2 * This file is part of the GROMACS molecular simulation package.
4 * Copyright (c) 2014,2015,2016,2017,2018 by the GROMACS development team.
5 * Copyright (c) 2019,2020, by the GROMACS development team, led by
6 * Mark Abraham, David van der Spoel, Berk Hess, and Erik Lindahl,
7 * and including many others, as listed in the AUTHORS file in the
8 * top-level source directory and at http://www.gromacs.org.
10 * GROMACS is free software; you can redistribute it and/or
11 * modify it under the terms of the GNU Lesser General Public License
12 * as published by the Free Software Foundation; either version 2.1
13 * of the License, or (at your option) any later version.
15 * GROMACS is distributed in the hope that it will be useful,
16 * but WITHOUT ANY WARRANTY; without even the implied warranty of
17 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
18 * Lesser General Public License for more details.
20 * You should have received a copy of the GNU Lesser General Public
21 * License along with GROMACS; if not, see
22 * http://www.gnu.org/licenses, or write to the Free Software Foundation,
23 * Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
25 * If you want to redistribute modifications to GROMACS, please
26 * consider that scientific software is very special. Version
27 * control is crucial - bugs must be traceable. We will be happy to
28 * consider code for inclusion in the official distribution, but
29 * derived work must not be called official GROMACS. Details are found
30 * in the README & COPYING files - if they are missing, get the
31 * official version at http://www.gromacs.org.
33 * To help us fund GROMACS development, we humbly ask that you cite
34 * the research papers on the package. Check out http://www.gromacs.org.
37 #ifndef GMX_SIMD_IMPL_X86_AVX_512_DEFINITIONS_H
38 #define GMX_SIMD_IMPL_X86_AVX_512_DEFINITIONS_H
40 // A couple of implementation notes:
42 // - We avoid using the reduce calls provided by the intel compiler, since they
43 // are not instructions but software routines, and not implemented on gcc.
44 // - gcc-4.9 through 5.2 do not provide int2mask and mask2int. However, since both
45 // compilers are fine when we just use hexadecimal literals, we use our
46 // own routines avx512Mask2Int() & avx512Int2Mask() that simply return
47 // the argument. If some compiler needs a different solution in the future,
48 // just write some conditional code for these routines.
49 // - gcc does not provide _mm512_trunc_ps(), and all alternative ops where we
50 // can specify a rounding mode also appear to be buggy. We conditionally work
51 // around this by using conversions to/from integer instead.
54 #define GMX_SIMD_HAVE_FLOAT 1
55 #define GMX_SIMD_HAVE_DOUBLE 1
56 #define GMX_SIMD_HAVE_LOADU 1
57 #define GMX_SIMD_HAVE_STOREU 1
58 #define GMX_SIMD_HAVE_LOGICAL 1
59 #define GMX_SIMD_HAVE_FMA 1
60 #define GMX_SIMD_HAVE_FINT32_EXTRACT 0
61 #define GMX_SIMD_HAVE_FINT32_LOGICAL 1
62 #define GMX_SIMD_HAVE_FINT32_ARITHMETICS 1
63 // Technically it is straightforward to emulate extract on AVX-512 through
64 // memory operations, but when applied to 16 elements as part of a table lookup
65 // it will be faster to just store the entire vector once, so we avoid setting it.
66 #define GMX_SIMD_HAVE_DINT32_EXTRACT 0
67 #define GMX_SIMD_HAVE_DINT32_LOGICAL 1
68 #define GMX_SIMD_HAVE_DINT32_ARITHMETICS 1
69 #define GMX_SIMD_HAVE_NATIVE_COPYSIGN_FLOAT 1
70 #define GMX_SIMD_HAVE_NATIVE_RSQRT_ITER_FLOAT 0
71 #define GMX_SIMD_HAVE_NATIVE_RCP_ITER_FLOAT 0
72 #define GMX_SIMD_HAVE_NATIVE_LOG_FLOAT 0
73 #define GMX_SIMD_HAVE_NATIVE_EXP2_FLOAT 0
74 #define GMX_SIMD_HAVE_NATIVE_EXP_FLOAT 0
75 #define GMX_SIMD_HAVE_NATIVE_COPYSIGN_DOUBLE 1
76 #define GMX_SIMD_HAVE_NATIVE_RSQRT_ITER_DOUBLE 0
77 #define GMX_SIMD_HAVE_NATIVE_RCP_ITER_DOUBLE 0
78 #define GMX_SIMD_HAVE_NATIVE_LOG_DOUBLE 0
79 #define GMX_SIMD_HAVE_NATIVE_EXP2_DOUBLE 0
80 #define GMX_SIMD_HAVE_NATIVE_EXP_DOUBLE 0
81 #define GMX_SIMD_HAVE_GATHER_LOADU_BYSIMDINT_TRANSPOSE_FLOAT 1
82 #define GMX_SIMD_HAVE_GATHER_LOADU_BYSIMDINT_TRANSPOSE_DOUBLE 1
83 #define GMX_SIMD_HAVE_HSIMD_UTIL_FLOAT 1
84 #define GMX_SIMD_HAVE_HSIMD_UTIL_DOUBLE 1
85 #define GMX_SIMD_HAVE_4NSIMD_UTIL_FLOAT 1
86 #define GMX_SIMD_HAVE_4NSIMD_UTIL_DOUBLE 1
88 #define GMX_SIMD4_HAVE_FLOAT 1
89 #define GMX_SIMD4_HAVE_DOUBLE 1
91 // Implementation details
92 #define GMX_SIMD_FLOAT_WIDTH 16
93 #define GMX_SIMD_DOUBLE_WIDTH 8
94 #define GMX_SIMD_FINT32_WIDTH 16
95 #define GMX_SIMD_DINT32_WIDTH 8
96 #define GMX_SIMD4_WIDTH 4
97 #define GMX_SIMD_ALIGNMENT 64 // Bytes (16*single or 8*double)
98 #define GMX_SIMD_RSQRT_BITS 14
99 #define GMX_SIMD_RCP_BITS 14
101 #endif // GMX_SIMD_IMPL_X86_AVX_512_DEFINITIONS_H