2 * This file is part of the GROMACS molecular simulation package.
4 * Copyright (c) 1991-2000, University of Groningen, The Netherlands.
5 * Copyright (c) 2001-2004, The GROMACS development team.
6 * Copyright (c) 2011,2012,2013,2014,2015, 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.
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.
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.
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.
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.
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.
45 #include "thread_mpi/threads.h"
47 #include "gromacs/domdec/domdec.h"
48 #include "gromacs/domdec/domdec_network.h"
49 #include "gromacs/ewald/pme-load-balancing.h"
50 #include "gromacs/ewald/pme.h"
51 #include "gromacs/fileio/filenm.h"
52 #include "gromacs/fileio/mdoutf.h"
53 #include "gromacs/fileio/trajectory_writing.h"
54 #include "gromacs/fileio/trx.h"
55 #include "gromacs/fileio/trxio.h"
56 #include "gromacs/imd/imd.h"
57 #include "gromacs/legacyheaders/constr.h"
58 #include "gromacs/legacyheaders/ebin.h"
59 #include "gromacs/legacyheaders/force.h"
60 #include "gromacs/legacyheaders/md_logging.h"
61 #include "gromacs/legacyheaders/md_support.h"
62 #include "gromacs/legacyheaders/mdatoms.h"
63 #include "gromacs/legacyheaders/mdebin.h"
64 #include "gromacs/legacyheaders/mdrun.h"
65 #include "gromacs/legacyheaders/network.h"
66 #include "gromacs/legacyheaders/nrnb.h"
67 #include "gromacs/legacyheaders/ns.h"
68 #include "gromacs/legacyheaders/shellfc.h"
69 #include "gromacs/legacyheaders/sighandler.h"
70 #include "gromacs/legacyheaders/sim_util.h"
71 #include "gromacs/legacyheaders/tgroup.h"
72 #include "gromacs/legacyheaders/typedefs.h"
73 #include "gromacs/legacyheaders/update.h"
74 #include "gromacs/legacyheaders/vcm.h"
75 #include "gromacs/legacyheaders/vsite.h"
76 #include "gromacs/legacyheaders/types/commrec.h"
77 #include "gromacs/legacyheaders/types/constr.h"
78 #include "gromacs/legacyheaders/types/enums.h"
79 #include "gromacs/legacyheaders/types/fcdata.h"
80 #include "gromacs/legacyheaders/types/force_flags.h"
81 #include "gromacs/legacyheaders/types/forcerec.h"
82 #include "gromacs/legacyheaders/types/group.h"
83 #include "gromacs/legacyheaders/types/inputrec.h"
84 #include "gromacs/legacyheaders/types/interaction_const.h"
85 #include "gromacs/legacyheaders/types/mdatom.h"
86 #include "gromacs/legacyheaders/types/membedt.h"
87 #include "gromacs/legacyheaders/types/nrnb.h"
88 #include "gromacs/legacyheaders/types/oenv.h"
89 #include "gromacs/legacyheaders/types/shellfc.h"
90 #include "gromacs/legacyheaders/types/state.h"
91 #include "gromacs/listed-forces/manage-threading.h"
92 #include "gromacs/math/utilities.h"
93 #include "gromacs/math/vec.h"
94 #include "gromacs/math/vectypes.h"
95 #include "gromacs/mdlib/compute_io.h"
96 #include "gromacs/mdlib/mdrun_signalling.h"
97 #include "gromacs/mdlib/nb_verlet.h"
98 #include "gromacs/mdlib/nbnxn_gpu_data_mgmt.h"
99 #include "gromacs/pbcutil/mshift.h"
100 #include "gromacs/pbcutil/pbc.h"
101 #include "gromacs/pulling/pull.h"
102 #include "gromacs/swap/swapcoords.h"
103 #include "gromacs/timing/wallcycle.h"
104 #include "gromacs/timing/walltime_accounting.h"
105 #include "gromacs/topology/atoms.h"
106 #include "gromacs/topology/idef.h"
107 #include "gromacs/topology/mtop_util.h"
108 #include "gromacs/topology/topology.h"
109 #include "gromacs/utility/basedefinitions.h"
110 #include "gromacs/utility/cstringutil.h"
111 #include "gromacs/utility/fatalerror.h"
112 #include "gromacs/utility/real.h"
113 #include "gromacs/utility/smalloc.h"
120 #include "corewrap.h"
123 static void reset_all_counters(FILE *fplog, t_commrec *cr,
125 gmx_int64_t *step_rel, t_inputrec *ir,
126 gmx_wallcycle_t wcycle, t_nrnb *nrnb,
127 gmx_walltime_accounting_t walltime_accounting,
128 struct nonbonded_verlet_t *nbv)
130 char sbuf[STEPSTRSIZE];
132 /* Reset all the counters related to performance over the run */
133 md_print_warn(cr, fplog, "step %s: resetting all time and cycle counters\n",
134 gmx_step_str(step, sbuf));
136 nbnxn_gpu_reset_timings(nbv);
138 wallcycle_stop(wcycle, ewcRUN);
139 wallcycle_reset_all(wcycle);
140 if (DOMAINDECOMP(cr))
142 reset_dd_statistics_counters(cr->dd);
145 ir->init_step += *step_rel;
146 ir->nsteps -= *step_rel;
148 wallcycle_start(wcycle, ewcRUN);
149 walltime_accounting_start(walltime_accounting);
150 print_date_and_time(fplog, cr->nodeid, "Restarted time", gmx_gettime());
153 double do_md(FILE *fplog, t_commrec *cr, int nfile, const t_filenm fnm[],
154 const output_env_t oenv, gmx_bool bVerbose, gmx_bool bCompact,
156 gmx_vsite_t *vsite, gmx_constr_t constr,
157 int stepout, t_inputrec *ir,
158 gmx_mtop_t *top_global,
160 t_state *state_global,
162 t_nrnb *nrnb, gmx_wallcycle_t wcycle,
163 gmx_edsam_t ed, t_forcerec *fr,
164 int repl_ex_nst, int repl_ex_nex, int repl_ex_seed, gmx_membed_t membed,
165 real cpt_period, real max_hours,
168 gmx_walltime_accounting_t walltime_accounting)
170 gmx_mdoutf_t outf = NULL;
171 gmx_int64_t step, step_rel;
173 double t, t0, lam0[efptNR];
174 gmx_bool bGStatEveryStep, bGStat, bCalcVir, bCalcEner;
175 gmx_bool bNS, bNStList, bSimAnn, bStopCM, bRerunMD, bNotLastFrame = FALSE,
176 bFirstStep, bStateFromCP, bStateFromTPX, bInitStep, bLastStep,
177 bBornRadii, bStartingFromCpt;
178 gmx_bool bDoDHDL = FALSE, bDoFEP = FALSE, bDoExpanded = FALSE;
179 gmx_bool do_ene, do_log, do_verbose, bRerunWarnNoV = TRUE,
180 bForceUpdate = FALSE, bCPT;
181 gmx_bool bMasterState;
182 int force_flags, cglo_flags;
183 tensor force_vir, shake_vir, total_vir, tmp_vir, pres;
190 gmx_repl_ex_t repl_ex = NULL;
193 t_mdebin *mdebin = NULL;
194 t_state *state = NULL;
195 rvec *f_global = NULL;
196 gmx_enerdata_t *enerd;
198 gmx_global_stat_t gstat;
199 gmx_update_t upd = NULL;
200 t_graph *graph = NULL;
202 gmx_groups_t *groups;
203 gmx_ekindata_t *ekind;
204 gmx_shellfc_t shellfc;
205 int count, nconverged = 0;
207 gmx_bool bConverged = TRUE, bSumEkinhOld, bDoReplEx, bExchanged, bNeedRepartition;
208 gmx_bool bResetCountersHalfMaxH = FALSE;
209 gmx_bool bVV, bTemp, bPres, bTrotter;
210 gmx_bool bUpdateDoLR;
219 real saved_conserved_quantity = 0;
223 char sbuf[STEPSTRSIZE], sbuf2[STEPSTRSIZE];
224 int handled_stop_condition = gmx_stop_cond_none; /* compare to get_stop_condition*/
225 gmx_int64_t multisim_nsteps = -1; /* number of steps to do before first multisim
226 simulation stops. If equal to zero, don't
227 communicate any more between multisims.*/
228 /* PME load balancing data for GPU kernels */
229 pme_load_balancing_t *pme_loadbal = NULL;
231 gmx_bool bPMETuneTry = FALSE, bPMETuneRunning = FALSE;
234 gmx_bool bIMDstep = FALSE;
237 /* Temporary addition for FAHCORE checkpointing */
241 /* Check for special mdrun options */
242 bRerunMD = (Flags & MD_RERUN);
243 if (Flags & MD_RESETCOUNTERSHALFWAY)
247 /* Signal to reset the counters half the simulation steps. */
248 wcycle_set_reset_counters(wcycle, ir->nsteps/2);
250 /* Signal to reset the counters halfway the simulation time. */
251 bResetCountersHalfMaxH = (max_hours > 0);
254 /* md-vv uses averaged full step velocities for T-control
255 md-vv-avek uses averaged half step velocities for T-control (but full step ekin for P control)
256 md uses averaged half step kinetic energies to determine temperature unless defined otherwise by GMX_EKIN_AVE_VEL; */
258 bTrotter = (bVV && (IR_NPT_TROTTER(ir) || IR_NPH_TROTTER(ir) || IR_NVT_TROTTER(ir)));
262 /* Since we don't know if the frames read are related in any way,
263 * rebuild the neighborlist at every step.
266 ir->nstcalcenergy = 1;
270 check_ir_old_tpx_versions(cr, fplog, ir, top_global);
272 nstglobalcomm = check_nstglobalcomm(fplog, cr, nstglobalcomm, ir);
273 bGStatEveryStep = (nstglobalcomm == 1);
277 ir->nstxout_compressed = 0;
279 groups = &top_global->groups;
282 init_md(fplog, cr, ir, oenv, &t, &t0, state_global->lambda,
283 &(state_global->fep_state), lam0,
284 nrnb, top_global, &upd,
285 nfile, fnm, &outf, &mdebin,
286 force_vir, shake_vir, mu_tot, &bSimAnn, &vcm, Flags, wcycle);
288 clear_mat(total_vir);
290 /* Energy terms and groups */
292 init_enerdata(top_global->groups.grps[egcENER].nr, ir->fepvals->n_lambda,
294 if (DOMAINDECOMP(cr))
300 snew(f, top_global->natoms);
303 /* Kinetic energy data */
305 init_ekindata(fplog, top_global, &(ir->opts), ekind);
306 /* Copy the cos acceleration to the groups struct */
307 ekind->cosacc.cos_accel = ir->cos_accel;
309 gstat = global_stat_init(ir);
312 /* Check for polarizable models and flexible constraints */
313 shellfc = init_shell_flexcon(fplog,
314 top_global, n_flexible_constraints(constr),
315 (ir->bContinuation ||
316 (DOMAINDECOMP(cr) && !MASTER(cr))) ?
317 NULL : state_global->x);
318 if (shellfc && ir->nstcalcenergy != 1)
320 gmx_fatal(FARGS, "You have nstcalcenergy set to a value (%d) that is different from 1.\nThis is not supported in combinations with shell particles.\nPlease make a new tpr file.", ir->nstcalcenergy);
322 if (shellfc && DOMAINDECOMP(cr))
324 gmx_fatal(FARGS, "Shell particles are not implemented with domain decomposition, use a single rank");
326 if (shellfc && ir->eI == eiNM)
328 /* Currently shells don't work with Normal Modes */
329 gmx_fatal(FARGS, "Normal Mode analysis is not supported with shells.\nIf you'd like to help with adding support, we have an open discussion at http://redmine.gromacs.org/issues/879\n");
332 if (vsite && ir->eI == eiNM)
334 /* Currently virtual sites don't work with Normal Modes */
335 gmx_fatal(FARGS, "Normal Mode analysis is not supported with virtual sites.\nIf you'd like to help with adding support, we have an open discussion at http://redmine.gromacs.org/issues/879\n");
338 if (bRerunMD && fr->cutoff_scheme == ecutsVERLET && ir->opts.ngener > 1 && usingGpu(fr->nbv))
340 gmx_fatal(FARGS, "The Verlet scheme on GPUs does not support energy groups, so your rerun should probably use a .tpr file without energy groups, or mdrun -nb auto");
345 tMPI_Thread_mutex_lock(&deform_init_box_mutex);
346 set_deform_reference_box(upd,
347 deform_init_init_step_tpx,
348 deform_init_box_tpx);
349 tMPI_Thread_mutex_unlock(&deform_init_box_mutex);
353 double io = compute_io(ir, top_global->natoms, groups, mdebin->ebin->nener, 1);
354 if ((io > 2000) && MASTER(cr))
357 "\nWARNING: This run will generate roughly %.0f Mb of data\n\n",
362 if (DOMAINDECOMP(cr))
364 top = dd_init_local_top(top_global);
367 dd_init_local_state(cr->dd, state_global, state);
369 if (DDMASTER(cr->dd) && ir->nstfout)
371 snew(f_global, state_global->natoms);
376 top = gmx_mtop_generate_local_top(top_global, ir);
378 forcerec_set_excl_load(fr, top);
380 state = serial_init_local_state(state_global);
383 atoms2md(top_global, ir, 0, NULL, top_global->natoms, mdatoms);
387 set_vsite_top(vsite, top, mdatoms, cr);
390 if (ir->ePBC != epbcNONE && !fr->bMolPBC)
392 graph = mk_graph(fplog, &(top->idef), 0, top_global->natoms, FALSE, FALSE);
397 make_local_shells(cr, mdatoms, shellfc);
400 setup_bonded_threading(fr, &top->idef);
403 /* Set up interactive MD (IMD) */
404 init_IMD(ir, cr, top_global, fplog, ir->nstcalcenergy, state_global->x,
405 nfile, fnm, oenv, imdport, Flags);
407 if (DOMAINDECOMP(cr))
409 /* Distribute the charge groups over the nodes from the master node */
410 dd_partition_system(fplog, ir->init_step, cr, TRUE, 1,
411 state_global, top_global, ir,
412 state, &f, mdatoms, top, fr,
413 vsite, shellfc, constr,
414 nrnb, wcycle, FALSE);
418 update_mdatoms(mdatoms, state->lambda[efptMASS]);
420 if (opt2bSet("-cpi", nfile, fnm))
422 bStateFromCP = gmx_fexist_master(opt2fn_master("-cpi", nfile, fnm, cr), cr);
426 bStateFromCP = FALSE;
431 init_expanded_ensemble(bStateFromCP, ir, &state->dfhist);
438 /* Update mdebin with energy history if appending to output files */
439 if (Flags & MD_APPENDFILES)
441 restore_energyhistory_from_state(mdebin, &state_global->enerhist);
445 /* We might have read an energy history from checkpoint,
446 * free the allocated memory and reset the counts.
448 done_energyhistory(&state_global->enerhist);
449 init_energyhistory(&state_global->enerhist);
452 /* Set the initial energy history in state by updating once */
453 update_energyhistory(&state_global->enerhist, mdebin);
456 /* Initialize constraints */
457 if (constr && !DOMAINDECOMP(cr))
459 set_constraints(constr, top, ir, mdatoms, cr);
462 if (repl_ex_nst > 0 && MASTER(cr))
464 repl_ex = init_replica_exchange(fplog, cr->ms, state_global, ir,
465 repl_ex_nst, repl_ex_nex, repl_ex_seed);
468 /* PME tuning is only supported with GPUs or PME nodes and not with rerun.
469 * PME tuning is not supported with PME only for LJ and not for Coulomb.
471 if ((Flags & MD_TUNEPME) &&
472 EEL_PME(fr->eeltype) &&
473 ( use_GPU(fr->nbv) || !(cr->duty & DUTY_PME)) &&
476 pme_loadbal_init(&pme_loadbal, ir, state->box, fr->ic, fr->pmedata);
478 if (cr->duty & DUTY_PME)
480 /* Start tuning right away, as we can't measure the load */
481 bPMETuneRunning = TRUE;
485 /* Separate PME nodes, we can measure the PP/PME load balance */
490 if (!ir->bContinuation && !bRerunMD)
492 if (mdatoms->cFREEZE && (state->flags & (1<<estV)))
494 /* Set the velocities of frozen particles to zero */
495 for (i = 0; i < mdatoms->homenr; i++)
497 for (m = 0; m < DIM; m++)
499 if (ir->opts.nFreeze[mdatoms->cFREEZE[i]][m])
509 /* Constrain the initial coordinates and velocities */
510 do_constrain_first(fplog, constr, ir, mdatoms, state,
515 /* Construct the virtual sites for the initial configuration */
516 construct_vsites(vsite, state->x, ir->delta_t, NULL,
517 top->idef.iparams, top->idef.il,
518 fr->ePBC, fr->bMolPBC, cr, state->box);
524 /* set free energy calculation frequency as the minimum
525 greatest common denominator of nstdhdl, nstexpanded, and repl_ex_nst*/
526 nstfep = ir->fepvals->nstdhdl;
529 nstfep = gmx_greatest_common_divisor(ir->fepvals->nstdhdl, nstfep);
533 nstfep = gmx_greatest_common_divisor(repl_ex_nst, nstfep);
536 /* I'm assuming we need global communication the first time! MRS */
537 cglo_flags = (CGLO_TEMPERATURE | CGLO_GSTAT
538 | ((ir->comm_mode != ecmNO) ? CGLO_STOPCM : 0)
539 | (bVV ? CGLO_PRESSURE : 0)
540 | (bVV ? CGLO_CONSTRAINT : 0)
541 | (bRerunMD ? CGLO_RERUNMD : 0)
542 | ((Flags & MD_READ_EKIN) ? CGLO_READEKIN : 0));
544 bSumEkinhOld = FALSE;
545 compute_globals(fplog, gstat, cr, ir, fr, ekind, state, state_global, mdatoms, nrnb, vcm,
546 NULL, enerd, force_vir, shake_vir, total_vir, pres, mu_tot,
547 constr, NULL, FALSE, state->box,
548 top_global, &bSumEkinhOld, cglo_flags);
549 if (ir->eI == eiVVAK)
551 /* a second call to get the half step temperature initialized as well */
552 /* we do the same call as above, but turn the pressure off -- internally to
553 compute_globals, this is recognized as a velocity verlet half-step
554 kinetic energy calculation. This minimized excess variables, but
555 perhaps loses some logic?*/
557 compute_globals(fplog, gstat, cr, ir, fr, ekind, state, state_global, mdatoms, nrnb, vcm,
558 NULL, enerd, force_vir, shake_vir, total_vir, pres, mu_tot,
559 constr, NULL, FALSE, state->box,
560 top_global, &bSumEkinhOld,
561 cglo_flags &~(CGLO_STOPCM | CGLO_PRESSURE));
564 /* Calculate the initial half step temperature, and save the ekinh_old */
565 if (!(Flags & MD_STARTFROMCPT))
567 for (i = 0; (i < ir->opts.ngtc); i++)
569 copy_mat(ekind->tcstat[i].ekinh, ekind->tcstat[i].ekinh_old);
574 enerd->term[F_TEMP] *= 2; /* result of averages being done over previous and current step,
575 and there is no previous step */
578 /* need to make an initiation call to get the Trotter variables set, as well as other constants for non-trotter
579 temperature control */
580 trotter_seq = init_npt_vars(ir, state, &MassQ, bTrotter);
584 if (constr && !ir->bContinuation && ir->eConstrAlg == econtLINCS)
587 "RMS relative constraint deviation after constraining: %.2e\n",
588 constr_rmsd(constr, FALSE));
590 if (EI_STATE_VELOCITY(ir->eI))
592 fprintf(fplog, "Initial temperature: %g K\n", enerd->term[F_TEMP]);
596 fprintf(stderr, "starting md rerun '%s', reading coordinates from"
597 " input trajectory '%s'\n\n",
598 *(top_global->name), opt2fn("-rerun", nfile, fnm));
601 fprintf(stderr, "Calculated time to finish depends on nsteps from "
602 "run input file,\nwhich may not correspond to the time "
603 "needed to process input trajectory.\n\n");
609 fprintf(stderr, "starting mdrun '%s'\n",
610 *(top_global->name));
613 sprintf(tbuf, "%8.1f", (ir->init_step+ir->nsteps)*ir->delta_t);
617 sprintf(tbuf, "%s", "infinite");
619 if (ir->init_step > 0)
621 fprintf(stderr, "%s steps, %s ps (continuing from step %s, %8.1f ps).\n",
622 gmx_step_str(ir->init_step+ir->nsteps, sbuf), tbuf,
623 gmx_step_str(ir->init_step, sbuf2),
624 ir->init_step*ir->delta_t);
628 fprintf(stderr, "%s steps, %s ps.\n",
629 gmx_step_str(ir->nsteps, sbuf), tbuf);
632 fprintf(fplog, "\n");
635 walltime_accounting_start(walltime_accounting);
636 wallcycle_start(wcycle, ewcRUN);
637 print_start(fplog, cr, walltime_accounting, "mdrun");
639 /* safest point to do file checkpointing is here. More general point would be immediately before integrator call */
641 chkpt_ret = fcCheckPointParallel( cr->nodeid,
645 gmx_fatal( 3, __FILE__, __LINE__, "Checkpoint error on step %d\n", 0 );
650 /***********************************************************
654 ************************************************************/
656 /* if rerunMD then read coordinates and velocities from input trajectory */
659 if (getenv("GMX_FORCE_UPDATE"))
667 bNotLastFrame = read_first_frame(oenv, &status,
668 opt2fn("-rerun", nfile, fnm),
669 &rerun_fr, TRX_NEED_X | TRX_READ_V);
670 if (rerun_fr.natoms != top_global->natoms)
673 "Number of atoms in trajectory (%d) does not match the "
674 "run input file (%d)\n",
675 rerun_fr.natoms, top_global->natoms);
677 if (ir->ePBC != epbcNONE)
681 gmx_fatal(FARGS, "Rerun trajectory frame step %d time %f does not contain a box, while pbc is used", rerun_fr.step, rerun_fr.time);
683 if (max_cutoff2(ir->ePBC, rerun_fr.box) < sqr(fr->rlistlong))
685 gmx_fatal(FARGS, "Rerun trajectory frame step %d time %f has too small box dimensions", rerun_fr.step, rerun_fr.time);
692 rerun_parallel_comm(cr, &rerun_fr, &bNotLastFrame);
695 if (ir->ePBC != epbcNONE)
697 /* Set the shift vectors.
698 * Necessary here when have a static box different from the tpr box.
700 calc_shifts(rerun_fr.box, fr->shift_vec);
704 /* loop over MD steps or if rerunMD to end of input trajectory */
706 /* Skip the first Nose-Hoover integration when we get the state from tpx */
707 bStateFromTPX = !bStateFromCP;
708 bInitStep = bFirstStep && (bStateFromTPX || bVV);
709 bStartingFromCpt = (Flags & MD_STARTFROMCPT) && bInitStep;
710 bSumEkinhOld = FALSE;
712 bNeedRepartition = FALSE;
714 init_global_signals(&gs, cr, ir, repl_ex_nst);
716 step = ir->init_step;
719 if (MULTISIM(cr) && (repl_ex_nst <= 0 ))
721 /* check how many steps are left in other sims */
722 multisim_nsteps = get_multisim_nsteps(cr, ir->nsteps);
726 /* and stop now if we should */
727 bLastStep = (bRerunMD || (ir->nsteps >= 0 && step_rel > ir->nsteps) ||
728 ((multisim_nsteps >= 0) && (step_rel >= multisim_nsteps )));
729 while (!bLastStep || (bRerunMD && bNotLastFrame))
732 wallcycle_start(wcycle, ewcSTEP);
738 step = rerun_fr.step;
739 step_rel = step - ir->init_step;
752 bLastStep = (step_rel == ir->nsteps);
753 t = t0 + step*ir->delta_t;
756 if (ir->efep != efepNO || ir->bSimTemp)
758 /* find and set the current lambdas. If rerunning, we either read in a state, or a lambda value,
759 requiring different logic. */
761 set_current_lambdas(step, ir->fepvals, bRerunMD, &rerun_fr, state_global, state, lam0);
762 bDoDHDL = do_per_step(step, ir->fepvals->nstdhdl);
763 bDoFEP = (do_per_step(step, nstfep) && (ir->efep != efepNO));
764 bDoExpanded = (do_per_step(step, ir->expandedvals->nstexpanded)
765 && (ir->bExpanded) && (step > 0) && (!bStartingFromCpt));
768 bDoReplEx = ((repl_ex_nst > 0) && (step > 0) && !bLastStep &&
769 do_per_step(step, repl_ex_nst));
773 update_annealing_target_temp(&(ir->opts), t);
778 if (!DOMAINDECOMP(cr) || MASTER(cr))
780 for (i = 0; i < state_global->natoms; i++)
782 copy_rvec(rerun_fr.x[i], state_global->x[i]);
786 for (i = 0; i < state_global->natoms; i++)
788 copy_rvec(rerun_fr.v[i], state_global->v[i]);
793 for (i = 0; i < state_global->natoms; i++)
795 clear_rvec(state_global->v[i]);
799 fprintf(stderr, "\nWARNING: Some frames do not contain velocities.\n"
800 " Ekin, temperature and pressure are incorrect,\n"
801 " the virial will be incorrect when constraints are present.\n"
803 bRerunWarnNoV = FALSE;
807 copy_mat(rerun_fr.box, state_global->box);
808 copy_mat(state_global->box, state->box);
810 if (vsite && (Flags & MD_RERUN_VSITE))
812 if (DOMAINDECOMP(cr))
814 gmx_fatal(FARGS, "Vsite recalculation with -rerun is not implemented with domain decomposition, use a single rank");
818 /* Following is necessary because the graph may get out of sync
819 * with the coordinates if we only have every N'th coordinate set
821 mk_mshift(fplog, graph, fr->ePBC, state->box, state->x);
822 shift_self(graph, state->box, state->x);
824 construct_vsites(vsite, state->x, ir->delta_t, state->v,
825 top->idef.iparams, top->idef.il,
826 fr->ePBC, fr->bMolPBC, cr, state->box);
829 unshift_self(graph, state->box, state->x);
834 /* Stop Center of Mass motion */
835 bStopCM = (ir->comm_mode != ecmNO && do_per_step(step, ir->nstcomm));
839 /* for rerun MD always do Neighbour Searching */
840 bNS = (bFirstStep || ir->nstlist != 0);
845 /* Determine whether or not to do Neighbour Searching and LR */
846 bNStList = (ir->nstlist > 0 && step % ir->nstlist == 0);
848 bNS = (bFirstStep || bExchanged || bNeedRepartition || bNStList || bDoFEP);
851 /* check whether we should stop because another simulation has
855 if ( (multisim_nsteps >= 0) && (step_rel >= multisim_nsteps) &&
856 (multisim_nsteps != ir->nsteps) )
863 "Stopping simulation %d because another one has finished\n",
867 gs.sig[eglsCHKPT] = 1;
872 /* < 0 means stop at next step, > 0 means stop at next NS step */
873 if ( (gs.set[eglsSTOPCOND] < 0) ||
874 ( (gs.set[eglsSTOPCOND] > 0) && (bNStList || ir->nstlist == 0) ) )
879 /* Determine whether or not to update the Born radii if doing GB */
880 bBornRadii = bFirstStep;
881 if (ir->implicit_solvent && (step % ir->nstgbradii == 0))
886 do_log = do_per_step(step, ir->nstlog) || bFirstStep || bLastStep;
887 do_verbose = bVerbose &&
888 (step % stepout == 0 || bFirstStep || bLastStep);
890 if (bNS && !(bFirstStep && ir->bContinuation && !bRerunMD))
898 bMasterState = FALSE;
899 /* Correct the new box if it is too skewed */
900 if (DYNAMIC_BOX(*ir))
902 if (correct_box(fplog, step, state->box, graph))
907 if (DOMAINDECOMP(cr) && bMasterState)
909 dd_collect_state(cr->dd, state, state_global);
913 if (DOMAINDECOMP(cr))
915 /* Repartition the domain decomposition */
916 wallcycle_start(wcycle, ewcDOMDEC);
917 dd_partition_system(fplog, step, cr,
918 bMasterState, nstglobalcomm,
919 state_global, top_global, ir,
920 state, &f, mdatoms, top, fr,
921 vsite, shellfc, constr,
923 do_verbose && !bPMETuneRunning);
924 wallcycle_stop(wcycle, ewcDOMDEC);
928 if (MASTER(cr) && do_log)
930 print_ebin_header(fplog, step, t, state->lambda[efptFEP]); /* can we improve the information printed here? */
933 if (ir->efep != efepNO)
935 update_mdatoms(mdatoms, state->lambda[efptMASS]);
938 if ((bRerunMD && rerun_fr.bV) || bExchanged)
941 /* We need the kinetic energy at minus the half step for determining
942 * the full step kinetic energy and possibly for T-coupling.*/
943 /* This may not be quite working correctly yet . . . . */
944 compute_globals(fplog, gstat, cr, ir, fr, ekind, state, state_global, mdatoms, nrnb, vcm,
945 wcycle, enerd, NULL, NULL, NULL, NULL, mu_tot,
946 constr, NULL, FALSE, state->box,
947 top_global, &bSumEkinhOld,
948 CGLO_RERUNMD | CGLO_GSTAT | CGLO_TEMPERATURE);
950 clear_mat(force_vir);
952 /* We write a checkpoint at this MD step when:
953 * either at an NS step when we signalled through gs,
954 * or at the last step (but not when we do not want confout),
955 * but never at the first step or with rerun.
957 bCPT = (((gs.set[eglsCHKPT] && (bNS || ir->nstlist == 0)) ||
958 (bLastStep && (Flags & MD_CONFOUT))) &&
959 step > ir->init_step && !bRerunMD);
962 gs.set[eglsCHKPT] = 0;
965 /* Determine the energy and pressure:
966 * at nstcalcenergy steps and at energy output steps (set below).
968 if (EI_VV(ir->eI) && (!bInitStep))
970 /* for vv, the first half of the integration actually corresponds
971 to the previous step. bCalcEner is only required to be evaluated on the 'next' step,
972 but the virial needs to be calculated on both the current step and the 'next' step. Future
973 reorganization may be able to get rid of one of the bCalcVir=TRUE steps. */
975 bCalcEner = do_per_step(step-1, ir->nstcalcenergy);
976 bCalcVir = bCalcEner ||
977 (ir->epc != epcNO && (do_per_step(step, ir->nstpcouple) || do_per_step(step-1, ir->nstpcouple)));
981 bCalcEner = do_per_step(step, ir->nstcalcenergy);
982 bCalcVir = bCalcEner ||
983 (ir->epc != epcNO && do_per_step(step, ir->nstpcouple));
986 /* Do we need global communication ? */
987 bGStat = (bCalcVir || bCalcEner || bStopCM ||
988 do_per_step(step, nstglobalcomm) ||
989 (bVV && IR_NVT_TROTTER(ir) && do_per_step(step-1, nstglobalcomm)));
991 do_ene = (do_per_step(step, ir->nstenergy) || bLastStep);
993 if (do_ene || do_log || bDoReplEx)
1000 /* these CGLO_ options remain the same throughout the iteration */
1001 cglo_flags = ((bRerunMD ? CGLO_RERUNMD : 0) |
1002 (bGStat ? CGLO_GSTAT : 0)
1005 force_flags = (GMX_FORCE_STATECHANGED |
1006 ((DYNAMIC_BOX(*ir) || bRerunMD) ? GMX_FORCE_DYNAMICBOX : 0) |
1007 GMX_FORCE_ALLFORCES |
1009 (bCalcVir ? GMX_FORCE_VIRIAL : 0) |
1010 (bCalcEner ? GMX_FORCE_ENERGY : 0) |
1011 (bDoFEP ? GMX_FORCE_DHDL : 0)
1016 if (do_per_step(step, ir->nstcalclr))
1018 force_flags |= GMX_FORCE_DO_LR;
1024 /* Now is the time to relax the shells */
1025 count = relax_shell_flexcon(fplog, cr, bVerbose, step,
1026 ir, bNS, force_flags,
1029 state, f, force_vir, mdatoms,
1030 nrnb, wcycle, graph, groups,
1031 shellfc, fr, bBornRadii, t, mu_tot,
1033 mdoutf_get_fp_field(outf));
1043 /* The coordinates (x) are shifted (to get whole molecules)
1045 * This is parallellized as well, and does communication too.
1046 * Check comments in sim_util.c
1048 do_force(fplog, cr, ir, step, nrnb, wcycle, top, groups,
1049 state->box, state->x, &state->hist,
1050 f, force_vir, mdatoms, enerd, fcd,
1051 state->lambda, graph,
1052 fr, vsite, mu_tot, t, mdoutf_get_fp_field(outf), ed, bBornRadii,
1053 (bNS ? GMX_FORCE_NS : 0) | force_flags);
1056 if (bVV && !bStartingFromCpt && !bRerunMD)
1057 /* ############### START FIRST UPDATE HALF-STEP FOR VV METHODS############### */
1061 wallcycle_start(wcycle, ewcUPDATE);
1062 if (ir->eI == eiVV && bInitStep)
1064 /* if using velocity verlet with full time step Ekin,
1065 * take the first half step only to compute the
1066 * virial for the first step. From there,
1067 * revert back to the initial coordinates
1068 * so that the input is actually the initial step.
1070 snew(vbuf, state->natoms);
1071 copy_rvecn(state->v, vbuf, 0, state->natoms); /* should make this better for parallelizing? */
1075 /* this is for NHC in the Ekin(t+dt/2) version of vv */
1076 trotter_update(ir, step, ekind, enerd, state, total_vir, mdatoms, &MassQ, trotter_seq, ettTSEQ1);
1079 /* If we are using twin-range interactions where the long-range component
1080 * is only evaluated every nstcalclr>1 steps, we should do a special update
1081 * step to combine the long-range forces on these steps.
1082 * For nstcalclr=1 this is not done, since the forces would have been added
1083 * directly to the short-range forces already.
1085 * TODO Remove various aspects of VV+twin-range in master
1086 * branch, because VV integrators did not ever support
1087 * twin-range multiple time stepping with constraints.
1089 bUpdateDoLR = (fr->bTwinRange && do_per_step(step, ir->nstcalclr));
1091 update_coords(fplog, step, ir, mdatoms, state, fr->bMolPBC,
1092 f, bUpdateDoLR, fr->f_twin, bCalcVir ? &fr->vir_twin_constr : NULL, fcd,
1093 ekind, M, upd, bInitStep, etrtVELOCITY1,
1094 cr, nrnb, constr, &top->idef);
1096 if (!bRerunMD || rerun_fr.bV || bForceUpdate) /* Why is rerun_fr.bV here? Unclear. */
1098 wallcycle_stop(wcycle, ewcUPDATE);
1099 update_constraints(fplog, step, NULL, ir, mdatoms,
1100 state, fr->bMolPBC, graph, f,
1101 &top->idef, shake_vir,
1102 cr, nrnb, wcycle, upd, constr,
1104 wallcycle_start(wcycle, ewcUPDATE);
1105 if (bCalcVir && bUpdateDoLR && ir->nstcalclr > 1)
1107 /* Correct the virial for multiple time stepping */
1108 m_sub(shake_vir, fr->vir_twin_constr, shake_vir);
1113 /* Need to unshift here if a do_force has been
1114 called in the previous step */
1115 unshift_self(graph, state->box, state->x);
1117 /* if VV, compute the pressure and constraints */
1118 /* For VV2, we strictly only need this if using pressure
1119 * control, but we really would like to have accurate pressures
1121 * Think about ways around this in the future?
1122 * For now, keep this choice in comments.
1124 /*bPres = (ir->eI==eiVV || IR_NPT_TROTTER(ir)); */
1125 /*bTemp = ((ir->eI==eiVV &&(!bInitStep)) || (ir->eI==eiVVAK && IR_NPT_TROTTER(ir)));*/
1127 bTemp = ((ir->eI == eiVV && (!bInitStep)) || (ir->eI == eiVVAK));
1128 if (bCalcEner && ir->eI == eiVVAK)
1130 bSumEkinhOld = TRUE;
1132 /* for vv, the first half of the integration actually corresponds to the previous step.
1133 So we need information from the last step in the first half of the integration */
1134 if (bGStat || do_per_step(step-1, nstglobalcomm))
1136 wallcycle_stop(wcycle, ewcUPDATE);
1137 compute_globals(fplog, gstat, cr, ir, fr, ekind, state, state_global, mdatoms, nrnb, vcm,
1138 wcycle, enerd, force_vir, shake_vir, total_vir, pres, mu_tot,
1139 constr, NULL, FALSE, state->box,
1140 top_global, &bSumEkinhOld,
1143 | (bTemp ? CGLO_TEMPERATURE : 0)
1144 | (bPres ? CGLO_PRESSURE : 0)
1145 | (bPres ? CGLO_CONSTRAINT : 0)
1148 /* explanation of above:
1149 a) We compute Ekin at the full time step
1150 if 1) we are using the AveVel Ekin, and it's not the
1151 initial step, or 2) if we are using AveEkin, but need the full
1152 time step kinetic energy for the pressure (always true now, since we want accurate statistics).
1153 b) If we are using EkinAveEkin for the kinetic energy for the temperature control, we still feed in
1154 EkinAveVel because it's needed for the pressure */
1155 wallcycle_start(wcycle, ewcUPDATE);
1157 /* temperature scaling and pressure scaling to produce the extended variables at t+dt */
1162 m_add(force_vir, shake_vir, total_vir); /* we need the un-dispersion corrected total vir here */
1163 trotter_update(ir, step, ekind, enerd, state, total_vir, mdatoms, &MassQ, trotter_seq, ettTSEQ2);
1169 wallcycle_stop(wcycle, ewcUPDATE);
1170 /* We need the kinetic energy at minus the half step for determining
1171 * the full step kinetic energy and possibly for T-coupling.*/
1172 /* This may not be quite working correctly yet . . . . */
1173 compute_globals(fplog, gstat, cr, ir, fr, ekind, state, state_global, mdatoms, nrnb, vcm,
1174 wcycle, enerd, NULL, NULL, NULL, NULL, mu_tot,
1175 constr, NULL, FALSE, state->box,
1176 top_global, &bSumEkinhOld,
1177 CGLO_RERUNMD | CGLO_GSTAT | CGLO_TEMPERATURE);
1178 wallcycle_start(wcycle, ewcUPDATE);
1182 if (bTrotter && !bInitStep)
1184 copy_mat(shake_vir, state->svir_prev);
1185 copy_mat(force_vir, state->fvir_prev);
1186 if (IR_NVT_TROTTER(ir) && ir->eI == eiVV)
1188 /* update temperature and kinetic energy now that step is over - this is the v(t+dt) point */
1189 enerd->term[F_TEMP] = sum_ekin(&(ir->opts), ekind, NULL, (ir->eI == eiVV), FALSE);
1190 enerd->term[F_EKIN] = trace(ekind->ekin);
1193 /* if it's the initial step, we performed this first step just to get the constraint virial */
1194 if (ir->eI == eiVV && bInitStep)
1196 copy_rvecn(vbuf, state->v, 0, state->natoms);
1199 wallcycle_stop(wcycle, ewcUPDATE);
1202 /* compute the conserved quantity */
1205 saved_conserved_quantity = compute_conserved_from_auxiliary(ir, state, &MassQ);
1208 last_ekin = enerd->term[F_EKIN];
1210 if ((ir->eDispCorr != edispcEnerPres) && (ir->eDispCorr != edispcAllEnerPres))
1212 saved_conserved_quantity -= enerd->term[F_DISPCORR];
1214 /* sum up the foreign energy and dhdl terms for vv. currently done every step so that dhdl is correct in the .edr */
1215 if (ir->efep != efepNO && !bRerunMD)
1217 sum_dhdl(enerd, state->lambda, ir->fepvals);
1221 /* ######## END FIRST UPDATE STEP ############## */
1222 /* ######## If doing VV, we now have v(dt) ###### */
1225 /* perform extended ensemble sampling in lambda - we don't
1226 actually move to the new state before outputting
1227 statistics, but if performing simulated tempering, we
1228 do update the velocities and the tau_t. */
1230 lamnew = ExpandedEnsembleDynamics(fplog, ir, enerd, state, &MassQ, state->fep_state, &state->dfhist, step, state->v, mdatoms);
1231 /* history is maintained in state->dfhist, but state_global is what is sent to trajectory and log output */
1232 copy_df_history(&state_global->dfhist, &state->dfhist);
1235 /* Now we have the energies and forces corresponding to the
1236 * coordinates at time t. We must output all of this before
1239 do_md_trajectory_writing(fplog, cr, nfile, fnm, step, step_rel, t,
1240 ir, state, state_global, top_global, fr,
1241 outf, mdebin, ekind, f, f_global,
1243 bCPT, bRerunMD, bLastStep, (Flags & MD_CONFOUT),
1245 /* Check if IMD step and do IMD communication, if bIMD is TRUE. */
1246 bIMDstep = do_IMD(ir->bIMD, step, cr, bNS, state->box, state->x, ir, t, wcycle);
1248 /* kludge -- virial is lost with restart for NPT control. Must restart */
1249 if (bStartingFromCpt && bVV)
1251 copy_mat(state->svir_prev, shake_vir);
1252 copy_mat(state->fvir_prev, force_vir);
1255 elapsed_time = walltime_accounting_get_current_elapsed_time(walltime_accounting);
1257 /* Check whether everything is still allright */
1258 if (((int)gmx_get_stop_condition() > handled_stop_condition)
1259 #ifdef GMX_THREAD_MPI
1264 /* this is just make gs.sig compatible with the hack
1265 of sending signals around by MPI_Reduce with together with
1267 if (gmx_get_stop_condition() == gmx_stop_cond_next_ns)
1269 gs.sig[eglsSTOPCOND] = 1;
1271 if (gmx_get_stop_condition() == gmx_stop_cond_next)
1273 gs.sig[eglsSTOPCOND] = -1;
1275 /* < 0 means stop at next step, > 0 means stop at next NS step */
1279 "\n\nReceived the %s signal, stopping at the next %sstep\n\n",
1280 gmx_get_signal_name(),
1281 gs.sig[eglsSTOPCOND] == 1 ? "NS " : "");
1285 "\n\nReceived the %s signal, stopping at the next %sstep\n\n",
1286 gmx_get_signal_name(),
1287 gs.sig[eglsSTOPCOND] == 1 ? "NS " : "");
1289 handled_stop_condition = (int)gmx_get_stop_condition();
1291 else if (MASTER(cr) && (bNS || ir->nstlist <= 0) &&
1292 (max_hours > 0 && elapsed_time > max_hours*60.0*60.0*0.99) &&
1293 gs.sig[eglsSTOPCOND] == 0 && gs.set[eglsSTOPCOND] == 0)
1295 /* Signal to terminate the run */
1296 gs.sig[eglsSTOPCOND] = 1;
1299 fprintf(fplog, "\nStep %s: Run time exceeded %.3f hours, will terminate the run\n", gmx_step_str(step, sbuf), max_hours*0.99);
1301 fprintf(stderr, "\nStep %s: Run time exceeded %.3f hours, will terminate the run\n", gmx_step_str(step, sbuf), max_hours*0.99);
1304 if (bResetCountersHalfMaxH && MASTER(cr) &&
1305 elapsed_time > max_hours*60.0*60.0*0.495)
1307 gs.sig[eglsRESETCOUNTERS] = 1;
1310 /* In parallel we only have to check for checkpointing in steps
1311 * where we do global communication,
1312 * otherwise the other nodes don't know.
1314 if (MASTER(cr) && ((bGStat || !PAR(cr)) &&
1317 elapsed_time >= nchkpt*cpt_period*60.0)) &&
1318 gs.set[eglsCHKPT] == 0)
1320 gs.sig[eglsCHKPT] = 1;
1323 /* at the start of step, randomize or scale the velocities (trotter done elsewhere) */
1328 update_tcouple(step, ir, state, ekind, &MassQ, mdatoms);
1330 if (ETC_ANDERSEN(ir->etc)) /* keep this outside of update_tcouple because of the extra info required to pass */
1332 gmx_bool bIfRandomize;
1333 bIfRandomize = update_randomize_velocities(ir, step, cr, mdatoms, state, upd, constr);
1334 /* if we have constraints, we have to remove the kinetic energy parallel to the bonds */
1335 if (constr && bIfRandomize)
1337 update_constraints(fplog, step, NULL, ir, mdatoms,
1338 state, fr->bMolPBC, graph, f,
1339 &top->idef, tmp_vir,
1340 cr, nrnb, wcycle, upd, constr,
1345 /* ######### START SECOND UPDATE STEP ################# */
1346 /* Box is changed in update() when we do pressure coupling,
1347 * but we should still use the old box for energy corrections and when
1348 * writing it to the energy file, so it matches the trajectory files for
1349 * the same timestep above. Make a copy in a separate array.
1351 copy_mat(state->box, lastbox);
1355 if (!bRerunMD || rerun_fr.bV || bForceUpdate)
1357 wallcycle_start(wcycle, ewcUPDATE);
1358 /* UPDATE PRESSURE VARIABLES IN TROTTER FORMULATION WITH CONSTRAINTS */
1361 trotter_update(ir, step, ekind, enerd, state, total_vir, mdatoms, &MassQ, trotter_seq, ettTSEQ3);
1362 /* We can only do Berendsen coupling after we have summed
1363 * the kinetic energy or virial. Since the happens
1364 * in global_state after update, we should only do it at
1365 * step % nstlist = 1 with bGStatEveryStep=FALSE.
1370 update_tcouple(step, ir, state, ekind, &MassQ, mdatoms);
1371 update_pcouple(fplog, step, ir, state, pcoupl_mu, M, bInitStep);
1376 bUpdateDoLR = (fr->bTwinRange && do_per_step(step, ir->nstcalclr));
1378 /* velocity half-step update */
1379 update_coords(fplog, step, ir, mdatoms, state, fr->bMolPBC, f,
1380 bUpdateDoLR, fr->f_twin, bCalcVir ? &fr->vir_twin_constr : NULL, fcd,
1381 ekind, M, upd, FALSE, etrtVELOCITY2,
1382 cr, nrnb, constr, &top->idef);
1385 /* Above, initialize just copies ekinh into ekin,
1386 * it doesn't copy position (for VV),
1387 * and entire integrator for MD.
1390 if (ir->eI == eiVVAK)
1392 /* We probably only need md->homenr, not state->natoms */
1393 if (state->natoms > cbuf_nalloc)
1395 cbuf_nalloc = state->natoms;
1396 srenew(cbuf, cbuf_nalloc);
1398 copy_rvecn(state->x, cbuf, 0, state->natoms);
1400 bUpdateDoLR = (fr->bTwinRange && do_per_step(step, ir->nstcalclr));
1402 update_coords(fplog, step, ir, mdatoms, state, fr->bMolPBC, f,
1403 bUpdateDoLR, fr->f_twin, bCalcVir ? &fr->vir_twin_constr : NULL, fcd,
1404 ekind, M, upd, bInitStep, etrtPOSITION, cr, nrnb, constr, &top->idef);
1405 wallcycle_stop(wcycle, ewcUPDATE);
1407 update_constraints(fplog, step, &dvdl_constr, ir, mdatoms, state,
1408 fr->bMolPBC, graph, f,
1409 &top->idef, shake_vir,
1410 cr, nrnb, wcycle, upd, constr,
1413 if (bCalcVir && bUpdateDoLR && ir->nstcalclr > 1)
1415 /* Correct the virial for multiple time stepping */
1416 m_sub(shake_vir, fr->vir_twin_constr, shake_vir);
1419 if (ir->eI == eiVVAK)
1421 /* erase F_EKIN and F_TEMP here? */
1422 /* just compute the kinetic energy at the half step to perform a trotter step */
1423 compute_globals(fplog, gstat, cr, ir, fr, ekind, state, state_global, mdatoms, nrnb, vcm,
1424 wcycle, enerd, force_vir, shake_vir, total_vir, pres, mu_tot,
1425 constr, NULL, FALSE, lastbox,
1426 top_global, &bSumEkinhOld,
1427 cglo_flags | CGLO_TEMPERATURE
1429 wallcycle_start(wcycle, ewcUPDATE);
1430 trotter_update(ir, step, ekind, enerd, state, total_vir, mdatoms, &MassQ, trotter_seq, ettTSEQ4);
1431 /* now we know the scaling, we can compute the positions again again */
1432 copy_rvecn(cbuf, state->x, 0, state->natoms);
1434 bUpdateDoLR = (fr->bTwinRange && do_per_step(step, ir->nstcalclr));
1436 update_coords(fplog, step, ir, mdatoms, state, fr->bMolPBC, f,
1437 bUpdateDoLR, fr->f_twin, bCalcVir ? &fr->vir_twin_constr : NULL, fcd,
1438 ekind, M, upd, bInitStep, etrtPOSITION, cr, nrnb, constr, &top->idef);
1439 wallcycle_stop(wcycle, ewcUPDATE);
1441 /* do we need an extra constraint here? just need to copy out of state->v to upd->xp? */
1442 /* are the small terms in the shake_vir here due
1443 * to numerical errors, or are they important
1444 * physically? I'm thinking they are just errors, but not completely sure.
1445 * For now, will call without actually constraining, constr=NULL*/
1446 update_constraints(fplog, step, NULL, ir, mdatoms,
1447 state, fr->bMolPBC, graph, f,
1448 &top->idef, tmp_vir,
1449 cr, nrnb, wcycle, upd, NULL,
1454 /* this factor or 2 correction is necessary
1455 because half of the constraint force is removed
1456 in the vv step, so we have to double it. See
1457 the Redmine issue #1255. It is not yet clear
1458 if the factor of 2 is exact, or just a very
1459 good approximation, and this will be
1460 investigated. The next step is to see if this
1461 can be done adding a dhdl contribution from the
1462 rattle step, but this is somewhat more
1463 complicated with the current code. Will be
1464 investigated, hopefully for 4.6.3. However,
1465 this current solution is much better than
1466 having it completely wrong.
1468 enerd->term[F_DVDL_CONSTR] += 2*dvdl_constr;
1472 enerd->term[F_DVDL_CONSTR] += dvdl_constr;
1477 /* Need to unshift here */
1478 unshift_self(graph, state->box, state->x);
1483 wallcycle_start(wcycle, ewcVSITECONSTR);
1486 shift_self(graph, state->box, state->x);
1488 construct_vsites(vsite, state->x, ir->delta_t, state->v,
1489 top->idef.iparams, top->idef.il,
1490 fr->ePBC, fr->bMolPBC, cr, state->box);
1494 unshift_self(graph, state->box, state->x);
1496 wallcycle_stop(wcycle, ewcVSITECONSTR);
1499 /* ############## IF NOT VV, Calculate globals HERE ############ */
1500 /* With Leap-Frog we can skip compute_globals at
1501 * non-communication steps, but we need to calculate
1502 * the kinetic energy one step before communication.
1504 if (bGStat || (!EI_VV(ir->eI) && do_per_step(step+1, nstglobalcomm)))
1506 compute_globals(fplog, gstat, cr, ir, fr, ekind, state, state_global, mdatoms, nrnb, vcm,
1507 wcycle, enerd, force_vir, shake_vir, total_vir, pres, mu_tot,
1509 (step_rel % gs.nstms == 0) &&
1510 (multisim_nsteps < 0 || (step_rel < multisim_nsteps)),
1512 top_global, &bSumEkinhOld,
1514 | (!EI_VV(ir->eI) || bRerunMD ? CGLO_ENERGY : 0)
1515 | (!EI_VV(ir->eI) && bStopCM ? CGLO_STOPCM : 0)
1516 | (!EI_VV(ir->eI) ? CGLO_TEMPERATURE : 0)
1517 | (!EI_VV(ir->eI) || bRerunMD ? CGLO_PRESSURE : 0)
1522 /* ############# END CALC EKIN AND PRESSURE ################# */
1524 /* Note: this is OK, but there are some numerical precision issues with using the convergence of
1525 the virial that should probably be addressed eventually. state->veta has better properies,
1526 but what we actually need entering the new cycle is the new shake_vir value. Ideally, we could
1527 generate the new shake_vir, but test the veta value for convergence. This will take some thought. */
1529 if (ir->efep != efepNO && (!bVV || bRerunMD))
1531 /* Sum up the foreign energy and dhdl terms for md and sd.
1532 Currently done every step so that dhdl is correct in the .edr */
1533 sum_dhdl(enerd, state->lambda, ir->fepvals);
1535 update_box(fplog, step, ir, mdatoms, state, f,
1536 pcoupl_mu, nrnb, upd);
1538 /* ################# END UPDATE STEP 2 ################# */
1539 /* #### We now have r(t+dt) and v(t+dt/2) ############# */
1541 /* The coordinates (x) were unshifted in update */
1544 /* We will not sum ekinh_old,
1545 * so signal that we still have to do it.
1547 bSumEkinhOld = TRUE;
1550 /* ######### BEGIN PREPARING EDR OUTPUT ########### */
1552 /* use the directly determined last velocity, not actually the averaged half steps */
1553 if (bTrotter && ir->eI == eiVV)
1555 enerd->term[F_EKIN] = last_ekin;
1557 enerd->term[F_ETOT] = enerd->term[F_EPOT] + enerd->term[F_EKIN];
1561 enerd->term[F_ECONSERVED] = enerd->term[F_ETOT] + saved_conserved_quantity;
1565 enerd->term[F_ECONSERVED] = enerd->term[F_ETOT] + compute_conserved_from_auxiliary(ir, state, &MassQ);
1567 /* ######### END PREPARING EDR OUTPUT ########### */
1572 gmx_bool do_dr, do_or;
1574 if (fplog && do_log && bDoExpanded)
1576 /* only needed if doing expanded ensemble */
1577 PrintFreeEnergyInfoToFile(fplog, ir->fepvals, ir->expandedvals, ir->bSimTemp ? ir->simtempvals : NULL,
1578 &state_global->dfhist, state->fep_state, ir->nstlog, step);
1580 if (!(bStartingFromCpt && (EI_VV(ir->eI))))
1584 upd_mdebin(mdebin, bDoDHDL, TRUE,
1585 t, mdatoms->tmass, enerd, state,
1586 ir->fepvals, ir->expandedvals, lastbox,
1587 shake_vir, force_vir, total_vir, pres,
1588 ekind, mu_tot, constr);
1592 upd_mdebin_step(mdebin);
1595 do_dr = do_per_step(step, ir->nstdisreout);
1596 do_or = do_per_step(step, ir->nstorireout);
1598 print_ebin(mdoutf_get_fp_ene(outf), do_ene, do_dr, do_or, do_log ? fplog : NULL,
1600 eprNORMAL, bCompact, mdebin, fcd, groups, &(ir->opts));
1604 pull_print_output(ir->pull, step, t);
1607 if (do_per_step(step, ir->nstlog))
1609 if (fflush(fplog) != 0)
1611 gmx_fatal(FARGS, "Cannot flush logfile - maybe you are out of disk space?");
1617 /* Have to do this part _after_ outputting the logfile and the edr file */
1618 /* Gets written into the state at the beginning of next loop*/
1619 state->fep_state = lamnew;
1621 /* Print the remaining wall clock time for the run */
1622 if (MULTIMASTER(cr) && (do_verbose || gmx_got_usr_signal()) && !bPMETuneRunning)
1626 fprintf(stderr, "\n");
1628 print_time(stderr, walltime_accounting, step, ir, cr);
1631 /* Ion/water position swapping.
1632 * Not done in last step since trajectory writing happens before this call
1633 * in the MD loop and exchanges would be lost anyway. */
1634 bNeedRepartition = FALSE;
1635 if ((ir->eSwapCoords != eswapNO) && (step > 0) && !bLastStep &&
1636 do_per_step(step, ir->swap->nstswap))
1638 bNeedRepartition = do_swapcoords(cr, step, t, ir, wcycle,
1639 bRerunMD ? rerun_fr.x : state->x,
1640 bRerunMD ? rerun_fr.box : state->box,
1641 top_global, MASTER(cr) && bVerbose, bRerunMD);
1643 if (bNeedRepartition && DOMAINDECOMP(cr))
1645 dd_collect_state(cr->dd, state, state_global);
1649 /* Replica exchange */
1653 bExchanged = replica_exchange(fplog, cr, repl_ex,
1654 state_global, enerd,
1658 if ( (bExchanged || bNeedRepartition) && DOMAINDECOMP(cr) )
1660 dd_partition_system(fplog, step, cr, TRUE, 1,
1661 state_global, top_global, ir,
1662 state, &f, mdatoms, top, fr,
1663 vsite, shellfc, constr,
1664 nrnb, wcycle, FALSE);
1669 bStartingFromCpt = FALSE;
1671 /* ####### SET VARIABLES FOR NEXT ITERATION IF THEY STILL NEED IT ###### */
1672 /* With all integrators, except VV, we need to retain the pressure
1673 * at the current step for coupling at the next step.
1675 if ((state->flags & (1<<estPRES_PREV)) &&
1677 (ir->nstpcouple > 0 && step % ir->nstpcouple == 0)))
1679 /* Store the pressure in t_state for pressure coupling
1680 * at the next MD step.
1682 copy_mat(pres, state->pres_prev);
1685 /* ####### END SET VARIABLES FOR NEXT ITERATION ###### */
1687 if ( (membed != NULL) && (!bLastStep) )
1689 rescale_membed(step_rel, membed, state_global->x);
1696 /* read next frame from input trajectory */
1697 bNotLastFrame = read_next_frame(oenv, status, &rerun_fr);
1702 rerun_parallel_comm(cr, &rerun_fr, &bNotLastFrame);
1706 if (!bRerunMD || !rerun_fr.bStep)
1708 /* increase the MD step number */
1713 cycles = wallcycle_stop(wcycle, ewcSTEP);
1714 if (DOMAINDECOMP(cr) && wcycle)
1716 dd_cycles_add(cr->dd, cycles, ddCyclStep);
1719 if (bPMETuneRunning || bPMETuneTry)
1721 /* PME grid + cut-off optimization with GPUs or PME nodes */
1723 /* Count the total cycles over the last steps */
1724 cycles_pmes += cycles;
1726 /* We can only switch cut-off at NS steps */
1727 if (step % ir->nstlist == 0)
1729 /* PME grid + cut-off optimization with GPUs or PME nodes */
1732 if (DDMASTER(cr->dd))
1734 /* PME node load is too high, start tuning */
1735 bPMETuneRunning = (dd_pme_f_ratio(cr->dd) >= 1.05);
1737 dd_bcast(cr->dd, sizeof(gmx_bool), &bPMETuneRunning);
1739 if (bPMETuneRunning &&
1740 use_GPU(fr->nbv) && DOMAINDECOMP(cr) &&
1741 !(cr->duty & DUTY_PME))
1743 /* Lock DLB=auto to off (does nothing when DLB=yes/no).
1744 * With GPUs + separate PME ranks, we don't want DLB.
1745 * This could happen when we scan coarse grids and
1746 * it would then never be turned off again.
1747 * This would hurt performance at the final, optimal
1748 * grid spacing, where DLB almost never helps.
1749 * Also, DLB can limit the cut-off for PME tuning.
1751 dd_dlb_set_lock(cr->dd, TRUE);
1754 if (bPMETuneRunning || step_rel > ir->nstlist*50)
1756 bPMETuneTry = FALSE;
1759 if (bPMETuneRunning)
1761 /* init_step might not be a multiple of nstlist,
1762 * but the first cycle is always skipped anyhow.
1765 pme_load_balance(pme_loadbal, cr,
1766 (bVerbose && MASTER(cr)) ? stderr : NULL,
1768 ir, state, cycles_pmes,
1769 fr->ic, fr->nbv, &fr->pmedata,
1772 /* Update constants in forcerec/inputrec to keep them in sync with fr->ic */
1773 fr->ewaldcoeff_q = fr->ic->ewaldcoeff_q;
1774 fr->ewaldcoeff_lj = fr->ic->ewaldcoeff_lj;
1775 fr->rlist = fr->ic->rlist;
1776 fr->rlistlong = fr->ic->rlistlong;
1777 fr->rcoulomb = fr->ic->rcoulomb;
1778 fr->rvdw = fr->ic->rvdw;
1780 if (ir->eDispCorr != edispcNO)
1782 calc_enervirdiff(NULL, ir->eDispCorr, fr);
1785 if (!bPMETuneRunning &&
1787 dd_dlb_is_locked(cr->dd))
1789 /* Unlock the DLB=auto, DLB is allowed to activate
1790 * (but we don't expect it to activate in most cases).
1792 dd_dlb_set_lock(cr->dd, FALSE);
1799 if (step_rel == wcycle_get_reset_counters(wcycle) ||
1800 gs.set[eglsRESETCOUNTERS] != 0)
1802 /* Reset all the counters related to performance over the run */
1803 reset_all_counters(fplog, cr, step, &step_rel, ir, wcycle, nrnb, walltime_accounting,
1804 use_GPU(fr->nbv) ? fr->nbv : NULL);
1805 wcycle_set_reset_counters(wcycle, -1);
1806 if (!(cr->duty & DUTY_PME))
1808 /* Tell our PME node to reset its counters */
1809 gmx_pme_send_resetcounters(cr, step);
1811 /* Correct max_hours for the elapsed time */
1812 max_hours -= elapsed_time/(60.0*60.0);
1813 bResetCountersHalfMaxH = FALSE;
1814 gs.set[eglsRESETCOUNTERS] = 0;
1817 /* If bIMD is TRUE, the master updates the IMD energy record and sends positions to VMD client */
1818 IMD_prep_energies_send_positions(ir->bIMD && MASTER(cr), bIMDstep, ir->imd, enerd, step, bCalcEner, wcycle);
1821 /* End of main MD loop */
1824 /* Closing TNG files can include compressing data. Therefore it is good to do that
1825 * before stopping the time measurements. */
1826 mdoutf_tng_close(outf);
1828 /* Stop measuring walltime */
1829 walltime_accounting_end(walltime_accounting);
1831 if (bRerunMD && MASTER(cr))
1836 if (!(cr->duty & DUTY_PME))
1838 /* Tell the PME only node to finish */
1839 gmx_pme_send_finish(cr);
1844 if (ir->nstcalcenergy > 0 && !bRerunMD)
1846 print_ebin(mdoutf_get_fp_ene(outf), FALSE, FALSE, FALSE, fplog, step, t,
1847 eprAVER, FALSE, mdebin, fcd, groups, &(ir->opts));
1854 if (pme_loadbal != NULL)
1856 pme_loadbal_done(pme_loadbal, cr, fplog,
1860 if (shellfc && fplog)
1862 fprintf(fplog, "Fraction of iterations that converged: %.2f %%\n",
1863 (nconverged*100.0)/step_rel);
1864 fprintf(fplog, "Average number of force evaluations per MD step: %.2f\n\n",
1868 if (repl_ex_nst > 0 && MASTER(cr))
1870 print_replica_exchange_statistics(fplog, repl_ex);
1873 /* IMD cleanup, if bIMD is TRUE. */
1874 IMD_finalize(ir->bIMD, ir->imd);
1876 walltime_accounting_set_nsteps_done(walltime_accounting, step_rel);