src/gromacs/mdlib/force.cpp

   1 /*
   2  * This file is part of the GROMACS molecular simulation package.
   3  *
   4  * Copyright (c) 1991-2000, University of Groningen, The Netherlands.
   5  * Copyright (c) 2001-2004, The GROMACS development team.
   6  * Copyright (c) 2013,2014,2015,2016, 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 #include "gmxpre.h"
  38
  39 #include "force.h"
  40
  41 #include "config.h"
  42
  43 #include <assert.h>
  44 #include <math.h>
  45 #include <string.h>
  46
  47 #include "gromacs/domdec/domdec.h"
  48 #include "gromacs/domdec/domdec_struct.h"
  49 #include "gromacs/ewald/ewald.h"
  50 #include "gromacs/ewald/long-range-correction.h"
  51 #include "gromacs/ewald/pme.h"
  52 #include "gromacs/gmxlib/network.h"
  53 #include "gromacs/gmxlib/nrnb.h"
  54 #include "gromacs/gmxlib/nonbonded/nonbonded.h"
  55 #include "gromacs/listed-forces/listed-forces.h"
  56 #include "gromacs/math/vec.h"
  57 #include "gromacs/math/vecdump.h"
  58 #include "gromacs/mdlib/forcerec-threading.h"
  59 #include "gromacs/mdlib/genborn.h"
  60 #include "gromacs/mdlib/mdrun.h"
  61 #include "gromacs/mdlib/ns.h"
  62 #include "gromacs/mdlib/qmmm.h"
  63 #include "gromacs/mdtypes/commrec.h"
  64 #include "gromacs/mdtypes/inputrec.h"
  65 #include "gromacs/mdtypes/md_enums.h"
  66 #include "gromacs/pbcutil/ishift.h"
  67 #include "gromacs/pbcutil/mshift.h"
  68 #include "gromacs/pbcutil/pbc.h"
  69 #include "gromacs/timing/wallcycle.h"
  70 #include "gromacs/utility/cstringutil.h"
  71 #include "gromacs/utility/exceptions.h"
  72 #include "gromacs/utility/fatalerror.h"
  73 #include "gromacs/utility/smalloc.h"
  74
  75 void ns(FILE              *fp,
  76         t_forcerec        *fr,
  77         matrix             box,
  78         gmx_groups_t      *groups,
  79         gmx_localtop_t    *top,
  80         t_mdatoms         *md,
  81         t_commrec         *cr,
  82         t_nrnb            *nrnb,
  83         gmx_bool           bFillGrid)
  84 {
  85     int     nsearch;
  86
  87
  88     if (!fr->ns->nblist_initialized)
  89     {
  90         init_neighbor_list(fp, fr, md->homenr);
  91     }
  92
  93     nsearch = search_neighbours(fp, fr, box, top, groups, cr, nrnb, md,
  94                                 bFillGrid);
  95     if (debug)
  96     {
  97         fprintf(debug, "nsearch = %d\n", nsearch);
  98     }
  99
 100     /* Check whether we have to do dynamic load balancing */
 101     /*if ((nsb->nstDlb > 0) && (mod(step,nsb->nstDlb) == 0))
 102        count_nb(cr,nsb,&(top->blocks[ebCGS]),nns,fr->nlr,
 103        &(top->idef),opts->ngener);
 104      */
 105     if (fr->ns->dump_nl > 0)
 106     {
 107         dump_nblist(fp, cr, fr, fr->ns->dump_nl);
 108     }
 109 }
 110
 111 static void reduce_thread_energies(tensor vir_q, tensor vir_lj,
 112                                    real *Vcorr_q, real *Vcorr_lj,
 113                                    real *dvdl_q, real *dvdl_lj,
 114                                    int nthreads,
 115                                    ewald_corr_thread_t *ewc_t)
 116 {
 117     int t;
 118
 119     for (t = 1; t < nthreads; t++)
 120     {
 121         *Vcorr_q  += ewc_t[t].Vcorr_q;
 122         *Vcorr_lj += ewc_t[t].Vcorr_lj;
 123         *dvdl_q   += ewc_t[t].dvdl[efptCOUL];
 124         *dvdl_lj  += ewc_t[t].dvdl[efptVDW];
 125         m_add(vir_q, ewc_t[t].vir_q, vir_q);
 126         m_add(vir_lj, ewc_t[t].vir_lj, vir_lj);
 127     }
 128 }
 129
 130 void do_force_lowlevel(t_forcerec *fr,      t_inputrec *ir,
 131                        t_idef     *idef,    t_commrec  *cr,
 132                        t_nrnb     *nrnb,    gmx_wallcycle_t wcycle,
 133                        t_mdatoms  *md,
 134                        rvec       x[],      history_t  *hist,
 135                        rvec       f[],
 136                        gmx_enerdata_t *enerd,
 137                        t_fcdata   *fcd,
 138                        gmx_localtop_t *top,
 139                        gmx_genborn_t *born,
 140                        gmx_bool       bBornRadii,
 141                        matrix     box,
 142                        t_lambda   *fepvals,
 143                        real       *lambda,
 144                        t_graph    *graph,
 145                        t_blocka   *excl,
 146                        rvec       mu_tot[],
 147                        int        flags,
 148                        float      *cycles_pme)
 149 {
 150     int         i, j;
 151     int         donb_flags;
 152     gmx_bool    bSB;
 153     int         pme_flags;
 154     matrix      boxs;
 155     rvec        box_size;
 156     t_pbc       pbc;
 157     real        dvdl_dum[efptNR], dvdl_nb[efptNR];
 158
 159 #if GMX_MPI
 160     double  t0 = 0.0, t1, t2, t3; /* time measurement for coarse load balancing */
 161 #endif
 162
 163     set_pbc(&pbc, fr->ePBC, box);
 164
 165     /* reset free energy components */
 166     for (i = 0; i < efptNR; i++)
 167     {
 168         dvdl_nb[i]  = 0;
 169         dvdl_dum[i] = 0;
 170     }
 171
 172     /* Reset box */
 173     for (i = 0; (i < DIM); i++)
 174     {
 175         box_size[i] = box[i][i];
 176     }
 177
 178     /* do QMMM first if requested */
 179     if (fr->bQMMM)
 180     {
 181         enerd->term[F_EQM] = calculate_QMMM(cr, x, f, fr);
 182     }
 183
 184     /* Call the short range functions all in one go. */
 185
 186 #if GMX_MPI
 187     /*#define TAKETIME ((cr->npmenodes) && (fr->timesteps < 12))*/
 188 #define TAKETIME FALSE
 189     if (TAKETIME)
 190     {
 191         MPI_Barrier(cr->mpi_comm_mygroup);
 192         t0 = MPI_Wtime();
 193     }
 194 #endif
 195
 196     if (ir->nwall)
 197     {
 198         /* foreign lambda component for walls */
 199         real dvdl_walls = do_walls(ir, fr, box, md, x, f, lambda[efptVDW],
 200                                    enerd->grpp.ener[egLJSR], nrnb);
 201         enerd->dvdl_lin[efptVDW] += dvdl_walls;
 202     }
 203
 204     /* If doing GB, reset dvda and calculate the Born radii */
 205     if (ir->implicit_solvent)
 206     {
 207         wallcycle_sub_start(wcycle, ewcsNONBONDED);
 208
 209         for (i = 0; i < born->nr; i++)
 210         {
 211             fr->dvda[i] = 0;
 212         }
 213
 214         if (bBornRadii)
 215         {
 216             calc_gb_rad(cr, fr, ir, top, x, fr->gblist, born, md, nrnb);
 217         }
 218
 219         wallcycle_sub_stop(wcycle, ewcsNONBONDED);
 220     }
 221
 222     where();
 223     /* We only do non-bonded calculation with group scheme here, the verlet
 224      * calls are done from do_force_cutsVERLET(). */
 225     if (fr->cutoff_scheme == ecutsGROUP && (flags & GMX_FORCE_NONBONDED))
 226     {
 227         donb_flags = 0;
 228         /* Add short-range interactions */
 229         donb_flags |= GMX_NONBONDED_DO_SR;
 230
 231         /* Currently all group scheme kernels always calculate (shift-)forces */
 232         if (flags & GMX_FORCE_FORCES)
 233         {
 234             donb_flags |= GMX_NONBONDED_DO_FORCE;
 235         }
 236         if (flags & GMX_FORCE_VIRIAL)
 237         {
 238             donb_flags |= GMX_NONBONDED_DO_SHIFTFORCE;
 239         }
 240         if (flags & GMX_FORCE_ENERGY)
 241         {
 242             donb_flags |= GMX_NONBONDED_DO_POTENTIAL;
 243         }
 244
 245         wallcycle_sub_start(wcycle, ewcsNONBONDED);
 246         do_nonbonded(fr, x, f, md, excl,
 247                      &enerd->grpp, nrnb,
 248                      lambda, dvdl_nb, -1, -1, donb_flags);
 249
 250         /* If we do foreign lambda and we have soft-core interactions
 251          * we have to recalculate the (non-linear) energies contributions.
 252          */
 253         if (fepvals->n_lambda > 0 && (flags & GMX_FORCE_DHDL) && fepvals->sc_alpha != 0)
 254         {
 255             for (i = 0; i < enerd->n_lambda; i++)
 256             {
 257                 real lam_i[efptNR];
 258
 259                 for (j = 0; j < efptNR; j++)
 260                 {
 261                     lam_i[j] = (i == 0 ? lambda[j] : fepvals->all_lambda[j][i-1]);
 262                 }
 263                 reset_foreign_enerdata(enerd);
 264                 do_nonbonded(fr, x, f, md, excl,
 265                              &(enerd->foreign_grpp), nrnb,
 266                              lam_i, dvdl_dum, -1, -1,
 267                              (donb_flags & ~GMX_NONBONDED_DO_FORCE) | GMX_NONBONDED_DO_FOREIGNLAMBDA);
 268                 sum_epot(&(enerd->foreign_grpp), enerd->foreign_term);
 269                 enerd->enerpart_lambda[i] += enerd->foreign_term[F_EPOT];
 270             }
 271         }
 272         wallcycle_sub_stop(wcycle, ewcsNONBONDED);
 273         where();
 274     }
 275
 276     /* If we are doing GB, calculate bonded forces and apply corrections
 277      * to the solvation forces */
 278     /* MRS: Eventually, many need to include free energy contribution here! */
 279     if (ir->implicit_solvent)
 280     {
 281         wallcycle_sub_start(wcycle, ewcsLISTED);
 282         calc_gb_forces(cr, md, born, top, x, f, fr, idef,
 283                        ir->gb_algorithm, ir->sa_algorithm, nrnb, &pbc, graph, enerd);
 284         wallcycle_sub_stop(wcycle, ewcsLISTED);
 285     }
 286
 287 #if GMX_MPI
 288     if (TAKETIME)
 289     {
 290         t1          = MPI_Wtime();
 291         fr->t_fnbf += t1-t0;
 292     }
 293 #endif
 294
 295     if (fepvals->sc_alpha != 0)
 296     {
 297         enerd->dvdl_nonlin[efptVDW] += dvdl_nb[efptVDW];
 298     }
 299     else
 300     {
 301         enerd->dvdl_lin[efptVDW] += dvdl_nb[efptVDW];
 302     }
 303
 304     if (fepvals->sc_alpha != 0)
 305
 306     /* even though coulomb part is linear, we already added it, beacuse we
 307        need to go through the vdw calculation anyway */
 308     {
 309         enerd->dvdl_nonlin[efptCOUL] += dvdl_nb[efptCOUL];
 310     }
 311     else
 312     {
 313         enerd->dvdl_lin[efptCOUL] += dvdl_nb[efptCOUL];
 314     }
 315
 316     if (debug)
 317     {
 318         pr_rvecs(debug, 0, "fshift after SR", fr->fshift, SHIFTS);
 319     }
 320
 321     /* Shift the coordinates. Must be done before listed forces and PPPM,
 322      * but is also necessary for SHAKE and update, therefore it can NOT
 323      * go when no listed forces have to be evaluated.
 324      *
 325      * The shifting and PBC code is deliberately not timed, since with
 326      * the Verlet scheme it only takes non-zero time with triclinic
 327      * boxes, and even then the time is around a factor of 100 less
 328      * than the next smallest counter.
 329      */
 330
 331
 332     /* Here sometimes we would not need to shift with NBFonly,
 333      * but we do so anyhow for consistency of the returned coordinates.
 334      */
 335     if (graph)
 336     {
 337         shift_self(graph, box, x);
 338         if (TRICLINIC(box))
 339         {
 340             inc_nrnb(nrnb, eNR_SHIFTX, 2*graph->nnodes);
 341         }
 342         else
 343         {
 344             inc_nrnb(nrnb, eNR_SHIFTX, graph->nnodes);
 345         }
 346     }
 347     /* Check whether we need to do listed interactions or correct for exclusions */
 348     if (fr->bMolPBC &&
 349         ((flags & GMX_FORCE_LISTED)
 350          || EEL_RF(fr->eeltype) || EEL_FULL(fr->eeltype) || EVDW_PME(fr->vdwtype)))
 351     {
 352         /* TODO There are no electrostatics methods that require this
 353            transformation, when using the Verlet scheme, so update the
 354            above conditional. */
 355         /* Since all atoms are in the rectangular or triclinic unit-cell,
 356          * only single box vector shifts (2 in x) are required.
 357          */
 358         set_pbc_dd(&pbc, fr->ePBC, DOMAINDECOMP(cr) ? cr->dd->nc : nullptr,
 359                    TRUE, box);
 360     }
 361
 362     do_force_listed(wcycle, box, ir->fepvals, cr->ms,
 363                     idef, (const rvec *) x, hist, f, fr,
 364                     &pbc, graph, enerd, nrnb, lambda, md, fcd,
 365                     DOMAINDECOMP(cr) ? cr->dd->gatindex : NULL,
 366                     flags);
 367
 368     where();
 369
 370     *cycles_pme = 0;
 371     clear_mat(fr->vir_el_recip);
 372     clear_mat(fr->vir_lj_recip);
 373
 374     /* Do long-range electrostatics and/or LJ-PME, including related short-range
 375      * corrections.
 376      */
 377     if (EEL_FULL(fr->eeltype) || EVDW_PME(fr->vdwtype))
 378     {
 379         int  status            = 0;
 380         real Vlr_q             = 0, Vlr_lj = 0, Vcorr_q = 0, Vcorr_lj = 0;
 381         real dvdl_long_range_q = 0, dvdl_long_range_lj = 0;
 382
 383         bSB = (ir->nwall == 2);
 384         if (bSB)
 385         {
 386             copy_mat(box, boxs);
 387             svmul(ir->wall_ewald_zfac, boxs[ZZ], boxs[ZZ]);
 388             box_size[ZZ] *= ir->wall_ewald_zfac;
 389         }
 390
 391         if (EEL_PME_EWALD(fr->eeltype) || EVDW_PME(fr->vdwtype))
 392         {
 393             real dvdl_long_range_correction_q   = 0;
 394             real dvdl_long_range_correction_lj  = 0;
 395             /* With the Verlet scheme exclusion forces are calculated
 396              * in the non-bonded kernel.
 397              */
 398             /* The TPI molecule does not have exclusions with the rest
 399              * of the system and no intra-molecular PME grid
 400              * contributions will be calculated in
 401              * gmx_pme_calc_energy.
 402              */
 403             if ((ir->cutoff_scheme == ecutsGROUP && fr->n_tpi == 0) ||
 404                 ir->ewald_geometry != eewg3D ||
 405                 ir->epsilon_surface != 0)
 406             {
 407                 int nthreads, t;
 408
 409                 wallcycle_sub_start(wcycle, ewcsEWALD_CORRECTION);
 410
 411                 if (fr->n_tpi > 0)
 412                 {
 413                     gmx_fatal(FARGS, "TPI with PME currently only works in a 3D geometry with tin-foil boundary conditions");
 414                 }
 415
 416                 nthreads = fr->nthread_ewc;
 417 #pragma omp parallel for num_threads(nthreads) schedule(static)
 418                 for (t = 0; t < nthreads; t++)
 419                 {
 420                     try
 421                     {
 422                         tensor *vir_q, *vir_lj;
 423                         real   *Vcorrt_q, *Vcorrt_lj, *dvdlt_q, *dvdlt_lj;
 424                         if (t == 0)
 425                         {
 426                             vir_q     = &fr->vir_el_recip;
 427                             vir_lj    = &fr->vir_lj_recip;
 428                             Vcorrt_q  = &Vcorr_q;
 429                             Vcorrt_lj = &Vcorr_lj;
 430                             dvdlt_q   = &dvdl_long_range_correction_q;
 431                             dvdlt_lj  = &dvdl_long_range_correction_lj;
 432                         }
 433                         else
 434                         {
 435                             vir_q     = &fr->ewc_t[t].vir_q;
 436                             vir_lj    = &fr->ewc_t[t].vir_lj;
 437                             Vcorrt_q  = &fr->ewc_t[t].Vcorr_q;
 438                             Vcorrt_lj = &fr->ewc_t[t].Vcorr_lj;
 439                             dvdlt_q   = &fr->ewc_t[t].dvdl[efptCOUL];
 440                             dvdlt_lj  = &fr->ewc_t[t].dvdl[efptVDW];
 441                             clear_mat(*vir_q);
 442                             clear_mat(*vir_lj);
 443                         }
 444                         *dvdlt_q  = 0;
 445                         *dvdlt_lj = 0;
 446
 447                         /* Threading is only supported with the Verlet cut-off
 448                          * scheme and then only single particle forces (no
 449                          * exclusion forces) are calculated, so we can store
 450                          * the forces in the normal, single fr->f_novirsum array.
 451                          */
 452                         ewald_LRcorrection(md->homenr, cr, nthreads, t, fr,
 453                                            md->chargeA, md->chargeB,
 454                                            md->sqrt_c6A, md->sqrt_c6B,
 455                                            md->sigmaA, md->sigmaB,
 456                                            md->sigma3A, md->sigma3B,
 457                                            md->nChargePerturbed || md->nTypePerturbed,
 458                                            ir->cutoff_scheme != ecutsVERLET,
 459                                            excl, x, bSB ? boxs : box, mu_tot,
 460                                            ir->ewald_geometry,
 461                                            ir->epsilon_surface,
 462                                            fr->f_novirsum, *vir_q, *vir_lj,
 463                                            Vcorrt_q, Vcorrt_lj,
 464                                            lambda[efptCOUL], lambda[efptVDW],
 465                                            dvdlt_q, dvdlt_lj);
 466                     }
 467                     GMX_CATCH_ALL_AND_EXIT_WITH_FATAL_ERROR;
 468                 }
 469                 if (nthreads > 1)
 470                 {
 471                     reduce_thread_energies(fr->vir_el_recip, fr->vir_lj_recip,
 472                                            &Vcorr_q, &Vcorr_lj,
 473                                            &dvdl_long_range_correction_q,
 474                                            &dvdl_long_range_correction_lj,
 475                                            nthreads, fr->ewc_t);
 476                 }
 477                 wallcycle_sub_stop(wcycle, ewcsEWALD_CORRECTION);
 478             }
 479
 480             if (EEL_PME_EWALD(fr->eeltype) && fr->n_tpi == 0)
 481             {
 482                 /* This is not in a subcounter because it takes a
 483                    negligible and constant-sized amount of time */
 484                 Vcorr_q += ewald_charge_correction(cr, fr, lambda[efptCOUL], box,
 485                                                    &dvdl_long_range_correction_q,
 486                                                    fr->vir_el_recip);
 487             }
 488
 489             enerd->dvdl_lin[efptCOUL] += dvdl_long_range_correction_q;
 490             enerd->dvdl_lin[efptVDW]  += dvdl_long_range_correction_lj;
 491
 492             if ((EEL_PME(fr->eeltype) || EVDW_PME(fr->vdwtype)) && (cr->duty & DUTY_PME))
 493             {
 494                 /* Do reciprocal PME for Coulomb and/or LJ. */
 495                 assert(fr->n_tpi >= 0);
 496                 if (fr->n_tpi == 0 || (flags & GMX_FORCE_STATECHANGED))
 497                 {
 498                     pme_flags = GMX_PME_SPREAD | GMX_PME_SOLVE;
 499                     if (EEL_PME(fr->eeltype))
 500                     {
 501                         pme_flags     |= GMX_PME_DO_COULOMB;
 502                     }
 503                     if (EVDW_PME(fr->vdwtype))
 504                     {
 505                         pme_flags |= GMX_PME_DO_LJ;
 506                     }
 507                     if (flags & GMX_FORCE_FORCES)
 508                     {
 509                         pme_flags |= GMX_PME_CALC_F;
 510                     }
 511                     if (flags & GMX_FORCE_VIRIAL)
 512                     {
 513                         pme_flags |= GMX_PME_CALC_ENER_VIR;
 514                     }
 515                     if (fr->n_tpi > 0)
 516                     {
 517                         /* We don't calculate f, but we do want the potential */
 518                         pme_flags |= GMX_PME_CALC_POT;
 519                     }
 520                     wallcycle_start(wcycle, ewcPMEMESH);
 521                     status = gmx_pme_do(fr->pmedata,
 522                                         0, md->homenr - fr->n_tpi,
 523                                         x, fr->f_novirsum,
 524                                         md->chargeA, md->chargeB,
 525                                         md->sqrt_c6A, md->sqrt_c6B,
 526                                         md->sigmaA, md->sigmaB,
 527                                         bSB ? boxs : box, cr,
 528                                         DOMAINDECOMP(cr) ? dd_pme_maxshift_x(cr->dd) : 0,
 529                                         DOMAINDECOMP(cr) ? dd_pme_maxshift_y(cr->dd) : 0,
 530                                         nrnb, wcycle,
 531                                         fr->vir_el_recip, fr->ewaldcoeff_q,
 532                                         fr->vir_lj_recip, fr->ewaldcoeff_lj,
 533                                         &Vlr_q, &Vlr_lj,
 534                                         lambda[efptCOUL], lambda[efptVDW],
 535                                         &dvdl_long_range_q, &dvdl_long_range_lj, pme_flags);
 536                     *cycles_pme = wallcycle_stop(wcycle, ewcPMEMESH);
 537                     if (status != 0)
 538                     {
 539                         gmx_fatal(FARGS, "Error %d in reciprocal PME routine", status);
 540                     }
 541                     /* We should try to do as little computation after
 542                      * this as possible, because parallel PME synchronizes
 543                      * the nodes, so we want all load imbalance of the
 544                      * rest of the force calculation to be before the PME
 545                      * call.  DD load balancing is done on the whole time
 546                      * of the force call (without PME).
 547                      */
 548                 }
 549                 if (fr->n_tpi > 0)
 550                 {
 551                     if (EVDW_PME(ir->vdwtype))
 552                     {
 553
 554                         gmx_fatal(FARGS, "Test particle insertion not implemented with LJ-PME");
 555                     }
 556                     /* Determine the PME grid energy of the test molecule
 557                      * with the PME grid potential of the other charges.
 558                      */
 559                     gmx_pme_calc_energy(fr->pmedata, fr->n_tpi,
 560                                         x + md->homenr - fr->n_tpi,
 561                                         md->chargeA + md->homenr - fr->n_tpi,
 562                                         &Vlr_q);
 563                 }
 564             }
 565         }
 566
 567         if (!EEL_PME(fr->eeltype) && EEL_PME_EWALD(fr->eeltype))
 568         {
 569             Vlr_q = do_ewald(ir, x, fr->f_novirsum,
 570                              md->chargeA, md->chargeB,
 571                              box_size, cr, md->homenr,
 572                              fr->vir_el_recip, fr->ewaldcoeff_q,
 573                              lambda[efptCOUL], &dvdl_long_range_q, fr->ewald_table);
 574         }
 575
 576         /* Note that with separate PME nodes we get the real energies later */
 577         enerd->dvdl_lin[efptCOUL] += dvdl_long_range_q;
 578         enerd->dvdl_lin[efptVDW]  += dvdl_long_range_lj;
 579         enerd->term[F_COUL_RECIP]  = Vlr_q + Vcorr_q;
 580         enerd->term[F_LJ_RECIP]    = Vlr_lj + Vcorr_lj;
 581         if (debug)
 582         {
 583             fprintf(debug, "Vlr_q = %g, Vcorr_q = %g, Vlr_corr_q = %g\n",
 584                     Vlr_q, Vcorr_q, enerd->term[F_COUL_RECIP]);
 585             pr_rvecs(debug, 0, "vir_el_recip after corr", fr->vir_el_recip, DIM);
 586             pr_rvecs(debug, 0, "fshift after LR Corrections", fr->fshift, SHIFTS);
 587             fprintf(debug, "Vlr_lj: %g, Vcorr_lj = %g, Vlr_corr_lj = %g\n",
 588                     Vlr_lj, Vcorr_lj, enerd->term[F_LJ_RECIP]);
 589             pr_rvecs(debug, 0, "vir_lj_recip after corr", fr->vir_lj_recip, DIM);
 590         }
 591     }
 592     else
 593     {
 594         /* Is there a reaction-field exclusion correction needed?
 595          * With the Verlet scheme, exclusion forces are calculated
 596          * in the non-bonded kernel.
 597          */
 598         if (ir->cutoff_scheme != ecutsVERLET && EEL_RF(fr->eeltype))
 599         {
 600             real dvdl_rf_excl      = 0;
 601             enerd->term[F_RF_EXCL] =
 602                 RF_excl_correction(fr, graph, md, excl, x, f,
 603                                    fr->fshift, &pbc, lambda[efptCOUL], &dvdl_rf_excl);
 604
 605             enerd->dvdl_lin[efptCOUL] += dvdl_rf_excl;
 606         }
 607     }
 608     where();
 609
 610     if (debug)
 611     {
 612         print_nrnb(debug, nrnb);
 613     }
 614
 615 #if GMX_MPI
 616     if (TAKETIME)
 617     {
 618         t2 = MPI_Wtime();
 619         MPI_Barrier(cr->mpi_comm_mygroup);
 620         t3          = MPI_Wtime();
 621         fr->t_wait += t3-t2;
 622         if (fr->timesteps == 11)
 623         {
 624             char buf[22];
 625             fprintf(stderr, "* PP load balancing info: rank %d, step %s, rel wait time=%3.0f%% , load string value: %7.2f\n",
 626                     cr->nodeid, gmx_step_str(fr->timesteps, buf),
 627                     100*fr->t_wait/(fr->t_wait+fr->t_fnbf),
 628                     (fr->t_fnbf+fr->t_wait)/fr->t_fnbf);
 629         }
 630         fr->timesteps++;
 631     }
 632 #endif
 633
 634     if (debug)
 635     {
 636         pr_rvecs(debug, 0, "fshift after bondeds", fr->fshift, SHIFTS);
 637     }
 638
 639 }
 640
 641 void init_enerdata(int ngener, int n_lambda, gmx_enerdata_t *enerd)
 642 {
 643     int i, n2;
 644
 645     for (i = 0; i < F_NRE; i++)
 646     {
 647         enerd->term[i]         = 0;
 648         enerd->foreign_term[i] = 0;
 649     }
 650
 651
 652     for (i = 0; i < efptNR; i++)
 653     {
 654         enerd->dvdl_lin[i]     = 0;
 655         enerd->dvdl_nonlin[i]  = 0;
 656     }
 657
 658     n2 = ngener*ngener;
 659     if (debug)
 660     {
 661         fprintf(debug, "Creating %d sized group matrix for energies\n", n2);
 662     }
 663     enerd->grpp.nener         = n2;
 664     enerd->foreign_grpp.nener = n2;
 665     for (i = 0; (i < egNR); i++)
 666     {
 667         snew(enerd->grpp.ener[i], n2);
 668         snew(enerd->foreign_grpp.ener[i], n2);
 669     }
 670
 671     if (n_lambda)
 672     {
 673         enerd->n_lambda = 1 + n_lambda;
 674         snew(enerd->enerpart_lambda, enerd->n_lambda);
 675     }
 676     else
 677     {
 678         enerd->n_lambda = 0;
 679     }
 680 }
 681
 682 void destroy_enerdata(gmx_enerdata_t *enerd)
 683 {
 684     int i;
 685
 686     for (i = 0; (i < egNR); i++)
 687     {
 688         sfree(enerd->grpp.ener[i]);
 689     }
 690
 691     for (i = 0; (i < egNR); i++)
 692     {
 693         sfree(enerd->foreign_grpp.ener[i]);
 694     }
 695
 696     if (enerd->n_lambda)
 697     {
 698         sfree(enerd->enerpart_lambda);
 699     }
 700 }
 701
 702 static real sum_v(int n, real v[])
 703 {
 704     real t;
 705     int  i;
 706
 707     t = 0.0;
 708     for (i = 0; (i < n); i++)
 709     {
 710         t = t + v[i];
 711     }
 712
 713     return t;
 714 }
 715
 716 void sum_epot(gmx_grppairener_t *grpp, real *epot)
 717 {
 718     int i;
 719
 720     /* Accumulate energies */
 721     epot[F_COUL_SR]  = sum_v(grpp->nener, grpp->ener[egCOULSR]);
 722     epot[F_LJ]       = sum_v(grpp->nener, grpp->ener[egLJSR]);
 723     epot[F_LJ14]     = sum_v(grpp->nener, grpp->ener[egLJ14]);
 724     epot[F_COUL14]   = sum_v(grpp->nener, grpp->ener[egCOUL14]);
 725     /* We have already added 1-2,1-3, and 1-4 terms to F_GBPOL */
 726     epot[F_GBPOL]   += sum_v(grpp->nener, grpp->ener[egGB]);
 727
 728 /* lattice part of LR doesnt belong to any group
 729  * and has been added earlier
 730  */
 731     epot[F_BHAM]     = sum_v(grpp->nener, grpp->ener[egBHAMSR]);
 732
 733     epot[F_EPOT] = 0;
 734     for (i = 0; (i < F_EPOT); i++)
 735     {
 736         if (i != F_DISRESVIOL && i != F_ORIRESDEV)
 737         {
 738             epot[F_EPOT] += epot[i];
 739         }
 740     }
 741 }
 742
 743 void sum_dhdl(gmx_enerdata_t *enerd, real *lambda, t_lambda *fepvals)
 744 {
 745     int    i, j, index;
 746     double dlam;
 747
 748     enerd->dvdl_lin[efptVDW] += enerd->term[F_DVDL_VDW];  /* include dispersion correction */
 749     enerd->term[F_DVDL]       = 0.0;
 750     for (i = 0; i < efptNR; i++)
 751     {
 752         if (fepvals->separate_dvdl[i])
 753         {
 754             /* could this be done more readably/compactly? */
 755             switch (i)
 756             {
 757                 case (efptMASS):
 758                     index = F_DKDL;
 759                     break;
 760                 case (efptCOUL):
 761                     index = F_DVDL_COUL;
 762                     break;
 763                 case (efptVDW):
 764                     index = F_DVDL_VDW;
 765                     break;
 766                 case (efptBONDED):
 767                     index = F_DVDL_BONDED;
 768                     break;
 769                 case (efptRESTRAINT):
 770                     index = F_DVDL_RESTRAINT;
 771                     break;
 772                 default:
 773                     index = F_DVDL;
 774                     break;
 775             }
 776             enerd->term[index] = enerd->dvdl_lin[i] + enerd->dvdl_nonlin[i];
 777             if (debug)
 778             {
 779                 fprintf(debug, "dvdl-%s[%2d]: %f: non-linear %f + linear %f\n",
 780                         efpt_names[i], i, enerd->term[index], enerd->dvdl_nonlin[i], enerd->dvdl_lin[i]);
 781             }
 782         }
 783         else
 784         {
 785             enerd->term[F_DVDL] += enerd->dvdl_lin[i] + enerd->dvdl_nonlin[i];
 786             if (debug)
 787             {
 788                 fprintf(debug, "dvd-%sl[%2d]: %f: non-linear %f + linear %f\n",
 789                         efpt_names[0], i, enerd->term[F_DVDL], enerd->dvdl_nonlin[i], enerd->dvdl_lin[i]);
 790             }
 791         }
 792     }
 793
 794     /* Notes on the foreign lambda free energy difference evaluation:
 795      * Adding the potential and ekin terms that depend linearly on lambda
 796      * as delta lam * dvdl to the energy differences is exact.
 797      * For the constraints this is not exact, but we have no other option
 798      * without literally changing the lengths and reevaluating the energies at each step.
 799      * (try to remedy this post 4.6 - MRS)
 800      */
 801     if (fepvals->separate_dvdl[efptBONDED])
 802     {
 803         enerd->term[F_DVDL_BONDED] += enerd->term[F_DVDL_CONSTR];
 804     }
 805     else
 806     {
 807         enerd->term[F_DVDL] += enerd->term[F_DVDL_CONSTR];
 808     }
 809     enerd->term[F_DVDL_CONSTR] = 0;
 810
 811     for (i = 0; i < fepvals->n_lambda; i++)
 812     {
 813         /* note we are iterating over fepvals here!
 814            For the current lam, dlam = 0 automatically,
 815            so we don't need to add anything to the
 816            enerd->enerpart_lambda[0] */
 817
 818         /* we don't need to worry about dvdl_lin contributions to dE at
 819            current lambda, because the contributions to the current
 820            lambda are automatically zeroed */
 821
 822         for (j = 0; j < efptNR; j++)
 823         {
 824             /* Note that this loop is over all dhdl components, not just the separated ones */
 825             dlam = (fepvals->all_lambda[j][i]-lambda[j]);
 826             enerd->enerpart_lambda[i+1] += dlam*enerd->dvdl_lin[j];
 827             if (debug)
 828             {
 829                 fprintf(debug, "enerdiff lam %g: (%15s), non-linear %f linear %f*%f\n",
 830                         fepvals->all_lambda[j][i], efpt_names[j],
 831                         (enerd->enerpart_lambda[i+1] - enerd->enerpart_lambda[0]),
 832                         dlam, enerd->dvdl_lin[j]);
 833             }
 834         }
 835     }
 836 }
 837
 838
 839 void reset_foreign_enerdata(gmx_enerdata_t *enerd)
 840 {
 841     int  i, j;
 842
 843     /* First reset all foreign energy components.  Foreign energies always called on
 844        neighbor search steps */
 845     for (i = 0; (i < egNR); i++)
 846     {
 847         for (j = 0; (j < enerd->grpp.nener); j++)
 848         {
 849             enerd->foreign_grpp.ener[i][j] = 0.0;
 850         }
 851     }
 852
 853     /* potential energy components */
 854     for (i = 0; (i <= F_EPOT); i++)
 855     {
 856         enerd->foreign_term[i] = 0.0;
 857     }
 858 }
 859
 860 void reset_enerdata(gmx_enerdata_t *enerd)
 861 {
 862     int      i, j;
 863
 864     /* First reset all energy components. */
 865     for (i = 0; (i < egNR); i++)
 866     {
 867         for (j = 0; (j < enerd->grpp.nener); j++)
 868         {
 869             enerd->grpp.ener[i][j] = 0.0;
 870         }
 871     }
 872     for (i = 0; i < efptNR; i++)
 873     {
 874         enerd->dvdl_lin[i]    = 0.0;
 875         enerd->dvdl_nonlin[i] = 0.0;
 876     }
 877
 878     /* Normal potential energy components */
 879     for (i = 0; (i <= F_EPOT); i++)
 880     {
 881         enerd->term[i] = 0.0;
 882     }
 883     enerd->term[F_DVDL]            = 0.0;
 884     enerd->term[F_DVDL_COUL]       = 0.0;
 885     enerd->term[F_DVDL_VDW]        = 0.0;
 886     enerd->term[F_DVDL_BONDED]     = 0.0;
 887     enerd->term[F_DVDL_RESTRAINT]  = 0.0;
 888     enerd->term[F_DKDL]            = 0.0;
 889     if (enerd->n_lambda > 0)
 890     {
 891         for (i = 0; i < enerd->n_lambda; i++)
 892         {
 893             enerd->enerpart_lambda[i] = 0.0;
 894         }
 895     }
 896     /* reset foreign energy data - separate function since we also call it elsewhere */
 897     reset_foreign_enerdata(enerd);
 898 }