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     gmx_enerdata_t ed_lam;
 174     double  clam_i,vlam_i;
 175     real    dvdl_dum[efptNR], dvdl, dvdl_nb[efptNR], lam_i[efptNR];
 176     real    dvdlsum;
 177
 178 #ifdef GMX_MPI
 179     double  t0=0.0,t1,t2,t3; /* time measurement for coarse load balancing */
 180 #endif
 181
 182 #define PRINT_SEPDVDL(s,v,dvdlambda) if (bSepDVDL) fprintf(fplog,sepdvdlformat,s,v,dvdlambda);
 183
 184     GMX_MPE_LOG(ev_force_start);
 185     set_pbc(&pbc,fr->ePBC,box);
 186
 187     /* reset free energy components */
 188     for (i=0;i<efptNR;i++)
 189     {
 190         dvdl_nb[i]  = 0;
 191         dvdl_dum[i] = 0;
 192     }
 193
 194     /* Reset box */
 195     for(i=0; (i<DIM); i++)
 196     {
 197         box_size[i]=box[i][i];
 198     }
 199
 200     bSepDVDL=(fr->bSepDVDL && do_per_step(step,ir->nstlog));
 201     debug_gmx();
 202
 203     /* do QMMM first if requested */
 204     if(fr->bQMMM)
 205     {
 206         enerd->term[F_EQM] = calculate_QMMM(cr,x,f,fr,md);
 207     }
 208
 209     if (bSepDVDL)
 210     {
 211         fprintf(fplog,"Step %s: non-bonded V and dVdl for node %d:\n",
 212                 gmx_step_str(step,buf),cr->nodeid);
 213     }
 214
 215     /* Call the short range functions all in one go. */
 216     GMX_MPE_LOG(ev_do_fnbf_start);
 217
 218 #ifdef GMX_MPI
 219     /*#define TAKETIME ((cr->npmenodes) && (fr->timesteps < 12))*/
 220 #define TAKETIME FALSE
 221     if (TAKETIME)
 222     {
 223         MPI_Barrier(cr->mpi_comm_mygroup);
 224         t0=MPI_Wtime();
 225     }
 226 #endif
 227
 228     if (ir->nwall)
 229     {
 230         /* foreign lambda component for walls */
 231         dvdl = do_walls(ir,fr,box,md,x,f,lambda[efptVDW],
 232                         enerd->grpp.ener[egLJSR],nrnb);
 233         PRINT_SEPDVDL("Walls",0.0,dvdl);
 234         enerd->dvdl_lin[efptVDW] += dvdl;
 235     }
 236
 237         /* If doing GB, reset dvda and calculate the Born radii */
 238         if (ir->implicit_solvent)
 239         {
 240         wallcycle_sub_start(wcycle, ewcsNONBONDED);
 241
 242                 for(i=0;i<born->nr;i++)
 243                 {
 244                         fr->dvda[i]=0;
 245                 }
 246
 247                 if(bBornRadii)
 248                 {
 249                         calc_gb_rad(cr,fr,ir,top,atype,x,&(fr->gblist),born,md,nrnb);
 250                 }
 251
 252         wallcycle_sub_stop(wcycle, ewcsNONBONDED);
 253         }
 254
 255     where();
 256     /* We only do non-bonded calculation with group scheme here, the verlet
 257      * calls are done from do_force_cutsVERLET(). */
 258     if (fr->cutoff_scheme == ecutsGROUP && (flags & GMX_FORCE_NONBONDED))
 259     {
 260         donb_flags = 0;
 261         /* Add short-range interactions */
 262         donb_flags |= GMX_NONBONDED_DO_SR;
 263
 264         if (flags & GMX_FORCE_FORCES)
 265         {
 266             donb_flags |= GMX_NONBONDED_DO_FORCE;
 267         }
 268         if (flags & GMX_FORCE_ENERGY)
 269         {
 270             donb_flags |= GMX_NONBONDED_DO_POTENTIAL;
 271         }
 272         if (flags & GMX_FORCE_DO_LR)
 273         {
 274             donb_flags |= GMX_NONBONDED_DO_LR;
 275         }
 276
 277         wallcycle_sub_start(wcycle, ewcsNONBONDED);
 278         do_nonbonded(cr,fr,x,f,f_longrange,md,excl,
 279                     &enerd->grpp,box_size,nrnb,
 280                     lambda,dvdl_nb,-1,-1,donb_flags);
 281         wallcycle_sub_stop(wcycle, ewcsNONBONDED);
 282     }
 283
 284     /* If we do foreign lambda and we have soft-core interactions
 285      * we have to recalculate the (non-linear) energies contributions.
 286      */
 287     if (fepvals->n_lambda > 0 && (flags & GMX_FORCE_DHDL) && fepvals->sc_alpha != 0)
 288     {
 289         wallcycle_sub_start(wcycle, ewcsNONBONDED);
 290         init_enerdata(mtop->groups.grps[egcENER].nr,fepvals->n_lambda,&ed_lam);
 291
 292         for(i=0; i<enerd->n_lambda; i++)
 293         {
 294             for (j=0;j<efptNR;j++)
 295             {
 296                 lam_i[j] = (i==0 ? lambda[j] : fepvals->all_lambda[j][i-1]);
 297             }
 298             reset_enerdata(&ir->opts,fr,TRUE,&ed_lam,FALSE);
 299             do_nonbonded(cr,fr,x,f,f_longrange,md,excl,
 300                          &(ed_lam.grpp), box_size,nrnb,
 301                          lam_i,dvdl_dum,-1,-1,
 302                          GMX_NONBONDED_DO_FOREIGNLAMBDA | GMX_NONBONDED_DO_SR);
 303             sum_epot(&ir->opts,&ed_lam);
 304             enerd->enerpart_lambda[i] += ed_lam.term[F_EPOT];
 305         }
 306         destroy_enerdata(&ed_lam);
 307         wallcycle_sub_stop(wcycle, ewcsNONBONDED);
 308     }
 309     where();
 310
 311         /* If we are doing GB, calculate bonded forces and apply corrections
 312          * to the solvation forces */
 313     /* MRS: Eventually, many need to include free energy contribution here! */
 314         if (ir->implicit_solvent)
 315     {
 316         wallcycle_sub_start(wcycle, ewcsBONDED);
 317                 calc_gb_forces(cr,md,born,top,atype,x,f,fr,idef,
 318                        ir->gb_algorithm,ir->sa_algorithm,nrnb,bBornRadii,&pbc,graph,enerd);
 319         wallcycle_sub_stop(wcycle, ewcsBONDED);
 320     }
 321
 322 #ifdef GMX_MPI
 323     if (TAKETIME)
 324     {
 325         t1=MPI_Wtime();
 326         fr->t_fnbf += t1-t0;
 327     }
 328 #endif
 329
 330     if (fepvals->sc_alpha!=0)
 331     {
 332         enerd->dvdl_nonlin[efptVDW] += dvdl_nb[efptVDW];
 333     }
 334     else
 335     {
 336         enerd->dvdl_lin[efptVDW] += dvdl_nb[efptVDW];
 337     }
 338
 339     if (fepvals->sc_alpha!=0)
 340
 341         /* even though coulomb part is linear, we already added it, beacuse we
 342            need to go through the vdw calculation anyway */
 343     {
 344         enerd->dvdl_nonlin[efptCOUL] += dvdl_nb[efptCOUL];
 345     }
 346     else
 347     {
 348         enerd->dvdl_lin[efptCOUL] += dvdl_nb[efptCOUL];
 349     }
 350
 351     Vsr = 0;
 352     if (bSepDVDL)
 353     {
 354         for(i=0; i<enerd->grpp.nener; i++)
 355         {
 356             Vsr +=
 357                 (fr->bBHAM ?
 358                  enerd->grpp.ener[egBHAMSR][i] :
 359                  enerd->grpp.ener[egLJSR][i])
 360                 + enerd->grpp.ener[egCOULSR][i] + enerd->grpp.ener[egGB][i];
 361         }
 362         dvdlsum = dvdl_nb[efptVDW] + dvdl_nb[efptCOUL];
 363         PRINT_SEPDVDL("VdW and Coulomb SR particle-p.",Vsr,dvdlsum);
 364     }
 365     debug_gmx();
 366
 367     GMX_MPE_LOG(ev_do_fnbf_finish);
 368
 369     if (debug)
 370     {
 371         pr_rvecs(debug,0,"fshift after SR",fr->fshift,SHIFTS);
 372     }
 373
 374     /* Shift the coordinates. Must be done before bonded forces and PPPM,
 375      * but is also necessary for SHAKE and update, therefore it can NOT
 376      * go when no bonded forces have to be evaluated.
 377      */
 378
 379     /* Here sometimes we would not need to shift with NBFonly,
 380      * but we do so anyhow for consistency of the returned coordinates.
 381      */
 382     if (graph)
 383     {
 384         shift_self(graph,box,x);
 385         if (TRICLINIC(box))
 386         {
 387             inc_nrnb(nrnb,eNR_SHIFTX,2*graph->nnodes);
 388         }
 389         else
 390         {
 391             inc_nrnb(nrnb,eNR_SHIFTX,graph->nnodes);
 392         }
 393     }
 394     /* Check whether we need to do bondeds or correct for exclusions */
 395     if (fr->bMolPBC &&
 396         ((flags & GMX_FORCE_BONDED)
 397          || EEL_RF(fr->eeltype) || EEL_FULL(fr->eeltype)))
 398     {
 399         /* Since all atoms are in the rectangular or triclinic unit-cell,
 400          * only single box vector shifts (2 in x) are required.
 401          */
 402         set_pbc_dd(&pbc,fr->ePBC,cr->dd,TRUE,box);
 403     }
 404     debug_gmx();
 405
 406     if (flags & GMX_FORCE_BONDED)
 407     {
 408         GMX_MPE_LOG(ev_calc_bonds_start);
 409
 410         wallcycle_sub_start(wcycle, ewcsBONDED);
 411         calc_bonds(fplog,cr->ms,
 412                    idef,x,hist,f,fr,&pbc,graph,enerd,nrnb,lambda,md,fcd,
 413                    DOMAINDECOMP(cr) ? cr->dd->gatindex : NULL, atype, born,
 414                    flags,
 415                    fr->bSepDVDL && do_per_step(step,ir->nstlog),step);
 416
 417         /* Check if we have to determine energy differences
 418          * at foreign lambda's.
 419          */
 420         if (fepvals->n_lambda > 0 && (flags & GMX_FORCE_DHDL) &&
 421             idef->ilsort != ilsortNO_FE)
 422         {
 423             if (idef->ilsort != ilsortFE_SORTED)
 424             {
 425                 gmx_incons("The bonded interactions are not sorted for free energy");
 426             }
 427             init_enerdata(mtop->groups.grps[egcENER].nr,fepvals->n_lambda,&ed_lam);
 428
 429             for(i=0; i<enerd->n_lambda; i++)
 430             {
 431                 reset_enerdata(&ir->opts,fr,TRUE,&ed_lam,FALSE);
 432                 for (j=0;j<efptNR;j++)
 433                 {
 434                     lam_i[j] = (i==0 ? lambda[j] : fepvals->all_lambda[j][i-1]);
 435                 }
 436                 calc_bonds_lambda(fplog,idef,x,fr,&pbc,graph,&ed_lam,nrnb,lam_i,md,
 437                                   fcd,DOMAINDECOMP(cr) ? cr->dd->gatindex : NULL);
 438                 sum_epot(&ir->opts,&ed_lam);
 439                 enerd->enerpart_lambda[i] += ed_lam.term[F_EPOT];
 440             }
 441             destroy_enerdata(&ed_lam);
 442         }
 443         debug_gmx();
 444         GMX_MPE_LOG(ev_calc_bonds_finish);
 445         wallcycle_sub_stop(wcycle, ewcsBONDED);
 446     }
 447
 448     where();
 449
 450     *cycles_pme = 0;
 451     if (EEL_FULL(fr->eeltype))
 452     {
 453         bSB = (ir->nwall == 2);
 454         if (bSB)
 455         {
 456             copy_mat(box,boxs);
 457             svmul(ir->wall_ewald_zfac,boxs[ZZ],boxs[ZZ]);
 458             box_size[ZZ] *= ir->wall_ewald_zfac;
 459         }
 460
 461         clear_mat(fr->vir_el_recip);
 462
 463         if (fr->bEwald)
 464         {
 465             Vcorr = 0;
 466             dvdl  = 0;
 467
 468             /* With the Verlet scheme exclusion forces are calculated
 469              * in the non-bonded kernel.
 470              */
 471             /* The TPI molecule does not have exclusions with the rest
 472              * of the system and no intra-molecular PME grid contributions
 473              * will be calculated in gmx_pme_calc_energy.
 474              */
 475             if ((ir->cutoff_scheme == ecutsGROUP && fr->n_tpi == 0) ||
 476                 ir->ewald_geometry != eewg3D ||
 477                 ir->epsilon_surface != 0)
 478             {
 479                 int nthreads,t;
 480
 481                 wallcycle_sub_start(wcycle, ewcsEWALD_CORRECTION);
 482
 483                 if (fr->n_tpi > 0)
 484                 {
 485                     gmx_fatal(FARGS,"TPI with PME currently only works in a 3D geometry with tin-foil boundary conditions");
 486                 }
 487
 488                 nthreads = gmx_omp_nthreads_get(emntBonded);
 489 #pragma omp parallel for num_threads(nthreads) schedule(static)
 490                 for(t=0; t<nthreads; t++)
 491                 {
 492                     int s,e,i;
 493                     rvec *fnv;
 494                     tensor *vir;
 495                     real *Vcorrt,*dvdlt;
 496                     if (t == 0)
 497                     {
 498                         fnv    = fr->f_novirsum;
 499                         vir    = &fr->vir_el_recip;
 500                         Vcorrt = &Vcorr;
 501                         dvdlt  = &dvdl;
 502                     }
 503                     else
 504                     {
 505                         fnv    = fr->f_t[t].f;
 506                         vir    = &fr->f_t[t].vir;
 507                         Vcorrt = &fr->f_t[t].Vcorr;
 508                         dvdlt  = &fr->f_t[t].dvdl[efptCOUL];
 509                         for(i=0; i<fr->natoms_force; i++)
 510                         {
 511                             clear_rvec(fnv[i]);
 512                         }
 513                         clear_mat(*vir);
 514                     }
 515                     *dvdlt = 0;
 516                     *Vcorrt =
 517                         ewald_LRcorrection(fplog,
 518                                            fr->excl_load[t],fr->excl_load[t+1],
 519                                            cr,t,fr,
 520                                            md->chargeA,
 521                                            md->nChargePerturbed ? md->chargeB : NULL,
 522                                            ir->cutoff_scheme != ecutsVERLET,
 523                                            excl,x,bSB ? boxs : box,mu_tot,
 524                                            ir->ewald_geometry,
 525                                            ir->epsilon_surface,
 526                                            fnv,*vir,
 527                                            lambda[efptCOUL],dvdlt);
 528                 }
 529                 if (nthreads > 1)
 530                 {
 531                     reduce_thread_forces(fr->natoms_force,fr->f_novirsum,
 532                                          fr->vir_el_recip,
 533                                          &Vcorr,efptCOUL,&dvdl,
 534                                          nthreads,fr->f_t);
 535                 }
 536
 537                 wallcycle_sub_stop(wcycle, ewcsEWALD_CORRECTION);
 538             }
 539
 540             if (fr->n_tpi == 0)
 541             {
 542                 Vcorr += ewald_charge_correction(cr,fr,lambda[efptCOUL],box,
 543                                                  &dvdl,fr->vir_el_recip);
 544             }
 545
 546             PRINT_SEPDVDL("Ewald excl./charge/dip. corr.",Vcorr,dvdl);
 547             enerd->dvdl_lin[efptCOUL] += dvdl;
 548         }
 549
 550         status = 0;
 551         Vlr  = 0;
 552         dvdl = 0;
 553         switch (fr->eeltype)
 554         {
 555         case eelPME:
 556         case eelPMESWITCH:
 557         case eelPMEUSER:
 558         case eelPMEUSERSWITCH:
 559         case eelP3M_AD:
 560             if (cr->duty & DUTY_PME)
 561             {
 562                 assert(fr->n_tpi >= 0);
 563                 if (fr->n_tpi == 0 || (flags & GMX_FORCE_STATECHANGED))
 564                 {
 565                     pme_flags = GMX_PME_SPREAD_Q | GMX_PME_SOLVE;
 566                     if (flags & GMX_FORCE_FORCES)
 567                     {
 568                         pme_flags |= GMX_PME_CALC_F;
 569                     }
 570                     if (flags & (GMX_FORCE_VIRIAL | GMX_FORCE_ENERGY))
 571                     {
 572                         pme_flags |= GMX_PME_CALC_ENER_VIR;
 573                     }
 574                     if (fr->n_tpi > 0)
 575                     {
 576                         /* We don't calculate f, but we do want the potential */
 577                         pme_flags |= GMX_PME_CALC_POT;
 578                     }
 579                     wallcycle_start(wcycle,ewcPMEMESH);
 580                     status = gmx_pme_do(fr->pmedata,
 581                                         md->start,md->homenr - fr->n_tpi,
 582                                         x,fr->f_novirsum,
 583                                         md->chargeA,md->chargeB,
 584                                         bSB ? boxs : box,cr,
 585                                         DOMAINDECOMP(cr) ? dd_pme_maxshift_x(cr->dd) : 0,
 586                                         DOMAINDECOMP(cr) ? dd_pme_maxshift_y(cr->dd) : 0,
 587                                         nrnb,wcycle,
 588                                         fr->vir_el_recip,fr->ewaldcoeff,
 589                                         &Vlr,lambda[efptCOUL],&dvdl,
 590                                         pme_flags);
 591                     *cycles_pme = wallcycle_stop(wcycle,ewcPMEMESH);
 592
 593                     /* We should try to do as little computation after
 594                      * this as possible, because parallel PME synchronizes
 595                      * the nodes, so we want all load imbalance of the rest
 596                      * of the force calculation to be before the PME call.
 597                      * DD load balancing is done on the whole time of
 598                      * the force call (without PME).
 599                      */
 600                 }
 601                 if (fr->n_tpi > 0)
 602                 {
 603                     /* Determine the PME grid energy of the test molecule
 604                      * with the PME grid potential of the other charges.
 605                      */
 606                     gmx_pme_calc_energy(fr->pmedata,fr->n_tpi,
 607                                         x + md->homenr - fr->n_tpi,
 608                                         md->chargeA + md->homenr - fr->n_tpi,
 609                                         &Vlr);
 610                 }
 611                 PRINT_SEPDVDL("PME mesh",Vlr,dvdl);
 612             }
 613             break;
 614         case eelEWALD:
 615             Vlr = do_ewald(fplog,FALSE,ir,x,fr->f_novirsum,
 616                            md->chargeA,md->chargeB,
 617                            box_size,cr,md->homenr,
 618                            fr->vir_el_recip,fr->ewaldcoeff,
 619                            lambda[efptCOUL],&dvdl,fr->ewald_table);
 620             PRINT_SEPDVDL("Ewald long-range",Vlr,dvdl);
 621             break;
 622         default:
 623             gmx_fatal(FARGS,"No such electrostatics method implemented %s",
 624                       eel_names[fr->eeltype]);
 625         }
 626         if (status != 0)
 627         {
 628             gmx_fatal(FARGS,"Error %d in long range electrostatics routine %s",
 629                       status,EELTYPE(fr->eeltype));
 630                 }
 631         /* Note that with separate PME nodes we get the real energies later */
 632         enerd->dvdl_lin[efptCOUL] += dvdl;
 633         enerd->term[F_COUL_RECIP] = Vlr + Vcorr;
 634         if (debug)
 635         {
 636             fprintf(debug,"Vlr = %g, Vcorr = %g, Vlr_corr = %g\n",
 637                     Vlr,Vcorr,enerd->term[F_COUL_RECIP]);
 638             pr_rvecs(debug,0,"vir_el_recip after corr",fr->vir_el_recip,DIM);
 639             pr_rvecs(debug,0,"fshift after LR Corrections",fr->fshift,SHIFTS);
 640         }
 641     }
 642     else
 643     {
 644         if (EEL_RF(fr->eeltype))
 645         {
 646             /* With the Verlet scheme exclusion forces are calculated
 647              * in the non-bonded kernel.
 648              */
 649             if (ir->cutoff_scheme != ecutsVERLET && fr->eeltype != eelRF_NEC)
 650             {
 651                 dvdl = 0;
 652                 enerd->term[F_RF_EXCL] =
 653                     RF_excl_correction(fplog,fr,graph,md,excl,x,f,
 654                                        fr->fshift,&pbc,lambda[efptCOUL],&dvdl);
 655             }
 656
 657             enerd->dvdl_lin[efptCOUL] += dvdl;
 658             PRINT_SEPDVDL("RF exclusion correction",
 659                           enerd->term[F_RF_EXCL],dvdl);
 660         }
 661     }
 662     where();
 663     debug_gmx();
 664
 665     if (debug)
 666     {
 667         print_nrnb(debug,nrnb);
 668     }
 669     debug_gmx();
 670
 671 #ifdef GMX_MPI
 672     if (TAKETIME)
 673     {
 674         t2=MPI_Wtime();
 675         MPI_Barrier(cr->mpi_comm_mygroup);
 676         t3=MPI_Wtime();
 677         fr->t_wait += t3-t2;
 678         if (fr->timesteps == 11)
 679         {
 680             fprintf(stderr,"* PP load balancing info: node %d, step %s, rel wait time=%3.0f%% , load string value: %7.2f\n",
 681                     cr->nodeid, gmx_step_str(fr->timesteps,buf),
 682                     100*fr->t_wait/(fr->t_wait+fr->t_fnbf),
 683                     (fr->t_fnbf+fr->t_wait)/fr->t_fnbf);
 684         }
 685         fr->timesteps++;
 686     }
 687 #endif
 688
 689     if (debug)
 690     {
 691         pr_rvecs(debug,0,"fshift after bondeds",fr->fshift,SHIFTS);
 692     }
 693
 694     GMX_MPE_LOG(ev_force_finish);
 695
 696 }
 697
 698 void init_enerdata(int ngener,int n_lambda,gmx_enerdata_t *enerd)
 699 {
 700     int i,n2;
 701
 702     for(i=0; i<F_NRE; i++)
 703     {
 704         enerd->term[i] = 0;
 705     }
 706
 707
 708     for(i=0; i<efptNR; i++) {
 709         enerd->dvdl_lin[i]  = 0;
 710         enerd->dvdl_nonlin[i]  = 0;
 711     }
 712
 713     n2=ngener*ngener;
 714     if (debug)
 715     {
 716         fprintf(debug,"Creating %d sized group matrix for energies\n",n2);
 717     }
 718     enerd->grpp.nener = n2;
 719     for(i=0; (i<egNR); i++)
 720     {
 721         snew(enerd->grpp.ener[i],n2);
 722     }
 723
 724     if (n_lambda)
 725     {
 726         enerd->n_lambda = 1 + n_lambda;
 727         snew(enerd->enerpart_lambda,enerd->n_lambda);
 728     }
 729     else
 730     {
 731         enerd->n_lambda = 0;
 732     }
 733 }
 734
 735 void destroy_enerdata(gmx_enerdata_t *enerd)
 736 {
 737     int i;
 738
 739     for(i=0; (i<egNR); i++)
 740     {
 741         sfree(enerd->grpp.ener[i]);
 742     }
 743
 744     if (enerd->n_lambda)
 745     {
 746         sfree(enerd->enerpart_lambda);
 747     }
 748 }
 749
 750 static real sum_v(int n,real v[])
 751 {
 752   real t;
 753   int  i;
 754
 755   t = 0.0;
 756   for(i=0; (i<n); i++)
 757     t = t + v[i];
 758
 759   return t;
 760 }
 761
 762 void sum_epot(t_grpopts *opts,gmx_enerdata_t *enerd)
 763 {
 764   gmx_grppairener_t *grpp;
 765   real *epot;
 766   int i;
 767
 768   grpp = &enerd->grpp;
 769   epot = enerd->term;
 770
 771   /* Accumulate energies */
 772   epot[F_COUL_SR]  = sum_v(grpp->nener,grpp->ener[egCOULSR]);
 773   epot[F_LJ]       = sum_v(grpp->nener,grpp->ener[egLJSR]);
 774   epot[F_LJ14]     = sum_v(grpp->nener,grpp->ener[egLJ14]);
 775   epot[F_COUL14]   = sum_v(grpp->nener,grpp->ener[egCOUL14]);
 776   epot[F_COUL_LR]  = sum_v(grpp->nener,grpp->ener[egCOULLR]);
 777   epot[F_LJ_LR]    = sum_v(grpp->nener,grpp->ener[egLJLR]);
 778   /* We have already added 1-2,1-3, and 1-4 terms to F_GBPOL */
 779   epot[F_GBPOL]   += sum_v(grpp->nener,grpp->ener[egGB]);
 780
 781 /* lattice part of LR doesnt belong to any group
 782  * and has been added earlier
 783  */
 784   epot[F_BHAM]     = sum_v(grpp->nener,grpp->ener[egBHAMSR]);
 785   epot[F_BHAM_LR]  = sum_v(grpp->nener,grpp->ener[egBHAMLR]);
 786
 787   epot[F_EPOT] = 0;
 788   for(i=0; (i<F_EPOT); i++)
 789   {
 790       if (i != F_DISRESVIOL && i != F_ORIRESDEV)
 791       {
 792           epot[F_EPOT] += epot[i];
 793       }
 794   }
 795 }
 796
 797 void sum_dhdl(gmx_enerdata_t *enerd, real *lambda, t_lambda *fepvals)
 798 {
 799     int i,j,index;
 800     double dlam;
 801
 802     enerd->dvdl_lin[efptVDW] += enerd->term[F_DVDL_VDW];  /* include dispersion correction */
 803     enerd->term[F_DVDL] = 0.0;
 804     for (i=0;i<efptNR;i++)
 805     {
 806         if (fepvals->separate_dvdl[i])
 807         {
 808             /* could this be done more readably/compactly? */
 809             switch (i) {
 810             case (efptCOUL):
 811                 index = F_DVDL_COUL;
 812                 break;
 813             case (efptVDW):
 814                 index = F_DVDL_VDW;
 815                 break;
 816             case (efptBONDED):
 817                 index = F_DVDL_BONDED;
 818                 break;
 819             case (efptRESTRAINT):
 820                 index = F_DVDL_RESTRAINT;
 821                 break;
 822             case (efptMASS):
 823                 index = F_DKDL;
 824                 break;
 825             default:
 826                 index = F_DVDL;
 827                 break;
 828             }
 829             enerd->term[index] = enerd->dvdl_lin[i] + enerd->dvdl_nonlin[i];
 830             if (debug)
 831             {
 832                 fprintf(debug,"dvdl-%s[%2d]: %f: non-linear %f + linear %f\n",
 833                         efpt_names[i],i,enerd->term[index],enerd->dvdl_nonlin[i],enerd->dvdl_lin[i]);
 834             }
 835         }
 836         else
 837         {
 838             enerd->term[F_DVDL] += enerd->dvdl_lin[i] + enerd->dvdl_nonlin[i];
 839             if (debug)
 840             {
 841                 fprintf(debug,"dvd-%sl[%2d]: %f: non-linear %f + linear %f\n",
 842                         efpt_names[0],i,enerd->term[F_DVDL],enerd->dvdl_nonlin[i],enerd->dvdl_lin[i]);
 843             }
 844         }
 845     }
 846
 847     /* Notes on the foreign lambda free energy difference evaluation:
 848      * Adding the potential and ekin terms that depend linearly on lambda
 849      * as delta lam * dvdl to the energy differences is exact.
 850      * For the constraints this is not exact, but we have no other option
 851      * without literally changing the lengths and reevaluating the energies at each step.
 852      * (try to remedy this post 4.6 - MRS)
 853      * For the non-bonded LR term we assume that the soft-core (if present)
 854      * no longer affects the energy beyond the short-range cut-off,
 855      * which is a very good approximation (except for exotic settings).
 856      * (investigate how to overcome this post 4.6 - MRS)
 857      */
 858
 859     for(i=0; i<fepvals->n_lambda; i++)
 860     {                                         /* note we are iterating over fepvals here!
 861                                                  For the current lam, dlam = 0 automatically,
 862                                                  so we don't need to add anything to the
 863                                                  enerd->enerpart_lambda[0] */
 864
 865         /* we don't need to worry about dvdl contributions to the current lambda, because
 866            it's automatically zero */
 867
 868         /* first kinetic energy term */
 869         dlam = (fepvals->all_lambda[efptMASS][i] - lambda[efptMASS]);
 870
 871         enerd->enerpart_lambda[i+1] += enerd->term[F_DKDL]*dlam;
 872
 873         for (j=0;j<efptNR;j++)
 874         {
 875             if (j==efptMASS) {continue;} /* no other mass term to worry about */
 876
 877             dlam = (fepvals->all_lambda[j][i]-lambda[j]);
 878             enerd->enerpart_lambda[i+1] += dlam*enerd->dvdl_lin[j];
 879             if (debug)
 880             {
 881                 fprintf(debug,"enerdiff lam %g: (%15s), non-linear %f linear %f*%f\n",
 882                         fepvals->all_lambda[j][i],efpt_names[j],
 883                         (enerd->enerpart_lambda[i+1] - enerd->enerpart_lambda[0]),
 884                         dlam,enerd->dvdl_lin[j]);
 885             }
 886         }
 887     }
 888 }
 889
 890 void reset_enerdata(t_grpopts *opts,
 891                     t_forcerec *fr,gmx_bool bNS,
 892                     gmx_enerdata_t *enerd,
 893                     gmx_bool bMaster)
 894 {
 895     gmx_bool bKeepLR;
 896     int  i,j;
 897
 898     /* First reset all energy components, except for the long range terms
 899      * on the master at non neighbor search steps, since the long range
 900      * terms have already been summed at the last neighbor search step.
 901      */
 902     bKeepLR = (fr->bTwinRange && !bNS);
 903     for(i=0; (i<egNR); i++) {
 904         if (!(bKeepLR && bMaster && (i == egCOULLR || i == egLJLR))) {
 905             for(j=0; (j<enerd->grpp.nener); j++)
 906                 enerd->grpp.ener[i][j] = 0.0;
 907         }
 908     }
 909     for (i=0;i<efptNR;i++)
 910     {
 911         enerd->dvdl_lin[i]    = 0.0;
 912         enerd->dvdl_nonlin[i] = 0.0;
 913     }
 914
 915     /* Normal potential energy components */
 916     for(i=0; (i<=F_EPOT); i++) {
 917         enerd->term[i] = 0.0;
 918     }
 919     /* Initialize the dVdlambda term with the long range contribution */
 920     /* Initialize the dvdl term with the long range contribution */
 921     enerd->term[F_DVDL]            = 0.0;
 922     enerd->term[F_DVDL_COUL]       = 0.0;
 923     enerd->term[F_DVDL_VDW]        = 0.0;
 924     enerd->term[F_DVDL_BONDED]     = 0.0;
 925     enerd->term[F_DVDL_RESTRAINT]  = 0.0;
 926     enerd->term[F_DKDL]            = 0.0;
 927     if (enerd->n_lambda > 0)
 928     {
 929         for(i=0; i<enerd->n_lambda; i++)
 930         {
 931             enerd->enerpart_lambda[i] = 0.0;
 932         }
 933     }
 934 }