Mercurial > hg > sv-dependency-builds

annotate src/fftw-3.3.5/dft/scalar/codelets/n1_4.c @ 84:08ae793730bd

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
Add null config files
author Chris Cannam
date Mon, 02 Mar 2020 14:03:47 +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 16:35:50 EDT 2016 */
Chris@42 23
Chris@42 24 #include "codelet-dft.h"
Chris@42 25
Chris@42 26 #ifdef HAVE_FMA
Chris@42 27
Chris@42 28 /* Generated by: ../../../genfft/gen_notw.native -fma -reorder-insns -schedule-for-pipeline -compact -variables 4 -pipeline-latency 4 -n 4 -name n1_4 -include n.h */
Chris@42 29
Chris@42 30 /*
Chris@42 31 * This function contains 16 FP additions, 0 FP multiplications,
Chris@42 32 * (or, 16 additions, 0 multiplications, 0 fused multiply/add),
Chris@42 33 * 13 stack variables, 0 constants, and 16 memory accesses
Chris@42 34 */
Chris@42 35 #include "n.h"
Chris@42 36
Chris@42 37 static void n1_4(const R *ri, const R *ii, R *ro, R *io, stride is, stride os, INT v, INT ivs, INT ovs)
Chris@42 38 {
Chris@42 39 {
Chris@42 40 INT i;
Chris@42 41 for (i = v; i > 0; i = i - 1, ri = ri + ivs, ii = ii + ivs, ro = ro + ovs, io = io + ovs, MAKE_VOLATILE_STRIDE(16, is), MAKE_VOLATILE_STRIDE(16, os)) {
Chris@42 42 E T4, Tb, T3, Tf, T9, T5, Tc, Td;
Chris@42 43 {
Chris@42 44 E T1, T2, T7, T8;
Chris@42 45 T1 = ri[0];
Chris@42 46 T2 = ri[WS(is, 2)];
Chris@42 47 T7 = ii[0];
Chris@42 48 T8 = ii[WS(is, 2)];
Chris@42 49 T4 = ri[WS(is, 1)];
Chris@42 50 Tb = T1 - T2;
Chris@42 51 T3 = T1 + T2;
Chris@42 52 Tf = T7 + T8;
Chris@42 53 T9 = T7 - T8;
Chris@42 54 T5 = ri[WS(is, 3)];
Chris@42 55 Tc = ii[WS(is, 1)];
Chris@42 56 Td = ii[WS(is, 3)];
Chris@42 57 }
Chris@42 58 {
Chris@42 59 E T6, Ta, Te, Tg;
Chris@42 60 T6 = T4 + T5;
Chris@42 61 Ta = T4 - T5;
Chris@42 62 Te = Tc - Td;
Chris@42 63 Tg = Tc + Td;
Chris@42 64 io[WS(os, 3)] = Ta + T9;
Chris@42 65 io[WS(os, 1)] = T9 - Ta;
Chris@42 66 ro[0] = T3 + T6;
Chris@42 67 ro[WS(os, 2)] = T3 - T6;
Chris@42 68 io[0] = Tf + Tg;
Chris@42 69 io[WS(os, 2)] = Tf - Tg;
Chris@42 70 ro[WS(os, 3)] = Tb - Te;
Chris@42 71 ro[WS(os, 1)] = Tb + Te;
Chris@42 72 }
Chris@42 73 }
Chris@42 74 }
Chris@42 75 }
Chris@42 76
Chris@42 77 static const kdft_desc desc = { 4, "n1_4", {16, 0, 0, 0}, &GENUS, 0, 0, 0, 0 };
Chris@42 78
Chris@42 79 void X(codelet_n1_4) (planner *p) {
Chris@42 80 X(kdft_register) (p, n1_4, &desc);
Chris@42 81 }
Chris@42 82
Chris@42 83 #else /* HAVE_FMA */
Chris@42 84
Chris@42 85 /* Generated by: ../../../genfft/gen_notw.native -compact -variables 4 -pipeline-latency 4 -n 4 -name n1_4 -include n.h */
Chris@42 86
Chris@42 87 /*
Chris@42 88 * This function contains 16 FP additions, 0 FP multiplications,
Chris@42 89 * (or, 16 additions, 0 multiplications, 0 fused multiply/add),
Chris@42 90 * 13 stack variables, 0 constants, and 16 memory accesses
Chris@42 91 */
Chris@42 92 #include "n.h"
Chris@42 93
Chris@42 94 static void n1_4(const R *ri, const R *ii, R *ro, R *io, stride is, stride os, INT v, INT ivs, INT ovs)
Chris@42 95 {
Chris@42 96 {
Chris@42 97 INT i;
Chris@42 98 for (i = v; i > 0; i = i - 1, ri = ri + ivs, ii = ii + ivs, ro = ro + ovs, io = io + ovs, MAKE_VOLATILE_STRIDE(16, is), MAKE_VOLATILE_STRIDE(16, os)) {
Chris@42 99 E T3, Tb, T9, Tf, T6, Ta, Te, Tg;
Chris@42 100 {
Chris@42 101 E T1, T2, T7, T8;
Chris@42 102 T1 = ri[0];
Chris@42 103 T2 = ri[WS(is, 2)];
Chris@42 104 T3 = T1 + T2;
Chris@42 105 Tb = T1 - T2;
Chris@42 106 T7 = ii[0];
Chris@42 107 T8 = ii[WS(is, 2)];
Chris@42 108 T9 = T7 - T8;
Chris@42 109 Tf = T7 + T8;
Chris@42 110 }
Chris@42 111 {
Chris@42 112 E T4, T5, Tc, Td;
Chris@42 113 T4 = ri[WS(is, 1)];
Chris@42 114 T5 = ri[WS(is, 3)];
Chris@42 115 T6 = T4 + T5;
Chris@42 116 Ta = T4 - T5;
Chris@42 117 Tc = ii[WS(is, 1)];
Chris@42 118 Td = ii[WS(is, 3)];
Chris@42 119 Te = Tc - Td;
Chris@42 120 Tg = Tc + Td;
Chris@42 121 }
Chris@42 122 ro[WS(os, 2)] = T3 - T6;
Chris@42 123 io[WS(os, 2)] = Tf - Tg;
Chris@42 124 ro[0] = T3 + T6;
Chris@42 125 io[0] = Tf + Tg;
Chris@42 126 io[WS(os, 1)] = T9 - Ta;
Chris@42 127 ro[WS(os, 1)] = Tb + Te;
Chris@42 128 io[WS(os, 3)] = Ta + T9;
Chris@42 129 ro[WS(os, 3)] = Tb - Te;
Chris@42 130 }
Chris@42 131 }
Chris@42 132 }
Chris@42 133
Chris@42 134 static const kdft_desc desc = { 4, "n1_4", {16, 0, 0, 0}, &GENUS, 0, 0, 0, 0 };
Chris@42 135
Chris@42 136 void X(codelet_n1_4) (planner *p) {
Chris@42 137 X(kdft_register) (p, n1_4, &desc);
Chris@42 138 }
Chris@42 139
Chris@42 140 #endif /* HAVE_FMA */