src/gromacs/gmxlib/nonbonded/nb_generic.c

   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) 2012, 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 #ifdef HAVE_CONFIG_H
  38 #include <config.h>
  39 #endif
  40
  41 #include <math.h>
  42
  43 #include "types/simple.h"
  44 #include "vec.h"
  45 #include "typedefs.h"
  46 #include "nb_generic.h"
  47 #include "nrnb.h"
  48
  49 #include "nonbonded.h"
  50 #include "nb_kernel.h"
  51
  52
  53 void
  54 gmx_nb_generic_kernel(t_nblist *                nlist,
  55                       rvec *                    xx,
  56                       rvec *                    ff,
  57                       t_forcerec *              fr,
  58                       t_mdatoms *               mdatoms,
  59                       nb_kernel_data_t *        kernel_data,
  60                       t_nrnb *                  nrnb)
  61 {
  62     int           nri, ntype, table_nelements, ielec, ivdw;
  63     real          facel, gbtabscale;
  64     int           n, ii, is3, ii3, k, nj0, nj1, jnr, j3, ggid, nnn, n0;
  65     real          shX, shY, shZ;
  66     real          fscal, felec, fvdw, velec, vvdw, tx, ty, tz;
  67     real          rinvsq;
  68     real          iq;
  69     real          qq, vctot;
  70     int           nti, nvdwparam;
  71     int           tj;
  72     real          rt, r, eps, eps2, Y, F, Geps, Heps2, VV, FF, Fp, fijD, fijR;
  73     real          rinvsix;
  74     real          vvdwtot;
  75     real          vvdw_rep, vvdw_disp;
  76     real          ix, iy, iz, fix, fiy, fiz;
  77     real          jx, jy, jz;
  78     real          dx, dy, dz, rsq, rinv;
  79     real          c6, c12, cexp1, cexp2, br;
  80     real *        charge;
  81     real *        shiftvec;
  82     real *        vdwparam;
  83     int *         shift;
  84     int *         type;
  85     real *        fshift;
  86     real *        velecgrp;
  87     real *        vvdwgrp;
  88     real          tabscale;
  89     real *        VFtab;
  90     real *        x;
  91     real *        f;
  92     int           ewitab;
  93     real          ewtabscale, eweps, sh_ewald, ewrt, ewtabhalfspace;
  94     real *        ewtab;
  95     real          rcoulomb2, rvdw, rvdw2, sh_invrc6;
  96     real          rcutoff, rcutoff2;
  97     real          rswitch_elec, rswitch_vdw, d, d2, sw, dsw, rinvcorr;
  98     real          elec_swV3, elec_swV4, elec_swV5, elec_swF2, elec_swF3, elec_swF4;
  99     real          vdw_swV3, vdw_swV4, vdw_swV5, vdw_swF2, vdw_swF3, vdw_swF4;
 100     gmx_bool      bExactElecCutoff, bExactVdwCutoff, bExactCutoff;
 101
 102     x                   = xx[0];
 103     f                   = ff[0];
 104     ielec               = nlist->ielec;
 105     ivdw                = nlist->ivdw;
 106
 107     fshift              = fr->fshift[0];
 108     velecgrp            = kernel_data->energygrp_elec;
 109     vvdwgrp             = kernel_data->energygrp_vdw;
 110     tabscale            = kernel_data->table_elec_vdw->scale;
 111     VFtab               = kernel_data->table_elec_vdw->data;
 112
 113     sh_ewald            = fr->ic->sh_ewald;
 114     ewtab               = fr->ic->tabq_coul_FDV0;
 115     ewtabscale          = fr->ic->tabq_scale;
 116     ewtabhalfspace      = 0.5/ewtabscale;
 117
 118     rcoulomb2           = fr->rcoulomb*fr->rcoulomb;
 119     rvdw                = fr->rvdw;
 120     rvdw2               = rvdw*rvdw;
 121     sh_invrc6           = fr->ic->sh_invrc6;
 122
 123     if (fr->coulomb_modifier == eintmodPOTSWITCH)
 124     {
 125         d               = fr->rcoulomb-fr->rcoulomb_switch;
 126         elec_swV3       = -10.0/(d*d*d);
 127         elec_swV4       =  15.0/(d*d*d*d);
 128         elec_swV5       =  -6.0/(d*d*d*d*d);
 129         elec_swF2       = -30.0/(d*d*d);
 130         elec_swF3       =  60.0/(d*d*d*d);
 131         elec_swF4       = -30.0/(d*d*d*d*d);
 132     }
 133     else
 134     {
 135         /* Avoid warnings from stupid compilers (looking at you, Clang!) */
 136         elec_swV3 = elec_swV4 = elec_swV5 = elec_swF2 = elec_swF3 = elec_swF4 = 0.0;
 137     }
 138     if (fr->vdw_modifier == eintmodPOTSWITCH)
 139     {
 140         d               = fr->rvdw-fr->rvdw_switch;
 141         vdw_swV3        = -10.0/(d*d*d);
 142         vdw_swV4        =  15.0/(d*d*d*d);
 143         vdw_swV5        =  -6.0/(d*d*d*d*d);
 144         vdw_swF2        = -30.0/(d*d*d);
 145         vdw_swF3        =  60.0/(d*d*d*d);
 146         vdw_swF4        = -30.0/(d*d*d*d*d);
 147     }
 148     else
 149     {
 150         /* Avoid warnings from stupid compilers (looking at you, Clang!) */
 151         vdw_swV3 = vdw_swV4 = vdw_swV5 = vdw_swF2 = vdw_swF3 = vdw_swF4 = 0.0;
 152     }
 153
 154     bExactElecCutoff    = (fr->coulomb_modifier != eintmodNONE) || fr->eeltype == eelRF_ZERO;
 155     bExactVdwCutoff     = (fr->vdw_modifier != eintmodNONE);
 156     bExactCutoff        = bExactElecCutoff || bExactVdwCutoff;
 157
 158     if (bExactCutoff)
 159     {
 160         rcutoff  = ( fr->rcoulomb > fr->rvdw ) ? fr->rcoulomb : fr->rvdw;
 161         rcutoff2 = rcutoff*rcutoff;
 162     }
 163     else
 164     {
 165         /* Fix warnings for stupid compilers */
 166         rcutoff = rcutoff2 = 1e30;
 167     }
 168
 169     /* avoid compiler warnings for cases that cannot happen */
 170     nnn                 = 0;
 171     eps                 = 0.0;
 172     eps2                = 0.0;
 173
 174     /* 3 VdW parameters for buckingham, otherwise 2 */
 175     nvdwparam           = (ivdw == GMX_NBKERNEL_VDW_BUCKINGHAM) ? 3 : 2;
 176     table_nelements     = 12;
 177
 178     charge              = mdatoms->chargeA;
 179     type                = mdatoms->typeA;
 180     facel               = fr->epsfac;
 181     shiftvec            = fr->shift_vec[0];
 182     vdwparam            = fr->nbfp;
 183     ntype               = fr->ntype;
 184
 185     for (n = 0; (n < nlist->nri); n++)
 186     {
 187         is3              = 3*nlist->shift[n];
 188         shX              = shiftvec[is3];
 189         shY              = shiftvec[is3+1];
 190         shZ              = shiftvec[is3+2];
 191         nj0              = nlist->jindex[n];
 192         nj1              = nlist->jindex[n+1];
 193         ii               = nlist->iinr[n];
 194         ii3              = 3*ii;
 195         ix               = shX + x[ii3+0];
 196         iy               = shY + x[ii3+1];
 197         iz               = shZ + x[ii3+2];
 198         iq               = facel*charge[ii];
 199         nti              = nvdwparam*ntype*type[ii];
 200         vctot            = 0;
 201         vvdwtot          = 0;
 202         fix              = 0;
 203         fiy              = 0;
 204         fiz              = 0;
 205
 206         for (k = nj0; (k < nj1); k++)
 207         {
 208             jnr              = nlist->jjnr[k];
 209             j3               = 3*jnr;
 210             jx               = x[j3+0];
 211             jy               = x[j3+1];
 212             jz               = x[j3+2];
 213             dx               = ix - jx;
 214             dy               = iy - jy;
 215             dz               = iz - jz;
 216             rsq              = dx*dx+dy*dy+dz*dz;
 217             rinv             = gmx_invsqrt(rsq);
 218             rinvsq           = rinv*rinv;
 219             felec            = 0;
 220             fvdw             = 0;
 221             velec            = 0;
 222             vvdw             = 0;
 223
 224             if (bExactCutoff && rsq > rcutoff2)
 225             {
 226                 continue;
 227             }
 228
 229             if (ielec == GMX_NBKERNEL_ELEC_CUBICSPLINETABLE || ivdw == GMX_NBKERNEL_VDW_CUBICSPLINETABLE)
 230             {
 231                 r                = rsq*rinv;
 232                 rt               = r*tabscale;
 233                 n0               = rt;
 234                 eps              = rt-n0;
 235                 eps2             = eps*eps;
 236                 nnn              = table_nelements*n0;
 237             }
 238
 239             /* Coulomb interaction. ielec==0 means no interaction */
 240             if (ielec != GMX_NBKERNEL_ELEC_NONE)
 241             {
 242                 qq               = iq*charge[jnr];
 243
 244                 switch (ielec)
 245                 {
 246                     case GMX_NBKERNEL_ELEC_NONE:
 247                         break;
 248
 249                     case GMX_NBKERNEL_ELEC_COULOMB:
 250                         /* Vanilla cutoff coulomb */
 251                         velec            = qq*rinv;
 252                         felec            = velec*rinvsq;
 253                         break;
 254
 255                     case GMX_NBKERNEL_ELEC_REACTIONFIELD:
 256                         /* Reaction-field */
 257                         velec            = qq*(rinv+fr->k_rf*rsq-fr->c_rf);
 258                         felec            = qq*(rinv*rinvsq-2.0*fr->k_rf);
 259                         break;
 260
 261                     case GMX_NBKERNEL_ELEC_CUBICSPLINETABLE:
 262                         /* Tabulated coulomb */
 263                         Y                = VFtab[nnn];
 264                         F                = VFtab[nnn+1];
 265                         Geps             = eps*VFtab[nnn+2];
 266                         Heps2            = eps2*VFtab[nnn+3];
 267                         Fp               = F+Geps+Heps2;
 268                         VV               = Y+eps*Fp;
 269                         FF               = Fp+Geps+2.0*Heps2;
 270                         velec            = qq*VV;
 271                         felec            = -qq*FF*tabscale*rinv;
 272                         break;
 273
 274                     case GMX_NBKERNEL_ELEC_GENERALIZEDBORN:
 275                         /* GB */
 276                         gmx_fatal(FARGS, "Death & horror! GB generic interaction not implemented.\n");
 277                         break;
 278
 279                     case GMX_NBKERNEL_ELEC_EWALD:
 280                         ewrt             = rsq*rinv*ewtabscale;
 281                         ewitab           = ewrt;
 282                         eweps            = ewrt-ewitab;
 283                         ewitab           = 4*ewitab;
 284                         felec            = ewtab[ewitab]+eweps*ewtab[ewitab+1];
 285                         rinvcorr         = (fr->coulomb_modifier == eintmodPOTSHIFT) ? rinv-fr->ic->sh_ewald : rinv;
 286                         velec            = qq*(rinvcorr-(ewtab[ewitab+2]-ewtabhalfspace*eweps*(ewtab[ewitab]+felec)));
 287                         felec            = qq*rinv*(rinvsq-felec);
 288                         break;
 289
 290                     default:
 291                         gmx_fatal(FARGS, "Death & horror! No generic coulomb interaction for ielec=%d.\n", ielec);
 292                         break;
 293                 }
 294                 if (fr->coulomb_modifier == eintmodPOTSWITCH)
 295                 {
 296                     d                = rsq*rinv-fr->rcoulomb_switch;
 297                     d                = (d > 0.0) ? d : 0.0;
 298                     d2               = d*d;
 299                     sw               = 1.0+d2*d*(elec_swV3+d*(elec_swV4+d*elec_swV5));
 300                     dsw              = d2*(elec_swF2+d*(elec_swF3+d*elec_swF4));
 301                     /* Apply switch function. Note that felec=f/r since it will be multiplied
 302                      * by the i-j displacement vector. This means felec'=f'/r=-(v*sw)'/r=
 303                      * -(v'*sw+v*dsw)/r=-v'*sw/r-v*dsw/r=felec*sw-v*dsw/r
 304                      */
 305                     felec            = felec*sw - rinv*velec*dsw;
 306                     /* Once we have used velec to update felec we can modify velec too */
 307                     velec           *= sw;
 308                 }
 309                 if (bExactElecCutoff)
 310                 {
 311                     felec            = (rsq <= rcoulomb2) ? felec : 0.0;
 312                     velec            = (rsq <= rcoulomb2) ? velec : 0.0;
 313                 }
 314                 vctot           += velec;
 315             } /* End of coulomb interactions */
 316
 317
 318             /* VdW interaction. ivdw==0 means no interaction */
 319             if (ivdw != GMX_NBKERNEL_VDW_NONE)
 320             {
 321                 tj               = nti+nvdwparam*type[jnr];
 322
 323                 switch (ivdw)
 324                 {
 325                     case GMX_NBKERNEL_VDW_NONE:
 326                         break;
 327
 328                     case GMX_NBKERNEL_VDW_LENNARDJONES:
 329                         /* Vanilla Lennard-Jones cutoff */
 330                         c6               = vdwparam[tj];
 331                         c12              = vdwparam[tj+1];
 332                         rinvsix          = rinvsq*rinvsq*rinvsq;
 333                         vvdw_disp        = c6*rinvsix;
 334                         vvdw_rep         = c12*rinvsix*rinvsix;
 335                         fvdw             = (vvdw_rep-vvdw_disp)*rinvsq;
 336                         if (fr->vdw_modifier == eintmodPOTSHIFT)
 337                         {
 338                             vvdw             = (vvdw_rep-c12*sh_invrc6*sh_invrc6)*(1.0/12.0)-(vvdw_disp-c6*sh_invrc6)*(1.0/6.0);
 339                         }
 340                         else
 341                         {
 342                             vvdw             = vvdw_rep/12.0-vvdw_disp/6.0;
 343                         }
 344                         break;
 345
 346                     case GMX_NBKERNEL_VDW_BUCKINGHAM:
 347                         /* Buckingham */
 348                         c6               = vdwparam[tj];
 349                         cexp1            = vdwparam[tj+1];
 350                         cexp2            = vdwparam[tj+2];
 351
 352                         rinvsix          = rinvsq*rinvsq*rinvsq;
 353                         vvdw_disp        = c6*rinvsix;
 354                         br               = cexp2*rsq*rinv;
 355                         vvdw_rep         = cexp1*exp(-br);
 356                         fvdw             = (br*vvdw_rep-vvdw_disp)*rinvsq;
 357                         if (fr->vdw_modifier == eintmodPOTSHIFT)
 358                         {
 359                             vvdw             = (vvdw_rep-cexp1*exp(-cexp2*rvdw))-(vvdw_disp-c6*sh_invrc6)/6.0;
 360                         }
 361                         else
 362                         {
 363                             vvdw             = vvdw_rep-vvdw_disp/6.0;
 364                         }
 365                         break;
 366
 367                     case GMX_NBKERNEL_VDW_CUBICSPLINETABLE:
 368                         /* Tabulated VdW */
 369                         c6               = vdwparam[tj];
 370                         c12              = vdwparam[tj+1];
 371                         Y                = VFtab[nnn+4];
 372                         F                = VFtab[nnn+5];
 373                         Geps             = eps*VFtab[nnn+6];
 374                         Heps2            = eps2*VFtab[nnn+7];
 375                         Fp               = F+Geps+Heps2;
 376                         VV               = Y+eps*Fp;
 377                         FF               = Fp+Geps+2.0*Heps2;
 378                         vvdw_disp        = c6*VV;
 379                         fijD             = c6*FF;
 380                         Y                = VFtab[nnn+8];
 381                         F                = VFtab[nnn+9];
 382                         Geps             = eps*VFtab[nnn+10];
 383                         Heps2            = eps2*VFtab[nnn+11];
 384                         Fp               = F+Geps+Heps2;
 385                         VV               = Y+eps*Fp;
 386                         FF               = Fp+Geps+2.0*Heps2;
 387                         vvdw_rep         = c12*VV;
 388                         fijR             = c12*FF;
 389                         fvdw             = -(fijD+fijR)*tabscale*rinv;
 390                         vvdw             = vvdw_disp + vvdw_rep;
 391                         break;
 392
 393                     default:
 394                         gmx_fatal(FARGS, "Death & horror! No generic VdW interaction for ivdw=%d.\n", ivdw);
 395                         break;
 396                 }
 397                 if (fr->vdw_modifier == eintmodPOTSWITCH)
 398                 {
 399                     d                = rsq*rinv-fr->rvdw_switch;
 400                     d                = (d > 0.0) ? d : 0.0;
 401                     d2               = d*d;
 402                     sw               = 1.0+d2*d*(vdw_swV3+d*(vdw_swV4+d*vdw_swV5));
 403                     dsw              = d2*(vdw_swF2+d*(vdw_swF3+d*vdw_swF4));
 404                     /* See coulomb interaction for the force-switch formula */
 405                     fvdw             = fvdw*sw - rinv*vvdw*dsw;
 406                     vvdw            *= sw;
 407                 }
 408                 if (bExactVdwCutoff)
 409                 {
 410                     fvdw             = (rsq <= rvdw2) ? fvdw : 0.0;
 411                     vvdw             = (rsq <= rvdw2) ? vvdw : 0.0;
 412                 }
 413                 vvdwtot         += vvdw;
 414             } /* end VdW interactions */
 415
 416             fscal            = felec+fvdw;
 417
 418             tx               = fscal*dx;
 419             ty               = fscal*dy;
 420             tz               = fscal*dz;
 421             fix              = fix + tx;
 422             fiy              = fiy + ty;
 423             fiz              = fiz + tz;
 424             f[j3+0]          = f[j3+0] - tx;
 425             f[j3+1]          = f[j3+1] - ty;
 426             f[j3+2]          = f[j3+2] - tz;
 427         }
 428
 429         f[ii3+0]         = f[ii3+0] + fix;
 430         f[ii3+1]         = f[ii3+1] + fiy;
 431         f[ii3+2]         = f[ii3+2] + fiz;
 432         fshift[is3]      = fshift[is3]+fix;
 433         fshift[is3+1]    = fshift[is3+1]+fiy;
 434         fshift[is3+2]    = fshift[is3+2]+fiz;
 435         ggid             = nlist->gid[n];
 436         velecgrp[ggid]  += vctot;
 437         vvdwgrp[ggid]   += vvdwtot;
 438     }
 439     /* Estimate flops, average for generic kernel:
 440      * 12 flops per outer iteration
 441      * 50 flops per inner iteration
 442      */
 443     inc_nrnb(nrnb, eNR_NBKERNEL_GENERIC, nlist->nri*12 + nlist->jindex[n]*50);
 444 }