Mercurial > hg > sv-dependency-builds

annotate src/fftw-3.3.5/rdft/scalar/r2cb/r2cb_7.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 16:49:26 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 7 -name r2cb_7 -include r2cb.h */
Chris@42 29
Chris@42 30 /*
Chris@42 31 * This function contains 24 FP additions, 22 FP multiplications,
Chris@42 32 * (or, 2 additions, 0 multiplications, 22 fused multiply/add),
Chris@42 33 * 31 stack variables, 7 constants, and 14 memory accesses
Chris@42 34 */
Chris@42 35 #include "r2cb.h"
Chris@42 36
Chris@42 37 static void r2cb_7(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_949855824, +1.949855824363647214036263365987862434465571601);
Chris@42 40 DK(KP1_801937735, +1.801937735804838252472204639014890102331838324);
Chris@42 41 DK(KP2_000000000, +2.000000000000000000000000000000000000000000000);
Chris@42 42 DK(KP692021471, +0.692021471630095869627814897002069140197260599);
Chris@42 43 DK(KP801937735, +0.801937735804838252472204639014890102331838324);
Chris@42 44 DK(KP356895867, +0.356895867892209443894399510021300583399127187);
Chris@42 45 DK(KP554958132, +0.554958132087371191422194871006410481067288862);
Chris@42 46 {
Chris@42 47 INT i;
Chris@42 48 for (i = v; i > 0; i = i - 1, R0 = R0 + ovs, R1 = R1 + ovs, Cr = Cr + ivs, Ci = Ci + ivs, MAKE_VOLATILE_STRIDE(28, rs), MAKE_VOLATILE_STRIDE(28, csr), MAKE_VOLATILE_STRIDE(28, csi)) {
Chris@42 49 E Tn, Td, Tg, Ti, Tl, T8;
Chris@42 50 {
Chris@42 51 E T1, T9, Tb, Ta, T2, T4, Th, Tm, Tc, T3, Te;
Chris@42 52 T1 = Cr[0];
Chris@42 53 T9 = Ci[WS(csi, 2)];
Chris@42 54 Tb = Ci[WS(csi, 3)];
Chris@42 55 Ta = Ci[WS(csi, 1)];
Chris@42 56 T2 = Cr[WS(csr, 1)];
Chris@42 57 T4 = Cr[WS(csr, 3)];
Chris@42 58 Th = FMA(KP554958132, T9, Tb);
Chris@42 59 Tm = FMS(KP554958132, Ta, T9);
Chris@42 60 Tc = FMA(KP554958132, Tb, Ta);
Chris@42 61 T3 = Cr[WS(csr, 2)];
Chris@42 62 Te = FNMS(KP356895867, T2, T4);
Chris@42 63 Tn = FMA(KP801937735, Tm, Tb);
Chris@42 64 {
Chris@42 65 E Tf, Tk, T7, T5, Tj, T6;
Chris@42 66 Td = FMA(KP801937735, Tc, T9);
Chris@42 67 T5 = T2 + T3 + T4;
Chris@42 68 Tj = FNMS(KP356895867, T4, T3);
Chris@42 69 T6 = FNMS(KP356895867, T3, T2);
Chris@42 70 Tf = FNMS(KP692021471, Te, T3);
Chris@42 71 R0[0] = FMA(KP2_000000000, T5, T1);
Chris@42 72 Tk = FNMS(KP692021471, Tj, T2);
Chris@42 73 T7 = FNMS(KP692021471, T6, T4);
Chris@42 74 Tg = FNMS(KP1_801937735, Tf, T1);
Chris@42 75 Ti = FNMS(KP801937735, Th, Ta);
Chris@42 76 Tl = FNMS(KP1_801937735, Tk, T1);
Chris@42 77 T8 = FNMS(KP1_801937735, T7, T1);
Chris@42 78 }
Chris@42 79 }
Chris@42 80 R1[WS(rs, 2)] = FMA(KP1_949855824, Ti, Tg);
Chris@42 81 R0[WS(rs, 1)] = FNMS(KP1_949855824, Ti, Tg);
Chris@42 82 R0[WS(rs, 2)] = FMA(KP1_949855824, Tn, Tl);
Chris@42 83 R1[WS(rs, 1)] = FNMS(KP1_949855824, Tn, Tl);
Chris@42 84 R0[WS(rs, 3)] = FMA(KP1_949855824, Td, T8);
Chris@42 85 R1[0] = FNMS(KP1_949855824, Td, T8);
Chris@42 86 }
Chris@42 87 }
Chris@42 88 }
Chris@42 89
Chris@42 90 static const kr2c_desc desc = { 7, "r2cb_7", {2, 0, 22, 0}, &GENUS };
Chris@42 91
Chris@42 92 void X(codelet_r2cb_7) (planner *p) {
Chris@42 93 X(kr2c_register) (p, r2cb_7, &desc);
Chris@42 94 }
Chris@42 95
Chris@42 96 #else /* HAVE_FMA */
Chris@42 97
Chris@42 98 /* Generated by: ../../../genfft/gen_r2cb.native -compact -variables 4 -pipeline-latency 4 -sign 1 -n 7 -name r2cb_7 -include r2cb.h */
Chris@42 99
Chris@42 100 /*
Chris@42 101 * This function contains 24 FP additions, 19 FP multiplications,
Chris@42 102 * (or, 11 additions, 6 multiplications, 13 fused multiply/add),
Chris@42 103 * 21 stack variables, 7 constants, and 14 memory accesses
Chris@42 104 */
Chris@42 105 #include "r2cb.h"
Chris@42 106
Chris@42 107 static void r2cb_7(R *R0, R *R1, R *Cr, R *Ci, stride rs, stride csr, stride csi, INT v, INT ivs, INT ovs)
Chris@42 108 {
Chris@42 109 DK(KP2_000000000, +2.000000000000000000000000000000000000000000000);
Chris@42 110 DK(KP1_801937735, +1.801937735804838252472204639014890102331838324);
Chris@42 111 DK(KP445041867, +0.445041867912628808577805128993589518932711138);
Chris@42 112 DK(KP1_246979603, +1.246979603717467061050009768008479621264549462);
Chris@42 113 DK(KP867767478, +0.867767478235116240951536665696717509219981456);
Chris@42 114 DK(KP1_949855824, +1.949855824363647214036263365987862434465571601);
Chris@42 115 DK(KP1_563662964, +1.563662964936059617416889053348115500464669037);
Chris@42 116 {
Chris@42 117 INT i;
Chris@42 118 for (i = v; i > 0; i = i - 1, R0 = R0 + ovs, R1 = R1 + ovs, Cr = Cr + ivs, Ci = Ci + ivs, MAKE_VOLATILE_STRIDE(28, rs), MAKE_VOLATILE_STRIDE(28, csr), MAKE_VOLATILE_STRIDE(28, csi)) {
Chris@42 119 E T9, Td, Tb, T1, T4, T2, T3, T5, Tc, Ta, T6, T8, T7;
Chris@42 120 T6 = Ci[WS(csi, 2)];
Chris@42 121 T8 = Ci[WS(csi, 1)];
Chris@42 122 T7 = Ci[WS(csi, 3)];
Chris@42 123 T9 = FNMS(KP1_949855824, T7, KP1_563662964 * T6) - (KP867767478 * T8);
Chris@42 124 Td = FMA(KP867767478, T6, KP1_563662964 * T7) - (KP1_949855824 * T8);
Chris@42 125 Tb = FMA(KP1_563662964, T8, KP1_949855824 * T6) + (KP867767478 * T7);
Chris@42 126 T1 = Cr[0];
Chris@42 127 T4 = Cr[WS(csr, 3)];
Chris@42 128 T2 = Cr[WS(csr, 1)];
Chris@42 129 T3 = Cr[WS(csr, 2)];
Chris@42 130 T5 = FMA(KP1_246979603, T3, T1) + FNMA(KP445041867, T4, KP1_801937735 * T2);
Chris@42 131 Tc = FMA(KP1_246979603, T4, T1) + FNMA(KP1_801937735, T3, KP445041867 * T2);
Chris@42 132 Ta = FMA(KP1_246979603, T2, T1) + FNMA(KP1_801937735, T4, KP445041867 * T3);
Chris@42 133 R0[WS(rs, 2)] = T5 - T9;
Chris@42 134 R1[WS(rs, 1)] = T5 + T9;
Chris@42 135 R0[WS(rs, 1)] = Tc + Td;
Chris@42 136 R1[WS(rs, 2)] = Tc - Td;
Chris@42 137 R0[WS(rs, 3)] = Ta + Tb;
Chris@42 138 R1[0] = Ta - Tb;
Chris@42 139 R0[0] = FMA(KP2_000000000, T2 + T3 + T4, T1);
Chris@42 140 }
Chris@42 141 }
Chris@42 142 }
Chris@42 143
Chris@42 144 static const kr2c_desc desc = { 7, "r2cb_7", {11, 6, 13, 0}, &GENUS };
Chris@42 145
Chris@42 146 void X(codelet_r2cb_7) (planner *p) {
Chris@42 147 X(kr2c_register) (p, r2cb_7, &desc);
Chris@42 148 }
Chris@42 149
Chris@42 150 #endif /* HAVE_FMA */