Mercurial > hg > sv-dependency-builds

annotate src/fftw-3.3.5/rdft/scalar/r2cb/r2cb_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 16:49:26 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 8 -name r2cb_8 -include r2cb.h */
Chris@42 29
Chris@42 30 /*
Chris@42 31 * This function contains 20 FP additions, 12 FP multiplications,
Chris@42 32 * (or, 8 additions, 0 multiplications, 12 fused multiply/add),
Chris@42 33 * 19 stack variables, 2 constants, and 16 memory accesses
Chris@42 34 */
Chris@42 35 #include "r2cb.h"
Chris@42 36
Chris@42 37 static void r2cb_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(KP1_414213562, +1.414213562373095048801688724209698078569671875);
Chris@42 40 DK(KP2_000000000, +2.000000000000000000000000000000000000000000000);
Chris@42 41 {
Chris@42 42 INT i;
Chris@42 43 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@42 44 E Th, Tb, Tg, Ti;
Chris@42 45 {
Chris@42 46 E T4, Ta, Td, T9, T3, Tc, T8, Te;
Chris@42 47 T4 = Cr[WS(csr, 2)];
Chris@42 48 Ta = Ci[WS(csi, 2)];
Chris@42 49 {
Chris@42 50 E T1, T2, T6, T7;
Chris@42 51 T1 = Cr[0];
Chris@42 52 T2 = Cr[WS(csr, 4)];
Chris@42 53 T6 = Cr[WS(csr, 1)];
Chris@42 54 T7 = Cr[WS(csr, 3)];
Chris@42 55 Td = Ci[WS(csi, 1)];
Chris@42 56 T9 = T1 - T2;
Chris@42 57 T3 = T1 + T2;
Chris@42 58 Tc = T6 - T7;
Chris@42 59 T8 = T6 + T7;
Chris@42 60 Te = Ci[WS(csi, 3)];
Chris@42 61 }
Chris@42 62 {
Chris@42 63 E Tj, T5, Tk, Tf;
Chris@42 64 Tj = FNMS(KP2_000000000, T4, T3);
Chris@42 65 T5 = FMA(KP2_000000000, T4, T3);
Chris@42 66 Th = FMA(KP2_000000000, Ta, T9);
Chris@42 67 Tb = FNMS(KP2_000000000, Ta, T9);
Chris@42 68 Tk = Td - Te;
Chris@42 69 Tf = Td + Te;
Chris@42 70 R0[0] = FMA(KP2_000000000, T8, T5);
Chris@42 71 R0[WS(rs, 2)] = FNMS(KP2_000000000, T8, T5);
Chris@42 72 R0[WS(rs, 3)] = FMA(KP2_000000000, Tk, Tj);
Chris@42 73 R0[WS(rs, 1)] = FNMS(KP2_000000000, Tk, Tj);
Chris@42 74 Tg = Tc - Tf;
Chris@42 75 Ti = Tc + Tf;
Chris@42 76 }
Chris@42 77 }
Chris@42 78 R1[0] = FMA(KP1_414213562, Tg, Tb);
Chris@42 79 R1[WS(rs, 2)] = FNMS(KP1_414213562, Tg, Tb);
Chris@42 80 R1[WS(rs, 3)] = FMA(KP1_414213562, Ti, Th);
Chris@42 81 R1[WS(rs, 1)] = FNMS(KP1_414213562, Ti, Th);
Chris@42 82 }
Chris@42 83 }
Chris@42 84 }
Chris@42 85
Chris@42 86 static const kr2c_desc desc = { 8, "r2cb_8", {8, 0, 12, 0}, &GENUS };
Chris@42 87
Chris@42 88 void X(codelet_r2cb_8) (planner *p) {
Chris@42 89 X(kr2c_register) (p, r2cb_8, &desc);
Chris@42 90 }
Chris@42 91
Chris@42 92 #else /* HAVE_FMA */
Chris@42 93
Chris@42 94 /* Generated by: ../../../genfft/gen_r2cb.native -compact -variables 4 -pipeline-latency 4 -sign 1 -n 8 -name r2cb_8 -include r2cb.h */
Chris@42 95
Chris@42 96 /*
Chris@42 97 * This function contains 20 FP additions, 6 FP multiplications,
Chris@42 98 * (or, 20 additions, 6 multiplications, 0 fused multiply/add),
Chris@42 99 * 21 stack variables, 2 constants, and 16 memory accesses
Chris@42 100 */
Chris@42 101 #include "r2cb.h"
Chris@42 102
Chris@42 103 static void r2cb_8(R *R0, R *R1, R *Cr, R *Ci, stride rs, stride csr, stride csi, INT v, INT ivs, INT ovs)
Chris@42 104 {
Chris@42 105 DK(KP1_414213562, +1.414213562373095048801688724209698078569671875);
Chris@42 106 DK(KP2_000000000, +2.000000000000000000000000000000000000000000000);
Chris@42 107 {
Chris@42 108 INT i;
Chris@42 109 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@42 110 E T5, Tg, T3, Te, T9, Ti, Td, Tj, T6, Ta;
Chris@42 111 {
Chris@42 112 E T4, Tf, T1, T2;
Chris@42 113 T4 = Cr[WS(csr, 2)];
Chris@42 114 T5 = KP2_000000000 * T4;
Chris@42 115 Tf = Ci[WS(csi, 2)];
Chris@42 116 Tg = KP2_000000000 * Tf;
Chris@42 117 T1 = Cr[0];
Chris@42 118 T2 = Cr[WS(csr, 4)];
Chris@42 119 T3 = T1 + T2;
Chris@42 120 Te = T1 - T2;
Chris@42 121 {
Chris@42 122 E T7, T8, Tb, Tc;
Chris@42 123 T7 = Cr[WS(csr, 1)];
Chris@42 124 T8 = Cr[WS(csr, 3)];
Chris@42 125 T9 = KP2_000000000 * (T7 + T8);
Chris@42 126 Ti = T7 - T8;
Chris@42 127 Tb = Ci[WS(csi, 1)];
Chris@42 128 Tc = Ci[WS(csi, 3)];
Chris@42 129 Td = KP2_000000000 * (Tb - Tc);
Chris@42 130 Tj = Tb + Tc;
Chris@42 131 }
Chris@42 132 }
Chris@42 133 T6 = T3 + T5;
Chris@42 134 R0[WS(rs, 2)] = T6 - T9;
Chris@42 135 R0[0] = T6 + T9;
Chris@42 136 Ta = T3 - T5;
Chris@42 137 R0[WS(rs, 1)] = Ta - Td;
Chris@42 138 R0[WS(rs, 3)] = Ta + Td;
Chris@42 139 {
Chris@42 140 E Th, Tk, Tl, Tm;
Chris@42 141 Th = Te - Tg;
Chris@42 142 Tk = KP1_414213562 * (Ti - Tj);
Chris@42 143 R1[WS(rs, 2)] = Th - Tk;
Chris@42 144 R1[0] = Th + Tk;
Chris@42 145 Tl = Te + Tg;
Chris@42 146 Tm = KP1_414213562 * (Ti + Tj);
Chris@42 147 R1[WS(rs, 1)] = Tl - Tm;
Chris@42 148 R1[WS(rs, 3)] = Tl + Tm;
Chris@42 149 }
Chris@42 150 }
Chris@42 151 }
Chris@42 152 }
Chris@42 153
Chris@42 154 static const kr2c_desc desc = { 8, "r2cb_8", {20, 6, 0, 0}, &GENUS };
Chris@42 155
Chris@42 156 void X(codelet_r2cb_8) (planner *p) {
Chris@42 157 X(kr2c_register) (p, r2cb_8, &desc);
Chris@42 158 }
Chris@42 159
Chris@42 160 #endif /* HAVE_FMA */