Mercurial > hg > sv-dependency-builds

annotate src/fftw-3.3.8/rdft/scalar/r2cb/r2cb_6.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 08:07:28 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 6 -name r2cb_6 -include rdft/scalar/r2cb.h */
Chris@82 29
Chris@82 30 /*
Chris@82 31 * This function contains 14 FP additions, 6 FP multiplications,
Chris@82 32 * (or, 8 additions, 0 multiplications, 6 fused multiply/add),
Chris@82 33 * 17 stack variables, 2 constants, and 12 memory accesses
Chris@82 34 */
Chris@82 35 #include "rdft/scalar/r2cb.h"
Chris@82 36
Chris@82 37 static void r2cb_6(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_732050807, +1.732050807568877293527446341505872366942805254);
Chris@82 40 DK(KP2_000000000, +2.000000000000000000000000000000000000000000000);
Chris@82 41 {
Chris@82 42 INT i;
Chris@82 43 for (i = v; i > 0; i = i - 1, R0 = R0 + ovs, R1 = R1 + ovs, Cr = Cr + ivs, Ci = Ci + ivs, MAKE_VOLATILE_STRIDE(24, rs), MAKE_VOLATILE_STRIDE(24, csr), MAKE_VOLATILE_STRIDE(24, csi)) {
Chris@82 44 E T3, T7, Tc, Te, T6, T8, T1, T2, T9, Td;
Chris@82 45 T1 = Cr[0];
Chris@82 46 T2 = Cr[WS(csr, 3)];
Chris@82 47 T3 = T1 + T2;
Chris@82 48 T7 = T1 - T2;
Chris@82 49 {
Chris@82 50 E Ta, Tb, T4, T5;
Chris@82 51 Ta = Ci[WS(csi, 2)];
Chris@82 52 Tb = Ci[WS(csi, 1)];
Chris@82 53 Tc = Ta - Tb;
Chris@82 54 Te = Ta + Tb;
Chris@82 55 T4 = Cr[WS(csr, 2)];
Chris@82 56 T5 = Cr[WS(csr, 1)];
Chris@82 57 T6 = T4 + T5;
Chris@82 58 T8 = T5 - T4;
Chris@82 59 }
Chris@82 60 R0[0] = FMA(KP2_000000000, T6, T3);
Chris@82 61 R1[WS(rs, 1)] = FNMS(KP2_000000000, T8, T7);
Chris@82 62 T9 = T3 - T6;
Chris@82 63 R0[WS(rs, 2)] = FNMS(KP1_732050807, Tc, T9);
Chris@82 64 R0[WS(rs, 1)] = FMA(KP1_732050807, Tc, T9);
Chris@82 65 Td = T7 + T8;
Chris@82 66 R1[0] = FNMS(KP1_732050807, Te, Td);
Chris@82 67 R1[WS(rs, 2)] = FMA(KP1_732050807, Te, Td);
Chris@82 68 }
Chris@82 69 }
Chris@82 70 }
Chris@82 71
Chris@82 72 static const kr2c_desc desc = { 6, "r2cb_6", {8, 0, 6, 0}, &GENUS };
Chris@82 73
Chris@82 74 void X(codelet_r2cb_6) (planner *p) {
Chris@82 75 X(kr2c_register) (p, r2cb_6, &desc);
Chris@82 76 }
Chris@82 77
Chris@82 78 #else
Chris@82 79
Chris@82 80 /* Generated by: ../../../genfft/gen_r2cb.native -compact -variables 4 -pipeline-latency 4 -sign 1 -n 6 -name r2cb_6 -include rdft/scalar/r2cb.h */
Chris@82 81
Chris@82 82 /*
Chris@82 83 * This function contains 14 FP additions, 4 FP multiplications,
Chris@82 84 * (or, 12 additions, 2 multiplications, 2 fused multiply/add),
Chris@82 85 * 17 stack variables, 2 constants, and 12 memory accesses
Chris@82 86 */
Chris@82 87 #include "rdft/scalar/r2cb.h"
Chris@82 88
Chris@82 89 static void r2cb_6(R *R0, R *R1, R *Cr, R *Ci, stride rs, stride csr, stride csi, INT v, INT ivs, INT ovs)
Chris@82 90 {
Chris@82 91 DK(KP2_000000000, +2.000000000000000000000000000000000000000000000);
Chris@82 92 DK(KP1_732050807, +1.732050807568877293527446341505872366942805254);
Chris@82 93 {
Chris@82 94 INT i;
Chris@82 95 for (i = v; i > 0; i = i - 1, R0 = R0 + ovs, R1 = R1 + ovs, Cr = Cr + ivs, Ci = Ci + ivs, MAKE_VOLATILE_STRIDE(24, rs), MAKE_VOLATILE_STRIDE(24, csr), MAKE_VOLATILE_STRIDE(24, csi)) {
Chris@82 96 E T3, T7, Tc, Te, T6, T8, T1, T2, T9, Td;
Chris@82 97 T1 = Cr[0];
Chris@82 98 T2 = Cr[WS(csr, 3)];
Chris@82 99 T3 = T1 - T2;
Chris@82 100 T7 = T1 + T2;
Chris@82 101 {
Chris@82 102 E Ta, Tb, T4, T5;
Chris@82 103 Ta = Ci[WS(csi, 2)];
Chris@82 104 Tb = Ci[WS(csi, 1)];
Chris@82 105 Tc = KP1_732050807 * (Ta - Tb);
Chris@82 106 Te = KP1_732050807 * (Ta + Tb);
Chris@82 107 T4 = Cr[WS(csr, 2)];
Chris@82 108 T5 = Cr[WS(csr, 1)];
Chris@82 109 T6 = T4 - T5;
Chris@82 110 T8 = T4 + T5;
Chris@82 111 }
Chris@82 112 R1[WS(rs, 1)] = FMA(KP2_000000000, T6, T3);
Chris@82 113 R0[0] = FMA(KP2_000000000, T8, T7);
Chris@82 114 T9 = T7 - T8;
Chris@82 115 R0[WS(rs, 2)] = T9 - Tc;
Chris@82 116 R0[WS(rs, 1)] = T9 + Tc;
Chris@82 117 Td = T3 - T6;
Chris@82 118 R1[0] = Td - Te;
Chris@82 119 R1[WS(rs, 2)] = Td + Te;
Chris@82 120 }
Chris@82 121 }
Chris@82 122 }
Chris@82 123
Chris@82 124 static const kr2c_desc desc = { 6, "r2cb_6", {12, 2, 2, 0}, &GENUS };
Chris@82 125
Chris@82 126 void X(codelet_r2cb_6) (planner *p) {
Chris@82 127 X(kr2c_register) (p, r2cb_6, &desc);
Chris@82 128 }
Chris@82 129
Chris@82 130 #endif