Mercurial > hg > sv-dependency-builds

annotate src/fftw-3.3.5/rdft/scalar/r2cf/r2cfII_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 16:47:24 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_r2cf.native -fma -reorder-insns -schedule-for-pipeline -compact -variables 4 -pipeline-latency 4 -n 8 -name r2cfII_8 -dft-II -include r2cfII.h */
Chris@42 29
Chris@42 30 /*
Chris@42 31 * This function contains 22 FP additions, 16 FP multiplications,
Chris@42 32 * (or, 6 additions, 0 multiplications, 16 fused multiply/add),
Chris@42 33 * 22 stack variables, 3 constants, and 16 memory accesses
Chris@42 34 */
Chris@42 35 #include "r2cfII.h"
Chris@42 36
Chris@42 37 static void r2cfII_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(KP923879532, +0.923879532511286756128183189396788286822416626);
Chris@42 40 DK(KP707106781, +0.707106781186547524400844362104849039284835938);
Chris@42 41 DK(KP414213562, +0.414213562373095048801688724209698078569671875);
Chris@42 42 {
Chris@42 43 INT i;
Chris@42 44 for (i = v; i > 0; i = i - 1, R0 = R0 + ivs, R1 = R1 + ivs, Cr = Cr + ovs, Ci = Ci + ovs, MAKE_VOLATILE_STRIDE(32, rs), MAKE_VOLATILE_STRIDE(32, csr), MAKE_VOLATILE_STRIDE(32, csi)) {
Chris@42 45 E Te, T8, Td, T5, Tj, Tl, Tf, Tb;
Chris@42 46 {
Chris@42 47 E T1, Th, T9, Ti, T4, Ta;
Chris@42 48 T1 = R0[0];
Chris@42 49 Th = R0[WS(rs, 2)];
Chris@42 50 {
Chris@42 51 E T2, T3, T6, T7;
Chris@42 52 T2 = R0[WS(rs, 1)];
Chris@42 53 T3 = R0[WS(rs, 3)];
Chris@42 54 T6 = R1[0];
Chris@42 55 T7 = R1[WS(rs, 2)];
Chris@42 56 T9 = R1[WS(rs, 3)];
Chris@42 57 Ti = T2 + T3;
Chris@42 58 T4 = T2 - T3;
Chris@42 59 Te = FMA(KP414213562, T6, T7);
Chris@42 60 T8 = FNMS(KP414213562, T7, T6);
Chris@42 61 Ta = R1[WS(rs, 1)];
Chris@42 62 }
Chris@42 63 Td = FNMS(KP707106781, T4, T1);
Chris@42 64 T5 = FMA(KP707106781, T4, T1);
Chris@42 65 Tj = FMA(KP707106781, Ti, Th);
Chris@42 66 Tl = FNMS(KP707106781, Ti, Th);
Chris@42 67 Tf = FMA(KP414213562, T9, Ta);
Chris@42 68 Tb = FMS(KP414213562, Ta, T9);
Chris@42 69 }
Chris@42 70 {
Chris@42 71 E Tk, Tg, Tc, Tm;
Chris@42 72 Tk = Te + Tf;
Chris@42 73 Tg = Te - Tf;
Chris@42 74 Tc = T8 + Tb;
Chris@42 75 Tm = Tb - T8;
Chris@42 76 Cr[WS(csr, 1)] = FMA(KP923879532, Tg, Td);
Chris@42 77 Cr[WS(csr, 2)] = FNMS(KP923879532, Tg, Td);
Chris@42 78 Ci[WS(csi, 3)] = FNMS(KP923879532, Tk, Tj);
Chris@42 79 Ci[0] = -(FMA(KP923879532, Tk, Tj));
Chris@42 80 Ci[WS(csi, 1)] = FMA(KP923879532, Tm, Tl);
Chris@42 81 Ci[WS(csi, 2)] = FMS(KP923879532, Tm, Tl);
Chris@42 82 Cr[0] = FMA(KP923879532, Tc, T5);
Chris@42 83 Cr[WS(csr, 3)] = FNMS(KP923879532, Tc, T5);
Chris@42 84 }
Chris@42 85 }
Chris@42 86 }
Chris@42 87 }
Chris@42 88
Chris@42 89 static const kr2c_desc desc = { 8, "r2cfII_8", {6, 0, 16, 0}, &GENUS };
Chris@42 90
Chris@42 91 void X(codelet_r2cfII_8) (planner *p) {
Chris@42 92 X(kr2c_register) (p, r2cfII_8, &desc);
Chris@42 93 }
Chris@42 94
Chris@42 95 #else /* HAVE_FMA */
Chris@42 96
Chris@42 97 /* Generated by: ../../../genfft/gen_r2cf.native -compact -variables 4 -pipeline-latency 4 -n 8 -name r2cfII_8 -dft-II -include r2cfII.h */
Chris@42 98
Chris@42 99 /*
Chris@42 100 * This function contains 22 FP additions, 10 FP multiplications,
Chris@42 101 * (or, 18 additions, 6 multiplications, 4 fused multiply/add),
Chris@42 102 * 18 stack variables, 3 constants, and 16 memory accesses
Chris@42 103 */
Chris@42 104 #include "r2cfII.h"
Chris@42 105
Chris@42 106 static void r2cfII_8(R *R0, R *R1, R *Cr, R *Ci, stride rs, stride csr, stride csi, INT v, INT ivs, INT ovs)
Chris@42 107 {
Chris@42 108 DK(KP382683432, +0.382683432365089771728459984030398866761344562);
Chris@42 109 DK(KP923879532, +0.923879532511286756128183189396788286822416626);
Chris@42 110 DK(KP707106781, +0.707106781186547524400844362104849039284835938);
Chris@42 111 {
Chris@42 112 INT i;
Chris@42 113 for (i = v; i > 0; i = i - 1, R0 = R0 + ivs, R1 = R1 + ivs, Cr = Cr + ovs, Ci = Ci + ovs, MAKE_VOLATILE_STRIDE(32, rs), MAKE_VOLATILE_STRIDE(32, csr), MAKE_VOLATILE_STRIDE(32, csi)) {
Chris@42 114 E T1, Tj, T4, Ti, T8, Te, Tb, Tf, T2, T3;
Chris@42 115 T1 = R0[0];
Chris@42 116 Tj = R0[WS(rs, 2)];
Chris@42 117 T2 = R0[WS(rs, 1)];
Chris@42 118 T3 = R0[WS(rs, 3)];
Chris@42 119 T4 = KP707106781 * (T2 - T3);
Chris@42 120 Ti = KP707106781 * (T2 + T3);
Chris@42 121 {
Chris@42 122 E T6, T7, T9, Ta;
Chris@42 123 T6 = R1[0];
Chris@42 124 T7 = R1[WS(rs, 2)];
Chris@42 125 T8 = FNMS(KP382683432, T7, KP923879532 * T6);
Chris@42 126 Te = FMA(KP382683432, T6, KP923879532 * T7);
Chris@42 127 T9 = R1[WS(rs, 1)];
Chris@42 128 Ta = R1[WS(rs, 3)];
Chris@42 129 Tb = FNMS(KP923879532, Ta, KP382683432 * T9);
Chris@42 130 Tf = FMA(KP923879532, T9, KP382683432 * Ta);
Chris@42 131 }
Chris@42 132 {
Chris@42 133 E T5, Tc, Th, Tk;
Chris@42 134 T5 = T1 + T4;
Chris@42 135 Tc = T8 + Tb;
Chris@42 136 Cr[WS(csr, 3)] = T5 - Tc;
Chris@42 137 Cr[0] = T5 + Tc;
Chris@42 138 Th = Te + Tf;
Chris@42 139 Tk = Ti + Tj;
Chris@42 140 Ci[0] = -(Th + Tk);
Chris@42 141 Ci[WS(csi, 3)] = Tk - Th;
Chris@42 142 }
Chris@42 143 {
Chris@42 144 E Td, Tg, Tl, Tm;
Chris@42 145 Td = T1 - T4;
Chris@42 146 Tg = Te - Tf;
Chris@42 147 Cr[WS(csr, 2)] = Td - Tg;
Chris@42 148 Cr[WS(csr, 1)] = Td + Tg;
Chris@42 149 Tl = Tb - T8;
Chris@42 150 Tm = Tj - Ti;
Chris@42 151 Ci[WS(csi, 2)] = Tl - Tm;
Chris@42 152 Ci[WS(csi, 1)] = Tl + Tm;
Chris@42 153 }
Chris@42 154 }
Chris@42 155 }
Chris@42 156 }
Chris@42 157
Chris@42 158 static const kr2c_desc desc = { 8, "r2cfII_8", {18, 6, 4, 0}, &GENUS };
Chris@42 159
Chris@42 160 void X(codelet_r2cfII_8) (planner *p) {
Chris@42 161 X(kr2c_register) (p, r2cfII_8, &desc);
Chris@42 162 }
Chris@42 163
Chris@42 164 #endif /* HAVE_FMA */