Mercurial > hg > sv-dependency-builds

annotate src/fftw-3.3.5/rdft/scalar/r2cb/r2cbIII_8.c @ 56:af97cad61ff0

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
Add updated build of PortAudio for OSX
author Chris Cannam <cannam@all-day-breakfast.com>
date Tue, 03 Jan 2017 15:10:52 +0000
parents 2cd0e3b3e1fd
children
rev   line source
Chris@42 1 /*
Chris@42 2 * Copyright (c) 2003, 2007-14 Matteo Frigo
Chris@42 3 * Copyright (c) 2003, 2007-14 Massachusetts Institute of Technology
Chris@42 4 *
Chris@42 5 * This program is free software; you can redistribute it and/or modify
Chris@42 6 * it under the terms of the GNU General Public License as published by
Chris@42 7 * the Free Software Foundation; either version 2 of the License, or
Chris@42 8 * (at your option) any later version.
Chris@42 9 *
Chris@42 10 * This program is distributed in the hope that it will be useful,
Chris@42 11 * but WITHOUT ANY WARRANTY; without even the implied warranty of
Chris@42 12 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
Chris@42 13 * GNU General Public License for more details.
Chris@42 14 *
Chris@42 15 * You should have received a copy of the GNU General Public License
Chris@42 16 * along with this program; if not, write to the Free Software
Chris@42 17 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
Chris@42 18 *
Chris@42 19 */
Chris@42 20
Chris@42 21 /* This file was automatically generated --- DO NOT EDIT */
Chris@42 22 /* Generated on Sat Jul 30 16:50:38 EDT 2016 */
Chris@42 23
Chris@42 24 #include "codelet-rdft.h"
Chris@42 25
Chris@42 26 #ifdef HAVE_FMA
Chris@42 27
Chris@42 28 /* Generated by: ../../../genfft/gen_r2cb.native -fma -reorder-insns -schedule-for-pipeline -compact -variables 4 -pipeline-latency 4 -sign 1 -n 8 -name r2cbIII_8 -dft-III -include r2cbIII.h */
Chris@42 29
Chris@42 30 /*
Chris@42 31 * This function contains 22 FP additions, 12 FP multiplications,
Chris@42 32 * (or, 18 additions, 8 multiplications, 4 fused multiply/add),
Chris@42 33 * 23 stack variables, 4 constants, and 16 memory accesses
Chris@42 34 */
Chris@42 35 #include "r2cbIII.h"
Chris@42 36
Chris@42 37 static void r2cbIII_8(R *R0, R *R1, R *Cr, R *Ci, stride rs, stride csr, stride csi, INT v, INT ivs, INT ovs)
Chris@42 38 {
Chris@42 39 DK(KP414213562, +0.414213562373095048801688724209698078569671875);
Chris@42 40 DK(KP1_847759065, +1.847759065022573512256366378793576573644833252);
Chris@42 41 DK(KP1_414213562, +1.414213562373095048801688724209698078569671875);
Chris@42 42 DK(KP2_000000000, +2.000000000000000000000000000000000000000000000);
Chris@42 43 {
Chris@42 44 INT i;
Chris@42 45 for (i = v; i > 0; i = i - 1, R0 = R0 + ovs, R1 = R1 + ovs, Cr = Cr + ivs, Ci = Ci + ivs, MAKE_VOLATILE_STRIDE(32, rs), MAKE_VOLATILE_STRIDE(32, csr), MAKE_VOLATILE_STRIDE(32, csi)) {
Chris@42 46 E T4, T7, T3, Tl, Tf, T5, T8, T9, T6, Tc;
Chris@42 47 {
Chris@42 48 E T1, T2, Td, Te;
Chris@42 49 T1 = Cr[0];
Chris@42 50 T2 = Cr[WS(csr, 3)];
Chris@42 51 Td = Ci[0];
Chris@42 52 Te = Ci[WS(csi, 3)];
Chris@42 53 T4 = Cr[WS(csr, 2)];
Chris@42 54 T7 = T1 - T2;
Chris@42 55 T3 = T1 + T2;
Chris@42 56 Tl = Te - Td;
Chris@42 57 Tf = Td + Te;
Chris@42 58 T5 = Cr[WS(csr, 1)];
Chris@42 59 T8 = Ci[WS(csi, 2)];
Chris@42 60 T9 = Ci[WS(csi, 1)];
Chris@42 61 }
Chris@42 62 T6 = T4 + T5;
Chris@42 63 Tc = T4 - T5;
Chris@42 64 {
Chris@42 65 E Ta, Tk, Tg, Th;
Chris@42 66 Ta = T8 + T9;
Chris@42 67 Tk = T8 - T9;
Chris@42 68 Tg = Tc + Tf;
Chris@42 69 Th = Tc - Tf;
Chris@42 70 {
Chris@42 71 E Tj, Tm, Tb, Ti;
Chris@42 72 Tj = T3 - T6;
Chris@42 73 R0[0] = KP2_000000000 * (T3 + T6);
Chris@42 74 Tm = Tk + Tl;
Chris@42 75 R0[WS(rs, 2)] = KP2_000000000 * (Tl - Tk);
Chris@42 76 Tb = T7 - Ta;
Chris@42 77 Ti = T7 + Ta;
Chris@42 78 R0[WS(rs, 3)] = KP1_414213562 * (Tm - Tj);
Chris@42 79 R0[WS(rs, 1)] = KP1_414213562 * (Tj + Tm);
Chris@42 80 R1[WS(rs, 3)] = -(KP1_847759065 * (FNMS(KP414213562, Th, Ti)));
Chris@42 81 R1[WS(rs, 1)] = KP1_847759065 * (FMA(KP414213562, Ti, Th));
Chris@42 82 R1[WS(rs, 2)] = -(KP1_847759065 * (FMA(KP414213562, Tb, Tg)));
Chris@42 83 R1[0] = KP1_847759065 * (FNMS(KP414213562, Tg, Tb));
Chris@42 84 }
Chris@42 85 }
Chris@42 86 }
Chris@42 87 }
Chris@42 88 }
Chris@42 89
Chris@42 90 static const kr2c_desc desc = { 8, "r2cbIII_8", {18, 8, 4, 0}, &GENUS };
Chris@42 91
Chris@42 92 void X(codelet_r2cbIII_8) (planner *p) {
Chris@42 93 X(kr2c_register) (p, r2cbIII_8, &desc);
Chris@42 94 }
Chris@42 95
Chris@42 96 #else /* HAVE_FMA */
Chris@42 97
Chris@42 98 /* Generated by: ../../../genfft/gen_r2cb.native -compact -variables 4 -pipeline-latency 4 -sign 1 -n 8 -name r2cbIII_8 -dft-III -include r2cbIII.h */
Chris@42 99
Chris@42 100 /*
Chris@42 101 * This function contains 22 FP additions, 12 FP multiplications,
Chris@42 102 * (or, 18 additions, 8 multiplications, 4 fused multiply/add),
Chris@42 103 * 19 stack variables, 4 constants, and 16 memory accesses
Chris@42 104 */
Chris@42 105 #include "r2cbIII.h"
Chris@42 106
Chris@42 107 static void r2cbIII_8(R *R0, R *R1, R *Cr, R *Ci, stride rs, stride csr, stride csi, INT v, INT ivs, INT ovs)
Chris@42 108 {
Chris@42 109 DK(KP1_414213562, +1.414213562373095048801688724209698078569671875);
Chris@42 110 DK(KP765366864, +0.765366864730179543456919968060797733522689125);
Chris@42 111 DK(KP1_847759065, +1.847759065022573512256366378793576573644833252);
Chris@42 112 DK(KP2_000000000, +2.000000000000000000000000000000000000000000000);
Chris@42 113 {
Chris@42 114 INT i;
Chris@42 115 for (i = v; i > 0; i = i - 1, R0 = R0 + ovs, R1 = R1 + ovs, Cr = Cr + ivs, Ci = Ci + ivs, MAKE_VOLATILE_STRIDE(32, rs), MAKE_VOLATILE_STRIDE(32, csr), MAKE_VOLATILE_STRIDE(32, csi)) {
Chris@42 116 E T3, T7, Tf, Tl, T6, Tc, Ta, Tk, Tb, Tg;
Chris@42 117 {
Chris@42 118 E T1, T2, Td, Te;
Chris@42 119 T1 = Cr[0];
Chris@42 120 T2 = Cr[WS(csr, 3)];
Chris@42 121 T3 = T1 + T2;
Chris@42 122 T7 = T1 - T2;
Chris@42 123 Td = Ci[0];
Chris@42 124 Te = Ci[WS(csi, 3)];
Chris@42 125 Tf = Td + Te;
Chris@42 126 Tl = Te - Td;
Chris@42 127 }
Chris@42 128 {
Chris@42 129 E T4, T5, T8, T9;
Chris@42 130 T4 = Cr[WS(csr, 2)];
Chris@42 131 T5 = Cr[WS(csr, 1)];
Chris@42 132 T6 = T4 + T5;
Chris@42 133 Tc = T4 - T5;
Chris@42 134 T8 = Ci[WS(csi, 2)];
Chris@42 135 T9 = Ci[WS(csi, 1)];
Chris@42 136 Ta = T8 + T9;
Chris@42 137 Tk = T8 - T9;
Chris@42 138 }
Chris@42 139 R0[0] = KP2_000000000 * (T3 + T6);
Chris@42 140 R0[WS(rs, 2)] = KP2_000000000 * (Tl - Tk);
Chris@42 141 Tb = T7 - Ta;
Chris@42 142 Tg = Tc + Tf;
Chris@42 143 R1[0] = FNMS(KP765366864, Tg, KP1_847759065 * Tb);
Chris@42 144 R1[WS(rs, 2)] = -(FMA(KP765366864, Tb, KP1_847759065 * Tg));
Chris@42 145 {
Chris@42 146 E Th, Ti, Tj, Tm;
Chris@42 147 Th = T7 + Ta;
Chris@42 148 Ti = Tc - Tf;
Chris@42 149 R1[WS(rs, 1)] = FMA(KP765366864, Th, KP1_847759065 * Ti);
Chris@42 150 R1[WS(rs, 3)] = FNMS(KP1_847759065, Th, KP765366864 * Ti);
Chris@42 151 Tj = T3 - T6;
Chris@42 152 Tm = Tk + Tl;
Chris@42 153 R0[WS(rs, 1)] = KP1_414213562 * (Tj + Tm);
Chris@42 154 R0[WS(rs, 3)] = KP1_414213562 * (Tm - Tj);
Chris@42 155 }
Chris@42 156 }
Chris@42 157 }
Chris@42 158 }
Chris@42 159
Chris@42 160 static const kr2c_desc desc = { 8, "r2cbIII_8", {18, 8, 4, 0}, &GENUS };
Chris@42 161
Chris@42 162 void X(codelet_r2cbIII_8) (planner *p) {
Chris@42 163 X(kr2c_register) (p, r2cbIII_8, &desc);
Chris@42 164 }
Chris@42 165
Chris@42 166 #endif /* HAVE_FMA */