src/mdlib/force.c

   1 /* -*- mode: c; tab-width: 4; indent-tabs-mode: nil; c-basic-offset: 4; c-file-style: "stroustrup"; -*-
   2  *
   3  *
   4  *                This source code is part of
   5  *
   6  *                 G   R   O   M   A   C   S
   7  *
   8  *          GROningen MAchine for Chemical Simulations
   9  *
  10  *                        VERSION 3.2.0
  11  * Written by David van der Spoel, Erik Lindahl, Berk Hess, and others.
  12  * Copyright (c) 1991-2000, University of Groningen, The Netherlands.
  13  * Copyright (c) 2001-2004, The GROMACS development team,
  14  * check out http://www.gromacs.org for more information.
  15
  16  * This program is free software; you can redistribute it and/or
  17  * modify it under the terms of the GNU General Public License
  18  * as published by the Free Software Foundation; either version 2
  19  * of the License, or (at your option) any later version.
  20  *
  21  * If you want to redistribute modifications, please consider that
  22  * scientific software is very special. Version control is crucial -
  23  * bugs must be traceable. We will be happy to consider code for
  24  * inclusion in the official distribution, but derived work must not
  25  * be called official GROMACS. Details are found in the README & COPYING
  26  * files - if they are missing, get the official version at www.gromacs.org.
  27  *
  28  * To help us fund GROMACS development, we humbly ask that you cite
  29  * the papers on the package - you can find them in the top README file.
  30  *
  31  * For more info, check our website at http://www.gromacs.org
  32  *
  33  * And Hey:
  34  * GROwing Monsters And Cloning Shrimps
  35  */
  36 #ifdef HAVE_CONFIG_H
  37 #include <config.h>
  38 #endif
  39
  40 #include <math.h>
  41 #include <string.h>
  42 #include <assert.h>
  43 #include "sysstuff.h"
  44 #include "typedefs.h"
  45 #include "macros.h"
  46 #include "smalloc.h"
  47 #include "macros.h"
  48 #include "physics.h"
  49 #include "force.h"
  50 #include "nonbonded.h"
  51 #include "names.h"
  52 #include "network.h"
  53 #include "pbc.h"
  54 #include "ns.h"
  55 #include "nrnb.h"
  56 #include "bondf.h"
  57 #include "mshift.h"
  58 #include "txtdump.h"
  59 #include "coulomb.h"
  60 #include "pme.h"
  61 #include "mdrun.h"
  62 #include "domdec.h"
  63 #include "partdec.h"
  64 #include "qmmm.h"
  65 #include "mpelogging.h"
  66 #include "gmx_omp_nthreads.h"
  67
  68 void ns(FILE *fp,
  69         t_forcerec *fr,
  70         rvec       x[],
  71         matrix     box,
  72         gmx_groups_t *groups,
  73         t_grpopts  *opts,
  74         gmx_localtop_t *top,
  75         t_mdatoms  *md,
  76         t_commrec  *cr,
  77         t_nrnb     *nrnb,
  78         real       *lambda,
  79         real       *dvdlambda,
  80         gmx_grppairener_t *grppener,
  81         gmx_bool       bFillGrid,
  82         gmx_bool       bDoLongRangeNS)
  83 {
  84   char   *ptr;
  85   int    nsearch;
  86
  87   GMX_MPE_LOG(ev_ns_start);
  88   if (!fr->ns.nblist_initialized)
  89   {
  90       init_neighbor_list(fp, fr, md->homenr);
  91   }
  92
  93   if (fr->bTwinRange)
  94     fr->nlr=0;
  95
  96     nsearch = search_neighbours(fp,fr,x,box,top,groups,cr,nrnb,md,
  97                                 lambda,dvdlambda,grppener,
  98                                 bFillGrid,bDoLongRangeNS,TRUE);
  99   if (debug)
 100     fprintf(debug,"nsearch = %d\n",nsearch);
 101
 102   /* Check whether we have to do dynamic load balancing */
 103   /*if ((nsb->nstDlb > 0) && (mod(step,nsb->nstDlb) == 0))
 104     count_nb(cr,nsb,&(top->blocks[ebCGS]),nns,fr->nlr,
 105     &(top->idef),opts->ngener);
 106   */
 107   if (fr->ns.dump_nl > 0)
 108     dump_nblist(fp,cr,fr,fr->ns.dump_nl);
 109
 110   GMX_MPE_LOG(ev_ns_finish);
 111 }
 112
 113 static void reduce_thread_forces(int n,rvec *f,
 114                                  tensor vir,
 115                                  real *Vcorr,
 116                                  int efpt_ind,real *dvdl,
 117                                  int nthreads,f_thread_t *f_t)
 118 {
 119     int t,i;
 120
 121     /* This reduction can run over any number of threads */
 122 #pragma omp parallel for num_threads(gmx_omp_nthreads_get(emntBonded)) private(t) schedule(static)
 123     for(i=0; i<n; i++)
 124     {
 125         for(t=1; t<nthreads; t++)
 126         {
 127             rvec_inc(f[i],f_t[t].f[i]);
 128         }
 129     }
 130     for(t=1; t<nthreads; t++)
 131     {
 132         *Vcorr += f_t[t].Vcorr;
 133         *dvdl  += f_t[t].dvdl[efpt_ind];
 134         m_add(vir,f_t[t].vir,vir);
 135     }
 136 }
 137
 138 void do_force_lowlevel(FILE       *fplog,   gmx_large_int_t step,
 139                        t_forcerec *fr,      t_inputrec *ir,
 140                        t_idef     *idef,    t_commrec  *cr,
 141                        t_nrnb     *nrnb,    gmx_wallcycle_t wcycle,
 142                        t_mdatoms  *md,
 143                        t_grpopts  *opts,
 144                        rvec       x[],      history_t  *hist,
 145                        rvec       f[],
 146                        rvec       f_longrange[],
 147                        gmx_enerdata_t *enerd,
 148                        t_fcdata   *fcd,
 149                        gmx_mtop_t     *mtop,
 150                        gmx_localtop_t *top,
 151                        gmx_genborn_t *born,
 152                        t_atomtypes *atype,
 153                        gmx_bool       bBornRadii,
 154                        matrix     box,
 155                        t_lambda   *fepvals,
 156                        real       *lambda,
 157                        t_graph    *graph,
 158                        t_blocka   *excl,
 159                        rvec       mu_tot[],
 160                        int        flags,
 161                        float      *cycles_pme)
 162 {
 163     int     i,j,status;
 164     int     donb_flags;
 165     gmx_bool    bDoEpot,bSepDVDL,bSB;
 166     int     pme_flags;
 167     matrix  boxs;
 168     rvec    box_size;
 169     real    Vsr,Vlr,Vcorr=0;
 170     t_pbc   pbc;
 171     real    dvdgb;
 172     char    buf[22];
 173     double  clam_i,vlam_i;
 174     real    dvdl_dum[efptNR], dvdl, dvdl_nb[efptNR], lam_i[efptNR];
 175     real    dvdlsum;
 176
 177 #ifdef GMX_MPI
 178     double  t0=0.0,t1,t2,t3; /* time measurement for coarse load balancing */
 179 #endif
 180
 181 #define PRINT_SEPDVDL(s,v,dvdlambda) if (bSepDVDL) fprintf(fplog,sepdvdlformat,s,v,dvdlambda);
 182
 183     GMX_MPE_LOG(ev_force_start);
 184     set_pbc(&pbc,fr->ePBC,box);
 185
 186     /* reset free energy components */
 187     for (i=0;i<efptNR;i++)
 188     {
 189         dvdl_nb[i]  = 0;
 190         dvdl_dum[i] = 0;
 191     }
 192
 193     /* Reset box */
 194     for(i=0; (i<DIM); i++)
 195     {
 196         box_size[i]=box[i][i];
 197     }
 198
 199     bSepDVDL=(fr->bSepDVDL && do_per_step(step,ir->nstlog));
 200     debug_gmx();
 201
 202     /* do QMMM first if requested */
 203     if(fr->bQMMM)
 204     {
 205         enerd->term[F_EQM] = calculate_QMMM(cr,x,f,fr,md);
 206     }
 207
 208     if (bSepDVDL)
 209     {
 210         fprintf(fplog,"Step %s: non-bonded V and dVdl for node %d:\n",
 211                 gmx_step_str(step,buf),cr->nodeid);
 212     }
 213
 214     /* Call the short range functions all in one go. */
 215     GMX_MPE_LOG(ev_do_fnbf_start);
 216
 217 #ifdef GMX_MPI
 218     /*#define TAKETIME ((cr->npmenodes) && (fr->timesteps < 12))*/
 219 #define TAKETIME FALSE
 220     if (TAKETIME)
 221     {
 222         MPI_Barrier(cr->mpi_comm_mygroup);
 223         t0=MPI_Wtime();
 224     }
 225 #endif
 226
 227     if (ir->nwall)
 228     {
 229         /* foreign lambda component for walls */
 230         dvdl = do_walls(ir,fr,box,md,x,f,lambda[efptVDW],
 231                         enerd->grpp.ener[egLJSR],nrnb);
 232         PRINT_SEPDVDL("Walls",0.0,dvdl);
 233         enerd->dvdl_lin[efptVDW] += dvdl;
 234     }
 235
 236         /* If doing GB, reset dvda and calculate the Born radii */
 237         if (ir->implicit_solvent)
 238         {
 239         wallcycle_sub_start(wcycle, ewcsNONBONDED);
 240
 241                 for(i=0;i<born->nr;i++)
 242                 {
 243                         fr->dvda[i]=0;
 244                 }
 245
 246                 if(bBornRadii)
 247                 {
 248                         calc_gb_rad(cr,fr,ir,top,atype,x,&(fr->gblist),born,md,nrnb);
 249                 }
 250
 251         wallcycle_sub_stop(wcycle, ewcsNONBONDED);
 252         }
 253
 254     where();
 255     /* We only do non-bonded calculation with group scheme here, the verlet
 256      * calls are done from do_force_cutsVERLET(). */
 257     if (fr->cutoff_scheme == ecutsGROUP && (flags & GMX_FORCE_NONBONDED))
 258     {
 259         donb_flags = 0;
 260         /* Add short-range interactions */
 261         donb_flags |= GMX_NONBONDED_DO_SR;
 262
 263         if (flags & GMX_FORCE_FORCES)
 264         {
 265             donb_flags |= GMX_NONBONDED_DO_FORCE;
 266         }
 267         if (flags & GMX_FORCE_ENERGY)
 268         {
 269             donb_flags |= GMX_NONBONDED_DO_POTENTIAL;
 270         }
 271         if (flags & GMX_FORCE_DO_LR)
 272         {
 273             donb_flags |= GMX_NONBONDED_DO_LR;
 274         }
 275
 276         wallcycle_sub_start(wcycle, ewcsNONBONDED);
 277         do_nonbonded(cr,fr,x,f,f_longrange,md,excl,
 278                     &enerd->grpp,box_size,nrnb,
 279                     lambda,dvdl_nb,-1,-1,donb_flags);
 280
 281         /* If we do foreign lambda and we have soft-core interactions
 282          * we have to recalculate the (non-linear) energies contributions.
 283          */
 284         if (fepvals->n_lambda > 0 && (flags & GMX_FORCE_DHDL) && fepvals->sc_alpha != 0)
 285         {
 286             for(i=0; i<enerd->n_lambda; i++)
 287             {
 288                 for (j=0;j<efptNR;j++)
 289                 {
 290                     lam_i[j] = (i==0 ? lambda[j] : fepvals->all_lambda[j][i-1]);
 291                 }
 292                 reset_foreign_enerdata(enerd);
 293                 do_nonbonded(cr,fr,x,f,f_longrange,md,excl,
 294                              &(enerd->foreign_grpp),box_size,nrnb,
 295                              lam_i,dvdl_dum,-1,-1,
 296                              (donb_flags & ~GMX_NONBONDED_DO_FORCE) | GMX_NONBONDED_DO_FOREIGNLAMBDA);
 297                 sum_epot(&ir->opts,&(enerd->foreign_grpp),enerd->foreign_term);
 298                 enerd->enerpart_lambda[i] += enerd->foreign_term[F_EPOT];
 299             }
 300         }
 301         wallcycle_sub_stop(wcycle, ewcsNONBONDED);
 302         where();
 303     }
 304
 305         /* If we are doing GB, calculate bonded forces and apply corrections
 306          * to the solvation forces */
 307     /* MRS: Eventually, many need to include free energy contribution here! */
 308         if (ir->implicit_solvent)
 309     {
 310         wallcycle_sub_start(wcycle, ewcsBONDED);
 311                 calc_gb_forces(cr,md,born,top,atype,x,f,fr,idef,
 312                        ir->gb_algorithm,ir->sa_algorithm,nrnb,bBornRadii,&pbc,graph,enerd);
 313         wallcycle_sub_stop(wcycle, ewcsBONDED);
 314     }
 315
 316 #ifdef GMX_MPI
 317     if (TAKETIME)
 318     {
 319         t1=MPI_Wtime();
 320         fr->t_fnbf += t1-t0;
 321     }
 322 #endif
 323
 324     if (fepvals->sc_alpha!=0)
 325     {
 326         enerd->dvdl_nonlin[efptVDW] += dvdl_nb[efptVDW];
 327     }
 328     else
 329     {
 330         enerd->dvdl_lin[efptVDW] += dvdl_nb[efptVDW];
 331     }
 332
 333     if (fepvals->sc_alpha!=0)
 334
 335         /* even though coulomb part is linear, we already added it, beacuse we
 336            need to go through the vdw calculation anyway */
 337     {
 338         enerd->dvdl_nonlin[efptCOUL] += dvdl_nb[efptCOUL];
 339     }
 340     else
 341     {
 342         enerd->dvdl_lin[efptCOUL] += dvdl_nb[efptCOUL];
 343     }
 344
 345     Vsr = 0;
 346     if (bSepDVDL)
 347     {
 348         for(i=0; i<enerd->grpp.nener; i++)
 349         {
 350             Vsr +=
 351                 (fr->bBHAM ?
 352                  enerd->grpp.ener[egBHAMSR][i] :
 353                  enerd->grpp.ener[egLJSR][i])
 354                 + enerd->grpp.ener[egCOULSR][i] + enerd->grpp.ener[egGB][i];
 355         }
 356         dvdlsum = dvdl_nb[efptVDW] + dvdl_nb[efptCOUL];
 357         PRINT_SEPDVDL("VdW and Coulomb SR particle-p.",Vsr,dvdlsum);
 358     }
 359     debug_gmx();
 360
 361     GMX_MPE_LOG(ev_do_fnbf_finish);
 362
 363     if (debug)
 364     {
 365         pr_rvecs(debug,0,"fshift after SR",fr->fshift,SHIFTS);
 366     }
 367
 368     /* Shift the coordinates. Must be done before bonded forces and PPPM,
 369      * but is also necessary for SHAKE and update, therefore it can NOT
 370      * go when no bonded forces have to be evaluated.
 371      */
 372
 373     /* Here sometimes we would not need to shift with NBFonly,
 374      * but we do so anyhow for consistency of the returned coordinates.
 375      */
 376     if (graph)
 377     {
 378         shift_self(graph,box,x);
 379         if (TRICLINIC(box))
 380         {
 381             inc_nrnb(nrnb,eNR_SHIFTX,2*graph->nnodes);
 382         }
 383         else
 384         {
 385             inc_nrnb(nrnb,eNR_SHIFTX,graph->nnodes);
 386         }
 387     }
 388     /* Check whether we need to do bondeds or correct for exclusions */
 389     if (fr->bMolPBC &&
 390         ((flags & GMX_FORCE_BONDED)
 391          || EEL_RF(fr->eeltype) || EEL_FULL(fr->eeltype)))
 392     {
 393         /* Since all atoms are in the rectangular or triclinic unit-cell,
 394          * only single box vector shifts (2 in x) are required.
 395          */
 396         set_pbc_dd(&pbc,fr->ePBC,cr->dd,TRUE,box);
 397     }
 398     debug_gmx();
 399
 400     if (flags & GMX_FORCE_BONDED)
 401     {
 402         GMX_MPE_LOG(ev_calc_bonds_start);
 403
 404         wallcycle_sub_start(wcycle, ewcsBONDED);
 405         calc_bonds(fplog,cr->ms,
 406                    idef,x,hist,f,fr,&pbc,graph,enerd,nrnb,lambda,md,fcd,
 407                    DOMAINDECOMP(cr) ? cr->dd->gatindex : NULL, atype, born,
 408                    flags,
 409                    fr->bSepDVDL && do_per_step(step,ir->nstlog),step);
 410
 411         /* Check if we have to determine energy differences
 412          * at foreign lambda's.
 413          */
 414         if (fepvals->n_lambda > 0 && (flags & GMX_FORCE_DHDL) &&
 415             idef->ilsort != ilsortNO_FE)
 416         {
 417             if (idef->ilsort != ilsortFE_SORTED)
 418             {
 419                 gmx_incons("The bonded interactions are not sorted for free energy");
 420             }
 421             for(i=0; i<enerd->n_lambda; i++)
 422             {
 423                 reset_foreign_enerdata(enerd);
 424                 for (j=0;j<efptNR;j++)
 425                 {
 426                     lam_i[j] = (i==0 ? lambda[j] : fepvals->all_lambda[j][i-1]);
 427                 }
 428                 calc_bonds_lambda(fplog,idef,x,fr,&pbc,graph,&(enerd->foreign_grpp),enerd->foreign_term,nrnb,lam_i,md,
 429                                   fcd,DOMAINDECOMP(cr) ? cr->dd->gatindex : NULL);
 430                 sum_epot(&ir->opts,&(enerd->foreign_grpp),enerd->foreign_term);
 431                 enerd->enerpart_lambda[i] += enerd->foreign_term[F_EPOT];
 432             }
 433         }
 434         debug_gmx();
 435         GMX_MPE_LOG(ev_calc_bonds_finish);
 436         wallcycle_sub_stop(wcycle, ewcsBONDED);
 437     }
 438
 439     where();
 440
 441     *cycles_pme = 0;
 442     if (EEL_FULL(fr->eeltype))
 443     {
 444         bSB = (ir->nwall == 2);
 445         if (bSB)
 446         {
 447             copy_mat(box,boxs);
 448             svmul(ir->wall_ewald_zfac,boxs[ZZ],boxs[ZZ]);
 449             box_size[ZZ] *= ir->wall_ewald_zfac;
 450         }
 451
 452         clear_mat(fr->vir_el_recip);
 453
 454         if (fr->bEwald)
 455         {
 456             Vcorr = 0;
 457             dvdl  = 0;
 458
 459             /* With the Verlet scheme exclusion forces are calculated
 460              * in the non-bonded kernel.
 461              */
 462             /* The TPI molecule does not have exclusions with the rest
 463              * of the system and no intra-molecular PME grid contributions
 464              * will be calculated in gmx_pme_calc_energy.
 465              */
 466             if ((ir->cutoff_scheme == ecutsGROUP && fr->n_tpi == 0) ||
 467                 ir->ewald_geometry != eewg3D ||
 468                 ir->epsilon_surface != 0)
 469             {
 470                 int nthreads,t;
 471
 472                 wallcycle_sub_start(wcycle, ewcsEWALD_CORRECTION);
 473
 474                 if (fr->n_tpi > 0)
 475                 {
 476                     gmx_fatal(FARGS,"TPI with PME currently only works in a 3D geometry with tin-foil boundary conditions");
 477                 }
 478
 479                 nthreads = gmx_omp_nthreads_get(emntBonded);
 480 #pragma omp parallel for num_threads(nthreads) schedule(static)
 481                 for(t=0; t<nthreads; t++)
 482                 {
 483                     int s,e,i;
 484                     rvec *fnv;
 485                     tensor *vir;
 486                     real *Vcorrt,*dvdlt;
 487                     if (t == 0)
 488                     {
 489                         fnv    = fr->f_novirsum;
 490                         vir    = &fr->vir_el_recip;
 491                         Vcorrt = &Vcorr;
 492                         dvdlt  = &dvdl;
 493                     }
 494                     else
 495                     {
 496                         fnv    = fr->f_t[t].f;
 497                         vir    = &fr->f_t[t].vir;
 498                         Vcorrt = &fr->f_t[t].Vcorr;
 499                         dvdlt  = &fr->f_t[t].dvdl[efptCOUL];
 500                         for(i=0; i<fr->natoms_force; i++)
 501                         {
 502                             clear_rvec(fnv[i]);
 503                         }
 504                         clear_mat(*vir);
 505                     }
 506                     *dvdlt = 0;
 507                     *Vcorrt =
 508                         ewald_LRcorrection(fplog,
 509                                            fr->excl_load[t],fr->excl_load[t+1],
 510                                            cr,t,fr,
 511                                            md->chargeA,
 512                                            md->nChargePerturbed ? md->chargeB : NULL,
 513                                            ir->cutoff_scheme != ecutsVERLET,
 514                                            excl,x,bSB ? boxs : box,mu_tot,
 515                                            ir->ewald_geometry,
 516                                            ir->epsilon_surface,
 517                                            fnv,*vir,
 518                                            lambda[efptCOUL],dvdlt);
 519                 }
 520                 if (nthreads > 1)
 521                 {
 522                     reduce_thread_forces(fr->natoms_force,fr->f_novirsum,
 523                                          fr->vir_el_recip,
 524                                          &Vcorr,efptCOUL,&dvdl,
 525                                          nthreads,fr->f_t);
 526                 }
 527
 528                 wallcycle_sub_stop(wcycle, ewcsEWALD_CORRECTION);
 529             }
 530
 531             if (fr->n_tpi == 0)
 532             {
 533                 Vcorr += ewald_charge_correction(cr,fr,lambda[efptCOUL],box,
 534                                                  &dvdl,fr->vir_el_recip);
 535             }
 536
 537             PRINT_SEPDVDL("Ewald excl./charge/dip. corr.",Vcorr,dvdl);
 538             enerd->dvdl_lin[efptCOUL] += dvdl;
 539         }
 540
 541         status = 0;
 542         Vlr  = 0;
 543         dvdl = 0;
 544         switch (fr->eeltype)
 545         {
 546         case eelPME:
 547         case eelPMESWITCH:
 548         case eelPMEUSER:
 549         case eelPMEUSERSWITCH:
 550         case eelP3M_AD:
 551             if (cr->duty & DUTY_PME)
 552             {
 553                 assert(fr->n_tpi >= 0);
 554                 if (fr->n_tpi == 0 || (flags & GMX_FORCE_STATECHANGED))
 555                 {
 556                     pme_flags = GMX_PME_SPREAD_Q | GMX_PME_SOLVE;
 557                     if (flags & GMX_FORCE_FORCES)
 558                     {
 559                         pme_flags |= GMX_PME_CALC_F;
 560                     }
 561                     if (flags & (GMX_FORCE_VIRIAL | GMX_FORCE_ENERGY))
 562                     {
 563                         pme_flags |= GMX_PME_CALC_ENER_VIR;
 564                     }
 565                     if (fr->n_tpi > 0)
 566                     {
 567                         /* We don't calculate f, but we do want the potential */
 568                         pme_flags |= GMX_PME_CALC_POT;
 569                     }
 570                     wallcycle_start(wcycle,ewcPMEMESH);
 571                     status = gmx_pme_do(fr->pmedata,
 572                                         md->start,md->homenr - fr->n_tpi,
 573                                         x,fr->f_novirsum,
 574                                         md->chargeA,md->chargeB,
 575                                         bSB ? boxs : box,cr,
 576                                         DOMAINDECOMP(cr) ? dd_pme_maxshift_x(cr->dd) : 0,
 577                                         DOMAINDECOMP(cr) ? dd_pme_maxshift_y(cr->dd) : 0,
 578                                         nrnb,wcycle,
 579                                         fr->vir_el_recip,fr->ewaldcoeff,
 580                                         &Vlr,lambda[efptCOUL],&dvdl,
 581                                         pme_flags);
 582                     *cycles_pme = wallcycle_stop(wcycle,ewcPMEMESH);
 583
 584                     /* We should try to do as little computation after
 585                      * this as possible, because parallel PME synchronizes
 586                      * the nodes, so we want all load imbalance of the rest
 587                      * of the force calculation to be before the PME call.
 588                      * DD load balancing is done on the whole time of
 589                      * the force call (without PME).
 590                      */
 591                 }
 592                 if (fr->n_tpi > 0)
 593                 {
 594                     /* Determine the PME grid energy of the test molecule
 595                      * with the PME grid potential of the other charges.
 596                      */
 597                     gmx_pme_calc_energy(fr->pmedata,fr->n_tpi,
 598                                         x + md->homenr - fr->n_tpi,
 599                                         md->chargeA + md->homenr - fr->n_tpi,
 600                                         &Vlr);
 601                 }
 602                 PRINT_SEPDVDL("PME mesh",Vlr,dvdl);
 603             }
 604             break;
 605         case eelEWALD:
 606             Vlr = do_ewald(fplog,FALSE,ir,x,fr->f_novirsum,
 607                            md->chargeA,md->chargeB,
 608                            box_size,cr,md->homenr,
 609                            fr->vir_el_recip,fr->ewaldcoeff,
 610                            lambda[efptCOUL],&dvdl,fr->ewald_table);
 611             PRINT_SEPDVDL("Ewald long-range",Vlr,dvdl);
 612             break;
 613         default:
 614             gmx_fatal(FARGS,"No such electrostatics method implemented %s",
 615                       eel_names[fr->eeltype]);
 616         }
 617         if (status != 0)
 618         {
 619             gmx_fatal(FARGS,"Error %d in long range electrostatics routine %s",
 620                       status,EELTYPE(fr->eeltype));
 621                 }
 622         /* Note that with separate PME nodes we get the real energies later */
 623         enerd->dvdl_lin[efptCOUL] += dvdl;
 624         enerd->term[F_COUL_RECIP] = Vlr + Vcorr;
 625         if (debug)
 626         {
 627             fprintf(debug,"Vlr = %g, Vcorr = %g, Vlr_corr = %g\n",
 628                     Vlr,Vcorr,enerd->term[F_COUL_RECIP]);
 629             pr_rvecs(debug,0,"vir_el_recip after corr",fr->vir_el_recip,DIM);
 630             pr_rvecs(debug,0,"fshift after LR Corrections",fr->fshift,SHIFTS);
 631         }
 632     }
 633     else
 634     {
 635         if (EEL_RF(fr->eeltype))
 636         {
 637             /* With the Verlet scheme exclusion forces are calculated
 638              * in the non-bonded kernel.
 639              */
 640             if (ir->cutoff_scheme != ecutsVERLET && fr->eeltype != eelRF_NEC)
 641             {
 642                 dvdl = 0;
 643                 enerd->term[F_RF_EXCL] =
 644                     RF_excl_correction(fplog,fr,graph,md,excl,x,f,
 645                                        fr->fshift,&pbc,lambda[efptCOUL],&dvdl);
 646             }
 647
 648             enerd->dvdl_lin[efptCOUL] += dvdl;
 649             PRINT_SEPDVDL("RF exclusion correction",
 650                           enerd->term[F_RF_EXCL],dvdl);
 651         }
 652     }
 653     where();
 654     debug_gmx();
 655
 656     if (debug)
 657     {
 658         print_nrnb(debug,nrnb);
 659     }
 660     debug_gmx();
 661
 662 #ifdef GMX_MPI
 663     if (TAKETIME)
 664     {
 665         t2=MPI_Wtime();
 666         MPI_Barrier(cr->mpi_comm_mygroup);
 667         t3=MPI_Wtime();
 668         fr->t_wait += t3-t2;
 669         if (fr->timesteps == 11)
 670         {
 671             fprintf(stderr,"* PP load balancing info: node %d, step %s, rel wait time=%3.0f%% , load string value: %7.2f\n",
 672                     cr->nodeid, gmx_step_str(fr->timesteps,buf),
 673                     100*fr->t_wait/(fr->t_wait+fr->t_fnbf),
 674                     (fr->t_fnbf+fr->t_wait)/fr->t_fnbf);
 675         }
 676         fr->timesteps++;
 677     }
 678 #endif
 679
 680     if (debug)
 681     {
 682         pr_rvecs(debug,0,"fshift after bondeds",fr->fshift,SHIFTS);
 683     }
 684
 685     GMX_MPE_LOG(ev_force_finish);
 686
 687 }
 688
 689 void init_enerdata(int ngener,int n_lambda,gmx_enerdata_t *enerd)
 690 {
 691     int i,n2;
 692
 693     for(i=0; i<F_NRE; i++)
 694     {
 695         enerd->term[i] = 0;
 696         enerd->foreign_term[i] = 0;
 697     }
 698
 699
 700     for(i=0; i<efptNR; i++) {
 701         enerd->dvdl_lin[i]  = 0;
 702         enerd->dvdl_nonlin[i]  = 0;
 703     }
 704
 705     n2=ngener*ngener;
 706     if (debug)
 707     {
 708         fprintf(debug,"Creating %d sized group matrix for energies\n",n2);
 709     }
 710     enerd->grpp.nener = n2;
 711     enerd->foreign_grpp.nener = n2;
 712     for(i=0; (i<egNR); i++)
 713     {
 714         snew(enerd->grpp.ener[i],n2);
 715         snew(enerd->foreign_grpp.ener[i],n2);
 716     }
 717
 718     if (n_lambda)
 719     {
 720         enerd->n_lambda = 1 + n_lambda;
 721         snew(enerd->enerpart_lambda,enerd->n_lambda);
 722     }
 723     else
 724     {
 725         enerd->n_lambda = 0;
 726     }
 727 }
 728
 729 void destroy_enerdata(gmx_enerdata_t *enerd)
 730 {
 731     int i;
 732
 733     for(i=0; (i<egNR); i++)
 734     {
 735         sfree(enerd->grpp.ener[i]);
 736     }
 737
 738     for(i=0; (i<egNR); i++)
 739     {
 740         sfree(enerd->foreign_grpp.ener[i]);
 741     }
 742
 743     if (enerd->n_lambda)
 744     {
 745         sfree(enerd->enerpart_lambda);
 746     }
 747 }
 748
 749 static real sum_v(int n,real v[])
 750 {
 751   real t;
 752   int  i;
 753
 754   t = 0.0;
 755   for(i=0; (i<n); i++)
 756     t = t + v[i];
 757
 758   return t;
 759 }
 760
 761 void sum_epot(t_grpopts *opts, gmx_grppairener_t *grpp, real *epot)
 762 {
 763   int i;
 764
 765   /* Accumulate energies */
 766   epot[F_COUL_SR]  = sum_v(grpp->nener,grpp->ener[egCOULSR]);
 767   epot[F_LJ]       = sum_v(grpp->nener,grpp->ener[egLJSR]);
 768   epot[F_LJ14]     = sum_v(grpp->nener,grpp->ener[egLJ14]);
 769   epot[F_COUL14]   = sum_v(grpp->nener,grpp->ener[egCOUL14]);
 770   epot[F_COUL_LR]  = sum_v(grpp->nener,grpp->ener[egCOULLR]);
 771   epot[F_LJ_LR]    = sum_v(grpp->nener,grpp->ener[egLJLR]);
 772   /* We have already added 1-2,1-3, and 1-4 terms to F_GBPOL */
 773   epot[F_GBPOL]   += sum_v(grpp->nener,grpp->ener[egGB]);
 774
 775 /* lattice part of LR doesnt belong to any group
 776  * and has been added earlier
 777  */
 778   epot[F_BHAM]     = sum_v(grpp->nener,grpp->ener[egBHAMSR]);
 779   epot[F_BHAM_LR]  = sum_v(grpp->nener,grpp->ener[egBHAMLR]);
 780
 781   epot[F_EPOT] = 0;
 782   for(i=0; (i<F_EPOT); i++)
 783   {
 784       if (i != F_DISRESVIOL && i != F_ORIRESDEV)
 785       {
 786           epot[F_EPOT] += epot[i];
 787       }
 788   }
 789 }
 790
 791 void sum_dhdl(gmx_enerdata_t *enerd, real *lambda, t_lambda *fepvals)
 792 {
 793     int i,j,index;
 794     double dlam;
 795
 796     enerd->dvdl_lin[efptVDW] += enerd->term[F_DVDL_VDW];  /* include dispersion correction */
 797     enerd->term[F_DVDL] = 0.0;
 798     for (i=0;i<efptNR;i++)
 799     {
 800         if (fepvals->separate_dvdl[i])
 801         {
 802             /* could this be done more readably/compactly? */
 803             switch (i) {
 804             case (efptCOUL):
 805                 index = F_DVDL_COUL;
 806                 break;
 807             case (efptVDW):
 808                 index = F_DVDL_VDW;
 809                 break;
 810             case (efptBONDED):
 811                 index = F_DVDL_BONDED;
 812                 break;
 813             case (efptRESTRAINT):
 814                 index = F_DVDL_RESTRAINT;
 815                 break;
 816             case (efptMASS):
 817                 index = F_DKDL;
 818                 break;
 819             default:
 820                 index = F_DVDL;
 821                 break;
 822             }
 823             enerd->term[index] = enerd->dvdl_lin[i] + enerd->dvdl_nonlin[i];
 824             if (debug)
 825             {
 826                 fprintf(debug,"dvdl-%s[%2d]: %f: non-linear %f + linear %f\n",
 827                         efpt_names[i],i,enerd->term[index],enerd->dvdl_nonlin[i],enerd->dvdl_lin[i]);
 828             }
 829         }
 830         else
 831         {
 832             enerd->term[F_DVDL] += enerd->dvdl_lin[i] + enerd->dvdl_nonlin[i];
 833             if (debug)
 834             {
 835                 fprintf(debug,"dvd-%sl[%2d]: %f: non-linear %f + linear %f\n",
 836                         efpt_names[0],i,enerd->term[F_DVDL],enerd->dvdl_nonlin[i],enerd->dvdl_lin[i]);
 837             }
 838         }
 839     }
 840
 841     /* Notes on the foreign lambda free energy difference evaluation:
 842      * Adding the potential and ekin terms that depend linearly on lambda
 843      * as delta lam * dvdl to the energy differences is exact.
 844      * For the constraints this is not exact, but we have no other option
 845      * without literally changing the lengths and reevaluating the energies at each step.
 846      * (try to remedy this post 4.6 - MRS)
 847      * For the non-bonded LR term we assume that the soft-core (if present)
 848      * no longer affects the energy beyond the short-range cut-off,
 849      * which is a very good approximation (except for exotic settings).
 850      * (investigate how to overcome this post 4.6 - MRS)
 851      */
 852
 853     for(i=0; i<fepvals->n_lambda; i++)
 854     {                                         /* note we are iterating over fepvals here!
 855                                                  For the current lam, dlam = 0 automatically,
 856                                                  so we don't need to add anything to the
 857                                                  enerd->enerpart_lambda[0] */
 858
 859         /* we don't need to worry about dvdl contributions to the current lambda, because
 860            it's automatically zero */
 861
 862         /* first kinetic energy term */
 863         dlam = (fepvals->all_lambda[efptMASS][i] - lambda[efptMASS]);
 864
 865         enerd->enerpart_lambda[i+1] += enerd->term[F_DKDL]*dlam;
 866
 867         for (j=0;j<efptNR;j++)
 868         {
 869             if (j==efptMASS) {continue;} /* no other mass term to worry about */
 870
 871             dlam = (fepvals->all_lambda[j][i]-lambda[j]);
 872             enerd->enerpart_lambda[i+1] += dlam*enerd->dvdl_lin[j];
 873             if (debug)
 874             {
 875                 fprintf(debug,"enerdiff lam %g: (%15s), non-linear %f linear %f*%f\n",
 876                         fepvals->all_lambda[j][i],efpt_names[j],
 877                         (enerd->enerpart_lambda[i+1] - enerd->enerpart_lambda[0]),
 878                         dlam,enerd->dvdl_lin[j]);
 879             }
 880         }
 881     }
 882 }
 883
 884
 885 void reset_foreign_enerdata(gmx_enerdata_t *enerd)
 886 {
 887     int  i,j;
 888
 889     /* First reset all foreign energy components.  Foreign energies always called on
 890        neighbor search steps */
 891     for(i=0; (i<egNR); i++)
 892     {
 893         for(j=0; (j<enerd->grpp.nener); j++)
 894         {
 895             enerd->foreign_grpp.ener[i][j] = 0.0;
 896         }
 897     }
 898
 899     /* potential energy components */
 900     for(i=0; (i<=F_EPOT); i++)
 901     {
 902         enerd->foreign_term[i] = 0.0;
 903     }
 904 }
 905
 906 void reset_enerdata(t_grpopts *opts,
 907                     t_forcerec *fr,gmx_bool bNS,
 908                     gmx_enerdata_t *enerd,
 909                     gmx_bool bMaster)
 910 {
 911     gmx_bool bKeepLR;
 912     int  i,j;
 913
 914     /* First reset all energy components, except for the long range terms
 915      * on the master at non neighbor search steps, since the long range
 916      * terms have already been summed at the last neighbor search step.
 917      */
 918     bKeepLR = (fr->bTwinRange && !bNS);
 919     for(i=0; (i<egNR); i++) {
 920         if (!(bKeepLR && bMaster && (i == egCOULLR || i == egLJLR))) {
 921             for(j=0; (j<enerd->grpp.nener); j++)
 922                 enerd->grpp.ener[i][j] = 0.0;
 923         }
 924     }
 925     for (i=0;i<efptNR;i++)
 926     {
 927         enerd->dvdl_lin[i]    = 0.0;
 928         enerd->dvdl_nonlin[i] = 0.0;
 929     }
 930
 931     /* Normal potential energy components */
 932     for(i=0; (i<=F_EPOT); i++) {
 933         enerd->term[i] = 0.0;
 934     }
 935     /* Initialize the dVdlambda term with the long range contribution */
 936     /* Initialize the dvdl term with the long range contribution */
 937     enerd->term[F_DVDL]            = 0.0;
 938     enerd->term[F_DVDL_COUL]       = 0.0;
 939     enerd->term[F_DVDL_VDW]        = 0.0;
 940     enerd->term[F_DVDL_BONDED]     = 0.0;
 941     enerd->term[F_DVDL_RESTRAINT]  = 0.0;
 942     enerd->term[F_DKDL]            = 0.0;
 943     if (enerd->n_lambda > 0)
 944     {
 945         for(i=0; i<enerd->n_lambda; i++)
 946         {
 947             enerd->enerpart_lambda[i] = 0.0;
 948         }
 949     }
 950     /* reset foreign energy data - separate function since we also call it elsewhere */
 951     reset_foreign_enerdata(enerd);
 952 }