src/gromacs/ewald/pme-spline-work.cpp

   1 /*
   2  * This file is part of the GROMACS molecular simulation package.
   3  *
   4  * Copyright (c) 1991-2000, University of Groningen, The Netherlands.
   5  * Copyright (c) 2001-2004, The GROMACS development team.
   6  * Copyright (c) 2013,2014,2015, by the GROMACS development team, led by
   7  * Mark Abraham, David van der Spoel, Berk Hess, and Erik Lindahl,
   8  * and including many others, as listed in the AUTHORS file in the
   9  * top-level source directory and at http://www.gromacs.org.
  10  *
  11  * GROMACS is free software; you can redistribute it and/or
  12  * modify it under the terms of the GNU Lesser General Public License
  13  * as published by the Free Software Foundation; either version 2.1
  14  * of the License, or (at your option) any later version.
  15  *
  16  * GROMACS is distributed in the hope that it will be useful,
  17  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  18  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  19  * Lesser General Public License for more details.
  20  *
  21  * You should have received a copy of the GNU Lesser General Public
  22  * License along with GROMACS; if not, see
  23  * http://www.gnu.org/licenses, or write to the Free Software Foundation,
  24  * Inc., 51 Franklin Street, Fifth Floor, Boston, MA  02110-1301  USA.
  25  *
  26  * If you want to redistribute modifications to GROMACS, please
  27  * consider that scientific software is very special. Version
  28  * control is crucial - bugs must be traceable. We will be happy to
  29  * consider code for inclusion in the official distribution, but
  30  * derived work must not be called official GROMACS. Details are found
  31  * in the README & COPYING files - if they are missing, get the
  32  * official version at http://www.gromacs.org.
  33  *
  34  * To help us fund GROMACS development, we humbly ask that you cite
  35  * the research papers on the package. Check out http://www.gromacs.org.
  36  */
  37
  38 #include "gmxpre.h"
  39
  40 #include "pme-spline-work.h"
  41
  42 #include "gromacs/simd/simd.h"
  43 #include "gromacs/utility/alignedallocator.h"
  44 #include "gromacs/utility/basedefinitions.h"
  45 #include "gromacs/utility/real.h"
  46 #include "gromacs/utility/smalloc.h"
  47
  48 #include "pme-simd.h"
  49
  50 using namespace gmx; // TODO: Remove when this file is moved into gmx namespace
  51
  52 pme_spline_work *make_pme_spline_work(int gmx_unused order)
  53 {
  54     pme_spline_work *work;
  55
  56 #ifdef PME_SIMD4_SPREAD_GATHER
  57     GMX_ALIGNED(real, GMX_SIMD4_WIDTH)  tmp[GMX_SIMD4_WIDTH*2];
  58     Simd4Real        zero_S;
  59     Simd4Real        real_mask_S0, real_mask_S1;
  60     int              of, i;
  61
  62     work = new(internal::alignedMalloc(sizeof(pme_spline_work)))pme_spline_work;
  63
  64     zero_S = setZero();
  65
  66     /* Generate bit masks to mask out the unused grid entries,
  67      * as we only operate on order of the 8 grid entries that are
  68      * load into 2 SIMD registers.
  69      */
  70     for (of = 0; of < 2*GMX_SIMD4_WIDTH-(order-1); of++)
  71     {
  72         for (i = 0; i < 2*GMX_SIMD4_WIDTH; i++)
  73         {
  74             tmp[i] = (i >= of && i < of+order ? -1.0 : 1.0);
  75         }
  76         real_mask_S0      = load4(tmp);
  77         real_mask_S1      = load4(tmp+GMX_SIMD4_WIDTH);
  78         work->mask_S0[of] = (real_mask_S0 < zero_S);
  79         work->mask_S1[of] = (real_mask_S1 < zero_S);
  80     }
  81 #else
  82     work = nullptr;
  83 #endif
  84
  85     return work;
  86 }