Mercurial > hg > sv-dependency-builds
comparison src/fftw-3.3.3/dft/simd/common/t3fv_4.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 07:38:49 EST 2012 */ | |
| 23 | |
| 24 #include "codelet-dft.h" | |
| 25 | |
| 26 #ifdef HAVE_FMA | |
| 27 | |
| 28 /* Generated by: ../../../genfft/gen_twiddle_c.native -fma -reorder-insns -schedule-for-pipeline -simd -compact -variables 4 -pipeline-latency 8 -twiddle-log3 -precompute-twiddles -no-generate-bytw -n 4 -name t3fv_4 -include t3f.h */ | |
| 29 | |
| 30 /* | |
| 31 * This function contains 12 FP additions, 10 FP multiplications, | |
| 32 * (or, 10 additions, 8 multiplications, 2 fused multiply/add), | |
| 33 * 16 stack variables, 0 constants, and 8 memory accesses | |
| 34 */ | |
| 35 #include "t3f.h" | |
| 36 | |
| 37 static void t3fv_4(R *ri, R *ii, const R *W, stride rs, INT mb, INT me, INT ms) | |
| 38 { | |
| 39 { | |
| 40 INT m; | |
| 41 R *x; | |
| 42 x = ri; | |
| 43 for (m = mb, W = W + (mb * ((TWVL / VL) * 4)); m < me; m = m + VL, x = x + (VL * ms), W = W + (TWVL * 4), MAKE_VOLATILE_STRIDE(4, rs)) { | |
| 44 V T2, T3, T1, Ta, T5, T8; | |
| 45 T2 = LDW(&(W[0])); | |
| 46 T3 = LDW(&(W[TWVL * 2])); | |
| 47 T1 = LD(&(x[0]), ms, &(x[0])); | |
| 48 Ta = LD(&(x[WS(rs, 3)]), ms, &(x[WS(rs, 1)])); | |
| 49 T5 = LD(&(x[WS(rs, 2)]), ms, &(x[0])); | |
| 50 T8 = LD(&(x[WS(rs, 1)]), ms, &(x[WS(rs, 1)])); | |
| 51 { | |
| 52 V T4, Tb, T9, T6; | |
| 53 T4 = VZMULJ(T2, T3); | |
| 54 Tb = VZMULJ(T3, Ta); | |
| 55 T9 = VZMULJ(T2, T8); | |
| 56 T6 = VZMULJ(T4, T5); | |
| 57 { | |
| 58 V Tc, Te, T7, Td; | |
| 59 Tc = VSUB(T9, Tb); | |
| 60 Te = VADD(T9, Tb); | |
| 61 T7 = VSUB(T1, T6); | |
| 62 Td = VADD(T1, T6); | |
| 63 ST(&(x[0]), VADD(Td, Te), ms, &(x[0])); | |
| 64 ST(&(x[WS(rs, 2)]), VSUB(Td, Te), ms, &(x[0])); | |
| 65 ST(&(x[WS(rs, 3)]), VFMAI(Tc, T7), ms, &(x[WS(rs, 1)])); | |
| 66 ST(&(x[WS(rs, 1)]), VFNMSI(Tc, T7), ms, &(x[WS(rs, 1)])); | |
| 67 } | |
| 68 } | |
| 69 } | |
| 70 } | |
| 71 VLEAVE(); | |
| 72 } | |
| 73 | |
| 74 static const tw_instr twinstr[] = { | |
| 75 VTW(0, 1), | |
| 76 VTW(0, 3), | |
| 77 {TW_NEXT, VL, 0} | |
| 78 }; | |
| 79 | |
| 80 static const ct_desc desc = { 4, XSIMD_STRING("t3fv_4"), twinstr, &GENUS, {10, 8, 2, 0}, 0, 0, 0 }; | |
| 81 | |
| 82 void XSIMD(codelet_t3fv_4) (planner *p) { | |
| 83 X(kdft_dit_register) (p, t3fv_4, &desc); | |
| 84 } | |
| 85 #else /* HAVE_FMA */ | |
| 86 | |
| 87 /* Generated by: ../../../genfft/gen_twiddle_c.native -simd -compact -variables 4 -pipeline-latency 8 -twiddle-log3 -precompute-twiddles -no-generate-bytw -n 4 -name t3fv_4 -include t3f.h */ | |
| 88 | |
| 89 /* | |
| 90 * This function contains 12 FP additions, 8 FP multiplications, | |
| 91 * (or, 12 additions, 8 multiplications, 0 fused multiply/add), | |
| 92 * 16 stack variables, 0 constants, and 8 memory accesses | |
| 93 */ | |
| 94 #include "t3f.h" | |
| 95 | |
| 96 static void t3fv_4(R *ri, R *ii, const R *W, stride rs, INT mb, INT me, INT ms) | |
| 97 { | |
| 98 { | |
| 99 INT m; | |
| 100 R *x; | |
| 101 x = ri; | |
| 102 for (m = mb, W = W + (mb * ((TWVL / VL) * 4)); m < me; m = m + VL, x = x + (VL * ms), W = W + (TWVL * 4), MAKE_VOLATILE_STRIDE(4, rs)) { | |
| 103 V T2, T3, T4; | |
| 104 T2 = LDW(&(W[0])); | |
| 105 T3 = LDW(&(W[TWVL * 2])); | |
| 106 T4 = VZMULJ(T2, T3); | |
| 107 { | |
| 108 V T1, Tb, T6, T9, Ta, T5, T8; | |
| 109 T1 = LD(&(x[0]), ms, &(x[0])); | |
| 110 Ta = LD(&(x[WS(rs, 3)]), ms, &(x[WS(rs, 1)])); | |
| 111 Tb = VZMULJ(T3, Ta); | |
| 112 T5 = LD(&(x[WS(rs, 2)]), ms, &(x[0])); | |
| 113 T6 = VZMULJ(T4, T5); | |
| 114 T8 = LD(&(x[WS(rs, 1)]), ms, &(x[WS(rs, 1)])); | |
| 115 T9 = VZMULJ(T2, T8); | |
| 116 { | |
| 117 V T7, Tc, Td, Te; | |
| 118 T7 = VSUB(T1, T6); | |
| 119 Tc = VBYI(VSUB(T9, Tb)); | |
| 120 ST(&(x[WS(rs, 1)]), VSUB(T7, Tc), ms, &(x[WS(rs, 1)])); | |
| 121 ST(&(x[WS(rs, 3)]), VADD(T7, Tc), ms, &(x[WS(rs, 1)])); | |
| 122 Td = VADD(T1, T6); | |
| 123 Te = VADD(T9, Tb); | |
| 124 ST(&(x[WS(rs, 2)]), VSUB(Td, Te), ms, &(x[0])); | |
| 125 ST(&(x[0]), VADD(Td, Te), ms, &(x[0])); | |
| 126 } | |
| 127 } | |
| 128 } | |
| 129 } | |
| 130 VLEAVE(); | |
| 131 } | |
| 132 | |
| 133 static const tw_instr twinstr[] = { | |
| 134 VTW(0, 1), | |
| 135 VTW(0, 3), | |
| 136 {TW_NEXT, VL, 0} | |
| 137 }; | |
| 138 | |
| 139 static const ct_desc desc = { 4, XSIMD_STRING("t3fv_4"), twinstr, &GENUS, {12, 8, 0, 0}, 0, 0, 0 }; | |
| 140 | |
| 141 void XSIMD(codelet_t3fv_4) (planner *p) { | |
| 142 X(kdft_dit_register) (p, t3fv_4, &desc); | |
| 143 } | |
| 144 #endif /* HAVE_FMA */ |