2 * This file is part of the GROMACS molecular simulation package.
4 * Copyright (c) 2012,2013,2014 by the GROMACS development team, led by
5 * Mark Abraham, David van der Spoel, Berk Hess, and Erik Lindahl,
6 * and including many others, as listed in the AUTHORS file in the
7 * top-level source directory and at http://www.gromacs.org.
9 * GROMACS is free software; you can redistribute it and/or
10 * modify it under the terms of the GNU Lesser General Public License
11 * as published by the Free Software Foundation; either version 2.1
12 * of the License, or (at your option) any later version.
14 * GROMACS is distributed in the hope that it will be useful,
15 * but WITHOUT ANY WARRANTY; without even the implied warranty of
16 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
17 * Lesser General Public License for more details.
19 * You should have received a copy of the GNU Lesser General Public
20 * License along with GROMACS; if not, see
21 * http://www.gnu.org/licenses, or write to the Free Software Foundation,
22 * Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
24 * If you want to redistribute modifications to GROMACS, please
25 * consider that scientific software is very special. Version
26 * control is crucial - bugs must be traceable. We will be happy to
27 * consider code for inclusion in the official distribution, but
28 * derived work must not be called official GROMACS. Details are found
29 * in the README & COPYING files - if they are missing, get the
30 * official version at http://www.gromacs.org.
32 * To help us fund GROMACS development, we humbly ask that you cite
33 * the research papers on the package. Check out http://www.gromacs.org.
36 #ifndef _nbnxn_consts_h
37 #define _nbnxn_consts_h
44 /* The number of pair-search sub-cells per super-cell, used for GPU */
45 #define GPU_NSUBCELL_Z 2
46 #define GPU_NSUBCELL_Y 2
47 #define GPU_NSUBCELL_X 2
48 #define GPU_NSUBCELL (GPU_NSUBCELL_Z*GPU_NSUBCELL_Y*GPU_NSUBCELL_X)
49 /* In the non-bonded GPU kernel we operate on cluster-pairs, not cells.
50 * The number of cluster in a super-cluster matches the number of sub-cells
51 * in a pair-search cell, so we introduce a new name for the same value.
53 #define NBNXN_GPU_NCLUSTER_PER_SUPERCLUSTER GPU_NSUBCELL
55 /* With CPU kernels the i-cluster size is always 4 atoms.
56 * With x86 SIMD the j-cluster size can be 2, 4 or 8, otherwise 4.
58 #define NBNXN_CPU_CLUSTER_I_SIZE 4
60 #define NBNXN_CPU_CLUSTER_I_SIZE_2LOG 2
62 /* With GPU kernels the cluster size is 8 atoms */
63 #define NBNXN_GPU_CLUSTER_SIZE 8
65 /* With GPU kernels we group cluster pairs in 4 to optimize memory usage.
66 * To change this, also change nbnxn_cj4_t in include/types/nbnxn_pairlist.h.
68 #define NBNXN_GPU_JGROUP_SIZE 4
69 #define NBNXN_GPU_JGROUP_SIZE_2LOG 2
71 /* To avoid NaN when excluded atoms are at zero distance, we add a small
72 * number to r^2. NBNXN_AVOID_SING_R2_INC^-3 should fit in real.
75 #define NBNXN_AVOID_SING_R2_INC 1.0e-12f
77 /* The double prec. x86 SIMD kernels use a single prec. invsqrt, so > 1e-38 */
78 #define NBNXN_AVOID_SING_R2_INC 1.0e-36
81 /* Coulomb force table size chosen such that it fits along the non-bonded
82 parameters in the texture cache. */
83 #define GPU_EWALD_COULOMB_FORCE_TABLE_SIZE 1536
86 /* Strides for x/f with xyz and xyzq coordinate (and charge) storage */
89 /* Size of packs of x, y or z with SSE/AVX packed coords/forces */
92 /* Strides for a pack of 4 and 8 coordinates/forces */
93 #define STRIDE_P4 (DIM*PACK_X4)
94 #define STRIDE_P8 (DIM*PACK_X8)
96 /* Index of atom a into the SSE/AVX coordinate/force array */
97 #define X4_IND_A(a) (STRIDE_P4*((a) >> 2) + ((a) & (PACK_X4 - 1)))
98 #define X8_IND_A(a) (STRIDE_P8*((a) >> 3) + ((a) & (PACK_X8 - 1)))
101 /* Cluster-pair Interaction masks for 4xN and 2xNN kernels.
102 * Bit i*CJ_SIZE + j tells if atom i and j interact.
104 /* All interaction mask is the same for all kernels */
105 static const unsigned int NBNXN_INTERACTION_MASK_ALL = 0xffffffffU;
106 /* 4x4 kernel diagonal mask */
107 static const unsigned int NBNXN_INTERACTION_MASK_DIAG = 0x08ceU;
108 /* 4x2 kernel diagonal masks */
109 static const unsigned int NBNXN_INTERACTION_MASK_DIAG_J2_0 = 0x0002U;
110 static const unsigned int NBNXN_INTERACTION_MASK_DIAG_J2_1 = 0x002fU;
111 /* 4x8 kernel diagonal masks */
112 static const unsigned int NBNXN_INTERACTION_MASK_DIAG_J8_0 = 0xf0f8fcfeU;
113 static const unsigned int NBNXN_INTERACTION_MASK_DIAG_J8_1 = 0x0080c0e0U;