projects / LbmBenchmarkKernelsPublic.git / blame
| Commit | Line | Data |
|---|---|---|
| 10988083 MW |
1 | // -------------------------------------------------------------------------- |
| 2 | // | |
| 3 | // Copyright | |
| 4 | // Markus Wittmann, 2016-2017 | |
| 5 | // RRZE, University of Erlangen-Nuremberg, Germany | |
| 6 | // markus.wittmann -at- fau.de or hpc -at- rrze.fau.de | |
| 7 | // | |
| 8 | // Viktor Haag, 2016 | |
| 9 | // LSS, University of Erlangen-Nuremberg, Germany | |
| 10 | // | |
| 11 | // This file is part of the Lattice Boltzmann Benchmark Kernels (LbmBenchKernels). | |
| 12 | // | |
| 13 | // LbmBenchKernels is free software: you can redistribute it and/or modify | |
| 14 | // it under the terms of the GNU General Public License as published by | |
| 15 | // the Free Software Foundation, either version 3 of the License, or | |
| 16 | // (at your option) any later version. | |
| 17 | // | |
| 18 | // LbmBenchKernels is distributed in the hope that it will be useful, | |
| 19 | // but WITHOUT ANY WARRANTY; without even the implied warranty of | |
| 20 | // MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the | |
| 21 | // GNU General Public License for more details. | |
| 22 | // | |
| 23 | // You should have received a copy of the GNU General Public License | |
| 24 | // along with LbmBenchKernels. If not, see <http://www.gnu.org/licenses/>. | |
| 25 | // | |
| 26 | // -------------------------------------------------------------------------- | |
| 27 | #ifndef __VECTOR_H__ | |
| 28 | #define __VECTOR_H__ | |
| 29 | ||
| 30 | #if !defined(VECTOR_AVX) && !defined(VECTOR_SSE) | |
| 31 | #warning Defining VECTOR_AVX as no ISA extension was selected. | |
| 32 | #define VECTOR_AVX | |
| 33 | #endif | |
| 34 | ||
| 35 | #if defined(VECTOR_AVX) && defined(VECTOR_SSE) | |
| 36 | #error Only VECTOR_AVX or VECTOR_SSE can be defined at the same time. | |
| 37 | #endif | |
| 38 | ||
| 39 | #ifdef VECTOR_AVX | |
| 40 | ||
| 41 | #include <immintrin.h> | |
| 42 | // Vector size in double-precision floatin-point numbers. | |
| 43 | #define VSIZE 4 | |
| 44 | ||
| 45 | #define VPDFT __m256d | |
| 46 | ||
| 47 | #define VSET(scalar) _mm256_set1_pd(scalar) | |
| 48 | ||
| 49 | #define VLD(expr) _mm256_load_pd(expr) | |
| 50 | #define VLDU(expr) _mm256_loadu_pd(expr) | |
| 51 | ||
| 52 | #define VST(dst, src) _mm256_store_pd(dst, src) | |
| 53 | #define VSTU(dst, src) _mm256_storeu_pd(dst, src) | |
| 54 | #define VSTNT(dst, src) _mm256_stream_pd(dst, src) | |
| 55 | ||
| 56 | #define VMUL(a, b) _mm256_mul_pd(a, b) | |
| 57 | #define VADD(a, b) _mm256_add_pd(a, b) | |
| 58 | #define VSUB(a, b) _mm256_sub_pd(a, b) | |
| 59 | #endif | |
| 60 | ||
| 61 | #ifdef VECTOR_SSE | |
| 62 | #include <emmintrin.h> | |
| 63 | // Vector size in double-precision floatin-point numbers. | |
| 64 | #define VSIZE 2 | |
| 65 | ||
| 66 | #define VPDFT __m128d | |
| 67 | ||
| 68 | #define VSET(scalar) _mm_set1_pd(scalar) | |
| 69 | ||
| 70 | #define VLD(expr) _mm_load_pd(expr) | |
| 71 | #define VLDU(expr) _mm_loadu_pd(expr) | |
| 72 | ||
| 73 | #define VST(dst, src) _mm_store_pd(dst, src) | |
| 74 | #define VSTU(dst, src) _mm_storeu_pd(dst, src) | |
| 75 | #define VSTNT(dst, src) _mm_stream_pd(dst, src) | |
| 76 | ||
| 77 | #define VMUL(a, b) _mm_mul_pd(a, b) | |
| 78 | #define VADD(a, b) _mm_add_pd(a, b) | |
| 79 | #define VSUB(a, b) _mm_sub_pd(a, b) | |
| 80 | #endif | |
| 81 | ||
| 82 | ||
| 83 | #endif // __VECTOR_H__ |