2 * This file is part of the GROMACS molecular simulation package.
4 * Copyright (c) 2017,2018,2019, 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 * \brief Implements common internal types for different NBNXN GPU implementations
38 * \author Szilárd Páll <pall.szilard@gmail.com>
39 * \ingroup module_nbnxm
42 #ifndef GMX_MDLIB_NBNXN_GPU_COMMON_TYPES_H
43 #define GMX_MDLIB_NBNXN_GPU_COMMON_TYPES_H
47 #include "gromacs/nbnxm/pairlist.h"
49 #if GMX_GPU == GMX_GPU_OPENCL
50 #include "gromacs/gpu_utils/gpuregiontimer_ocl.h"
53 #if GMX_GPU == GMX_GPU_CUDA
54 #include "gromacs/gpu_utils/gpuregiontimer.cuh"
58 * \brief GPU region timers used for timing GPU kernels and H2D/D2H transfers.
60 * The two-sized arrays hold the local and non-local values and should always
61 * be indexed with eintLocal/eintNonlocal.
63 struct nbnxn_gpu_timers_t
65 GpuRegionTimer atdat; /**< timer for atom data transfer (every PS step) */
66 GpuRegionTimer nb_h2d[2]; /**< timer for x/q H2D transfers (l/nl, every step) */
67 GpuRegionTimer nb_d2h[2]; /**< timer for f D2H transfer (l/nl, every step) */
68 GpuRegionTimer pl_h2d[2]; /**< timer for pair-list H2D transfers (l/nl, every PS step) */
69 bool didPairlistH2D[2]; /**< true when a pair-list transfer has been done at this step */
70 GpuRegionTimer nb_k[2]; /**< timer for non-bonded kernels (l/nl, every step) */
71 GpuRegionTimer prune_k[2]; /**< timer for the 1st pass list pruning kernel (l/nl, every PS step) */
72 bool didPrune[2]; /**< true when we timed pruning and the timings need to be accounted for */
73 GpuRegionTimer rollingPrune_k[2]; /**< timer for rolling pruning kernels (l/nl, frequency depends on chunk size) */
74 bool didRollingPrune[2]; /**< true when we timed rolling pruning (at the previous step) and the timings need to be accounted for */
79 int na_c; /**< number of atoms per cluster */
81 int nsci; /**< size of sci, # of i clusters in the list */
82 int sci_nalloc; /**< allocation size of sci */
83 DeviceBuffer<nbnxn_sci_t> sci; /**< list of i-cluster ("super-clusters") */
85 int ncj4; /**< total # of 4*j clusters */
86 int cj4_nalloc; /**< allocation size of cj4 */
87 DeviceBuffer<nbnxn_cj4_t> cj4; /**< 4*j cluster list, contains j cluster number
88 and index into the i cluster list */
89 int nimask; /**< # of 4*j clusters * # of warps */
90 int imask_nalloc; /**< allocation size of imask */
91 DeviceBuffer<unsigned int> imask; /**< imask for 2 warps for each 4*j cluster group */
92 DeviceBuffer<nbnxn_excl_t> excl; /**< atom interaction bits */
93 int nexcl; /**< count for excl */
94 int excl_nalloc; /**< allocation size of excl */
96 /* parameter+variables for normal and rolling pruning */
97 bool haveFreshList; /**< true after search, indictes that initial pruning with outer prunning is needed */
98 int rollingPruningNumParts; /**< the number of parts/steps over which one cyle of roling pruning takes places */
99 int rollingPruningPart; /**< the next part to which the roling pruning needs to be applied */