src/gromacs/fft/parallel_3dfft.h

   1 /*
   2  * This file is part of the GROMACS molecular simulation package.
   3  *
   4  * Copyright (c) 1991-2005 David van der Spoel, Erik Lindahl, University of Groningen.
   5  * Copyright (c) 2013,2014,2017,2018, 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.
   9  *
  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.
  14  *
  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.
  19  *
  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.
  24  *
  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.
  32  *
  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.
  35  */
  36
  37 #ifndef GMX_FFT_PARALLEL_3DFFT_H
  38 #define GMX_FFT_PARALLEL_3DFFT_H
  39
  40 #include "gromacs/fft/fft.h"
  41 #include "gromacs/gpu_utils/hostallocator.h"
  42 #include "gromacs/math/gmxcomplex.h"
  43 #include "gromacs/timing/wallcycle.h"
  44 #include "gromacs/utility/basedefinitions.h"
  45 #include "gromacs/utility/gmxmpi.h"
  46 #include "gromacs/utility/real.h"
  47
  48 typedef struct gmx_parallel_3dfft *
  49     gmx_parallel_3dfft_t;
  50
  51
  52
  53 /*! \brief Initialize parallel MPI-based 3D-FFT.
  54  *
  55  *  This routine performs real-to-complex and complex-to-real parallel 3D FFTs,
  56  *  but not complex-to-complex.
  57  *
  58  *  The routine is optimized for small-to-medium size FFTs used for PME and
  59  *  PPPM algorithms, and do allocate extra workspace whenever it might improve
  60  *  performance.
  61  *
  62  *  \param pfft_setup     Pointer to parallel 3dfft setup structure, previously
  63  *                        allocated or with automatic storage.
  64  *  \param ndata          Number of grid cells in each direction
  65  *  \param real_data      Real data. Input for forward and output for backward.
  66  *  \param complex_data   Complex data.
  67  *  \param comm           MPI communicator for both parallelization axis.
  68  *                        Needs to be either initialized or MPI_NULL for
  69  *                        no parallelization in that axis.
  70  *  \param bReproducible  Try to avoid FFT timing optimizations and other stuff
  71  *                        that could make results differ for two runs with
  72  *                        identical input (reproducibility for debugging).
  73  *  \param nthreads       Run in parallel using n threads
  74  *  \param realGridAllocation  Whether to make real grid use allocation pinned for GPU transfers.
  75  *                             Only used in PME mixed CPU+GPU mode.
  76  *
  77  *  \return 0 or a standard error code.
  78  */
  79 int
  80     gmx_parallel_3dfft_init   (gmx_parallel_3dfft_t *    pfft_setup,
  81                                const ivec                      ndata,
  82                                real **real_data,
  83                                t_complex **complex_data,
  84                                MPI_Comm                  comm[2],
  85                                gmx_bool                  bReproducible,
  86                                int                       nthreads,
  87                                gmx::PinningPolicy realGridAllocation = gmx::PinningPolicy::CannotBePinned);
  88
  89
  90
  91
  92
  93 /*! \brief Get direct space grid index limits
  94  */
  95 int
  96 gmx_parallel_3dfft_real_limits(gmx_parallel_3dfft_t      pfft_setup,
  97                                ivec                      local_ndata,
  98                                ivec                      local_offset,
  99                                ivec                      local_size);
 100
 101
 102 /*! \brief Get reciprocal space grid index limits
 103  */
 104 int
 105 gmx_parallel_3dfft_complex_limits(gmx_parallel_3dfft_t      pfft_setup,
 106                                   ivec                      complex_order,
 107                                   ivec                      local_ndata,
 108                                   ivec                      local_offset,
 109                                   ivec                      local_size);
 110
 111
 112 int
 113 gmx_parallel_3dfft_execute(gmx_parallel_3dfft_t    pfft_setup,
 114                            enum gmx_fft_direction  dir,
 115                            int                     thread,
 116                            gmx_wallcycle_t         wcycle);
 117
 118
 119 /*! \brief Release all data in parallel fft setup
 120  *
 121  *  All temporary storage and FFT plans are released. The structure itself
 122  *  is not released, but the contents is invalid after this call.
 123  *
 124  *  \param pfft_setup Parallel 3dfft setup.
 125  *
 126  *  \return 0 or a standard error code.
 127  */
 128 int
 129 gmx_parallel_3dfft_destroy(gmx_parallel_3dfft_t    pfft_setup);
 130
 131 #endif