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, 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 "gromacs/legacyheaders/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/ewald/ewald.h"
  49 #include "gromacs/ewald/long-range-correction.h"
  50 #include "gromacs/ewald/pme.h"
  51 #include "gromacs/legacyheaders/gmx_omp_nthreads.h"
  52 #include "gromacs/legacyheaders/macros.h"
  53 #include "gromacs/legacyheaders/mdrun.h"
  54 #include "gromacs/legacyheaders/names.h"
  55 #include "gromacs/legacyheaders/network.h"
  56 #include "gromacs/legacyheaders/nonbonded.h"
  57 #include "gromacs/legacyheaders/nrnb.h"
  58 #include "gromacs/legacyheaders/ns.h"
  59 #include "gromacs/legacyheaders/qmmm.h"
  60 #include "gromacs/legacyheaders/txtdump.h"
  61 #include "gromacs/legacyheaders/typedefs.h"
  62 #include "gromacs/legacyheaders/types/commrec.h"
  63 #include "gromacs/listed-forces/listed-forces.h"
  64 #include "gromacs/math/vec.h"
  65 #include "gromacs/pbcutil/ishift.h"
  66 #include "gromacs/pbcutil/mshift.h"
  67 #include "gromacs/pbcutil/pbc.h"
  68 #include "gromacs/timing/wallcycle.h"
  69 #include "gromacs/utility/fatalerror.h"
  70 #include "gromacs/utility/smalloc.h"
  71
  72 void ns(FILE              *fp,
  73         t_forcerec        *fr,
  74         matrix             box,
  75         gmx_groups_t      *groups,
  76         gmx_localtop_t    *top,
  77         t_mdatoms         *md,
  78         t_commrec         *cr,
  79         t_nrnb            *nrnb,
  80         gmx_bool           bFillGrid,
  81         gmx_bool           bDoLongRangeNS)
  82 {
  83     int     nsearch;
  84
  85
  86     if (!fr->ns.nblist_initialized)
  87     {
  88         init_neighbor_list(fp, fr, md->homenr);
  89     }
  90
  91     if (fr->bTwinRange)
  92     {
  93         fr->nlr = 0;
  94     }
  95
  96     nsearch = search_neighbours(fp, fr, box, top, groups, cr, nrnb, md,
  97                                 bFillGrid, bDoLongRangeNS);
  98     if (debug)
  99     {
 100         fprintf(debug, "nsearch = %d\n", nsearch);
 101     }
 102
 103     /* Check whether we have to do dynamic load balancing */
 104     /*if ((nsb->nstDlb > 0) && (mod(step,nsb->nstDlb) == 0))
 105        count_nb(cr,nsb,&(top->blocks[ebCGS]),nns,fr->nlr,
 106        &(top->idef),opts->ngener);
 107      */
 108     if (fr->ns.dump_nl > 0)
 109     {
 110         dump_nblist(fp, cr, fr, fr->ns.dump_nl);
 111     }
 112 }
 113
 114 static void reduce_thread_forces(int n, rvec *f,
 115                                  tensor vir_q, tensor vir_lj,
 116                                  real *Vcorr_q, real *Vcorr_lj,
 117                                  real *dvdl_q, real *dvdl_lj,
 118                                  int nthreads, f_thread_t *f_t)
 119 {
 120     int t, i;
 121     int nthreads_loop gmx_unused;
 122
 123     // cppcheck-suppress unreadVariable
 124     nthreads_loop = gmx_omp_nthreads_get(emntBonded);
 125     /* This reduction can run over any number of threads */
 126 #pragma omp parallel for num_threads(nthreads_loop) private(t) schedule(static)
 127     for (i = 0; i < n; i++)
 128     {
 129         for (t = 1; t < nthreads; t++)
 130         {
 131             rvec_inc(f[i], f_t[t].f[i]);
 132         }
 133     }
 134     for (t = 1; t < nthreads; t++)
 135     {
 136         *Vcorr_q  += f_t[t].Vcorr_q;
 137         *Vcorr_lj += f_t[t].Vcorr_lj;
 138         *dvdl_q   += f_t[t].dvdl[efptCOUL];
 139         *dvdl_lj  += f_t[t].dvdl[efptVDW];
 140         m_add(vir_q, f_t[t].vir_q, vir_q);
 141         m_add(vir_lj, f_t[t].vir_lj, vir_lj);
 142     }
 143 }
 144
 145 void do_force_lowlevel(t_forcerec *fr,      t_inputrec *ir,
 146                        t_idef     *idef,    t_commrec  *cr,
 147                        t_nrnb     *nrnb,    gmx_wallcycle_t wcycle,
 148                        t_mdatoms  *md,
 149                        rvec       x[],      history_t  *hist,
 150                        rvec       f[],
 151                        rvec       f_longrange[],
 152                        gmx_enerdata_t *enerd,
 153                        t_fcdata   *fcd,
 154                        gmx_localtop_t *top,
 155                        gmx_genborn_t *born,
 156                        gmx_bool       bBornRadii,
 157                        matrix     box,
 158                        t_lambda   *fepvals,
 159                        real       *lambda,
 160                        t_graph    *graph,
 161                        t_blocka   *excl,
 162                        rvec       mu_tot[],
 163                        int        flags,
 164                        float      *cycles_pme)
 165 {
 166     int         i, j;
 167     int         donb_flags;
 168     gmx_bool    bSB;
 169     int         pme_flags;
 170     matrix      boxs;
 171     rvec        box_size;
 172     t_pbc       pbc;
 173     real        dvdl_dum[efptNR], dvdl_nb[efptNR];
 174
 175 #ifdef GMX_MPI
 176     double  t0 = 0.0, t1, t2, t3; /* time measurement for coarse load balancing */
 177 #endif
 178
 179     set_pbc(&pbc, fr->ePBC, box);
 180
 181     /* reset free energy components */
 182     for (i = 0; i < efptNR; i++)
 183     {
 184         dvdl_nb[i]  = 0;
 185         dvdl_dum[i] = 0;
 186     }
 187
 188     /* Reset box */
 189     for (i = 0; (i < DIM); i++)
 190     {
 191         box_size[i] = box[i][i];
 192     }
 193
 194     debug_gmx();
 195
 196     /* do QMMM first if requested */
 197     if (fr->bQMMM)
 198     {
 199         enerd->term[F_EQM] = calculate_QMMM(cr, x, f, fr);
 200     }
 201
 202     /* Call the short range functions all in one go. */
 203
 204 #ifdef GMX_MPI
 205     /*#define TAKETIME ((cr->npmenodes) && (fr->timesteps < 12))*/
 206 #define TAKETIME FALSE
 207     if (TAKETIME)
 208     {
 209         MPI_Barrier(cr->mpi_comm_mygroup);
 210         t0 = MPI_Wtime();
 211     }
 212 #endif
 213
 214     if (ir->nwall)
 215     {
 216         /* foreign lambda component for walls */
 217         real dvdl_walls = do_walls(ir, fr, box, md, x, f, lambda[efptVDW],
 218                                    enerd->grpp.ener[egLJSR], nrnb);
 219         enerd->dvdl_lin[efptVDW] += dvdl_walls;
 220     }
 221
 222     /* If doing GB, reset dvda and calculate the Born radii */
 223     if (ir->implicit_solvent)
 224     {
 225         wallcycle_sub_start(wcycle, ewcsNONBONDED);
 226
 227         for (i = 0; i < born->nr; i++)
 228         {
 229             fr->dvda[i] = 0;
 230         }
 231
 232         if (bBornRadii)
 233         {
 234             calc_gb_rad(cr, fr, ir, top, x, &(fr->gblist), born, md, nrnb);
 235         }
 236
 237         wallcycle_sub_stop(wcycle, ewcsNONBONDED);
 238     }
 239
 240     where();
 241     /* We only do non-bonded calculation with group scheme here, the verlet
 242      * calls are done from do_force_cutsVERLET(). */
 243     if (fr->cutoff_scheme == ecutsGROUP && (flags & GMX_FORCE_NONBONDED))
 244     {
 245         donb_flags = 0;
 246         /* Add short-range interactions */
 247         donb_flags |= GMX_NONBONDED_DO_SR;
 248
 249         /* Currently all group scheme kernels always calculate (shift-)forces */
 250         if (flags & GMX_FORCE_FORCES)
 251         {
 252             donb_flags |= GMX_NONBONDED_DO_FORCE;
 253         }
 254         if (flags & GMX_FORCE_VIRIAL)
 255         {
 256             donb_flags |= GMX_NONBONDED_DO_SHIFTFORCE;
 257         }
 258         if (flags & GMX_FORCE_ENERGY)
 259         {
 260             donb_flags |= GMX_NONBONDED_DO_POTENTIAL;
 261         }
 262         if (flags & GMX_FORCE_DO_LR)
 263         {
 264             donb_flags |= GMX_NONBONDED_DO_LR;
 265         }
 266
 267         wallcycle_sub_start(wcycle, ewcsNONBONDED);
 268         do_nonbonded(fr, x, f, f_longrange, md, excl,
 269                      &enerd->grpp, nrnb,
 270                      lambda, dvdl_nb, -1, -1, donb_flags);
 271
 272         /* If we do foreign lambda and we have soft-core interactions
 273          * we have to recalculate the (non-linear) energies contributions.
 274          */
 275         if (fepvals->n_lambda > 0 && (flags & GMX_FORCE_DHDL) && fepvals->sc_alpha != 0)
 276         {
 277             for (i = 0; i < enerd->n_lambda; i++)
 278             {
 279                 real lam_i[efptNR];
 280
 281                 for (j = 0; j < efptNR; j++)
 282                 {
 283                     lam_i[j] = (i == 0 ? lambda[j] : fepvals->all_lambda[j][i-1]);
 284                 }
 285                 reset_foreign_enerdata(enerd);
 286                 do_nonbonded(fr, x, f, f_longrange, md, excl,
 287                              &(enerd->foreign_grpp), nrnb,
 288                              lam_i, dvdl_dum, -1, -1,
 289                              (donb_flags & ~GMX_NONBONDED_DO_FORCE) | GMX_NONBONDED_DO_FOREIGNLAMBDA);
 290                 sum_epot(&(enerd->foreign_grpp), enerd->foreign_term);
 291                 enerd->enerpart_lambda[i] += enerd->foreign_term[F_EPOT];
 292             }
 293         }
 294         wallcycle_sub_stop(wcycle, ewcsNONBONDED);
 295         where();
 296     }
 297
 298     /* If we are doing GB, calculate bonded forces and apply corrections
 299      * to the solvation forces */
 300     /* MRS: Eventually, many need to include free energy contribution here! */
 301     if (ir->implicit_solvent)
 302     {
 303         wallcycle_sub_start(wcycle, ewcsLISTED);
 304         calc_gb_forces(cr, md, born, top, x, f, fr, idef,
 305                        ir->gb_algorithm, ir->sa_algorithm, nrnb, &pbc, graph, enerd);
 306         wallcycle_sub_stop(wcycle, ewcsLISTED);
 307     }
 308
 309 #ifdef GMX_MPI
 310     if (TAKETIME)
 311     {
 312         t1          = MPI_Wtime();
 313         fr->t_fnbf += t1-t0;
 314     }
 315 #endif
 316
 317     if (fepvals->sc_alpha != 0)
 318     {
 319         enerd->dvdl_nonlin[efptVDW] += dvdl_nb[efptVDW];
 320     }
 321     else
 322     {
 323         enerd->dvdl_lin[efptVDW] += dvdl_nb[efptVDW];
 324     }
 325
 326     if (fepvals->sc_alpha != 0)
 327
 328     /* even though coulomb part is linear, we already added it, beacuse we
 329        need to go through the vdw calculation anyway */
 330     {
 331         enerd->dvdl_nonlin[efptCOUL] += dvdl_nb[efptCOUL];
 332     }
 333     else
 334     {
 335         enerd->dvdl_lin[efptCOUL] += dvdl_nb[efptCOUL];
 336     }
 337
 338     debug_gmx();
 339
 340
 341     if (debug)
 342     {
 343         pr_rvecs(debug, 0, "fshift after SR", fr->fshift, SHIFTS);
 344     }
 345
 346     /* Shift the coordinates. Must be done before listed forces and PPPM,
 347      * but is also necessary for SHAKE and update, therefore it can NOT
 348      * go when no listed forces have to be evaluated.
 349      */
 350
 351     /* Here sometimes we would not need to shift with NBFonly,
 352      * but we do so anyhow for consistency of the returned coordinates.
 353      */
 354     if (graph)
 355     {
 356         shift_self(graph, box, x);
 357         if (TRICLINIC(box))
 358         {
 359             inc_nrnb(nrnb, eNR_SHIFTX, 2*graph->nnodes);
 360         }
 361         else
 362         {
 363             inc_nrnb(nrnb, eNR_SHIFTX, graph->nnodes);
 364         }
 365     }
 366     /* Check whether we need to do listed interactions or correct for exclusions */
 367     if (fr->bMolPBC &&
 368         ((flags & GMX_FORCE_LISTED)
 369          || EEL_RF(fr->eeltype) || EEL_FULL(fr->eeltype) || EVDW_PME(fr->vdwtype)))
 370     {
 371         /* Since all atoms are in the rectangular or triclinic unit-cell,
 372          * only single box vector shifts (2 in x) are required.
 373          */
 374         set_pbc_dd(&pbc, fr->ePBC, cr->dd, TRUE, box);
 375     }
 376     debug_gmx();
 377
 378     do_force_listed(wcycle, box, ir->fepvals, cr->ms,
 379                     idef, (const rvec *) x, hist, f, fr,
 380                     &pbc, graph, enerd, nrnb, lambda, md, fcd,
 381                     DOMAINDECOMP(cr) ? cr->dd->gatindex : NULL,
 382                     flags);
 383
 384     where();
 385
 386     *cycles_pme = 0;
 387     clear_mat(fr->vir_el_recip);
 388     clear_mat(fr->vir_lj_recip);
 389
 390     /* Do long-range electrostatics and/or LJ-PME, including related short-range
 391      * corrections.
 392      */
 393     if (EEL_FULL(fr->eeltype) || EVDW_PME(fr->vdwtype))
 394     {
 395         int  status            = 0;
 396         real Vlr_q             = 0, Vlr_lj = 0, Vcorr_q = 0, Vcorr_lj = 0;
 397         real dvdl_long_range_q = 0, dvdl_long_range_lj = 0;
 398
 399         bSB = (ir->nwall == 2);
 400         if (bSB)
 401         {
 402             copy_mat(box, boxs);
 403             svmul(ir->wall_ewald_zfac, boxs[ZZ], boxs[ZZ]);
 404             box_size[ZZ] *= ir->wall_ewald_zfac;
 405         }
 406
 407         if (EEL_PME_EWALD(fr->eeltype) || EVDW_PME(fr->vdwtype))
 408         {
 409             real dvdl_long_range_correction_q   = 0;
 410             real dvdl_long_range_correction_lj  = 0;
 411             /* With the Verlet scheme exclusion forces are calculated
 412              * in the non-bonded kernel.
 413              */
 414             /* The TPI molecule does not have exclusions with the rest
 415              * of the system and no intra-molecular PME grid
 416              * contributions will be calculated in
 417              * gmx_pme_calc_energy.
 418              */
 419             if ((ir->cutoff_scheme == ecutsGROUP && fr->n_tpi == 0) ||
 420                 ir->ewald_geometry != eewg3D ||
 421                 ir->epsilon_surface != 0)
 422             {
 423                 int nthreads, t;
 424
 425                 wallcycle_sub_start(wcycle, ewcsEWALD_CORRECTION);
 426
 427                 if (fr->n_tpi > 0)
 428                 {
 429                     gmx_fatal(FARGS, "TPI with PME currently only works in a 3D geometry with tin-foil boundary conditions");
 430                 }
 431
 432                 nthreads = gmx_omp_nthreads_get(emntBonded);
 433 #pragma omp parallel for num_threads(nthreads) schedule(static)
 434                 for (t = 0; t < nthreads; t++)
 435                 {
 436                     int     i;
 437                     rvec   *fnv;
 438                     tensor *vir_q, *vir_lj;
 439                     real   *Vcorrt_q, *Vcorrt_lj, *dvdlt_q, *dvdlt_lj;
 440                     if (t == 0)
 441                     {
 442                         fnv       = fr->f_novirsum;
 443                         vir_q     = &fr->vir_el_recip;
 444                         vir_lj    = &fr->vir_lj_recip;
 445                         Vcorrt_q  = &Vcorr_q;
 446                         Vcorrt_lj = &Vcorr_lj;
 447                         dvdlt_q   = &dvdl_long_range_correction_q;
 448                         dvdlt_lj  = &dvdl_long_range_correction_lj;
 449                     }
 450                     else
 451                     {
 452                         fnv       = fr->f_t[t].f;
 453                         vir_q     = &fr->f_t[t].vir_q;
 454                         vir_lj    = &fr->f_t[t].vir_lj;
 455                         Vcorrt_q  = &fr->f_t[t].Vcorr_q;
 456                         Vcorrt_lj = &fr->f_t[t].Vcorr_lj;
 457                         dvdlt_q   = &fr->f_t[t].dvdl[efptCOUL];
 458                         dvdlt_lj  = &fr->f_t[t].dvdl[efptVDW];
 459                         for (i = 0; i < fr->natoms_force; i++)
 460                         {
 461                             clear_rvec(fnv[i]);
 462                         }
 463                         clear_mat(*vir_q);
 464                         clear_mat(*vir_lj);
 465                     }
 466                     *dvdlt_q  = 0;
 467                     *dvdlt_lj = 0;
 468
 469                     ewald_LRcorrection(fr->excl_load[t], fr->excl_load[t+1],
 470                                        cr, t, fr,
 471                                        md->chargeA,
 472                                        md->nChargePerturbed ? md->chargeB : NULL,
 473                                        md->sqrt_c6A,
 474                                        md->nTypePerturbed ? md->sqrt_c6B : NULL,
 475                                        md->sigmaA,
 476                                        md->nTypePerturbed ? md->sigmaB : NULL,
 477                                        md->sigma3A,
 478                                        md->nTypePerturbed ? md->sigma3B : NULL,
 479                                        ir->cutoff_scheme != ecutsVERLET,
 480                                        excl, x, bSB ? boxs : box, mu_tot,
 481                                        ir->ewald_geometry,
 482                                        ir->epsilon_surface,
 483                                        fnv, *vir_q, *vir_lj,
 484                                        Vcorrt_q, Vcorrt_lj,
 485                                        lambda[efptCOUL], lambda[efptVDW],
 486                                        dvdlt_q, dvdlt_lj);
 487                 }
 488                 if (nthreads > 1)
 489                 {
 490                     reduce_thread_forces(fr->natoms_force, fr->f_novirsum,
 491                                          fr->vir_el_recip, fr->vir_lj_recip,
 492                                          &Vcorr_q, &Vcorr_lj,
 493                                          &dvdl_long_range_correction_q,
 494                                          &dvdl_long_range_correction_lj,
 495                                          nthreads, fr->f_t);
 496                 }
 497                 wallcycle_sub_stop(wcycle, ewcsEWALD_CORRECTION);
 498             }
 499
 500             if (EEL_PME_EWALD(fr->eeltype) && fr->n_tpi == 0)
 501             {
 502                 Vcorr_q += ewald_charge_correction(cr, fr, lambda[efptCOUL], box,
 503                                                    &dvdl_long_range_correction_q,
 504                                                    fr->vir_el_recip);
 505             }
 506
 507             enerd->dvdl_lin[efptCOUL] += dvdl_long_range_correction_q;
 508             enerd->dvdl_lin[efptVDW]  += dvdl_long_range_correction_lj;
 509
 510             if ((EEL_PME(fr->eeltype) || EVDW_PME(fr->vdwtype)) && (cr->duty & DUTY_PME))
 511             {
 512                 /* Do reciprocal PME for Coulomb and/or LJ. */
 513                 assert(fr->n_tpi >= 0);
 514                 if (fr->n_tpi == 0 || (flags & GMX_FORCE_STATECHANGED))
 515                 {
 516                     pme_flags = GMX_PME_SPREAD | GMX_PME_SOLVE;
 517                     if (EEL_PME(fr->eeltype))
 518                     {
 519                         pme_flags     |= GMX_PME_DO_COULOMB;
 520                     }
 521                     if (EVDW_PME(fr->vdwtype))
 522                     {
 523                         pme_flags |= GMX_PME_DO_LJ;
 524                     }
 525                     if (flags & GMX_FORCE_FORCES)
 526                     {
 527                         pme_flags |= GMX_PME_CALC_F;
 528                     }
 529                     if (flags & GMX_FORCE_VIRIAL)
 530                     {
 531                         pme_flags |= GMX_PME_CALC_ENER_VIR;
 532                     }
 533                     if (fr->n_tpi > 0)
 534                     {
 535                         /* We don't calculate f, but we do want the potential */
 536                         pme_flags |= GMX_PME_CALC_POT;
 537                     }
 538                     wallcycle_start(wcycle, ewcPMEMESH);
 539                     status = gmx_pme_do(fr->pmedata,
 540                                         0, md->homenr - fr->n_tpi,
 541                                         x, fr->f_novirsum,
 542                                         md->chargeA, md->chargeB,
 543                                         md->sqrt_c6A, md->sqrt_c6B,
 544                                         md->sigmaA, md->sigmaB,
 545                                         bSB ? boxs : box, cr,
 546                                         DOMAINDECOMP(cr) ? dd_pme_maxshift_x(cr->dd) : 0,
 547                                         DOMAINDECOMP(cr) ? dd_pme_maxshift_y(cr->dd) : 0,
 548                                         nrnb, wcycle,
 549                                         fr->vir_el_recip, fr->ewaldcoeff_q,
 550                                         fr->vir_lj_recip, fr->ewaldcoeff_lj,
 551                                         &Vlr_q, &Vlr_lj,
 552                                         lambda[efptCOUL], lambda[efptVDW],
 553                                         &dvdl_long_range_q, &dvdl_long_range_lj, pme_flags);
 554                     *cycles_pme = wallcycle_stop(wcycle, ewcPMEMESH);
 555                     if (status != 0)
 556                     {
 557                         gmx_fatal(FARGS, "Error %d in reciprocal PME routine", status);
 558                     }
 559                     /* We should try to do as little computation after
 560                      * this as possible, because parallel PME synchronizes
 561                      * the nodes, so we want all load imbalance of the
 562                      * rest of the force calculation to be before the PME
 563                      * call.  DD load balancing is done on the whole time
 564                      * of the force call (without PME).
 565                      */
 566                 }
 567                 if (fr->n_tpi > 0)
 568                 {
 569                     if (EVDW_PME(ir->vdwtype))
 570                     {
 571
 572                         gmx_fatal(FARGS, "Test particle insertion not implemented with LJ-PME");
 573                     }
 574                     /* Determine the PME grid energy of the test molecule
 575                      * with the PME grid potential of the other charges.
 576                      */
 577                     gmx_pme_calc_energy(fr->pmedata, fr->n_tpi,
 578                                         x + md->homenr - fr->n_tpi,
 579                                         md->chargeA + md->homenr - fr->n_tpi,
 580                                         &Vlr_q);
 581                 }
 582             }
 583         }
 584
 585         if (!EEL_PME(fr->eeltype) && EEL_PME_EWALD(fr->eeltype))
 586         {
 587             Vlr_q = do_ewald(ir, x, fr->f_novirsum,
 588                              md->chargeA, md->chargeB,
 589                              box_size, cr, md->homenr,
 590                              fr->vir_el_recip, fr->ewaldcoeff_q,
 591                              lambda[efptCOUL], &dvdl_long_range_q, fr->ewald_table);
 592         }
 593
 594         /* Note that with separate PME nodes we get the real energies later */
 595         enerd->dvdl_lin[efptCOUL] += dvdl_long_range_q;
 596         enerd->dvdl_lin[efptVDW]  += dvdl_long_range_lj;
 597         enerd->term[F_COUL_RECIP]  = Vlr_q + Vcorr_q;
 598         enerd->term[F_LJ_RECIP]    = Vlr_lj + Vcorr_lj;
 599         if (debug)
 600         {
 601             fprintf(debug, "Vlr_q = %g, Vcorr_q = %g, Vlr_corr_q = %g\n",
 602                     Vlr_q, Vcorr_q, enerd->term[F_COUL_RECIP]);
 603             pr_rvecs(debug, 0, "vir_el_recip after corr", fr->vir_el_recip, DIM);
 604             pr_rvecs(debug, 0, "fshift after LR Corrections", fr->fshift, SHIFTS);
 605             fprintf(debug, "Vlr_lj: %g, Vcorr_lj = %g, Vlr_corr_lj = %g\n",
 606                     Vlr_lj, Vcorr_lj, enerd->term[F_LJ_RECIP]);
 607             pr_rvecs(debug, 0, "vir_lj_recip after corr", fr->vir_lj_recip, DIM);
 608         }
 609     }
 610     else
 611     {
 612         /* Is there a reaction-field exclusion correction needed? */
 613         if (EEL_RF(fr->eeltype) && eelRF_NEC != fr->eeltype)
 614         {
 615             /* With the Verlet scheme, exclusion forces are calculated
 616              * in the non-bonded kernel.
 617              */
 618             if (ir->cutoff_scheme != ecutsVERLET)
 619             {
 620                 real dvdl_rf_excl      = 0;
 621                 enerd->term[F_RF_EXCL] =
 622                     RF_excl_correction(fr, graph, md, excl, x, f,
 623                                        fr->fshift, &pbc, lambda[efptCOUL], &dvdl_rf_excl);
 624
 625                 enerd->dvdl_lin[efptCOUL] += dvdl_rf_excl;
 626             }
 627         }
 628     }
 629     where();
 630     debug_gmx();
 631
 632     if (debug)
 633     {
 634         print_nrnb(debug, nrnb);
 635     }
 636     debug_gmx();
 637
 638 #ifdef GMX_MPI
 639     if (TAKETIME)
 640     {
 641         t2 = MPI_Wtime();
 642         MPI_Barrier(cr->mpi_comm_mygroup);
 643         t3          = MPI_Wtime();
 644         fr->t_wait += t3-t2;
 645         if (fr->timesteps == 11)
 646         {
 647             char buf[22];
 648             fprintf(stderr, "* PP load balancing info: rank %d, step %s, rel wait time=%3.0f%% , load string value: %7.2f\n",
 649                     cr->nodeid, gmx_step_str(fr->timesteps, buf),
 650                     100*fr->t_wait/(fr->t_wait+fr->t_fnbf),
 651                     (fr->t_fnbf+fr->t_wait)/fr->t_fnbf);
 652         }
 653         fr->timesteps++;
 654     }
 655 #endif
 656
 657     if (debug)
 658     {
 659         pr_rvecs(debug, 0, "fshift after bondeds", fr->fshift, SHIFTS);
 660     }
 661
 662 }
 663
 664 void init_enerdata(int ngener, int n_lambda, gmx_enerdata_t *enerd)
 665 {
 666     int i, n2;
 667
 668     for (i = 0; i < F_NRE; i++)
 669     {
 670         enerd->term[i]         = 0;
 671         enerd->foreign_term[i] = 0;
 672     }
 673
 674
 675     for (i = 0; i < efptNR; i++)
 676     {
 677         enerd->dvdl_lin[i]     = 0;
 678         enerd->dvdl_nonlin[i]  = 0;
 679     }
 680
 681     n2 = ngener*ngener;
 682     if (debug)
 683     {
 684         fprintf(debug, "Creating %d sized group matrix for energies\n", n2);
 685     }
 686     enerd->grpp.nener         = n2;
 687     enerd->foreign_grpp.nener = n2;
 688     for (i = 0; (i < egNR); i++)
 689     {
 690         snew(enerd->grpp.ener[i], n2);
 691         snew(enerd->foreign_grpp.ener[i], n2);
 692     }
 693
 694     if (n_lambda)
 695     {
 696         enerd->n_lambda = 1 + n_lambda;
 697         snew(enerd->enerpart_lambda, enerd->n_lambda);
 698     }
 699     else
 700     {
 701         enerd->n_lambda = 0;
 702     }
 703 }
 704
 705 void destroy_enerdata(gmx_enerdata_t *enerd)
 706 {
 707     int i;
 708
 709     for (i = 0; (i < egNR); i++)
 710     {
 711         sfree(enerd->grpp.ener[i]);
 712     }
 713
 714     for (i = 0; (i < egNR); i++)
 715     {
 716         sfree(enerd->foreign_grpp.ener[i]);
 717     }
 718
 719     if (enerd->n_lambda)
 720     {
 721         sfree(enerd->enerpart_lambda);
 722     }
 723 }
 724
 725 static real sum_v(int n, real v[])
 726 {
 727     real t;
 728     int  i;
 729
 730     t = 0.0;
 731     for (i = 0; (i < n); i++)
 732     {
 733         t = t + v[i];
 734     }
 735
 736     return t;
 737 }
 738
 739 void sum_epot(gmx_grppairener_t *grpp, real *epot)
 740 {
 741     int i;
 742
 743     /* Accumulate energies */
 744     epot[F_COUL_SR]  = sum_v(grpp->nener, grpp->ener[egCOULSR]);
 745     epot[F_LJ]       = sum_v(grpp->nener, grpp->ener[egLJSR]);
 746     epot[F_LJ14]     = sum_v(grpp->nener, grpp->ener[egLJ14]);
 747     epot[F_COUL14]   = sum_v(grpp->nener, grpp->ener[egCOUL14]);
 748     epot[F_COUL_LR]  = sum_v(grpp->nener, grpp->ener[egCOULLR]);
 749     epot[F_LJ_LR]    = sum_v(grpp->nener, grpp->ener[egLJLR]);
 750     /* We have already added 1-2,1-3, and 1-4 terms to F_GBPOL */
 751     epot[F_GBPOL]   += sum_v(grpp->nener, grpp->ener[egGB]);
 752
 753 /* lattice part of LR doesnt belong to any group
 754  * and has been added earlier
 755  */
 756     epot[F_BHAM]     = sum_v(grpp->nener, grpp->ener[egBHAMSR]);
 757     epot[F_BHAM_LR]  = sum_v(grpp->nener, grpp->ener[egBHAMLR]);
 758
 759     epot[F_EPOT] = 0;
 760     for (i = 0; (i < F_EPOT); i++)
 761     {
 762         if (i != F_DISRESVIOL && i != F_ORIRESDEV)
 763         {
 764             epot[F_EPOT] += epot[i];
 765         }
 766     }
 767 }
 768
 769 void sum_dhdl(gmx_enerdata_t *enerd, real *lambda, t_lambda *fepvals)
 770 {
 771     int    i, j, index;
 772     double dlam;
 773
 774     enerd->dvdl_lin[efptVDW] += enerd->term[F_DVDL_VDW];  /* include dispersion correction */
 775     enerd->term[F_DVDL]       = 0.0;
 776     for (i = 0; i < efptNR; i++)
 777     {
 778         if (fepvals->separate_dvdl[i])
 779         {
 780             /* could this be done more readably/compactly? */
 781             switch (i)
 782             {
 783                 case (efptMASS):
 784                     index = F_DKDL;
 785                     break;
 786                 case (efptCOUL):
 787                     index = F_DVDL_COUL;
 788                     break;
 789                 case (efptVDW):
 790                     index = F_DVDL_VDW;
 791                     break;
 792                 case (efptBONDED):
 793                     index = F_DVDL_BONDED;
 794                     break;
 795                 case (efptRESTRAINT):
 796                     index = F_DVDL_RESTRAINT;
 797                     break;
 798                 default:
 799                     index = F_DVDL;
 800                     break;
 801             }
 802             enerd->term[index] = enerd->dvdl_lin[i] + enerd->dvdl_nonlin[i];
 803             if (debug)
 804             {
 805                 fprintf(debug, "dvdl-%s[%2d]: %f: non-linear %f + linear %f\n",
 806                         efpt_names[i], i, enerd->term[index], enerd->dvdl_nonlin[i], enerd->dvdl_lin[i]);
 807             }
 808         }
 809         else
 810         {
 811             enerd->term[F_DVDL] += enerd->dvdl_lin[i] + enerd->dvdl_nonlin[i];
 812             if (debug)
 813             {
 814                 fprintf(debug, "dvd-%sl[%2d]: %f: non-linear %f + linear %f\n",
 815                         efpt_names[0], i, enerd->term[F_DVDL], enerd->dvdl_nonlin[i], enerd->dvdl_lin[i]);
 816             }
 817         }
 818     }
 819
 820     /* Notes on the foreign lambda free energy difference evaluation:
 821      * Adding the potential and ekin terms that depend linearly on lambda
 822      * as delta lam * dvdl to the energy differences is exact.
 823      * For the constraints this is not exact, but we have no other option
 824      * without literally changing the lengths and reevaluating the energies at each step.
 825      * (try to remedy this post 4.6 - MRS)
 826      * For the non-bonded LR term we assume that the soft-core (if present)
 827      * no longer affects the energy beyond the short-range cut-off,
 828      * which is a very good approximation (except for exotic settings).
 829      * (investigate how to overcome this post 4.6 - MRS)
 830      */
 831     if (fepvals->separate_dvdl[efptBONDED])
 832     {
 833         enerd->term[F_DVDL_BONDED] += enerd->term[F_DVDL_CONSTR];
 834     }
 835     else
 836     {
 837         enerd->term[F_DVDL] += enerd->term[F_DVDL_CONSTR];
 838     }
 839     enerd->term[F_DVDL_CONSTR] = 0;
 840
 841     for (i = 0; i < fepvals->n_lambda; i++)
 842     {
 843         /* note we are iterating over fepvals here!
 844            For the current lam, dlam = 0 automatically,
 845            so we don't need to add anything to the
 846            enerd->enerpart_lambda[0] */
 847
 848         /* we don't need to worry about dvdl_lin contributions to dE at
 849            current lambda, because the contributions to the current
 850            lambda are automatically zeroed */
 851
 852         for (j = 0; j < efptNR; j++)
 853         {
 854             /* Note that this loop is over all dhdl components, not just the separated ones */
 855             dlam = (fepvals->all_lambda[j][i]-lambda[j]);
 856             enerd->enerpart_lambda[i+1] += dlam*enerd->dvdl_lin[j];
 857             if (debug)
 858             {
 859                 fprintf(debug, "enerdiff lam %g: (%15s), non-linear %f linear %f*%f\n",
 860                         fepvals->all_lambda[j][i], efpt_names[j],
 861                         (enerd->enerpart_lambda[i+1] - enerd->enerpart_lambda[0]),
 862                         dlam, enerd->dvdl_lin[j]);
 863             }
 864         }
 865     }
 866 }
 867
 868
 869 void reset_foreign_enerdata(gmx_enerdata_t *enerd)
 870 {
 871     int  i, j;
 872
 873     /* First reset all foreign energy components.  Foreign energies always called on
 874        neighbor search steps */
 875     for (i = 0; (i < egNR); i++)
 876     {
 877         for (j = 0; (j < enerd->grpp.nener); j++)
 878         {
 879             enerd->foreign_grpp.ener[i][j] = 0.0;
 880         }
 881     }
 882
 883     /* potential energy components */
 884     for (i = 0; (i <= F_EPOT); i++)
 885     {
 886         enerd->foreign_term[i] = 0.0;
 887     }
 888 }
 889
 890 void reset_enerdata(t_forcerec *fr, gmx_bool bNS,
 891                     gmx_enerdata_t *enerd,
 892                     gmx_bool bMaster)
 893 {
 894     gmx_bool bKeepLR;
 895     int      i, j;
 896
 897     /* First reset all energy components, except for the long range terms
 898      * on the master at non neighbor search steps, since the long range
 899      * terms have already been summed at the last neighbor search step.
 900      */
 901     bKeepLR = (fr->bTwinRange && !bNS);
 902     for (i = 0; (i < egNR); i++)
 903     {
 904         if (!(bKeepLR && bMaster && (i == egCOULLR || i == egLJLR)))
 905         {
 906             for (j = 0; (j < enerd->grpp.nener); j++)
 907             {
 908                 enerd->grpp.ener[i][j] = 0.0;
 909             }
 910         }
 911     }
 912     for (i = 0; i < efptNR; i++)
 913     {
 914         enerd->dvdl_lin[i]    = 0.0;
 915         enerd->dvdl_nonlin[i] = 0.0;
 916     }
 917
 918     /* Normal potential energy components */
 919     for (i = 0; (i <= F_EPOT); i++)
 920     {
 921         enerd->term[i] = 0.0;
 922     }
 923     /* Initialize the dVdlambda term with the long range contribution */
 924     /* Initialize the dvdl term with the long range contribution */
 925     enerd->term[F_DVDL]            = 0.0;
 926     enerd->term[F_DVDL_COUL]       = 0.0;
 927     enerd->term[F_DVDL_VDW]        = 0.0;
 928     enerd->term[F_DVDL_BONDED]     = 0.0;
 929     enerd->term[F_DVDL_RESTRAINT]  = 0.0;
 930     enerd->term[F_DKDL]            = 0.0;
 931     if (enerd->n_lambda > 0)
 932     {
 933         for (i = 0; i < enerd->n_lambda; i++)
 934         {
 935             enerd->enerpart_lambda[i] = 0.0;
 936         }
 937     }
 938     /* reset foreign energy data - separate function since we also call it elsewhere */
 939     reset_foreign_enerdata(enerd);
 940 }