Mercurial > hg > sv-dependency-builds

diff src/fftw-3.3.3/dft/simd/common/q1bv_4.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/q1bv_4.c	Wed Mar 20 15:35:50 2013 +0000
@@ -0,0 +1,253 @@
+/*
+ * 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 07:39:33 EST 2012 */
+
+#include "codelet-dft.h"
+
+#ifdef HAVE_FMA
+
+/* Generated by: ../../../genfft/gen_twidsq_c.native -fma -reorder-insns -schedule-for-pipeline -simd -compact -variables 4 -pipeline-latency 8 -n 4 -dif -name q1bv_4 -include q1b.h -sign 1 */
+
+/*
+ * This function contains 44 FP additions, 32 FP multiplications,
+ * (or, 36 additions, 24 multiplications, 8 fused multiply/add),
+ * 38 stack variables, 0 constants, and 32 memory accesses
+ */
+#include "q1b.h"
+
+static void q1bv_4(R *ri, R *ii, const R *W, stride rs, stride vs, INT mb, INT me, INT ms)
+{
+     {
+	  INT m;
+	  R *x;
+	  x = ii;
+	  for (m = mb, W = W + (mb * ((TWVL / VL) * 6)); m < me; m = m + VL, x = x + (VL * ms), W = W + (TWVL * 6), MAKE_VOLATILE_STRIDE(8, rs), MAKE_VOLATILE_STRIDE(8, vs)) {
+	       V Tb, Tm, Tx, TI;
+	       {
+		    V Tc, T9, T3, TG, TA, TH, TD, Ta, T6, Td, Tn, To, Tq, Tr, Tf;
+		    V Tg;
+		    {
+			 V T1, T2, Ty, Tz, TB, TC, T4, T5;
+			 T1 = LD(&(x[0]), ms, &(x[0]));
+			 T2 = LD(&(x[WS(rs, 2)]), ms, &(x[0]));
+			 Ty = LD(&(x[WS(vs, 3)]), ms, &(x[WS(vs, 3)]));
+			 Tz = LD(&(x[WS(vs, 3) + WS(rs, 2)]), ms, &(x[WS(vs, 3)]));
+			 TB = LD(&(x[WS(vs, 3) + WS(rs, 1)]), ms, &(x[WS(vs, 3) + WS(rs, 1)]));
+			 TC = LD(&(x[WS(vs, 3) + WS(rs, 3)]), ms, &(x[WS(vs, 3) + WS(rs, 1)]));
+			 T4 = LD(&(x[WS(rs, 1)]), ms, &(x[WS(rs, 1)]));
+			 T5 = LD(&(x[WS(rs, 3)]), ms, &(x[WS(rs, 1)]));
+			 Tc = LD(&(x[WS(vs, 1)]), ms, &(x[WS(vs, 1)]));
+			 T9 = VADD(T1, T2);
+			 T3 = VSUB(T1, T2);
+			 TG = VADD(Ty, Tz);
+			 TA = VSUB(Ty, Tz);
+			 TH = VADD(TB, TC);
+			 TD = VSUB(TB, TC);
+			 Ta = VADD(T4, T5);
+			 T6 = VSUB(T4, T5);
+			 Td = LD(&(x[WS(vs, 1) + WS(rs, 2)]), ms, &(x[WS(vs, 1)]));
+			 Tn = LD(&(x[WS(vs, 2)]), ms, &(x[WS(vs, 2)]));
+			 To = LD(&(x[WS(vs, 2) + WS(rs, 2)]), ms, &(x[WS(vs, 2)]));
+			 Tq = LD(&(x[WS(vs, 2) + WS(rs, 1)]), ms, &(x[WS(vs, 2) + WS(rs, 1)]));
+			 Tr = LD(&(x[WS(vs, 2) + WS(rs, 3)]), ms, &(x[WS(vs, 2) + WS(rs, 1)]));
+			 Tf = LD(&(x[WS(vs, 1) + WS(rs, 1)]), ms, &(x[WS(vs, 1) + WS(rs, 1)]));
+			 Tg = LD(&(x[WS(vs, 1) + WS(rs, 3)]), ms, &(x[WS(vs, 1) + WS(rs, 1)]));
+		    }
+		    {
+			 V Tk, Te, Tv, Tp, Tw, Ts, Tl, Th, T7, TE, Tu, TF;
+			 ST(&(x[0]), VADD(T9, Ta), ms, &(x[0]));
+			 Tk = VADD(Tc, Td);
+			 Te = VSUB(Tc, Td);
+			 Tv = VADD(Tn, To);
+			 Tp = VSUB(Tn, To);
+			 Tw = VADD(Tq, Tr);
+			 Ts = VSUB(Tq, Tr);
+			 Tl = VADD(Tf, Tg);
+			 Th = VSUB(Tf, Tg);
+			 ST(&(x[WS(rs, 3)]), VADD(TG, TH), ms, &(x[WS(rs, 1)]));
+			 T7 = BYTW(&(W[TWVL * 4]), VFNMSI(T6, T3));
+			 TE = BYTW(&(W[TWVL * 4]), VFNMSI(TD, TA));
+			 {
+			      V Tt, Ti, Tj, T8;
+			      T8 = BYTW(&(W[0]), VFMAI(T6, T3));
+			      ST(&(x[WS(rs, 2)]), VADD(Tv, Tw), ms, &(x[0]));
+			      Tt = BYTW(&(W[TWVL * 4]), VFNMSI(Ts, Tp));
+			      ST(&(x[WS(rs, 1)]), VADD(Tk, Tl), ms, &(x[WS(rs, 1)]));
+			      Ti = BYTW(&(W[TWVL * 4]), VFNMSI(Th, Te));
+			      Tj = BYTW(&(W[0]), VFMAI(Th, Te));
+			      ST(&(x[WS(vs, 3)]), T7, ms, &(x[WS(vs, 3)]));
+			      ST(&(x[WS(vs, 3) + WS(rs, 3)]), TE, ms, &(x[WS(vs, 3) + WS(rs, 1)]));
+			      ST(&(x[WS(vs, 1)]), T8, ms, &(x[WS(vs, 1)]));
+			      Tu = BYTW(&(W[0]), VFMAI(Ts, Tp));
+			      ST(&(x[WS(vs, 3) + WS(rs, 2)]), Tt, ms, &(x[WS(vs, 3)]));
+			      TF = BYTW(&(W[0]), VFMAI(TD, TA));
+			      ST(&(x[WS(vs, 3) + WS(rs, 1)]), Ti, ms, &(x[WS(vs, 3) + WS(rs, 1)]));
+			      ST(&(x[WS(vs, 1) + WS(rs, 1)]), Tj, ms, &(x[WS(vs, 1) + WS(rs, 1)]));
+			 }
+			 Tb = BYTW(&(W[TWVL * 2]), VSUB(T9, Ta));
+			 Tm = BYTW(&(W[TWVL * 2]), VSUB(Tk, Tl));
+			 Tx = BYTW(&(W[TWVL * 2]), VSUB(Tv, Tw));
+			 ST(&(x[WS(vs, 1) + WS(rs, 2)]), Tu, ms, &(x[WS(vs, 1)]));
+			 TI = BYTW(&(W[TWVL * 2]), VSUB(TG, TH));
+			 ST(&(x[WS(vs, 1) + WS(rs, 3)]), TF, ms, &(x[WS(vs, 1) + WS(rs, 1)]));
+		    }
+	       }
+	       ST(&(x[WS(vs, 2)]), Tb, ms, &(x[WS(vs, 2)]));
+	       ST(&(x[WS(vs, 2) + WS(rs, 1)]), Tm, ms, &(x[WS(vs, 2) + WS(rs, 1)]));
+	       ST(&(x[WS(vs, 2) + WS(rs, 2)]), Tx, ms, &(x[WS(vs, 2)]));
+	       ST(&(x[WS(vs, 2) + WS(rs, 3)]), TI, ms, &(x[WS(vs, 2) + WS(rs, 1)]));
+	  }
+     }
+     VLEAVE();
+}
+
+static const tw_instr twinstr[] = {
+     VTW(0, 1),
+     VTW(0, 2),
+     VTW(0, 3),
+     {TW_NEXT, VL, 0}
+};
+
+static const ct_desc desc = { 4, XSIMD_STRING("q1bv_4"), twinstr, &GENUS, {36, 24, 8, 0}, 0, 0, 0 };
+
+void XSIMD(codelet_q1bv_4) (planner *p) {
+     X(kdft_difsq_register) (p, q1bv_4, &desc);
+}
+#else				/* HAVE_FMA */
+
+/* Generated by: ../../../genfft/gen_twidsq_c.native -simd -compact -variables 4 -pipeline-latency 8 -n 4 -dif -name q1bv_4 -include q1b.h -sign 1 */
+
+/*
+ * This function contains 44 FP additions, 24 FP multiplications,
+ * (or, 44 additions, 24 multiplications, 0 fused multiply/add),
+ * 22 stack variables, 0 constants, and 32 memory accesses
+ */
+#include "q1b.h"
+
+static void q1bv_4(R *ri, R *ii, const R *W, stride rs, stride vs, INT mb, INT me, INT ms)
+{
+     {
+	  INT m;
+	  R *x;
+	  x = ii;
+	  for (m = mb, W = W + (mb * ((TWVL / VL) * 6)); m < me; m = m + VL, x = x + (VL * ms), W = W + (TWVL * 6), MAKE_VOLATILE_STRIDE(8, rs), MAKE_VOLATILE_STRIDE(8, vs)) {
+	       V T3, T9, TA, TG, TD, TH, T6, Ta, Te, Tk, Tp, Tv, Ts, Tw, Th;
+	       V Tl;
+	       {
+		    V T1, T2, Ty, Tz;
+		    T1 = LD(&(x[0]), ms, &(x[0]));
+		    T2 = LD(&(x[WS(rs, 2)]), ms, &(x[0]));
+		    T3 = VSUB(T1, T2);
+		    T9 = VADD(T1, T2);
+		    Ty = LD(&(x[WS(vs, 3)]), ms, &(x[WS(vs, 3)]));
+		    Tz = LD(&(x[WS(vs, 3) + WS(rs, 2)]), ms, &(x[WS(vs, 3)]));
+		    TA = VSUB(Ty, Tz);
+		    TG = VADD(Ty, Tz);
+	       }
+	       {
+		    V TB, TC, T4, T5;
+		    TB = LD(&(x[WS(vs, 3) + WS(rs, 1)]), ms, &(x[WS(vs, 3) + WS(rs, 1)]));
+		    TC = LD(&(x[WS(vs, 3) + WS(rs, 3)]), ms, &(x[WS(vs, 3) + WS(rs, 1)]));
+		    TD = VBYI(VSUB(TB, TC));
+		    TH = VADD(TB, TC);
+		    T4 = LD(&(x[WS(rs, 1)]), ms, &(x[WS(rs, 1)]));
+		    T5 = LD(&(x[WS(rs, 3)]), ms, &(x[WS(rs, 1)]));
+		    T6 = VBYI(VSUB(T4, T5));
+		    Ta = VADD(T4, T5);
+	       }
+	       {
+		    V Tc, Td, Tn, To;
+		    Tc = LD(&(x[WS(vs, 1)]), ms, &(x[WS(vs, 1)]));
+		    Td = LD(&(x[WS(vs, 1) + WS(rs, 2)]), ms, &(x[WS(vs, 1)]));
+		    Te = VSUB(Tc, Td);
+		    Tk = VADD(Tc, Td);
+		    Tn = LD(&(x[WS(vs, 2)]), ms, &(x[WS(vs, 2)]));
+		    To = LD(&(x[WS(vs, 2) + WS(rs, 2)]), ms, &(x[WS(vs, 2)]));
+		    Tp = VSUB(Tn, To);
+		    Tv = VADD(Tn, To);
+	       }
+	       {
+		    V Tq, Tr, Tf, Tg;
+		    Tq = LD(&(x[WS(vs, 2) + WS(rs, 1)]), ms, &(x[WS(vs, 2) + WS(rs, 1)]));
+		    Tr = LD(&(x[WS(vs, 2) + WS(rs, 3)]), ms, &(x[WS(vs, 2) + WS(rs, 1)]));
+		    Ts = VBYI(VSUB(Tq, Tr));
+		    Tw = VADD(Tq, Tr);
+		    Tf = LD(&(x[WS(vs, 1) + WS(rs, 1)]), ms, &(x[WS(vs, 1) + WS(rs, 1)]));
+		    Tg = LD(&(x[WS(vs, 1) + WS(rs, 3)]), ms, &(x[WS(vs, 1) + WS(rs, 1)]));
+		    Th = VBYI(VSUB(Tf, Tg));
+		    Tl = VADD(Tf, Tg);
+	       }
+	       ST(&(x[0]), VADD(T9, Ta), ms, &(x[0]));
+	       ST(&(x[WS(rs, 1)]), VADD(Tk, Tl), ms, &(x[WS(rs, 1)]));
+	       ST(&(x[WS(rs, 2)]), VADD(Tv, Tw), ms, &(x[0]));
+	       ST(&(x[WS(rs, 3)]), VADD(TG, TH), ms, &(x[WS(rs, 1)]));
+	       {
+		    V T7, Ti, Tt, TE;
+		    T7 = BYTW(&(W[TWVL * 4]), VSUB(T3, T6));
+		    ST(&(x[WS(vs, 3)]), T7, ms, &(x[WS(vs, 3)]));
+		    Ti = BYTW(&(W[TWVL * 4]), VSUB(Te, Th));
+		    ST(&(x[WS(vs, 3) + WS(rs, 1)]), Ti, ms, &(x[WS(vs, 3) + WS(rs, 1)]));
+		    Tt = BYTW(&(W[TWVL * 4]), VSUB(Tp, Ts));
+		    ST(&(x[WS(vs, 3) + WS(rs, 2)]), Tt, ms, &(x[WS(vs, 3)]));
+		    TE = BYTW(&(W[TWVL * 4]), VSUB(TA, TD));
+		    ST(&(x[WS(vs, 3) + WS(rs, 3)]), TE, ms, &(x[WS(vs, 3) + WS(rs, 1)]));
+	       }
+	       {
+		    V T8, Tj, Tu, TF;
+		    T8 = BYTW(&(W[0]), VADD(T3, T6));
+		    ST(&(x[WS(vs, 1)]), T8, ms, &(x[WS(vs, 1)]));
+		    Tj = BYTW(&(W[0]), VADD(Te, Th));
+		    ST(&(x[WS(vs, 1) + WS(rs, 1)]), Tj, ms, &(x[WS(vs, 1) + WS(rs, 1)]));
+		    Tu = BYTW(&(W[0]), VADD(Tp, Ts));
+		    ST(&(x[WS(vs, 1) + WS(rs, 2)]), Tu, ms, &(x[WS(vs, 1)]));
+		    TF = BYTW(&(W[0]), VADD(TA, TD));
+		    ST(&(x[WS(vs, 1) + WS(rs, 3)]), TF, ms, &(x[WS(vs, 1) + WS(rs, 1)]));
+	       }
+	       {
+		    V Tb, Tm, Tx, TI;
+		    Tb = BYTW(&(W[TWVL * 2]), VSUB(T9, Ta));
+		    ST(&(x[WS(vs, 2)]), Tb, ms, &(x[WS(vs, 2)]));
+		    Tm = BYTW(&(W[TWVL * 2]), VSUB(Tk, Tl));
+		    ST(&(x[WS(vs, 2) + WS(rs, 1)]), Tm, ms, &(x[WS(vs, 2) + WS(rs, 1)]));
+		    Tx = BYTW(&(W[TWVL * 2]), VSUB(Tv, Tw));
+		    ST(&(x[WS(vs, 2) + WS(rs, 2)]), Tx, ms, &(x[WS(vs, 2)]));
+		    TI = BYTW(&(W[TWVL * 2]), VSUB(TG, TH));
+		    ST(&(x[WS(vs, 2) + WS(rs, 3)]), TI, ms, &(x[WS(vs, 2) + WS(rs, 1)]));
+	       }
+	  }
+     }
+     VLEAVE();
+}
+
+static const tw_instr twinstr[] = {
+     VTW(0, 1),
+     VTW(0, 2),
+     VTW(0, 3),
+     {TW_NEXT, VL, 0}
+};
+
+static const ct_desc desc = { 4, XSIMD_STRING("q1bv_4"), twinstr, &GENUS, {44, 24, 0, 0}, 0, 0, 0 };
+
+void XSIMD(codelet_q1bv_4) (planner *p) {
+     X(kdft_difsq_register) (p, q1bv_4, &desc);
+}
+#endif				/* HAVE_FMA */