src/mdlib/ewald.c

   1 /*
   2  *
   3  *                This source code is part of
   4  *
   5  *                 G   R   O   M   A   C   S
   6  *
   7  *          GROningen MAchine for Chemical Simulations
   8  *
   9  *                        VERSION 3.2.0
  10  * Written by David van der Spoel, Erik Lindahl, Berk Hess, and others.
  11  * Copyright (c) 1991-2000, University of Groningen, The Netherlands.
  12  * Copyright (c) 2001-2004, The GROMACS development team,
  13  * check out http://www.gromacs.org for more information.
  14
  15  * This program is free software; you can redistribute it and/or
  16  * modify it under the terms of the GNU General Public License
  17  * as published by the Free Software Foundation; either version 2
  18  * of the License, or (at your option) any later version.
  19  *
  20  * If you want to redistribute modifications, please consider that
  21  * scientific software is very special. Version control is crucial -
  22  * bugs must be traceable. We will be happy to consider code for
  23  * inclusion in the official distribution, but derived work must not
  24  * be called official GROMACS. Details are found in the README & COPYING
  25  * files - if they are missing, get the official version at www.gromacs.org.
  26  *
  27  * To help us fund GROMACS development, we humbly ask that you cite
  28  * the papers on the package - you can find them in the top README file.
  29  *
  30  * For more info, check our website at http://www.gromacs.org
  31  *
  32  * And Hey:
  33  * GROwing Monsters And Cloning Shrimps
  34  */
  35 #ifdef HAVE_CONFIG_H
  36 #include <config.h>
  37 #endif
  38
  39 #include <stdio.h>
  40 #include <math.h>
  41 #include "typedefs.h"
  42 #include "vec.h"
  43 #include "gmxcomplex.h"
  44 #include "smalloc.h"
  45 #include "futil.h"
  46 #include "fftgrid.h"
  47 #include "gmx_fatal.h"
  48 #include "physics.h"
  49 #include "coulomb.h"
  50
  51 #define TOL 2e-5
  52
  53 struct ewald_tab
  54 {
  55     int nx,ny,nz,kmax;
  56     cvec **eir;
  57     t_complex *tab_xy, *tab_qxyz;
  58 };
  59
  60
  61
  62 /* TODO: fix thread-safety */
  63
  64 /* the other routines are in complex.h */
  65 static t_complex conjmul(t_complex a,t_complex b)
  66 {
  67   t_complex c;
  68
  69   c.re = a.re*b.re + a.im*b.im;
  70   c.im = a.im*b.re - a.re*b.im;
  71
  72   return c;
  73 }
  74
  75
  76
  77
  78 static void tabulate_eir(int natom,rvec x[],int kmax,cvec **eir,rvec lll)
  79 {
  80   int  i,j,m;
  81
  82   if (kmax < 1) {
  83       printf("Go away! kmax = %d\n",kmax);
  84       exit(1);
  85   }
  86
  87   for(i=0; (i<natom); i++) {
  88     for(m=0; (m<3); m++) {
  89       eir[0][i][m].re = 1;
  90       eir[0][i][m].im = 0;
  91     }
  92
  93     for(m=0; (m<3); m++) {
  94       eir[1][i][m].re = cos(x[i][m]*lll[m]);
  95       eir[1][i][m].im = sin(x[i][m]*lll[m]);
  96     }
  97     for(j=2; (j<kmax); j++)
  98       for(m=0; (m<3); m++)
  99         eir[j][i][m] = cmul(eir[j-1][i][m],eir[1][i][m]);
 100   }
 101 }
 102
 103 void init_ewald_tab(ewald_tab_t *et, const t_commrec *cr, const t_inputrec *ir,
 104                     FILE *fp)
 105 {
 106     int n;
 107
 108     snew(*et, 1);
 109     if (fp)
 110         fprintf(fp,"Will do ordinary reciprocal space Ewald sum.\n");
 111
 112     (*et)->nx = ir->nkx+1;
 113     (*et)->ny = ir->nky+1;
 114     (*et)->nz = ir->nkz+1;
 115     (*et)->kmax = max((*et)->nx,max((*et)->ny,(*et)->nz));
 116     (*et)->eir = NULL;
 117     (*et)->tab_xy = NULL;
 118     (*et)->tab_qxyz = NULL;
 119 }
 120
 121
 122
 123 real do_ewald(FILE *log,       bool bVerbose,
 124               t_inputrec *ir,
 125               rvec x[],        rvec f[],
 126               real chargeA[],  real chargeB[],
 127               rvec box,
 128               t_commrec *cr,   int natoms,
 129               matrix lrvir,    real ewaldcoeff,
 130               real lambda,     real *dvdlambda,
 131               ewald_tab_t et)
 132 {
 133   real factor=-1.0/(4*ewaldcoeff*ewaldcoeff);
 134   real scaleRecip =4.0*M_PI/(box[XX]*box[YY]*box[ZZ])*ONE_4PI_EPS0/ir->epsilon_r; // 1/(Vol*e0) //
 135   real *charge,energy_AB[2],energy;
 136   rvec lll;
 137   int  lowiy,lowiz,ix,iy,iz,n,q;
 138   real tmp,cs,ss,ak,akv,mx,my,mz,m2,scale;
 139   bool bFreeEnergy;
 140
 141     if (cr != NULL)
 142     {
 143         if (cr->nnodes > 1 || cr->nthreads>1)
 144             gmx_fatal(FARGS,"No parallel Ewald. Use PME instead.\n");
 145     }
 146
 147
 148   if (!et->eir) /* allocate if we need to */
 149   {
 150       snew(et->eir,et->kmax);
 151       for(n=0;n<et->kmax;n++)
 152           snew(et->eir[n],natoms);
 153       snew(et->tab_xy,natoms);
 154       snew(et->tab_qxyz,natoms);
 155   }
 156
 157   bFreeEnergy = (ir->efep != efepNO);
 158
 159   clear_mat(lrvir);
 160
 161   calc_lll(box,lll);
 162   /* make tables for the structure factor parts */
 163   tabulate_eir(natoms,x,et->kmax,et->eir,lll);
 164
 165   for(q=0; q<(bFreeEnergy ? 2 : 1); q++) {
 166     if (!bFreeEnergy) {
 167       charge = chargeA;
 168       scale = 1.0;
 169     } else if (q==0) {
 170       charge = chargeA;
 171       scale = 1.0 - lambda;
 172     } else {
 173       charge = chargeB;
 174       scale = lambda;
 175     }
 176     lowiy=0;
 177     lowiz=1;
 178     energy_AB[q]=0;
 179     for(ix=0;ix<et->nx;ix++) {
 180       mx=ix*lll[XX];
 181       for(iy=lowiy;iy<et->ny;iy++) {
 182         my=iy*lll[YY];
 183         if(iy>=0)
 184           for(n=0;n<natoms;n++)
 185             et->tab_xy[n]=cmul(et->eir[ix][n][XX],et->eir[iy][n][YY]);
 186         else
 187           for(n=0;n<natoms;n++)
 188             et->tab_xy[n]=conjmul(et->eir[ix][n][XX],et->eir[-iy][n][YY]);
 189         for(iz=lowiz;iz<et->nz;iz++) {
 190           mz=iz*lll[ZZ];
 191           m2=mx*mx+my*my+mz*mz;
 192           ak=exp(m2*factor)/m2;
 193           akv=2.0*ak*(1.0/m2-factor);
 194           if(iz>=0)
 195             for(n=0;n<natoms;n++)
 196               et->tab_qxyz[n]=rcmul(charge[n],cmul(et->tab_xy[n],
 197                                                    et->eir[iz][n][ZZ]));
 198           else
 199             for(n=0;n<natoms;n++)
 200               et->tab_qxyz[n]=rcmul(charge[n],conjmul(et->tab_xy[n],
 201                                                       et->eir[-iz][n][ZZ]));
 202
 203           cs=ss=0;
 204           for(n=0;n<natoms;n++) {
 205             cs+=et->tab_qxyz[n].re;
 206             ss+=et->tab_qxyz[n].im;
 207           }
 208           energy_AB[q]+=ak*(cs*cs+ss*ss);
 209           tmp=scale*akv*(cs*cs+ss*ss);
 210           lrvir[XX][XX]-=tmp*mx*mx;
 211           lrvir[XX][YY]-=tmp*mx*my;
 212           lrvir[XX][ZZ]-=tmp*mx*mz;
 213           lrvir[YY][YY]-=tmp*my*my;
 214           lrvir[YY][ZZ]-=tmp*my*mz;
 215           lrvir[ZZ][ZZ]-=tmp*mz*mz;
 216           for(n=0;n<natoms;n++) {
 217             /*tmp=scale*ak*(cs*tab_qxyz[n].im-ss*tab_qxyz[n].re);*/
 218             tmp=scale*ak*(cs*et->tab_qxyz[n].im-ss*et->tab_qxyz[n].re);
 219             f[n][XX]+=tmp*mx*2*scaleRecip;
 220             f[n][YY]+=tmp*my*2*scaleRecip;
 221             f[n][ZZ]+=tmp*mz*2*scaleRecip;
 222 #if 0
 223             f[n][XX]+=tmp*mx;
 224             f[n][YY]+=tmp*my;
 225             f[n][ZZ]+=tmp*mz;
 226 #endif
 227           }
 228           lowiz=1-et->nz;
 229         }
 230         lowiy=1-et->ny;
 231       }
 232     }
 233   }
 234
 235   if (!bFreeEnergy) {
 236     energy = energy_AB[0];
 237   } else {
 238     energy = (1.0 - lambda)*energy_AB[0] + lambda*energy_AB[1];
 239     *dvdlambda += scaleRecip*(energy_AB[1] - energy_AB[0]);
 240   }
 241
 242   lrvir[XX][XX]=-0.5*scaleRecip*(lrvir[XX][XX]+energy);
 243   lrvir[XX][YY]=-0.5*scaleRecip*(lrvir[XX][YY]);
 244   lrvir[XX][ZZ]=-0.5*scaleRecip*(lrvir[XX][ZZ]);
 245   lrvir[YY][YY]=-0.5*scaleRecip*(lrvir[YY][YY]+energy);
 246   lrvir[YY][ZZ]=-0.5*scaleRecip*(lrvir[YY][ZZ]);
 247   lrvir[ZZ][ZZ]=-0.5*scaleRecip*(lrvir[ZZ][ZZ]+energy);
 248
 249   lrvir[YY][XX]=lrvir[XX][YY];
 250   lrvir[ZZ][XX]=lrvir[XX][ZZ];
 251   lrvir[ZZ][YY]=lrvir[YY][ZZ];
 252
 253   energy*=scaleRecip;
 254
 255   return energy;
 256 }