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) 2013,2014,2015,2016,2017,2018, 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 <string.h>
  45
  46 #include <cmath>
  47
  48 #include "gromacs/domdec/domdec.h"
  49 #include "gromacs/domdec/domdec_struct.h"
  50 #include "gromacs/ewald/ewald.h"
  51 #include "gromacs/ewald/long-range-correction.h"
  52 #include "gromacs/ewald/pme.h"
  53 #include "gromacs/gmxlib/network.h"
  54 #include "gromacs/gmxlib/nrnb.h"
  55 #include "gromacs/gmxlib/nonbonded/nonbonded.h"
  56 #include "gromacs/listed-forces/listed-forces.h"
  57 #include "gromacs/math/vec.h"
  58 #include "gromacs/math/vecdump.h"
  59 #include "gromacs/mdlib/force_flags.h"
  60 #include "gromacs/mdlib/forcerec-threading.h"
  61 #include "gromacs/mdlib/mdrun.h"
  62 #include "gromacs/mdlib/ns.h"
  63 #include "gromacs/mdlib/qmmm.h"
  64 #include "gromacs/mdlib/rf_util.h"
  65 #include "gromacs/mdlib/wall.h"
  66 #include "gromacs/mdtypes/commrec.h"
  67 #include "gromacs/mdtypes/forceoutput.h"
  68 #include "gromacs/mdtypes/inputrec.h"
  69 #include "gromacs/mdtypes/md_enums.h"
  70 #include "gromacs/pbcutil/ishift.h"
  71 #include "gromacs/pbcutil/mshift.h"
  72 #include "gromacs/pbcutil/pbc.h"
  73 #include "gromacs/timing/wallcycle.h"
  74 #include "gromacs/utility/cstringutil.h"
  75 #include "gromacs/utility/exceptions.h"
  76 #include "gromacs/utility/fatalerror.h"
  77 #include "gromacs/utility/smalloc.h"
  78
  79 void ns(FILE               *fp,
  80         t_forcerec         *fr,
  81         matrix              box,
  82         const gmx_groups_t *groups,
  83         gmx_localtop_t     *top,
  84         const t_mdatoms    *md,
  85         const t_commrec    *cr,
  86         t_nrnb             *nrnb,
  87         gmx_bool            bFillGrid)
  88 {
  89     int     nsearch;
  90
  91
  92     if (!fr->ns->nblist_initialized)
  93     {
  94         init_neighbor_list(fp, fr, md->homenr);
  95     }
  96
  97     nsearch = search_neighbours(fp, fr, box, top, groups, cr, nrnb, md,
  98                                 bFillGrid);
  99     if (debug)
 100     {
 101         fprintf(debug, "nsearch = %d\n", nsearch);
 102     }
 103
 104     /* Check whether we have to do dynamic load balancing */
 105     /*if ((nsb->nstDlb > 0) && (mod(step,nsb->nstDlb) == 0))
 106        count_nb(cr,nsb,&(top->blocks[ebCGS]),nns,fr->nlr,
 107        &(top->idef),opts->ngener);
 108      */
 109     if (fr->ns->dump_nl > 0)
 110     {
 111         dump_nblist(fp, cr, fr, fr->ns->dump_nl);
 112     }
 113 }
 114
 115 static void clearEwaldThreadOutput(ewald_corr_thread_t *ewc_t)
 116 {
 117     ewc_t->Vcorr_q        = 0;
 118     ewc_t->Vcorr_lj       = 0;
 119     ewc_t->dvdl[efptCOUL] = 0;
 120     ewc_t->dvdl[efptVDW]  = 0;
 121     clear_mat(ewc_t->vir_q);
 122     clear_mat(ewc_t->vir_lj);
 123 }
 124
 125 static void reduceEwaldThreadOuput(int nthreads, ewald_corr_thread_t *ewc_t)
 126 {
 127     ewald_corr_thread_t &dest = ewc_t[0];
 128
 129     for (int t = 1; t < nthreads; t++)
 130     {
 131         dest.Vcorr_q        += ewc_t[t].Vcorr_q;
 132         dest.Vcorr_lj       += ewc_t[t].Vcorr_lj;
 133         dest.dvdl[efptCOUL] += ewc_t[t].dvdl[efptCOUL];
 134         dest.dvdl[efptVDW]  += ewc_t[t].dvdl[efptVDW];
 135         m_add(dest.vir_q,  ewc_t[t].vir_q,  dest.vir_q);
 136         m_add(dest.vir_lj, ewc_t[t].vir_lj, dest.vir_lj);
 137     }
 138 }
 139
 140 void do_force_lowlevel(t_forcerec           *fr,
 141                        const t_inputrec     *ir,
 142                        const t_idef         *idef,
 143                        const t_commrec      *cr,
 144                        const gmx_multisim_t *ms,
 145                        t_nrnb               *nrnb,
 146                        gmx_wallcycle_t       wcycle,
 147                        const t_mdatoms      *md,
 148                        rvec                  x[],
 149                        history_t            *hist,
 150                        rvec                 *forceForUseWithShiftForces,
 151                        gmx::ForceWithVirial *forceWithVirial,
 152                        gmx_enerdata_t       *enerd,
 153                        t_fcdata             *fcd,
 154                        matrix                box,
 155                        t_lambda             *fepvals,
 156                        real                 *lambda,
 157                        const t_graph        *graph,
 158                        const t_blocka       *excl,
 159                        rvec                  mu_tot[],
 160                        int                   flags,
 161                        float                *cycles_pme)
 162 {
 163     int         i, j;
 164     int         donb_flags;
 165     int         pme_flags;
 166     t_pbc       pbc;
 167     real        dvdl_dum[efptNR], dvdl_nb[efptNR];
 168
 169 #if GMX_MPI
 170     double  t0 = 0.0, t1, t2, t3; /* time measurement for coarse load balancing */
 171 #endif
 172
 173     set_pbc(&pbc, fr->ePBC, box);
 174
 175     /* reset free energy components */
 176     for (i = 0; i < efptNR; i++)
 177     {
 178         dvdl_nb[i]  = 0;
 179         dvdl_dum[i] = 0;
 180     }
 181
 182     /* do QMMM first if requested */
 183     if (fr->bQMMM)
 184     {
 185         enerd->term[F_EQM] = calculate_QMMM(cr, forceForUseWithShiftForces, fr);
 186     }
 187
 188     /* Call the short range functions all in one go. */
 189
 190 #if GMX_MPI
 191     /*#define TAKETIME ((cr->npmenodes) && (fr->timesteps < 12))*/
 192 #define TAKETIME FALSE
 193     if (TAKETIME)
 194     {
 195         MPI_Barrier(cr->mpi_comm_mygroup);
 196         t0 = MPI_Wtime();
 197     }
 198 #endif
 199
 200     if (ir->nwall)
 201     {
 202         /* foreign lambda component for walls */
 203         real dvdl_walls = do_walls(ir, fr, box, md, x, forceForUseWithShiftForces, lambda[efptVDW],
 204                                    enerd->grpp.ener[egLJSR], nrnb);
 205         enerd->dvdl_lin[efptVDW] += dvdl_walls;
 206     }
 207
 208     /* We only do non-bonded calculation with group scheme here, the verlet
 209      * calls are done from do_force_cutsVERLET(). */
 210     if (fr->cutoff_scheme == ecutsGROUP && (flags & GMX_FORCE_NONBONDED))
 211     {
 212         donb_flags = 0;
 213         /* Add short-range interactions */
 214         donb_flags |= GMX_NONBONDED_DO_SR;
 215
 216         /* Currently all group scheme kernels always calculate (shift-)forces */
 217         if (flags & GMX_FORCE_FORCES)
 218         {
 219             donb_flags |= GMX_NONBONDED_DO_FORCE;
 220         }
 221         if (flags & GMX_FORCE_VIRIAL)
 222         {
 223             donb_flags |= GMX_NONBONDED_DO_SHIFTFORCE;
 224         }
 225         if (flags & GMX_FORCE_ENERGY)
 226         {
 227             donb_flags |= GMX_NONBONDED_DO_POTENTIAL;
 228         }
 229
 230         wallcycle_sub_start(wcycle, ewcsNONBONDED);
 231         do_nonbonded(fr, x, forceForUseWithShiftForces, md, excl,
 232                      &enerd->grpp, nrnb,
 233                      lambda, dvdl_nb, -1, -1, donb_flags);
 234
 235         /* If we do foreign lambda and we have soft-core interactions
 236          * we have to recalculate the (non-linear) energies contributions.
 237          */
 238         if (fepvals->n_lambda > 0 && (flags & GMX_FORCE_DHDL) && fepvals->sc_alpha != 0)
 239         {
 240             for (i = 0; i < enerd->n_lambda; i++)
 241             {
 242                 real lam_i[efptNR];
 243
 244                 for (j = 0; j < efptNR; j++)
 245                 {
 246                     lam_i[j] = (i == 0 ? lambda[j] : fepvals->all_lambda[j][i-1]);
 247                 }
 248                 reset_foreign_enerdata(enerd);
 249                 do_nonbonded(fr, x, forceForUseWithShiftForces, md, excl,
 250                              &(enerd->foreign_grpp), nrnb,
 251                              lam_i, dvdl_dum, -1, -1,
 252                              (donb_flags & ~GMX_NONBONDED_DO_FORCE) | GMX_NONBONDED_DO_FOREIGNLAMBDA);
 253                 sum_epot(&(enerd->foreign_grpp), enerd->foreign_term);
 254                 enerd->enerpart_lambda[i] += enerd->foreign_term[F_EPOT];
 255             }
 256         }
 257         wallcycle_sub_stop(wcycle, ewcsNONBONDED);
 258     }
 259
 260 #if GMX_MPI
 261     if (TAKETIME)
 262     {
 263         t1          = MPI_Wtime();
 264         fr->t_fnbf += t1-t0;
 265     }
 266 #endif
 267
 268     if (fepvals->sc_alpha != 0)
 269     {
 270         enerd->dvdl_nonlin[efptVDW] += dvdl_nb[efptVDW];
 271     }
 272     else
 273     {
 274         enerd->dvdl_lin[efptVDW] += dvdl_nb[efptVDW];
 275     }
 276
 277     if (fepvals->sc_alpha != 0)
 278
 279     /* even though coulomb part is linear, we already added it, beacuse we
 280        need to go through the vdw calculation anyway */
 281     {
 282         enerd->dvdl_nonlin[efptCOUL] += dvdl_nb[efptCOUL];
 283     }
 284     else
 285     {
 286         enerd->dvdl_lin[efptCOUL] += dvdl_nb[efptCOUL];
 287     }
 288
 289     if (debug)
 290     {
 291         pr_rvecs(debug, 0, "fshift after SR", fr->fshift, SHIFTS);
 292     }
 293
 294     /* Shift the coordinates. Must be done before listed forces and PPPM,
 295      * but is also necessary for SHAKE and update, therefore it can NOT
 296      * go when no listed forces have to be evaluated.
 297      *
 298      * The shifting and PBC code is deliberately not timed, since with
 299      * the Verlet scheme it only takes non-zero time with triclinic
 300      * boxes, and even then the time is around a factor of 100 less
 301      * than the next smallest counter.
 302      */
 303
 304
 305     /* Here sometimes we would not need to shift with NBFonly,
 306      * but we do so anyhow for consistency of the returned coordinates.
 307      */
 308     if (graph)
 309     {
 310         shift_self(graph, box, x);
 311         if (TRICLINIC(box))
 312         {
 313             inc_nrnb(nrnb, eNR_SHIFTX, 2*graph->nnodes);
 314         }
 315         else
 316         {
 317             inc_nrnb(nrnb, eNR_SHIFTX, graph->nnodes);
 318         }
 319     }
 320     /* Check whether we need to do listed interactions or correct for exclusions */
 321     if (fr->bMolPBC &&
 322         ((flags & GMX_FORCE_LISTED)
 323          || EEL_RF(fr->ic->eeltype) || EEL_FULL(fr->ic->eeltype) || EVDW_PME(fr->ic->vdwtype)))
 324     {
 325         /* TODO There are no electrostatics methods that require this
 326            transformation, when using the Verlet scheme, so update the
 327            above conditional. */
 328         /* Since all atoms are in the rectangular or triclinic unit-cell,
 329          * only single box vector shifts (2 in x) are required.
 330          */
 331         set_pbc_dd(&pbc, fr->ePBC, DOMAINDECOMP(cr) ? cr->dd->nc : nullptr,
 332                    TRUE, box);
 333     }
 334
 335     do_force_listed(wcycle, box, ir->fepvals, cr, ms,
 336                     idef, (const rvec *) x, hist,
 337                     forceForUseWithShiftForces, forceWithVirial,
 338                     fr, &pbc, graph, enerd, nrnb, lambda, md, fcd,
 339                     DOMAINDECOMP(cr) ? cr->dd->globalAtomIndices.data() : nullptr,
 340                     flags);
 341
 342
 343     *cycles_pme = 0;
 344
 345     /* Do long-range electrostatics and/or LJ-PME, including related short-range
 346      * corrections.
 347      */
 348     if (EEL_FULL(fr->ic->eeltype) || EVDW_PME(fr->ic->vdwtype))
 349     {
 350         int  status            = 0;
 351         real Vlr_q             = 0, Vlr_lj = 0;
 352
 353         /* We reduce all virial, dV/dlambda and energy contributions, except
 354          * for the reciprocal energies (Vlr_q, Vlr_lj) into the same struct.
 355          */
 356         ewald_corr_thread_t &ewaldOutput = fr->ewc_t[0];
 357         clearEwaldThreadOutput(&ewaldOutput);
 358
 359         if (EEL_PME_EWALD(fr->ic->eeltype) || EVDW_PME(fr->ic->vdwtype))
 360         {
 361             /* With the Verlet scheme exclusion forces are calculated
 362              * in the non-bonded kernel.
 363              */
 364             /* The TPI molecule does not have exclusions with the rest
 365              * of the system and no intra-molecular PME grid
 366              * contributions will be calculated in
 367              * gmx_pme_calc_energy.
 368              */
 369             if ((ir->cutoff_scheme == ecutsGROUP && fr->n_tpi == 0) ||
 370                 ir->ewald_geometry != eewg3D ||
 371                 ir->epsilon_surface != 0)
 372             {
 373                 int nthreads, t;
 374
 375                 wallcycle_sub_start(wcycle, ewcsEWALD_CORRECTION);
 376
 377                 if (fr->n_tpi > 0)
 378                 {
 379                     gmx_fatal(FARGS, "TPI with PME currently only works in a 3D geometry with tin-foil boundary conditions");
 380                 }
 381
 382                 nthreads = fr->nthread_ewc;
 383 #pragma omp parallel for num_threads(nthreads) schedule(static)
 384                 for (t = 0; t < nthreads; t++)
 385                 {
 386                     try
 387                     {
 388                         ewald_corr_thread_t &ewc_t = fr->ewc_t[t];
 389                         if (t > 0)
 390                         {
 391                             clearEwaldThreadOutput(&ewc_t);
 392                         }
 393
 394                         /* Threading is only supported with the Verlet cut-off
 395                          * scheme and then only single particle forces (no
 396                          * exclusion forces) are calculated, so we can store
 397                          * the forces in the normal, single forceWithVirial->force_ array.
 398                          */
 399                         ewald_LRcorrection(md->homenr, cr, nthreads, t, fr, ir,
 400                                            md->chargeA, md->chargeB,
 401                                            md->sqrt_c6A, md->sqrt_c6B,
 402                                            md->sigmaA, md->sigmaB,
 403                                            md->sigma3A, md->sigma3B,
 404                                            md->nChargePerturbed || md->nTypePerturbed,
 405                                            ir->cutoff_scheme != ecutsVERLET,
 406                                            excl, x, box, mu_tot,
 407                                            ir->ewald_geometry,
 408                                            ir->epsilon_surface,
 409                                            as_rvec_array(forceWithVirial->force_.data()),
 410                                            ewc_t.vir_q, ewc_t.vir_lj,
 411                                            &ewc_t.Vcorr_q, &ewc_t.Vcorr_lj,
 412                                            lambda[efptCOUL], lambda[efptVDW],
 413                                            &ewc_t.dvdl[efptCOUL], &ewc_t.dvdl[efptVDW]);
 414                     }
 415                     GMX_CATCH_ALL_AND_EXIT_WITH_FATAL_ERROR;
 416                 }
 417                 if (nthreads > 1)
 418                 {
 419                     reduceEwaldThreadOuput(nthreads, fr->ewc_t);
 420                 }
 421                 wallcycle_sub_stop(wcycle, ewcsEWALD_CORRECTION);
 422             }
 423
 424             if (EEL_PME_EWALD(fr->ic->eeltype) && fr->n_tpi == 0)
 425             {
 426                 /* This is not in a subcounter because it takes a
 427                    negligible and constant-sized amount of time */
 428                 ewaldOutput.Vcorr_q +=
 429                     ewald_charge_correction(cr, fr, lambda[efptCOUL], box,
 430                                             &ewaldOutput.dvdl[efptCOUL],
 431                                             ewaldOutput.vir_q);
 432             }
 433
 434             if ((EEL_PME(fr->ic->eeltype) || EVDW_PME(fr->ic->vdwtype)) &&
 435                 thisRankHasDuty(cr, DUTY_PME) && (pme_run_mode(fr->pmedata) == PmeRunMode::CPU))
 436             {
 437                 /* Do reciprocal PME for Coulomb and/or LJ. */
 438                 assert(fr->n_tpi >= 0);
 439                 if (fr->n_tpi == 0 || (flags & GMX_FORCE_STATECHANGED))
 440                 {
 441                     pme_flags = GMX_PME_SPREAD | GMX_PME_SOLVE;
 442
 443                     if (flags & GMX_FORCE_FORCES)
 444                     {
 445                         pme_flags |= GMX_PME_CALC_F;
 446                     }
 447                     if (flags & GMX_FORCE_VIRIAL)
 448                     {
 449                         pme_flags |= GMX_PME_CALC_ENER_VIR;
 450                     }
 451                     if (fr->n_tpi > 0)
 452                     {
 453                         /* We don't calculate f, but we do want the potential */
 454                         pme_flags |= GMX_PME_CALC_POT;
 455                     }
 456
 457                     /* With domain decomposition we close the CPU side load
 458                      * balancing region here, because PME does global
 459                      * communication that acts as a global barrier.
 460                      */
 461                     if (DOMAINDECOMP(cr))
 462                     {
 463                         ddCloseBalanceRegionCpu(cr->dd);
 464                     }
 465
 466                     wallcycle_start(wcycle, ewcPMEMESH);
 467                     status = gmx_pme_do(fr->pmedata,
 468                                         0, md->homenr - fr->n_tpi,
 469                                         x,
 470                                         as_rvec_array(forceWithVirial->force_.data()),
 471                                         md->chargeA, md->chargeB,
 472                                         md->sqrt_c6A, md->sqrt_c6B,
 473                                         md->sigmaA, md->sigmaB,
 474                                         box, cr,
 475                                         DOMAINDECOMP(cr) ? dd_pme_maxshift_x(cr->dd) : 0,
 476                                         DOMAINDECOMP(cr) ? dd_pme_maxshift_y(cr->dd) : 0,
 477                                         nrnb, wcycle,
 478                                         ewaldOutput.vir_q, ewaldOutput.vir_lj,
 479                                         &Vlr_q, &Vlr_lj,
 480                                         lambda[efptCOUL], lambda[efptVDW],
 481                                         &ewaldOutput.dvdl[efptCOUL],
 482                                         &ewaldOutput.dvdl[efptVDW],
 483                                         pme_flags);
 484                     *cycles_pme = wallcycle_stop(wcycle, ewcPMEMESH);
 485                     if (status != 0)
 486                     {
 487                         gmx_fatal(FARGS, "Error %d in reciprocal PME routine", status);
 488                     }
 489
 490                     /* We should try to do as little computation after
 491                      * this as possible, because parallel PME synchronizes
 492                      * the nodes, so we want all load imbalance of the
 493                      * rest of the force calculation to be before the PME
 494                      * call.  DD load balancing is done on the whole time
 495                      * of the force call (without PME).
 496                      */
 497                 }
 498                 if (fr->n_tpi > 0)
 499                 {
 500                     if (EVDW_PME(ir->vdwtype))
 501                     {
 502
 503                         gmx_fatal(FARGS, "Test particle insertion not implemented with LJ-PME");
 504                     }
 505                     /* Determine the PME grid energy of the test molecule
 506                      * with the PME grid potential of the other charges.
 507                      */
 508                     gmx_pme_calc_energy(fr->pmedata, fr->n_tpi,
 509                                         x + md->homenr - fr->n_tpi,
 510                                         md->chargeA + md->homenr - fr->n_tpi,
 511                                         &Vlr_q);
 512                 }
 513             }
 514         }
 515
 516         if (!EEL_PME(fr->ic->eeltype) && EEL_PME_EWALD(fr->ic->eeltype))
 517         {
 518             Vlr_q = do_ewald(ir, x, as_rvec_array(forceWithVirial->force_.data()),
 519                              md->chargeA, md->chargeB,
 520                              box, cr, md->homenr,
 521                              ewaldOutput.vir_q, fr->ic->ewaldcoeff_q,
 522                              lambda[efptCOUL], &ewaldOutput.dvdl[efptCOUL],
 523                              fr->ewald_table);
 524         }
 525
 526         /* Note that with separate PME nodes we get the real energies later */
 527         forceWithVirial->addVirialContribution(ewaldOutput.vir_q);
 528         forceWithVirial->addVirialContribution(ewaldOutput.vir_lj);
 529         enerd->dvdl_lin[efptCOUL] += ewaldOutput.dvdl[efptCOUL];
 530         enerd->dvdl_lin[efptVDW]  += ewaldOutput.dvdl[efptVDW];
 531         enerd->term[F_COUL_RECIP]  = Vlr_q + ewaldOutput.Vcorr_q;
 532         enerd->term[F_LJ_RECIP]    = Vlr_lj + ewaldOutput.Vcorr_lj;
 533
 534         if (debug)
 535         {
 536             fprintf(debug, "Vlr_q = %g, Vcorr_q = %g, Vlr_corr_q = %g\n",
 537                     Vlr_q, ewaldOutput.Vcorr_q, enerd->term[F_COUL_RECIP]);
 538             pr_rvecs(debug, 0, "vir_el_recip after corr", ewaldOutput.vir_q, DIM);
 539             pr_rvecs(debug, 0, "fshift after LR Corrections", fr->fshift, SHIFTS);
 540             fprintf(debug, "Vlr_lj: %g, Vcorr_lj = %g, Vlr_corr_lj = %g\n",
 541                     Vlr_lj, ewaldOutput.Vcorr_lj, enerd->term[F_LJ_RECIP]);
 542             pr_rvecs(debug, 0, "vir_lj_recip after corr", ewaldOutput.vir_lj, DIM);
 543         }
 544     }
 545     else
 546     {
 547         /* Is there a reaction-field exclusion correction needed?
 548          * With the Verlet scheme, exclusion forces are calculated
 549          * in the non-bonded kernel.
 550          */
 551         if (ir->cutoff_scheme != ecutsVERLET && EEL_RF(fr->ic->eeltype))
 552         {
 553             real dvdl_rf_excl      = 0;
 554             enerd->term[F_RF_EXCL] =
 555                 RF_excl_correction(fr, graph, md, excl, DOMAINDECOMP(cr),
 556                                    x, forceForUseWithShiftForces,
 557                                    fr->fshift, &pbc, lambda[efptCOUL], &dvdl_rf_excl);
 558
 559             enerd->dvdl_lin[efptCOUL] += dvdl_rf_excl;
 560         }
 561     }
 562
 563     if (debug)
 564     {
 565         print_nrnb(debug, nrnb);
 566     }
 567
 568 #if GMX_MPI
 569     if (TAKETIME)
 570     {
 571         t2 = MPI_Wtime();
 572         MPI_Barrier(cr->mpi_comm_mygroup);
 573         t3          = MPI_Wtime();
 574         fr->t_wait += t3-t2;
 575         if (fr->timesteps == 11)
 576         {
 577             char buf[22];
 578             fprintf(stderr, "* PP load balancing info: rank %d, step %s, rel wait time=%3.0f%% , load string value: %7.2f\n",
 579                     cr->nodeid, gmx_step_str(fr->timesteps, buf),
 580                     100*fr->t_wait/(fr->t_wait+fr->t_fnbf),
 581                     (fr->t_fnbf+fr->t_wait)/fr->t_fnbf);
 582         }
 583         fr->timesteps++;
 584     }
 585 #endif
 586
 587     if (debug)
 588     {
 589         pr_rvecs(debug, 0, "fshift after bondeds", fr->fshift, SHIFTS);
 590     }
 591
 592 }
 593
 594 void init_enerdata(int ngener, int n_lambda, gmx_enerdata_t *enerd)
 595 {
 596     int i, n2;
 597
 598     for (i = 0; i < F_NRE; i++)
 599     {
 600         enerd->term[i]         = 0;
 601         enerd->foreign_term[i] = 0;
 602     }
 603
 604
 605     for (i = 0; i < efptNR; i++)
 606     {
 607         enerd->dvdl_lin[i]     = 0;
 608         enerd->dvdl_nonlin[i]  = 0;
 609     }
 610
 611     n2 = ngener*ngener;
 612     if (debug)
 613     {
 614         fprintf(debug, "Creating %d sized group matrix for energies\n", n2);
 615     }
 616     enerd->grpp.nener         = n2;
 617     enerd->foreign_grpp.nener = n2;
 618     for (i = 0; (i < egNR); i++)
 619     {
 620         snew(enerd->grpp.ener[i], n2);
 621         snew(enerd->foreign_grpp.ener[i], n2);
 622     }
 623
 624     if (n_lambda)
 625     {
 626         enerd->n_lambda = 1 + n_lambda;
 627         snew(enerd->enerpart_lambda, enerd->n_lambda);
 628     }
 629     else
 630     {
 631         enerd->n_lambda = 0;
 632     }
 633 }
 634
 635 void destroy_enerdata(gmx_enerdata_t *enerd)
 636 {
 637     int i;
 638
 639     for (i = 0; (i < egNR); i++)
 640     {
 641         sfree(enerd->grpp.ener[i]);
 642     }
 643
 644     for (i = 0; (i < egNR); i++)
 645     {
 646         sfree(enerd->foreign_grpp.ener[i]);
 647     }
 648
 649     if (enerd->n_lambda)
 650     {
 651         sfree(enerd->enerpart_lambda);
 652     }
 653 }
 654
 655 static real sum_v(int n, const real v[])
 656 {
 657     real t;
 658     int  i;
 659
 660     t = 0.0;
 661     for (i = 0; (i < n); i++)
 662     {
 663         t = t + v[i];
 664     }
 665
 666     return t;
 667 }
 668
 669 void sum_epot(gmx_grppairener_t *grpp, real *epot)
 670 {
 671     int i;
 672
 673     /* Accumulate energies */
 674     epot[F_COUL_SR]  = sum_v(grpp->nener, grpp->ener[egCOULSR]);
 675     epot[F_LJ]       = sum_v(grpp->nener, grpp->ener[egLJSR]);
 676     epot[F_LJ14]     = sum_v(grpp->nener, grpp->ener[egLJ14]);
 677     epot[F_COUL14]   = sum_v(grpp->nener, grpp->ener[egCOUL14]);
 678
 679 /* lattice part of LR doesnt belong to any group
 680  * and has been added earlier
 681  */
 682     epot[F_BHAM]     = sum_v(grpp->nener, grpp->ener[egBHAMSR]);
 683
 684     epot[F_EPOT] = 0;
 685     for (i = 0; (i < F_EPOT); i++)
 686     {
 687         if (i != F_DISRESVIOL && i != F_ORIRESDEV)
 688         {
 689             epot[F_EPOT] += epot[i];
 690         }
 691     }
 692 }
 693
 694 void sum_dhdl(gmx_enerdata_t *enerd, gmx::ArrayRef<const real> lambda, t_lambda *fepvals)
 695 {
 696     int    index;
 697     double dlam;
 698
 699     enerd->dvdl_lin[efptVDW] += enerd->term[F_DVDL_VDW];  /* include dispersion correction */
 700     enerd->term[F_DVDL]       = 0.0;
 701     for (int i = 0; i < efptNR; i++)
 702     {
 703         if (fepvals->separate_dvdl[i])
 704         {
 705             /* could this be done more readably/compactly? */
 706             switch (i)
 707             {
 708                 case (efptMASS):
 709                     index = F_DKDL;
 710                     break;
 711                 case (efptCOUL):
 712                     index = F_DVDL_COUL;
 713                     break;
 714                 case (efptVDW):
 715                     index = F_DVDL_VDW;
 716                     break;
 717                 case (efptBONDED):
 718                     index = F_DVDL_BONDED;
 719                     break;
 720                 case (efptRESTRAINT):
 721                     index = F_DVDL_RESTRAINT;
 722                     break;
 723                 default:
 724                     index = F_DVDL;
 725                     break;
 726             }
 727             enerd->term[index] = enerd->dvdl_lin[i] + enerd->dvdl_nonlin[i];
 728             if (debug)
 729             {
 730                 fprintf(debug, "dvdl-%s[%2d]: %f: non-linear %f + linear %f\n",
 731                         efpt_names[i], i, enerd->term[index], enerd->dvdl_nonlin[i], enerd->dvdl_lin[i]);
 732             }
 733         }
 734         else
 735         {
 736             enerd->term[F_DVDL] += enerd->dvdl_lin[i] + enerd->dvdl_nonlin[i];
 737             if (debug)
 738             {
 739                 fprintf(debug, "dvd-%sl[%2d]: %f: non-linear %f + linear %f\n",
 740                         efpt_names[0], i, enerd->term[F_DVDL], enerd->dvdl_nonlin[i], enerd->dvdl_lin[i]);
 741             }
 742         }
 743     }
 744
 745     /* Notes on the foreign lambda free energy difference evaluation:
 746      * Adding the potential and ekin terms that depend linearly on lambda
 747      * as delta lam * dvdl to the energy differences is exact.
 748      * For the constraints this is not exact, but we have no other option
 749      * without literally changing the lengths and reevaluating the energies at each step.
 750      * (try to remedy this post 4.6 - MRS)
 751      */
 752     if (fepvals->separate_dvdl[efptBONDED])
 753     {
 754         enerd->term[F_DVDL_BONDED] += enerd->term[F_DVDL_CONSTR];
 755     }
 756     else
 757     {
 758         enerd->term[F_DVDL] += enerd->term[F_DVDL_CONSTR];
 759     }
 760     enerd->term[F_DVDL_CONSTR] = 0;
 761
 762     for (int i = 0; i < fepvals->n_lambda; i++)
 763     {
 764         /* note we are iterating over fepvals here!
 765            For the current lam, dlam = 0 automatically,
 766            so we don't need to add anything to the
 767            enerd->enerpart_lambda[0] */
 768
 769         /* we don't need to worry about dvdl_lin contributions to dE at
 770            current lambda, because the contributions to the current
 771            lambda are automatically zeroed */
 772
 773         for (size_t j = 0; j < lambda.size(); j++)
 774         {
 775             /* Note that this loop is over all dhdl components, not just the separated ones */
 776             dlam = (fepvals->all_lambda[j][i] - lambda[j]);
 777             enerd->enerpart_lambda[i+1] += dlam*enerd->dvdl_lin[j];
 778             if (debug)
 779             {
 780                 fprintf(debug, "enerdiff lam %g: (%15s), non-linear %f linear %f*%f\n",
 781                         fepvals->all_lambda[j][i], efpt_names[j],
 782                         (enerd->enerpart_lambda[i+1] - enerd->enerpart_lambda[0]),
 783                         dlam, enerd->dvdl_lin[j]);
 784             }
 785         }
 786     }
 787 }
 788
 789
 790 void reset_foreign_enerdata(gmx_enerdata_t *enerd)
 791 {
 792     int  i, j;
 793
 794     /* First reset all foreign energy components.  Foreign energies always called on
 795        neighbor search steps */
 796     for (i = 0; (i < egNR); i++)
 797     {
 798         for (j = 0; (j < enerd->grpp.nener); j++)
 799         {
 800             enerd->foreign_grpp.ener[i][j] = 0.0;
 801         }
 802     }
 803
 804     /* potential energy components */
 805     for (i = 0; (i <= F_EPOT); i++)
 806     {
 807         enerd->foreign_term[i] = 0.0;
 808     }
 809 }
 810
 811 void reset_enerdata(gmx_enerdata_t *enerd)
 812 {
 813     int      i, j;
 814
 815     /* First reset all energy components. */
 816     for (i = 0; (i < egNR); i++)
 817     {
 818         for (j = 0; (j < enerd->grpp.nener); j++)
 819         {
 820             enerd->grpp.ener[i][j] = 0.0;
 821         }
 822     }
 823     for (i = 0; i < efptNR; i++)
 824     {
 825         enerd->dvdl_lin[i]    = 0.0;
 826         enerd->dvdl_nonlin[i] = 0.0;
 827     }
 828
 829     /* Normal potential energy components */
 830     for (i = 0; (i <= F_EPOT); i++)
 831     {
 832         enerd->term[i] = 0.0;
 833     }
 834     enerd->term[F_DVDL]            = 0.0;
 835     enerd->term[F_DVDL_COUL]       = 0.0;
 836     enerd->term[F_DVDL_VDW]        = 0.0;
 837     enerd->term[F_DVDL_BONDED]     = 0.0;
 838     enerd->term[F_DVDL_RESTRAINT]  = 0.0;
 839     enerd->term[F_DKDL]            = 0.0;
 840     if (enerd->n_lambda > 0)
 841     {
 842         for (i = 0; i < enerd->n_lambda; i++)
 843         {
 844             enerd->enerpart_lambda[i] = 0.0;
 845         }
 846     }
 847     /* reset foreign energy data - separate function since we also call it elsewhere */
 848     reset_foreign_enerdata(enerd);
 849 }