559016994, +0.559016994374947424102293417182819058860154590);
+     DVK(KP250000000, +0.250000000000000000000000000000000000000000000);
+     DVK(KP618033988, +0.618033988749894848204586834365638117720309180);
+     DVK(KP951056516, +0.951056516295153572116439333379382143405698634);
+     {
+	  INT m;
+	  R *x;
+	  x = ii;
+	  for (m = mb, W = W + (mb * ((TWVL / VL) * 18)); m < me; m = m + VL, x = x + (VL * ms), W = W + (TWVL * 18), MAKE_VOLATILE_STRIDE(10, rs)) {
+	       V Td, TA, T4, Ta, Tk, TE, Tp, TF, TB, T9, T1, T2, Tb;
+	       T1 = LD(&(x[0]), ms, &(x[0]));
+	       T2 = LD(&(x[WS(rs, 5)]), ms, &(x[WS(rs, 1)]));
+	       {
+		    V Tg, Tn, Ti, Tl;
+		    Tg = LD(&(x[WS(rs, 4)]), ms, &(x[0]));
+		    Tn = LD(&(x[WS(rs, 1)]), ms, &(x[WS(rs, 1)]));
+		    Ti = LD(&(x[WS(rs, 9)]), ms, &(x[WS(rs, 1)]));
+		    Tl = LD(&(x[WS(rs, 6)]), ms, &(x[0]));
+		    {
+			 V T6, T8, T5, Tc;
+			 T5 = LD(&(x[WS(rs, 2)]), ms, &(x[0]));
+			 Tc = LD(&(x[WS(rs, 3)]), ms, &(x[WS(rs, 1)]));
+			 {
+			      V T3, Th, To, Tj, Tm, T7;
+			      T7 = LD(&(x[WS(rs, 7)]), ms, &(x[WS(rs, 1)]));
+			      T3 = BYTW(&(W[TWVL * 8]), T2);
+			      Th = BYTW(&(W[TWVL * 6]), Tg);
+			      To = BYTW(&(W[0]), Tn);
+			      Tj = BYTW(&(W[TWVL * 16]), Ti);
+			      Tm = BYTW(&(W[TWVL * 10]), Tl);
+			      T6 = BYTW(&(W[TWVL * 2]), T5);
+			      Td = BYTW(&(W[TWVL * 4]), Tc);
+			      T8 = BYTW(&(W[TWVL * 12]), T7);
+			      TA = VADD(T1, T3);
+			      T4 = VSUB(T1, T3);
+			      Ta = LD(&(x[WS(rs, 8)]), ms, &(x[0]));
+			      Tk = VSUB(Th, Tj);
+			      TE = VADD(Th, Tj);
+			      Tp = VSUB(Tm, To);
+			      TF = VADD(Tm, To);
+			 }
+			 TB = VADD(T6, T8);
+			 T9 = VSUB(T6, T8);
+		    }
+	       }
+	       Tb = BYTW(&(W[TWVL * 14]), Ta);
+	       {
+		    V TL, TG, Tw, Tq, TC, Te;
+		    TL = VSUB(TE, TF);
+		    TG = VADD(TE, TF);
+		    Tw = VSUB(Tk, Tp);
+		    Tq = VADD(Tk, Tp);
+		    TC = VADD(Tb, Td);
+		    Te = VSUB(Tb, Td);
+		    {
+			 V TM, TD, Tv, Tf;
+			 TM = VSUB(TB, TC);
+			 TD = VADD(TB, TC);
+			 Tv = VSUB(T9, Te);
+			 Tf = VADD(T9, Te);
+			 {
+			      V TP, TN, TH, TJ, Tz, Tx, Tr, Tt, TI, Ts;
+			      TP = VMUL(LDK(KP951056516), VFMA(LDK(KP618033988), TL, TM));
+			      TN = VMUL(LDK(KP951056516), VFNMS(LDK(KP618033988), TM, TL));
+			      TH = VADD(TD, TG);
+			      TJ = VSUB(TD, TG);
+			      Tz = VMUL(LDK(KP951056516), VFNMS(LDK(KP618033988), Tv, Tw));
+			      Tx = VMUL(LDK(KP951056516), VFMA(LDK(KP618033988), Tw, Tv));
+			      Tr = VADD(Tf, Tq);
+			      Tt = VSUB(Tf, Tq);
+			      ST(&(x[0]), VADD(TA, TH), ms, &(x[0]));
+			      TI = VFNMS(LDK(KP250000000), TH, TA);
+			      ST(&(x[WS(rs, 5)]), VADD(T4, Tr), ms, &(x[WS(rs, 1)]));
+			      Ts = VFNMS(LDK(KP250000000), Tr, T4);
+			      {
+				   V TK, TO, Tu, Ty;
+				   TK = VFNMS(LDK(KP559016994), TJ, TI);
+				   TO = VFMA(LDK(KP559016994), TJ, TI);
+				   Tu = VFMA(LDK(KP559016994), Tt, Ts);
+				   Ty = VFNMS(LDK(KP559016994), Tt, Ts);
+				   ST(&(x[WS(rs, 8)]), VFMAI(TN, TK), ms, &(x[0]));
+				   ST(&(x[WS(rs, 2)]), VFNMSI(TN, TK), ms, &(x[0]));
+				   ST(&(x[WS(rs, 6)]), VFMAI(TP, TO), ms, &(x[0]));
+				   ST(&(x[WS(rs, 4)]), VFNMSI(TP, TO), ms, &(x[0]));
+				   ST(&(x[WS(rs, 9)]), VFNMSI(Tx, Tu), ms, &(x[WS(rs, 1)]));
+				   ST(&(x[WS(rs, 1)]), VFMAI(Tx, Tu), ms, &(x[WS(rs, 1)]));
+				   ST(&(x[WS(rs, 7)]), VFNMSI(Tz, Ty), ms, &(x[WS(rs, 1)]));
+				   ST(&(x[WS(rs, 3)]), VFMAI(Tz, Ty), ms, &(x[WS(rs, 1)]));
+			      }
+			 }
+		    }
+	       }
+	  }
+     }
+     VLEAVE();
+}
+
+static const tw_instr twinstr[] = {
+     VTW(0, 1),
+     VTW(0, 2),
+     VTW(0, 3),
+     VTW(0, 4),
+     VTW(0, 5),
+     VTW(0, 6),
+     VTW(0, 7),
+     VTW(0, 8),
+     VTW(0, 9),
+     {TW_NEXT, VL, 0}
+};
+
+static const ct_desc desc = { 10, XSIMD_STRING("t1bv_10"), twinstr, &GENUS, {33, 22, 18, 0}, 0, 0, 0 };
+
+void XSIMD(codelet_t1bv_10) (planner *p) {
+     X(kdft_dit_register) (p, t1bv_10, &desc);
+}
+#else				/* HAVE_FMA */
+
+/* Generated by: ../../../genfft/gen_twiddle_c.native -simd -compact -variables 4 -pipeline-latency 8 -n 10 -name t1bv_10 -include t1b.h -sign 1 */
+
+/*
+ * This function contains 51 FP additions, 30 FP multiplications,
+ * (or, 45 additions, 24 multiplications, 6 fused multiply/add),
+ * 32 stack variables, 4 constants, and 20 memory accesses
+ */
+#include "t1b.h"
+
+static void t1bv_10(R *ri, R *ii, const R *W, stride rs, INT mb, INT me, INT ms)
+{
+     DVK(KP587785252, +0.587785252292473129168705954639072768597652438);
+     DVK(KP951056516, +0.951056516295153572116439333379382143405698634);
+     DVK(KP250000000, +0.250000000000000000000000000000000000000000000);
+     DVK(KP559016994, +0.559016994374947424102293417182819058860154590);
+     {
+	  INT m;
+	  R *x;
+	  x = ii;
+	  for (m = mb, W = W + (mb * ((TWVL / VL) * 18)); m < me; m = m + VL, x = x + (VL * ms), W = W + (TWVL * 18), MAKE_VOLATILE_STRIDE(10, rs)) {
+	       V Tu, TH, Tg, Tl, Tp, TD, TE, TJ, T5, Ta, To, TA, TB, TI, Tr;
+	       V Tt, Ts;
+	       Tr = LD(&(x[0]), ms, &(x[0]));
+	       Ts = LD(&(x[WS(rs, 5)]), ms, &(x[WS(rs, 1)]));
+	       Tt = BYTW(&(W[TWVL * 8]), Ts);
+	       Tu = VSUB(Tr, Tt);
+	       TH = VADD(Tr, Tt);
+	       {
+		    V Td, Tk, Tf, Ti;
+		    {
+			 V Tc, Tj, Te, Th;
+			 Tc = LD(&(x[WS(rs, 4)]), ms, &(x[0]));
+			 Td = BYTW(&(W[TWVL * 6]), Tc);
+			 Tj = LD(&(x[WS(rs, 1)]), ms, &(x[WS(rs, 1)]));
+			 Tk = BYTW(&(W[0]), Tj);
+			 Te = LD(&(x[WS(rs, 9)]), ms, &(x[WS(rs, 1)]));
+			 Tf = BYTW(&(W[TWVL * 16]), Te);
+			 Th = LD(&(x[WS(rs, 6)]), ms, &(x[0]));
+			 Ti = BYTW(&(W[TWVL * 10]), Th);
+		    }
+		    Tg = VSUB(Td, Tf);
+		    Tl = VSUB(Ti, Tk);
+		    Tp = VADD(Tg, Tl);
+		    TD = VADD(Td, Tf);
+		    TE = VADD(Ti, Tk);
+		    TJ = VADD(TD, TE);
+	       }
+	       {
+		    V T2, T9, T4, T7;
+		    {
+			 V T1, T8, T3, T6;
+			 T1 = LD(&(x[WS(rs, 2)]), ms, &(x[0]));
+			 T2 = BYTW(&(W[TWVL * 2]), T1);
+			 T8 = LD(&(x[WS(rs, 3)]), ms, &(x[WS(rs, 1)]));
+			 T9 = BYTW(&(W[TWVL * 4]), T8);
+			 T3 = LD(&(x[WS(rs, 7)]), ms, &(x[WS(rs, 1)]));
+			 T4 = BYTW(&(W[TWVL * 12]), T3);
+			 T6 = LD(&(x[WS(rs, 8)]), ms, &(x[0]));
+			 T7 = BYTW(&(W[TWVL * 14]), T6);
+		    }
+		    T5 = VSUB(T2, T4);
+		    Ta = VSUB(T7, T9);
+		    To = VADD(T5, Ta);
+		    TA = VADD(T2, T4);
+		    TB = VADD(T7, T9);
+		    TI = VADD(TA, TB);
+	       }
+	       {
+		    V Tq, Tv, Tw, Tn, Tz, Tb, Tm, Ty, Tx;
+		    Tq = VMUL(LDK(KP559016994), VSUB(To, Tp));
+		    Tv = VADD(To, Tp);
+		    Tw = VFNMS(LDK(KP250000000), Tv, Tu);
+		    Tb = VSUB(T5, Ta);
+		    Tm = VSUB(Tg, Tl);
+		    Tn = VBYI(VFMA(LDK(KP951056516), Tb, VMUL(LDK(KP587785252), Tm)));
+		    Tz = VBYI(VFNMS(LDK(KP951056516), Tm, VMUL(LDK(KP587785252), Tb)));
+		    ST(&(x[WS(rs, 5)]), VADD(Tu, Tv), ms, &(x[WS(rs, 1)]));
+		    Ty = VSUB(Tw, Tq);
+		    ST(&(x[WS(rs, 3)]), VSUB(Ty, Tz), ms, &(x[WS(rs, 1)]));
+		    ST(&(x[WS(rs, 7)]), VADD(Tz, Ty), ms, &(x[WS(rs, 1)]));
+		    Tx = VADD(Tq, Tw);
+		    ST(&(x[WS(rs, 1)]), VADD(Tn, Tx), ms, &(x[WS(rs, 1)]));
+		    ST(&(x[WS(rs, 9)]), VSUB(Tx, Tn), ms, &(x[WS(rs, 1)]));
+	       }
+	       {
+		    V TM, TK, TL, TG, TP, TC, TF, TO, TN;
+		    TM = VMUL(LDK(KP559016994), VSUB(TI, TJ));
+		    TK = VADD(TI, TJ);
+		    TL = VFNMS(LDK(KP250000000), TK, TH);
+		    TC = VSUB(TA, TB);
+		    TF = VSUB(TD, TE);
+		    TG = VBYI(VFNMS(LDK(KP951056516), TF, VMUL(LDK(KP587785252), TC)));
+		    TP = VBYI(VFMA(LDK(KP951056516), TC, VMUL(LDK(KP587785252), TF)));
+		    ST(&(x[0]), VADD(TH, TK), ms, &(x[0]));
+		    TO = VADD(TM, TL);
+		    ST(&(x[WS(rs, 4)]), VSUB(TO, TP), ms, &(x[0]));
+		    ST(&(x[WS(rs, 6)]), VADD(TP, TO), ms, &(x[0]));
+		    TN = VSUB(TL, TM);
+		    ST(&(x[WS(rs, 2)]), VADD(TG, TN), ms, &(x[0]));
+		    ST(&(x[WS(rs, 8)]), VSUB(TN, TG), ms, &(x[0]));
+	       }
+	  }
+     }
+     VLEAVE();
+}
+
+static const tw_instr twinstr[] = {
+     VTW(0, 1),
+     VTW(0, 2),
+     VTW(0, 3),
+     VTW(0, 4),
+     VTW(0, 5),
+     VTW(0, 6),
+     VTW(0, 7),
+     VTW(0, 8),
+     VTW(0, 9),
+     {TW_NEXT, VL, 0}
+};
+
+static const ct_desc desc = { 10, XSIMD_STRING("t1bv_10"), twinstr, &GENUS, {45, 24, 6, 0}, 0, 0, 0 };
+
+void XSIMD(codelet_t1bv_10) (planner *p) {
+     X(kdft_dit_register) (p, t1bv_10, &desc);
+}
+#endif				/* HAVE_FMA */