Mercurial > hg > sv-dependency-builds

annotate src/fftw-3.3.5/rdft/scalar/r2r/e10_8.c @ 83:ae30d91d2ffe

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
Replace these with versions built using an older toolset (so as to avoid ABI compatibilities when linking on Ubuntu 14.04 for packaging purposes)
author Chris Cannam
date Fri, 07 Feb 2020 11:51:13 +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 16:52:38 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_r2r.native -fma -reorder-insns -schedule-for-pipeline -compact -variables 4 -pipeline-latency 4 -redft10 -n 8 -name e10_8 -include r2r.h */
Chris@42 29
Chris@42 30 /*
Chris@42 31 * This function contains 26 FP additions, 18 FP multiplications,
Chris@42 32 * (or, 16 additions, 8 multiplications, 10 fused multiply/add),
Chris@42 33 * 28 stack variables, 9 constants, and 16 memory accesses
Chris@42 34 */
Chris@42 35 #include "r2r.h"
Chris@42 36
Chris@42 37 static void e10_8(const R *I, R *O, stride is, stride os, INT v, INT ivs, INT ovs)
Chris@42 38 {
Chris@42 39 DK(KP668178637, +0.668178637919298919997757686523080761552472251);
Chris@42 40 DK(KP1_662939224, +1.662939224605090474157576755235811513477121624);
Chris@42 41 DK(KP198912367, +0.198912367379658006911597622644676228597850501);
Chris@42 42 DK(KP1_961570560, +1.961570560806460898252364472268478073947867462);
Chris@42 43 DK(KP707106781, +0.707106781186547524400844362104849039284835938);
Chris@42 44 DK(KP1_414213562, +1.414213562373095048801688724209698078569671875);
Chris@42 45 DK(KP2_000000000, +2.000000000000000000000000000000000000000000000);
Chris@42 46 DK(KP414213562, +0.414213562373095048801688724209698078569671875);
Chris@42 47 DK(KP1_847759065, +1.847759065022573512256366378793576573644833252);
Chris@42 48 {
Chris@42 49 INT i;
Chris@42 50 for (i = v; i > 0; i = i - 1, I = I + ivs, O = O + ovs, MAKE_VOLATILE_STRIDE(16, is), MAKE_VOLATILE_STRIDE(16, os)) {
Chris@42 51 E T3, Te, Tl, Tp, Tm, T6, Tn, T9;
Chris@42 52 {
Chris@42 53 E T4, Tj, Tk, T5, T7, T8;
Chris@42 54 {
Chris@42 55 E T1, T2, Tc, Td;
Chris@42 56 T1 = I[0];
Chris@42 57 T2 = I[WS(is, 7)];
Chris@42 58 Tc = I[WS(is, 4)];
Chris@42 59 Td = I[WS(is, 3)];
Chris@42 60 T4 = I[WS(is, 2)];
Chris@42 61 Tj = T1 + T2;
Chris@42 62 T3 = T1 - T2;
Chris@42 63 Tk = Tc + Td;
Chris@42 64 Te = Tc - Td;
Chris@42 65 T5 = I[WS(is, 5)];
Chris@42 66 T7 = I[WS(is, 1)];
Chris@42 67 T8 = I[WS(is, 6)];
Chris@42 68 }
Chris@42 69 Tl = Tj - Tk;
Chris@42 70 Tp = Tj + Tk;
Chris@42 71 Tm = T4 + T5;
Chris@42 72 T6 = T4 - T5;
Chris@42 73 Tn = T7 + T8;
Chris@42 74 T9 = T7 - T8;
Chris@42 75 }
Chris@42 76 {
Chris@42 77 E Tg, Ti, Tb, Th;
Chris@42 78 {
Chris@42 79 E Tq, To, Ta, Tf;
Chris@42 80 Tq = Tm + Tn;
Chris@42 81 To = Tm - Tn;
Chris@42 82 Ta = T6 + T9;
Chris@42 83 Tf = T6 - T9;
Chris@42 84 O[WS(os, 6)] = KP1_847759065 * (FMA(KP414213562, Tl, To));
Chris@42 85 O[WS(os, 2)] = KP1_847759065 * (FNMS(KP414213562, To, Tl));
Chris@42 86 O[0] = KP2_000000000 * (Tp + Tq);
Chris@42 87 O[WS(os, 4)] = KP1_414213562 * (Tp - Tq);
Chris@42 88 Tg = FNMS(KP707106781, Tf, Te);
Chris@42 89 Ti = FMA(KP707106781, Tf, Te);
Chris@42 90 Tb = FNMS(KP707106781, Ta, T3);
Chris@42 91 Th = FMA(KP707106781, Ta, T3);
Chris@42 92 }
Chris@42 93 O[WS(os, 7)] = KP1_961570560 * (FMA(KP198912367, Th, Ti));
Chris@42 94 O[WS(os, 1)] = KP1_961570560 * (FNMS(KP198912367, Ti, Th));
Chris@42 95 O[WS(os, 5)] = -(KP1_662939224 * (FNMS(KP668178637, Tb, Tg)));
Chris@42 96 O[WS(os, 3)] = KP1_662939224 * (FMA(KP668178637, Tg, Tb));
Chris@42 97 }
Chris@42 98 }
Chris@42 99 }
Chris@42 100 }
Chris@42 101
Chris@42 102 static const kr2r_desc desc = { 8, "e10_8", {16, 8, 10, 0}, &GENUS, REDFT10 };
Chris@42 103
Chris@42 104 void X(codelet_e10_8) (planner *p) {
Chris@42 105 X(kr2r_register) (p, e10_8, &desc);
Chris@42 106 }
Chris@42 107
Chris@42 108 #else /* HAVE_FMA */
Chris@42 109
Chris@42 110 /* Generated by: ../../../genfft/gen_r2r.native -compact -variables 4 -pipeline-latency 4 -redft10 -n 8 -name e10_8 -include r2r.h */
Chris@42 111
Chris@42 112 /*
Chris@42 113 * This function contains 26 FP additions, 16 FP multiplications,
Chris@42 114 * (or, 20 additions, 10 multiplications, 6 fused multiply/add),
Chris@42 115 * 28 stack variables, 9 constants, and 16 memory accesses
Chris@42 116 */
Chris@42 117 #include "r2r.h"
Chris@42 118
Chris@42 119 static void e10_8(const R *I, R *O, stride is, stride os, INT v, INT ivs, INT ovs)
Chris@42 120 {
Chris@42 121 DK(KP765366864, +0.765366864730179543456919968060797733522689125);
Chris@42 122 DK(KP1_847759065, +1.847759065022573512256366378793576573644833252);
Chris@42 123 DK(KP390180644, +0.390180644032256535696569736954044481855383236);
Chris@42 124 DK(KP1_961570560, +1.961570560806460898252364472268478073947867462);
Chris@42 125 DK(KP2_000000000, +2.000000000000000000000000000000000000000000000);
Chris@42 126 DK(KP1_414213562, +1.414213562373095048801688724209698078569671875);
Chris@42 127 DK(KP1_111140466, +1.111140466039204449485661627897065748749874382);
Chris@42 128 DK(KP1_662939224, +1.662939224605090474157576755235811513477121624);
Chris@42 129 DK(KP707106781, +0.707106781186547524400844362104849039284835938);
Chris@42 130 {
Chris@42 131 INT i;
Chris@42 132 for (i = v; i > 0; i = i - 1, I = I + ivs, O = O + ovs, MAKE_VOLATILE_STRIDE(16, is), MAKE_VOLATILE_STRIDE(16, os)) {
Chris@42 133 E T3, Tj, Tf, Tk, Ta, Tn, Tc, Tm;
Chris@42 134 {
Chris@42 135 E T1, T2, Td, Te;
Chris@42 136 T1 = I[0];
Chris@42 137 T2 = I[WS(is, 7)];
Chris@42 138 T3 = T1 - T2;
Chris@42 139 Tj = T1 + T2;
Chris@42 140 Td = I[WS(is, 4)];
Chris@42 141 Te = I[WS(is, 3)];
Chris@42 142 Tf = Td - Te;
Chris@42 143 Tk = Td + Te;
Chris@42 144 {
Chris@42 145 E T4, T5, T6, T7, T8, T9;
Chris@42 146 T4 = I[WS(is, 2)];
Chris@42 147 T5 = I[WS(is, 5)];
Chris@42 148 T6 = T4 - T5;
Chris@42 149 T7 = I[WS(is, 1)];
Chris@42 150 T8 = I[WS(is, 6)];
Chris@42 151 T9 = T7 - T8;
Chris@42 152 Ta = KP707106781 * (T6 + T9);
Chris@42 153 Tn = T7 + T8;
Chris@42 154 Tc = KP707106781 * (T6 - T9);
Chris@42 155 Tm = T4 + T5;
Chris@42 156 }
Chris@42 157 }
Chris@42 158 {
Chris@42 159 E Tb, Tg, Tp, Tq;
Chris@42 160 Tb = T3 - Ta;
Chris@42 161 Tg = Tc - Tf;
Chris@42 162 O[WS(os, 3)] = FNMS(KP1_111140466, Tg, KP1_662939224 * Tb);
Chris@42 163 O[WS(os, 5)] = FMA(KP1_662939224, Tg, KP1_111140466 * Tb);
Chris@42 164 Tp = Tj + Tk;
Chris@42 165 Tq = Tm + Tn;
Chris@42 166 O[WS(os, 4)] = KP1_414213562 * (Tp - Tq);
Chris@42 167 O[0] = KP2_000000000 * (Tp + Tq);
Chris@42 168 }
Chris@42 169 {
Chris@42 170 E Th, Ti, Tl, To;
Chris@42 171 Th = T3 + Ta;
Chris@42 172 Ti = Tf + Tc;
Chris@42 173 O[WS(os, 1)] = FNMS(KP390180644, Ti, KP1_961570560 * Th);
Chris@42 174 O[WS(os, 7)] = FMA(KP1_961570560, Ti, KP390180644 * Th);
Chris@42 175 Tl = Tj - Tk;
Chris@42 176 To = Tm - Tn;
Chris@42 177 O[WS(os, 2)] = FNMS(KP765366864, To, KP1_847759065 * Tl);
Chris@42 178 O[WS(os, 6)] = FMA(KP765366864, Tl, KP1_847759065 * To);
Chris@42 179 }
Chris@42 180 }
Chris@42 181 }
Chris@42 182 }
Chris@42 183
Chris@42 184 static const kr2r_desc desc = { 8, "e10_8", {20, 10, 6, 0}, &GENUS, REDFT10 };
Chris@42 185
Chris@42 186 void X(codelet_e10_8) (planner *p) {
Chris@42 187 X(kr2r_register) (p, e10_8, &desc);
Chris@42 188 }
Chris@42 189
Chris@42 190 #endif /* HAVE_FMA */