2 * This file is part of the GROMACS molecular simulation package.
4 * Copyright (c) 2018, 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.
38 * Implements PmeGpuProgramImpl, which stores permanent PME GPU context-derived data,
39 * such as (compiled) kernel handles.
41 * \author Aleksei Iupinov <a.yupinov@gmail.com>
42 * \ingroup module_ewald
46 #include "pme-gpu-program-impl.h"
48 #include "pme-gpu-constants.h"
49 #include "pme-gpu-internal.h" // for GridOrdering enum
50 #include "pme-gpu-types-host.h"
52 // PME interpolation order
53 constexpr int c_pmeOrder = 4;
54 // These hardcoded spread/gather parameters refer to not-implemented PME GPU 2D decomposition in X/Y
55 constexpr bool c_wrapX = true;
56 constexpr bool c_wrapY = true;
58 //! PME CUDA kernels forward declarations. Kernels are documented in their respective files.
61 const bool computeSplines,
62 const bool spreadCharges,
66 void pme_spline_and_spread_kernel(const PmeGpuCudaKernelParams kernelParams);
68 // Add extern declarations to inform that there will be a definition
69 // provided in another translation unit.
71 void pme_spline_and_spread_kernel<c_pmeOrder, true, true, c_wrapX, c_wrapY>(const PmeGpuCudaKernelParams);
73 void pme_spline_and_spread_kernel<c_pmeOrder, true, false, c_wrapX, c_wrapY>(const PmeGpuCudaKernelParams);
75 void pme_spline_and_spread_kernel<c_pmeOrder, false, true, c_wrapX, c_wrapY>(const PmeGpuCudaKernelParams);
78 GridOrdering gridOrdering,
79 bool computeEnergyAndVirial
81 void pme_solve_kernel(const PmeGpuCudaKernelParams kernelParams);
83 // Add extern declarations to inform that there will be a definition
84 // provided in another translation unit.
86 void pme_solve_kernel<GridOrdering::XYZ, false>(const PmeGpuCudaKernelParams);
88 void pme_solve_kernel<GridOrdering::XYZ, true>(const PmeGpuCudaKernelParams);
90 void pme_solve_kernel<GridOrdering::YZX, false>(const PmeGpuCudaKernelParams);
92 void pme_solve_kernel<GridOrdering::YZX, true>(const PmeGpuCudaKernelParams);
96 const bool overwriteForces,
100 void pme_gather_kernel(const PmeGpuCudaKernelParams kernelParams);
102 // Add extern declarations to inform that there will be a definition
103 // provided in another translation unit.
105 void pme_gather_kernel<c_pmeOrder, true, c_wrapX, c_wrapY>(const PmeGpuCudaKernelParams);
107 void pme_gather_kernel<c_pmeOrder, false, c_wrapX, c_wrapY>(const PmeGpuCudaKernelParams);
109 PmeGpuProgramImpl::PmeGpuProgramImpl(const gmx_device_info_t *)
112 warpSize = warp_size;
113 spreadWorkGroupSize = c_spreadMaxThreadsPerBlock;
114 solveMaxWorkGroupSize = c_solveMaxThreadsPerBlock;
115 gatherWorkGroupSize = c_gatherMaxThreadsPerBlock;
117 splineAndSpreadKernel = pme_spline_and_spread_kernel<c_pmeOrder, true, true, c_wrapX, c_wrapY>;
118 splineKernel = pme_spline_and_spread_kernel<c_pmeOrder, true, false, c_wrapX, c_wrapY>;
119 spreadKernel = pme_spline_and_spread_kernel<c_pmeOrder, false, true, c_wrapX, c_wrapY>;
120 gatherKernel = pme_gather_kernel<c_pmeOrder, true, c_wrapX, c_wrapY>;
121 gatherReduceWithInputKernel = pme_gather_kernel<c_pmeOrder, false, c_wrapX, c_wrapY>;
122 solveXYZKernel = pme_solve_kernel<GridOrdering::XYZ, false>;
123 solveXYZEnergyKernel = pme_solve_kernel<GridOrdering::XYZ, true>;
124 solveYZXKernel = pme_solve_kernel<GridOrdering::YZX, false>;
125 solveYZXEnergyKernel = pme_solve_kernel<GridOrdering::YZX, true>;
128 PmeGpuProgramImpl::~PmeGpuProgramImpl()