2 * This file is part of the GROMACS molecular simulation package.
4 * Copyright (c) 2021, 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.
37 * \brief Implements stub GPU 3D FFT routines for CPU-only builds
39 * \author Mark Abraham <mark.j.abraham@gmail.com>
40 * \author Gaurav Garg <gaugarg@nvidia.com>
46 #include "gpu_3dfft.h"
47 #include "gpu_3dfft_impl.h"
50 # include "gpu_3dfft_cufft.h"
52 # include "gpu_3dfft_ocl.h"
54 # include "gpu_3dfft_sycl.h"
55 # if GMX_SYCL_DPCPP && GMX_FFT_MKL
56 # include "gpu_3dfft_sycl_mkl.h"
59 # include "gpu_3dfft_sycl_rocfft.h"
64 # include "gpu_3dfft_heffte.h"
67 #include "gromacs/utility/arrayref.h"
68 #include "gromacs/utility/exceptions.h"
73 // [[noreturn]] attributes must be added in the common headers, so it's easier to silence the warning here
75 # pragma clang diagnostic push
76 # pragma clang diagnostic ignored "-Wmissing-noreturn"
79 #if (GMX_GPU_CUDA || GMX_GPU_OPENCL || GMX_GPU_SYCL)
81 Gpu3dFft::Gpu3dFft(FftBackend backend,
84 ArrayRef<const int> gridSizesInXForEachRank,
85 ArrayRef<const int> gridSizesInYForEachRank,
87 bool performOutOfPlaceFFT,
88 const DeviceContext& context,
89 const DeviceStream& pmeStream,
91 ivec realGridSizePadded,
92 ivec complexGridSizePadded,
93 DeviceBuffer<float>* realGrid,
94 DeviceBuffer<float>* complexGrid)
99 case FftBackend::Cufft:
100 impl_ = std::make_unique<Gpu3dFft::ImplCuFft>(allocateGrids,
102 gridSizesInXForEachRank,
103 gridSizesInYForEachRank,
105 performOutOfPlaceFFT,
110 complexGridSizePadded,
115 GMX_RELEASE_ASSERT(backend == FftBackend::HeFFTe_CUDA,
116 "Unsupported FFT backend requested");
118 # elif GMX_GPU_OPENCL
121 case FftBackend::Ocl:
122 impl_ = std::make_unique<Gpu3dFft::ImplOcl>(allocateGrids,
124 gridSizesInXForEachRank,
125 gridSizesInYForEachRank,
127 performOutOfPlaceFFT,
132 complexGridSizePadded,
136 default: GMX_THROW(InternalError("Unsupported FFT backend requested"));
141 # if GMX_SYCL_DPCPP && GMX_FFT_MKL
142 case FftBackend::SyclMkl:
143 impl_ = std::make_unique<Gpu3dFft::ImplSyclMkl>(allocateGrids,
145 gridSizesInXForEachRank,
146 gridSizesInYForEachRank,
148 performOutOfPlaceFFT,
153 complexGridSizePadded,
158 # if GMX_SYCL_HIPSYCL
159 case FftBackend::SyclRocfft:
160 impl_ = std::make_unique<Gpu3dFft::ImplSyclRocfft>(allocateGrids,
162 gridSizesInXForEachRank,
163 gridSizesInYForEachRank,
165 performOutOfPlaceFFT,
170 complexGridSizePadded,
175 case FftBackend::Sycl:
176 impl_ = std::make_unique<Gpu3dFft::ImplSycl>(allocateGrids,
178 gridSizesInXForEachRank,
179 gridSizesInYForEachRank,
181 performOutOfPlaceFFT,
186 complexGridSizePadded,
190 default: GMX_THROW(InternalError("Unsupported FFT backend requested"));
197 case FftBackend::HeFFTe_CUDA:
200 "HeFFTe_CUDA FFT backend is supported only with GROMACS compiled with CUDA");
201 GMX_RELEASE_ASSERT(heffte::backend::is_enabled<heffte::backend::cufft>::value,
202 "HeFFTe not compiled with CUDA support");
203 impl_ = std::make_unique<Gpu3dFft::ImplHeFfte<heffte::backend::cufft>>(
206 gridSizesInXForEachRank,
207 gridSizesInYForEachRank,
209 performOutOfPlaceFFT,
214 complexGridSizePadded,
219 default: GMX_RELEASE_ASSERT(impl_ != nullptr, "Unsupported FFT backend requested");
226 Gpu3dFft::Gpu3dFft(FftBackend /*backend */,
227 bool /*allocateGrids*/,
229 ArrayRef<const int> /*gridSizesInXForEachRank*/,
230 ArrayRef<const int> /*gridSizesInYForEachRank*/,
232 bool /*performOutOfPlaceFFT*/,
233 const DeviceContext& /*context*/,
234 const DeviceStream& /*pmeStream*/,
235 ivec /*realGridSize*/,
236 ivec /*realGridSizePadded*/,
237 ivec /*complexGridSizePadded*/,
238 DeviceBuffer<float>* /*realGrid*/,
239 DeviceBuffer<float>* /*complexGrid*/)
241 GMX_THROW(InternalError("Cannot run GPU routines in a CPU-only configuration"));
246 Gpu3dFft::~Gpu3dFft() = default;
248 void Gpu3dFft::perform3dFft(gmx_fft_direction dir, CommandEvent* timingEvent)
250 GMX_RELEASE_ASSERT(impl_ != nullptr, "Cannot run GPU routines in a CPU-only configuration");
251 impl_->perform3dFft(dir, timingEvent);
255 # pragma clang diagnostic pop