Mercurial > hg > sv-dependency-builds
comparison src/fftw-3.3.3/rdft/scalar/r2cf/r2cf_12.c @ 10:37bf6b4a2645
Add FFTW3
| author | Chris Cannam |
|---|---|
| date | Wed, 20 Mar 2013 15:35:50 +0000 |
| parents | |
| children |
comparison
equal
deleted
inserted
replaced
| 9:c0fb53affa76 | 10:37bf6b4a2645 |
|---|---|
| 1 /* | |
| 2 * Copyright (c) 2003, 2007-11 Matteo Frigo | |
| 3 * Copyright (c) 2003, 2007-11 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 Sun Nov 25 07:39:46 EST 2012 */ | |
| 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 12 -name r2cf_12 -include r2cf.h */ | |
| 29 | |
| 30 /* | |
| 31 * This function contains 38 FP additions, 10 FP multiplications, | |
| 32 * (or, 30 additions, 2 multiplications, 8 fused multiply/add), | |
| 33 * 31 stack variables, 2 constants, and 24 memory accesses | |
| 34 */ | |
| 35 #include "r2cf.h" | |
| 36 | |
| 37 static void r2cf_12(R *R0, R *R1, R *Cr, R *Ci, stride rs, stride csr, stride csi, INT v, INT ivs, INT ovs) | |
| 38 { | |
| 39 DK(KP866025403, +0.866025403784438646763723170752936183471402627); | |
| 40 DK(KP500000000, +0.500000000000000000000000000000000000000000000); | |
| 41 { | |
| 42 INT i; | |
| 43 for (i = v; i > 0; i = i - 1, R0 = R0 + ivs, R1 = R1 + ivs, Cr = Cr + ovs, Ci = Ci + ovs, MAKE_VOLATILE_STRIDE(48, rs), MAKE_VOLATILE_STRIDE(48, csr), MAKE_VOLATILE_STRIDE(48, csi)) { | |
| 44 E Tm, T6, Ty, Tp, T5, Tk, Tt, Tb, Tc, Td, T9, Tn; | |
| 45 { | |
| 46 E T1, Tg, Th, Ti, T4, T2, T3, T7, T8, Tj; | |
| 47 T1 = R0[0]; | |
| 48 T2 = R0[WS(rs, 2)]; | |
| 49 T3 = R0[WS(rs, 4)]; | |
| 50 Tg = R1[WS(rs, 1)]; | |
| 51 Th = R1[WS(rs, 3)]; | |
| 52 Ti = R1[WS(rs, 5)]; | |
| 53 T4 = T2 + T3; | |
| 54 Tm = T3 - T2; | |
| 55 T6 = R0[WS(rs, 3)]; | |
| 56 Ty = Ti - Th; | |
| 57 Tj = Th + Ti; | |
| 58 Tp = FNMS(KP500000000, T4, T1); | |
| 59 T5 = T1 + T4; | |
| 60 T7 = R0[WS(rs, 5)]; | |
| 61 Tk = FNMS(KP500000000, Tj, Tg); | |
| 62 Tt = Tg + Tj; | |
| 63 T8 = R0[WS(rs, 1)]; | |
| 64 Tb = R1[WS(rs, 4)]; | |
| 65 Tc = R1[0]; | |
| 66 Td = R1[WS(rs, 2)]; | |
| 67 T9 = T7 + T8; | |
| 68 Tn = T8 - T7; | |
| 69 } | |
| 70 { | |
| 71 E Te, Tz, To, TC; | |
| 72 Te = Tc + Td; | |
| 73 Tz = Td - Tc; | |
| 74 To = Tm - Tn; | |
| 75 TC = Tm + Tn; | |
| 76 { | |
| 77 E Ta, Tq, TA, TB; | |
| 78 Ta = T6 + T9; | |
| 79 Tq = FNMS(KP500000000, T9, T6); | |
| 80 TA = Ty - Tz; | |
| 81 TB = Ty + Tz; | |
| 82 { | |
| 83 E Tf, Tu, Tx, Tr; | |
| 84 Tf = FNMS(KP500000000, Te, Tb); | |
| 85 Tu = Tb + Te; | |
| 86 Tx = Tp - Tq; | |
| 87 Tr = Tp + Tq; | |
| 88 { | |
| 89 E Tv, Tw, Tl, Ts; | |
| 90 Tv = T5 + Ta; | |
| 91 Cr[WS(csr, 3)] = T5 - Ta; | |
| 92 Ci[WS(csi, 4)] = KP866025403 * (TC + TB); | |
| 93 Ci[WS(csi, 2)] = KP866025403 * (TB - TC); | |
| 94 Tw = Tt + Tu; | |
| 95 Ci[WS(csi, 3)] = Tt - Tu; | |
| 96 Tl = Tf - Tk; | |
| 97 Ts = Tk + Tf; | |
| 98 Cr[WS(csr, 1)] = FMA(KP866025403, TA, Tx); | |
| 99 Cr[WS(csr, 5)] = FNMS(KP866025403, TA, Tx); | |
| 100 Cr[0] = Tv + Tw; | |
| 101 Cr[WS(csr, 6)] = Tv - Tw; | |
| 102 Cr[WS(csr, 4)] = Tr + Ts; | |
| 103 Cr[WS(csr, 2)] = Tr - Ts; | |
| 104 Ci[WS(csi, 5)] = FNMS(KP866025403, To, Tl); | |
| 105 Ci[WS(csi, 1)] = FMA(KP866025403, To, Tl); | |
| 106 } | |
| 107 } | |
| 108 } | |
| 109 } | |
| 110 } | |
| 111 } | |
| 112 } | |
| 113 | |
| 114 static const kr2c_desc desc = { 12, "r2cf_12", {30, 2, 8, 0}, &GENUS }; | |
| 115 | |
| 116 void X(codelet_r2cf_12) (planner *p) { | |
| 117 X(kr2c_register) (p, r2cf_12, &desc); | |
| 118 } | |
| 119 | |
| 120 #else /* HAVE_FMA */ | |
| 121 | |
| 122 /* Generated by: ../../../genfft/gen_r2cf.native -compact -variables 4 -pipeline-latency 4 -n 12 -name r2cf_12 -include r2cf.h */ | |
| 123 | |
| 124 /* | |
| 125 * This function contains 38 FP additions, 8 FP multiplications, | |
| 126 * (or, 34 additions, 4 multiplications, 4 fused multiply/add), | |
| 127 * 21 stack variables, 2 constants, and 24 memory accesses | |
| 128 */ | |
| 129 #include "r2cf.h" | |
| 130 | |
| 131 static void r2cf_12(R *R0, R *R1, R *Cr, R *Ci, stride rs, stride csr, stride csi, INT v, INT ivs, INT ovs) | |
| 132 { | |
| 133 DK(KP866025403, +0.866025403784438646763723170752936183471402627); | |
| 134 DK(KP500000000, +0.500000000000000000000000000000000000000000000); | |
| 135 { | |
| 136 INT i; | |
| 137 for (i = v; i > 0; i = i - 1, R0 = R0 + ivs, R1 = R1 + ivs, Cr = Cr + ovs, Ci = Ci + ovs, MAKE_VOLATILE_STRIDE(48, rs), MAKE_VOLATILE_STRIDE(48, csr), MAKE_VOLATILE_STRIDE(48, csi)) { | |
| 138 E T5, Tp, Tb, Tn, Ty, Tt, Ta, Tq, Tc, Ti, Tz, Tu, Td, To; | |
| 139 { | |
| 140 E T1, T2, T3, T4; | |
| 141 T1 = R0[0]; | |
| 142 T2 = R0[WS(rs, 2)]; | |
| 143 T3 = R0[WS(rs, 4)]; | |
| 144 T4 = T2 + T3; | |
| 145 T5 = T1 + T4; | |
| 146 Tp = FNMS(KP500000000, T4, T1); | |
| 147 Tb = T3 - T2; | |
| 148 } | |
| 149 { | |
| 150 E Tj, Tk, Tl, Tm; | |
| 151 Tj = R1[WS(rs, 1)]; | |
| 152 Tk = R1[WS(rs, 3)]; | |
| 153 Tl = R1[WS(rs, 5)]; | |
| 154 Tm = Tk + Tl; | |
| 155 Tn = FNMS(KP500000000, Tm, Tj); | |
| 156 Ty = Tl - Tk; | |
| 157 Tt = Tj + Tm; | |
| 158 } | |
| 159 { | |
| 160 E T6, T7, T8, T9; | |
| 161 T6 = R0[WS(rs, 3)]; | |
| 162 T7 = R0[WS(rs, 5)]; | |
| 163 T8 = R0[WS(rs, 1)]; | |
| 164 T9 = T7 + T8; | |
| 165 Ta = T6 + T9; | |
| 166 Tq = FNMS(KP500000000, T9, T6); | |
| 167 Tc = T8 - T7; | |
| 168 } | |
| 169 { | |
| 170 E Te, Tf, Tg, Th; | |
| 171 Te = R1[WS(rs, 4)]; | |
| 172 Tf = R1[0]; | |
| 173 Tg = R1[WS(rs, 2)]; | |
| 174 Th = Tf + Tg; | |
| 175 Ti = FNMS(KP500000000, Th, Te); | |
| 176 Tz = Tg - Tf; | |
| 177 Tu = Te + Th; | |
| 178 } | |
| 179 Cr[WS(csr, 3)] = T5 - Ta; | |
| 180 Ci[WS(csi, 3)] = Tt - Tu; | |
| 181 Td = KP866025403 * (Tb - Tc); | |
| 182 To = Ti - Tn; | |
| 183 Ci[WS(csi, 1)] = Td + To; | |
| 184 Ci[WS(csi, 5)] = To - Td; | |
| 185 { | |
| 186 E Tx, TA, Tv, Tw; | |
| 187 Tx = Tp - Tq; | |
| 188 TA = KP866025403 * (Ty - Tz); | |
| 189 Cr[WS(csr, 5)] = Tx - TA; | |
| 190 Cr[WS(csr, 1)] = Tx + TA; | |
| 191 Tv = T5 + Ta; | |
| 192 Tw = Tt + Tu; | |
| 193 Cr[WS(csr, 6)] = Tv - Tw; | |
| 194 Cr[0] = Tv + Tw; | |
| 195 } | |
| 196 { | |
| 197 E Tr, Ts, TB, TC; | |
| 198 Tr = Tp + Tq; | |
| 199 Ts = Tn + Ti; | |
| 200 Cr[WS(csr, 2)] = Tr - Ts; | |
| 201 Cr[WS(csr, 4)] = Tr + Ts; | |
| 202 TB = Ty + Tz; | |
| 203 TC = Tb + Tc; | |
| 204 Ci[WS(csi, 2)] = KP866025403 * (TB - TC); | |
| 205 Ci[WS(csi, 4)] = KP866025403 * (TC + TB); | |
| 206 } | |
| 207 } | |
| 208 } | |
| 209 } | |
| 210 | |
| 211 static const kr2c_desc desc = { 12, "r2cf_12", {34, 4, 4, 0}, &GENUS }; | |
| 212 | |
| 213 void X(codelet_r2cf_12) (planner *p) { | |
| 214 X(kr2c_register) (p, r2cf_12, &desc); | |
| 215 } | |
| 216 | |
| 217 #endif /* HAVE_FMA */ |