Mercurial > hg > sv-dependency-builds

annotate src/fftw-3.3.8/rdft/scalar/r2cb/r2cbIII_8.c @ 82:d0c2a83c1364

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
Add FFTW 3.3.8 source, and a Linux build
author Chris Cannam
date Tue, 19 Nov 2019 14:52:55 +0000
parents
children
rev   line source
Chris@82 1 /*
Chris@82 2 * Copyright (c) 2003, 2007-14 Matteo Frigo
Chris@82 3 * Copyright (c) 2003, 2007-14 Massachusetts Institute of Technology
Chris@82 4 *
Chris@82 5 * This program is free software; you can redistribute it and/or modify
Chris@82 6 * it under the terms of the GNU General Public License as published by
Chris@82 7 * the Free Software Foundation; either version 2 of the License, or
Chris@82 8 * (at your option) any later version.
Chris@82 9 *
Chris@82 10 * This program is distributed in the hope that it will be useful,
Chris@82 11 * but WITHOUT ANY WARRANTY; without even the implied warranty of
Chris@82 12 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
Chris@82 13 * GNU General Public License for more details.
Chris@82 14 *
Chris@82 15 * You should have received a copy of the GNU General Public License
Chris@82 16 * along with this program; if not, write to the Free Software
Chris@82 17 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
Chris@82 18 *
Chris@82 19 */
Chris@82 20
Chris@82 21 /* This file was automatically generated --- DO NOT EDIT */
Chris@82 22 /* Generated on Thu May 24 08:07:43 EDT 2018 */
Chris@82 23
Chris@82 24 #include "rdft/codelet-rdft.h"
Chris@82 25
Chris@82 26 #if defined(ARCH_PREFERS_FMA) || defined(ISA_EXTENSION_PREFERS_FMA)
Chris@82 27
Chris@82 28 /* Generated by: ../../../genfft/gen_r2cb.native -fma -compact -variables 4 -pipeline-latency 4 -sign 1 -n 8 -name r2cbIII_8 -dft-III -include rdft/scalar/r2cbIII.h */
Chris@82 29
Chris@82 30 /*
Chris@82 31 * This function contains 22 FP additions, 12 FP multiplications,
Chris@82 32 * (or, 18 additions, 8 multiplications, 4 fused multiply/add),
Chris@82 33 * 19 stack variables, 4 constants, and 16 memory accesses
Chris@82 34 */
Chris@82 35 #include "rdft/scalar/r2cbIII.h"
Chris@82 36
Chris@82 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@82 38 {
Chris@82 39 DK(KP1_414213562, +1.414213562373095048801688724209698078569671875);
Chris@82 40 DK(KP414213562, +0.414213562373095048801688724209698078569671875);
Chris@82 41 DK(KP1_847759065, +1.847759065022573512256366378793576573644833252);
Chris@82 42 DK(KP2_000000000, +2.000000000000000000000000000000000000000000000);
Chris@82 43 {
Chris@82 44 INT i;
Chris@82 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@82 46 E T3, T7, Tf, Tl, T6, Tc, Ta, Tk, Tb, Tg;
Chris@82 47 {
Chris@82 48 E T1, T2, Td, Te;
Chris@82 49 T1 = Cr[0];
Chris@82 50 T2 = Cr[WS(csr, 3)];
Chris@82 51 T3 = T1 + T2;
Chris@82 52 T7 = T1 - T2;
Chris@82 53 Td = Ci[0];
Chris@82 54 Te = Ci[WS(csi, 3)];
Chris@82 55 Tf = Td + Te;
Chris@82 56 Tl = Te - Td;
Chris@82 57 }
Chris@82 58 {
Chris@82 59 E T4, T5, T8, T9;
Chris@82 60 T4 = Cr[WS(csr, 2)];
Chris@82 61 T5 = Cr[WS(csr, 1)];
Chris@82 62 T6 = T4 + T5;
Chris@82 63 Tc = T4 - T5;
Chris@82 64 T8 = Ci[WS(csi, 2)];
Chris@82 65 T9 = Ci[WS(csi, 1)];
Chris@82 66 Ta = T8 + T9;
Chris@82 67 Tk = T8 - T9;
Chris@82 68 }
Chris@82 69 R0[0] = KP2_000000000 * (T3 + T6);
Chris@82 70 R0[WS(rs, 2)] = KP2_000000000 * (Tl - Tk);
Chris@82 71 Tb = T7 - Ta;
Chris@82 72 Tg = Tc + Tf;
Chris@82 73 R1[0] = KP1_847759065 * (FNMS(KP414213562, Tg, Tb));
Chris@82 74 R1[WS(rs, 2)] = -(KP1_847759065 * (FMA(KP414213562, Tb, Tg)));
Chris@82 75 {
Chris@82 76 E Th, Ti, Tj, Tm;
Chris@82 77 Th = Tc - Tf;
Chris@82 78 Ti = T7 + Ta;
Chris@82 79 R1[WS(rs, 1)] = KP1_847759065 * (FMA(KP414213562, Ti, Th));
Chris@82 80 R1[WS(rs, 3)] = -(KP1_847759065 * (FNMS(KP414213562, Th, Ti)));
Chris@82 81 Tj = T3 - T6;
Chris@82 82 Tm = Tk + Tl;
Chris@82 83 R0[WS(rs, 1)] = KP1_414213562 * (Tj + Tm);
Chris@82 84 R0[WS(rs, 3)] = KP1_414213562 * (Tm - Tj);
Chris@82 85 }
Chris@82 86 }
Chris@82 87 }
Chris@82 88 }
Chris@82 89
Chris@82 90 static const kr2c_desc desc = { 8, "r2cbIII_8", {18, 8, 4, 0}, &GENUS };
Chris@82 91
Chris@82 92 void X(codelet_r2cbIII_8) (planner *p) {
Chris@82 93 X(kr2c_register) (p, r2cbIII_8, &desc);
Chris@82 94 }
Chris@82 95
Chris@82 96 #else
Chris@82 97
Chris@82 98 /* Generated by: ../../../genfft/gen_r2cb.native -compact -variables 4 -pipeline-latency 4 -sign 1 -n 8 -name r2cbIII_8 -dft-III -include rdft/scalar/r2cbIII.h */
Chris@82 99
Chris@82 100 /*
Chris@82 101 * This function contains 22 FP additions, 12 FP multiplications,
Chris@82 102 * (or, 18 additions, 8 multiplications, 4 fused multiply/add),
Chris@82 103 * 19 stack variables, 4 constants, and 16 memory accesses
Chris@82 104 */
Chris@82 105 #include "rdft/scalar/r2cbIII.h"
Chris@82 106
Chris@82 107 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@82 108 {
Chris@82 109 DK(KP1_414213562, +1.414213562373095048801688724209698078569671875);
Chris@82 110 DK(KP765366864, +0.765366864730179543456919968060797733522689125);
Chris@82 111 DK(KP1_847759065, +1.847759065022573512256366378793576573644833252);
Chris@82 112 DK(KP2_000000000, +2.000000000000000000000000000000000000000000000);
Chris@82 113 {
Chris@82 114 INT i;
Chris@82 115 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@82 116 E T3, T7, Tf, Tl, T6, Tc, Ta, Tk, Tb, Tg;
Chris@82 117 {
Chris@82 118 E T1, T2, Td, Te;
Chris@82 119 T1 = Cr[0];
Chris@82 120 T2 = Cr[WS(csr, 3)];
Chris@82 121 T3 = T1 + T2;
Chris@82 122 T7 = T1 - T2;
Chris@82 123 Td = Ci[0];
Chris@82 124 Te = Ci[WS(csi, 3)];
Chris@82 125 Tf = Td + Te;
Chris@82 126 Tl = Te - Td;
Chris@82 127 }
Chris@82 128 {
Chris@82 129 E T4, T5, T8, T9;
Chris@82 130 T4 = Cr[WS(csr, 2)];
Chris@82 131 T5 = Cr[WS(csr, 1)];
Chris@82 132 T6 = T4 + T5;
Chris@82 133 Tc = T4 - T5;
Chris@82 134 T8 = Ci[WS(csi, 2)];
Chris@82 135 T9 = Ci[WS(csi, 1)];
Chris@82 136 Ta = T8 + T9;
Chris@82 137 Tk = T8 - T9;
Chris@82 138 }
Chris@82 139 R0[0] = KP2_000000000 * (T3 + T6);
Chris@82 140 R0[WS(rs, 2)] = KP2_000000000 * (Tl - Tk);
Chris@82 141 Tb = T7 - Ta;
Chris@82 142 Tg = Tc + Tf;
Chris@82 143 R1[0] = FNMS(KP765366864, Tg, KP1_847759065 * Tb);
Chris@82 144 R1[WS(rs, 2)] = -(FMA(KP765366864, Tb, KP1_847759065 * Tg));
Chris@82 145 {
Chris@82 146 E Th, Ti, Tj, Tm;
Chris@82 147 Th = T7 + Ta;
Chris@82 148 Ti = Tc - Tf;
Chris@82 149 R1[WS(rs, 1)] = FMA(KP765366864, Th, KP1_847759065 * Ti);
Chris@82 150 R1[WS(rs, 3)] = FNMS(KP1_847759065, Th, KP765366864 * Ti);
Chris@82 151 Tj = T3 - T6;
Chris@82 152 Tm = Tk + Tl;
Chris@82 153 R0[WS(rs, 1)] = KP1_414213562 * (Tj + Tm);
Chris@82 154 R0[WS(rs, 3)] = KP1_414213562 * (Tm - Tj);
Chris@82 155 }
Chris@82 156 }
Chris@82 157 }
Chris@82 158 }
Chris@82 159
Chris@82 160 static const kr2c_desc desc = { 8, "r2cbIII_8", {18, 8, 4, 0}, &GENUS };
Chris@82 161
Chris@82 162 void X(codelet_r2cbIII_8) (planner *p) {
Chris@82 163 X(kr2c_register) (p, r2cbIII_8, &desc);
Chris@82 164 }
Chris@82 165
Chris@82 166 #endif