Mercurial > hg > js-dsp-test
comparison fft/fftw/fftw-3.3.4/rdft/scalar/r2cf/r2cf_10.c @ 19:26056e866c29
Add FFTW to comparison table
| author | Chris Cannam |
|---|---|
| date | Tue, 06 Oct 2015 13:08:39 +0100 |
| parents | |
| children |
comparison
equal
deleted
inserted
replaced
| 18:8db794ca3e0b | 19:26056e866c29 |
|---|---|
| 1 /* | |
| 2 * Copyright (c) 2003, 2007-14 Matteo Frigo | |
| 3 * Copyright (c) 2003, 2007-14 Massachusetts Institute of Technology | |
| 4 * | |
| 5 * This program is free software; you can redistribute it and/or modify | |
| 6 * it under the terms of the GNU General Public License as published by | |
| 7 * the Free Software Foundation; either version 2 of the License, or | |
| 8 * (at your option) any later version. | |
| 9 * | |
| 10 * This program is distributed in the hope that it will be useful, | |
| 11 * but WITHOUT ANY WARRANTY; without even the implied warranty of | |
| 12 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the | |
| 13 * GNU General Public License for more details. | |
| 14 * | |
| 15 * You should have received a copy of the GNU General Public License | |
| 16 * along with this program; if not, write to the Free Software | |
| 17 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA | |
| 18 * | |
| 19 */ | |
| 20 | |
| 21 /* This file was automatically generated --- DO NOT EDIT */ | |
| 22 /* Generated on Tue Mar 4 13:49:07 EST 2014 */ | |
| 23 | |
| 24 #include "codelet-rdft.h" | |
| 25 | |
| 26 #ifdef HAVE_FMA | |
| 27 | |
| 28 /* Generated by: ../../../genfft/gen_r2cf.native -fma -reorder-insns -schedule-for-pipeline -compact -variables 4 -pipeline-latency 4 -n 10 -name r2cf_10 -include r2cf.h */ | |
| 29 | |
| 30 /* | |
| 31 * This function contains 34 FP additions, 14 FP multiplications, | |
| 32 * (or, 24 additions, 4 multiplications, 10 fused multiply/add), | |
| 33 * 29 stack variables, 4 constants, and 20 memory accesses | |
| 34 */ | |
| 35 #include "r2cf.h" | |
| 36 | |
| 37 static void r2cf_10(R *R0, R *R1, R *Cr, R *Ci, stride rs, stride csr, stride csi, INT v, INT ivs, INT ovs) | |
| 38 { | |
| 39 DK(KP559016994, +0.559016994374947424102293417182819058860154590); | |
| 40 DK(KP250000000, +0.250000000000000000000000000000000000000000000); | |
| 41 DK(KP618033988, +0.618033988749894848204586834365638117720309180); | |
| 42 DK(KP951056516, +0.951056516295153572116439333379382143405698634); | |
| 43 { | |
| 44 INT i; | |
| 45 for (i = v; i > 0; i = i - 1, R0 = R0 + ivs, R1 = R1 + ivs, Cr = Cr + ovs, Ci = Ci + ovs, MAKE_VOLATILE_STRIDE(40, rs), MAKE_VOLATILE_STRIDE(40, csr), MAKE_VOLATILE_STRIDE(40, csi)) { | |
| 46 E Tt, T3, T7, Tq, T6, Tv, Tp, Tm, Th, T8, T1, T2, T9, Tr; | |
| 47 T1 = R0[0]; | |
| 48 T2 = R1[WS(rs, 2)]; | |
| 49 { | |
| 50 E Te, Tn, Td, Tf, Tb, Tc; | |
| 51 Tb = R0[WS(rs, 2)]; | |
| 52 Tc = R1[WS(rs, 4)]; | |
| 53 Te = R0[WS(rs, 3)]; | |
| 54 Tt = T1 + T2; | |
| 55 T3 = T1 - T2; | |
| 56 Tn = Tb + Tc; | |
| 57 Td = Tb - Tc; | |
| 58 Tf = R1[0]; | |
| 59 { | |
| 60 E T4, T5, To, Tg; | |
| 61 T4 = R0[WS(rs, 1)]; | |
| 62 T5 = R1[WS(rs, 3)]; | |
| 63 T7 = R0[WS(rs, 4)]; | |
| 64 To = Te + Tf; | |
| 65 Tg = Te - Tf; | |
| 66 Tq = T4 + T5; | |
| 67 T6 = T4 - T5; | |
| 68 Tv = Tn + To; | |
| 69 Tp = Tn - To; | |
| 70 Tm = Tg - Td; | |
| 71 Th = Td + Tg; | |
| 72 T8 = R1[WS(rs, 1)]; | |
| 73 } | |
| 74 } | |
| 75 T9 = T7 - T8; | |
| 76 Tr = T7 + T8; | |
| 77 { | |
| 78 E Ty, Tk, Tx, Tj, Tu, Ts; | |
| 79 Tu = Tq + Tr; | |
| 80 Ts = Tq - Tr; | |
| 81 { | |
| 82 E Ta, Tl, Tw, Ti; | |
| 83 Ta = T6 + T9; | |
| 84 Tl = T6 - T9; | |
| 85 Ci[WS(csi, 4)] = KP951056516 * (FMA(KP618033988, Tp, Ts)); | |
| 86 Ci[WS(csi, 2)] = KP951056516 * (FNMS(KP618033988, Ts, Tp)); | |
| 87 Ty = Tu - Tv; | |
| 88 Tw = Tu + Tv; | |
| 89 Ci[WS(csi, 3)] = KP951056516 * (FMA(KP618033988, Tl, Tm)); | |
| 90 Ci[WS(csi, 1)] = -(KP951056516 * (FNMS(KP618033988, Tm, Tl))); | |
| 91 Tk = Ta - Th; | |
| 92 Ti = Ta + Th; | |
| 93 Cr[0] = Tt + Tw; | |
| 94 Tx = FNMS(KP250000000, Tw, Tt); | |
| 95 Cr[WS(csr, 5)] = T3 + Ti; | |
| 96 Tj = FNMS(KP250000000, Ti, T3); | |
| 97 } | |
| 98 Cr[WS(csr, 4)] = FMA(KP559016994, Ty, Tx); | |
| 99 Cr[WS(csr, 2)] = FNMS(KP559016994, Ty, Tx); | |
| 100 Cr[WS(csr, 3)] = FNMS(KP559016994, Tk, Tj); | |
| 101 Cr[WS(csr, 1)] = FMA(KP559016994, Tk, Tj); | |
| 102 } | |
| 103 } | |
| 104 } | |
| 105 } | |
| 106 | |
| 107 static const kr2c_desc desc = { 10, "r2cf_10", {24, 4, 10, 0}, &GENUS }; | |
| 108 | |
| 109 void X(codelet_r2cf_10) (planner *p) { | |
| 110 X(kr2c_register) (p, r2cf_10, &desc); | |
| 111 } | |
| 112 | |
| 113 #else /* HAVE_FMA */ | |
| 114 | |
| 115 /* Generated by: ../../../genfft/gen_r2cf.native -compact -variables 4 -pipeline-latency 4 -n 10 -name r2cf_10 -include r2cf.h */ | |
| 116 | |
| 117 /* | |
| 118 * This function contains 34 FP additions, 12 FP multiplications, | |
| 119 * (or, 28 additions, 6 multiplications, 6 fused multiply/add), | |
| 120 * 26 stack variables, 4 constants, and 20 memory accesses | |
| 121 */ | |
| 122 #include "r2cf.h" | |
| 123 | |
| 124 static void r2cf_10(R *R0, R *R1, R *Cr, R *Ci, stride rs, stride csr, stride csi, INT v, INT ivs, INT ovs) | |
| 125 { | |
| 126 DK(KP250000000, +0.250000000000000000000000000000000000000000000); | |
| 127 DK(KP559016994, +0.559016994374947424102293417182819058860154590); | |
| 128 DK(KP951056516, +0.951056516295153572116439333379382143405698634); | |
| 129 DK(KP587785252, +0.587785252292473129168705954639072768597652438); | |
| 130 { | |
| 131 INT i; | |
| 132 for (i = v; i > 0; i = i - 1, R0 = R0 + ivs, R1 = R1 + ivs, Cr = Cr + ovs, Ci = Ci + ovs, MAKE_VOLATILE_STRIDE(40, rs), MAKE_VOLATILE_STRIDE(40, csr), MAKE_VOLATILE_STRIDE(40, csi)) { | |
| 133 E Ti, Tt, Ta, Tn, Td, To, Te, Tv, T3, Tq, T6, Tr, T7, Tu, Tg; | |
| 134 E Th; | |
| 135 Tg = R0[0]; | |
| 136 Th = R1[WS(rs, 2)]; | |
| 137 Ti = Tg - Th; | |
| 138 Tt = Tg + Th; | |
| 139 { | |
| 140 E T8, T9, Tb, Tc; | |
| 141 T8 = R0[WS(rs, 2)]; | |
| 142 T9 = R1[WS(rs, 4)]; | |
| 143 Ta = T8 - T9; | |
| 144 Tn = T8 + T9; | |
| 145 Tb = R0[WS(rs, 3)]; | |
| 146 Tc = R1[0]; | |
| 147 Td = Tb - Tc; | |
| 148 To = Tb + Tc; | |
| 149 } | |
| 150 Te = Ta + Td; | |
| 151 Tv = Tn + To; | |
| 152 { | |
| 153 E T1, T2, T4, T5; | |
| 154 T1 = R0[WS(rs, 1)]; | |
| 155 T2 = R1[WS(rs, 3)]; | |
| 156 T3 = T1 - T2; | |
| 157 Tq = T1 + T2; | |
| 158 T4 = R0[WS(rs, 4)]; | |
| 159 T5 = R1[WS(rs, 1)]; | |
| 160 T6 = T4 - T5; | |
| 161 Tr = T4 + T5; | |
| 162 } | |
| 163 T7 = T3 + T6; | |
| 164 Tu = Tq + Tr; | |
| 165 { | |
| 166 E Tl, Tm, Tf, Tj, Tk; | |
| 167 Tl = Td - Ta; | |
| 168 Tm = T3 - T6; | |
| 169 Ci[WS(csi, 1)] = FNMS(KP951056516, Tm, KP587785252 * Tl); | |
| 170 Ci[WS(csi, 3)] = FMA(KP587785252, Tm, KP951056516 * Tl); | |
| 171 Tf = KP559016994 * (T7 - Te); | |
| 172 Tj = T7 + Te; | |
| 173 Tk = FNMS(KP250000000, Tj, Ti); | |
| 174 Cr[WS(csr, 1)] = Tf + Tk; | |
| 175 Cr[WS(csr, 5)] = Ti + Tj; | |
| 176 Cr[WS(csr, 3)] = Tk - Tf; | |
| 177 } | |
| 178 { | |
| 179 E Tp, Ts, Ty, Tw, Tx; | |
| 180 Tp = Tn - To; | |
| 181 Ts = Tq - Tr; | |
| 182 Ci[WS(csi, 2)] = FNMS(KP587785252, Ts, KP951056516 * Tp); | |
| 183 Ci[WS(csi, 4)] = FMA(KP951056516, Ts, KP587785252 * Tp); | |
| 184 Ty = KP559016994 * (Tu - Tv); | |
| 185 Tw = Tu + Tv; | |
| 186 Tx = FNMS(KP250000000, Tw, Tt); | |
| 187 Cr[WS(csr, 2)] = Tx - Ty; | |
| 188 Cr[0] = Tt + Tw; | |
| 189 Cr[WS(csr, 4)] = Ty + Tx; | |
| 190 } | |
| 191 } | |
| 192 } | |
| 193 } | |
| 194 | |
| 195 static const kr2c_desc desc = { 10, "r2cf_10", {28, 6, 6, 0}, &GENUS }; | |
| 196 | |
| 197 void X(codelet_r2cf_10) (planner *p) { | |
| 198 X(kr2c_register) (p, r2cf_10, &desc); | |
| 199 } | |
| 200 | |
| 201 #endif /* HAVE_FMA */ |