Mercurial > hg > sv-dependency-builds

annotate src/fftw-3.3.5/rdft/scalar/r2cb/r2cbIII_10.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 16:50:41 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 10 -name r2cbIII_10 -dft-III -include r2cbIII.h */
Chris@42 29
Chris@42 30 /*
Chris@42 31 * This function contains 32 FP additions, 28 FP multiplications,
Chris@42 32 * (or, 14 additions, 10 multiplications, 18 fused multiply/add),
Chris@42 33 * 38 stack variables, 5 constants, and 20 memory accesses
Chris@42 34 */
Chris@42 35 #include "r2cbIII.h"
Chris@42 36
Chris@42 37 static void r2cbIII_10(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(KP951056516, +0.951056516295153572116439333379382143405698634);
Chris@42 40 DK(KP559016994, +0.559016994374947424102293417182819058860154590);
Chris@42 41 DK(KP250000000, +0.250000000000000000000000000000000000000000000);
Chris@42 42 DK(KP618033988, +0.618033988749894848204586834365638117720309180);
Chris@42 43 DK(KP2_000000000, +2.000000000000000000000000000000000000000000000);
Chris@42 44 {
Chris@42 45 INT i;
Chris@42 46 for (i = v; i > 0; i = i - 1, R0 = R0 + ovs, R1 = R1 + ovs, Cr = Cr + ivs, Ci = Ci + ivs, MAKE_VOLATILE_STRIDE(40, rs), MAKE_VOLATILE_STRIDE(40, csr), MAKE_VOLATILE_STRIDE(40, csi)) {
Chris@42 47 E Tq, Ti, Tk, Tu, Tw, Tp, Tb, Tj, Tr, Tv;
Chris@42 48 {
Chris@42 49 E T1, To, Ts, Tt, T8, Ta, Te, Tl, Tm, Th, Tn, T9;
Chris@42 50 T1 = Cr[WS(csr, 2)];
Chris@42 51 To = Ci[WS(csi, 2)];
Chris@42 52 {
Chris@42 53 E T2, T3, T5, T6;
Chris@42 54 T2 = Cr[WS(csr, 4)];
Chris@42 55 T3 = Cr[0];
Chris@42 56 T5 = Cr[WS(csr, 3)];
Chris@42 57 T6 = Cr[WS(csr, 1)];
Chris@42 58 {
Chris@42 59 E Tc, T4, T7, Td, Tf, Tg;
Chris@42 60 Tc = Ci[WS(csi, 3)];
Chris@42 61 Ts = T2 - T3;
Chris@42 62 T4 = T2 + T3;
Chris@42 63 Tt = T5 - T6;
Chris@42 64 T7 = T5 + T6;
Chris@42 65 Td = Ci[WS(csi, 1)];
Chris@42 66 Tf = Ci[WS(csi, 4)];
Chris@42 67 Tg = Ci[0];
Chris@42 68 T8 = T4 + T7;
Chris@42 69 Ta = T7 - T4;
Chris@42 70 Te = Tc - Td;
Chris@42 71 Tl = Tc + Td;
Chris@42 72 Tm = Tf + Tg;
Chris@42 73 Th = Tf - Tg;
Chris@42 74 }
Chris@42 75 }
Chris@42 76 R0[0] = KP2_000000000 * (T1 + T8);
Chris@42 77 Tn = Tl - Tm;
Chris@42 78 Tq = Tl + Tm;
Chris@42 79 Ti = FMA(KP618033988, Th, Te);
Chris@42 80 Tk = FNMS(KP618033988, Te, Th);
Chris@42 81 R1[WS(rs, 2)] = KP2_000000000 * (Tn - To);
Chris@42 82 T9 = FMS(KP250000000, T8, T1);
Chris@42 83 Tu = FMA(KP618033988, Tt, Ts);
Chris@42 84 Tw = FNMS(KP618033988, Ts, Tt);
Chris@42 85 Tp = FMA(KP250000000, Tn, To);
Chris@42 86 Tb = FNMS(KP559016994, Ta, T9);
Chris@42 87 Tj = FMA(KP559016994, Ta, T9);
Chris@42 88 }
Chris@42 89 Tr = FMA(KP559016994, Tq, Tp);
Chris@42 90 Tv = FNMS(KP559016994, Tq, Tp);
Chris@42 91 R0[WS(rs, 2)] = -(KP2_000000000 * (FNMS(KP951056516, Tk, Tj)));
Chris@42 92 R0[WS(rs, 3)] = KP2_000000000 * (FMA(KP951056516, Tk, Tj));
Chris@42 93 R0[WS(rs, 4)] = -(KP2_000000000 * (FNMS(KP951056516, Ti, Tb)));
Chris@42 94 R0[WS(rs, 1)] = KP2_000000000 * (FMA(KP951056516, Ti, Tb));
Chris@42 95 R1[WS(rs, 1)] = KP2_000000000 * (FMA(KP951056516, Tw, Tv));
Chris@42 96 R1[WS(rs, 3)] = KP2_000000000 * (FNMS(KP951056516, Tw, Tv));
Chris@42 97 R1[WS(rs, 4)] = -(KP2_000000000 * (FNMS(KP951056516, Tu, Tr)));
Chris@42 98 R1[0] = -(KP2_000000000 * (FMA(KP951056516, Tu, Tr)));
Chris@42 99 }
Chris@42 100 }
Chris@42 101 }
Chris@42 102
Chris@42 103 static const kr2c_desc desc = { 10, "r2cbIII_10", {14, 10, 18, 0}, &GENUS };
Chris@42 104
Chris@42 105 void X(codelet_r2cbIII_10) (planner *p) {
Chris@42 106 X(kr2c_register) (p, r2cbIII_10, &desc);
Chris@42 107 }
Chris@42 108
Chris@42 109 #else /* HAVE_FMA */
Chris@42 110
Chris@42 111 /* Generated by: ../../../genfft/gen_r2cb.native -compact -variables 4 -pipeline-latency 4 -sign 1 -n 10 -name r2cbIII_10 -dft-III -include r2cbIII.h */
Chris@42 112
Chris@42 113 /*
Chris@42 114 * This function contains 32 FP additions, 16 FP multiplications,
Chris@42 115 * (or, 26 additions, 10 multiplications, 6 fused multiply/add),
Chris@42 116 * 22 stack variables, 5 constants, and 20 memory accesses
Chris@42 117 */
Chris@42 118 #include "r2cbIII.h"
Chris@42 119
Chris@42 120 static void r2cbIII_10(R *R0, R *R1, R *Cr, R *Ci, stride rs, stride csr, stride csi, INT v, INT ivs, INT ovs)
Chris@42 121 {
Chris@42 122 DK(KP500000000, +0.500000000000000000000000000000000000000000000);
Chris@42 123 DK(KP1_902113032, +1.902113032590307144232878666758764286811397268);
Chris@42 124 DK(KP1_175570504, +1.175570504584946258337411909278145537195304875);
Chris@42 125 DK(KP2_000000000, +2.000000000000000000000000000000000000000000000);
Chris@42 126 DK(KP1_118033988, +1.118033988749894848204586834365638117720309180);
Chris@42 127 {
Chris@42 128 INT i;
Chris@42 129 for (i = v; i > 0; i = i - 1, R0 = R0 + ovs, R1 = R1 + ovs, Cr = Cr + ivs, Ci = Ci + ivs, MAKE_VOLATILE_STRIDE(40, rs), MAKE_VOLATILE_STRIDE(40, csr), MAKE_VOLATILE_STRIDE(40, csi)) {
Chris@42 130 E T1, To, T8, Tq, Ta, Tp, Te, Ts, Th, Tn;
Chris@42 131 T1 = Cr[WS(csr, 2)];
Chris@42 132 To = Ci[WS(csi, 2)];
Chris@42 133 {
Chris@42 134 E T2, T3, T4, T5, T6, T7;
Chris@42 135 T2 = Cr[WS(csr, 4)];
Chris@42 136 T3 = Cr[0];
Chris@42 137 T4 = T2 + T3;
Chris@42 138 T5 = Cr[WS(csr, 3)];
Chris@42 139 T6 = Cr[WS(csr, 1)];
Chris@42 140 T7 = T5 + T6;
Chris@42 141 T8 = T4 + T7;
Chris@42 142 Tq = T5 - T6;
Chris@42 143 Ta = KP1_118033988 * (T7 - T4);
Chris@42 144 Tp = T2 - T3;
Chris@42 145 }
Chris@42 146 {
Chris@42 147 E Tc, Td, Tm, Tf, Tg, Tl;
Chris@42 148 Tc = Ci[WS(csi, 4)];
Chris@42 149 Td = Ci[0];
Chris@42 150 Tm = Tc + Td;
Chris@42 151 Tf = Ci[WS(csi, 1)];
Chris@42 152 Tg = Ci[WS(csi, 3)];
Chris@42 153 Tl = Tg + Tf;
Chris@42 154 Te = Tc - Td;
Chris@42 155 Ts = KP1_118033988 * (Tl + Tm);
Chris@42 156 Th = Tf - Tg;
Chris@42 157 Tn = Tl - Tm;
Chris@42 158 }
Chris@42 159 R0[0] = KP2_000000000 * (T1 + T8);
Chris@42 160 R1[WS(rs, 2)] = KP2_000000000 * (Tn - To);
Chris@42 161 {
Chris@42 162 E Ti, Tj, Tb, Tk, T9;
Chris@42 163 Ti = FNMS(KP1_902113032, Th, KP1_175570504 * Te);
Chris@42 164 Tj = FMA(KP1_175570504, Th, KP1_902113032 * Te);
Chris@42 165 T9 = FNMS(KP2_000000000, T1, KP500000000 * T8);
Chris@42 166 Tb = T9 - Ta;
Chris@42 167 Tk = T9 + Ta;
Chris@42 168 R0[WS(rs, 1)] = Tb + Ti;
Chris@42 169 R0[WS(rs, 3)] = Tk + Tj;
Chris@42 170 R0[WS(rs, 4)] = Ti - Tb;
Chris@42 171 R0[WS(rs, 2)] = Tj - Tk;
Chris@42 172 }
Chris@42 173 {
Chris@42 174 E Tr, Tv, Tu, Tw, Tt;
Chris@42 175 Tr = FMA(KP1_902113032, Tp, KP1_175570504 * Tq);
Chris@42 176 Tv = FNMS(KP1_175570504, Tp, KP1_902113032 * Tq);
Chris@42 177 Tt = FMA(KP500000000, Tn, KP2_000000000 * To);
Chris@42 178 Tu = Ts + Tt;
Chris@42 179 Tw = Tt - Ts;
Chris@42 180 R1[0] = -(Tr + Tu);
Chris@42 181 R1[WS(rs, 3)] = Tw - Tv;
Chris@42 182 R1[WS(rs, 4)] = Tr - Tu;
Chris@42 183 R1[WS(rs, 1)] = Tv + Tw;
Chris@42 184 }
Chris@42 185 }
Chris@42 186 }
Chris@42 187 }
Chris@42 188
Chris@42 189 static const kr2c_desc desc = { 10, "r2cbIII_10", {26, 10, 6, 0}, &GENUS };
Chris@42 190
Chris@42 191 void X(codelet_r2cbIII_10) (planner *p) {
Chris@42 192 X(kr2c_register) (p, r2cbIII_10, &desc);
Chris@42 193 }
Chris@42 194
Chris@42 195 #endif /* HAVE_FMA */