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) 2011-2019,2020,2021, 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.
39 * \brief Implements the integrator for normal molecular dynamics simulations
41 * \author David van der Spoel <david.vanderspoel@icm.uu.se>
42 * \ingroup module_mdrun
55 #include "gromacs/applied_forces/awh/awh.h"
56 #include "gromacs/commandline/filenm.h"
57 #include "gromacs/domdec/collect.h"
58 #include "gromacs/domdec/dlbtiming.h"
59 #include "gromacs/domdec/domdec.h"
60 #include "gromacs/domdec/domdec_network.h"
61 #include "gromacs/domdec/domdec_struct.h"
62 #include "gromacs/domdec/gpuhaloexchange.h"
63 #include "gromacs/domdec/mdsetup.h"
64 #include "gromacs/domdec/partition.h"
65 #include "gromacs/essentialdynamics/edsam.h"
66 #include "gromacs/ewald/pme_load_balancing.h"
67 #include "gromacs/ewald/pme_pp.h"
68 #include "gromacs/fileio/trxio.h"
69 #include "gromacs/gmxlib/network.h"
70 #include "gromacs/gmxlib/nrnb.h"
71 #include "gromacs/gpu_utils/device_stream_manager.h"
72 #include "gromacs/gpu_utils/gpu_utils.h"
73 #include "gromacs/imd/imd.h"
74 #include "gromacs/listed_forces/listed_forces.h"
75 #include "gromacs/math/functions.h"
76 #include "gromacs/math/invertmatrix.h"
77 #include "gromacs/math/vec.h"
78 #include "gromacs/math/vectypes.h"
79 #include "gromacs/mdlib/checkpointhandler.h"
80 #include "gromacs/mdlib/compute_io.h"
81 #include "gromacs/mdlib/constr.h"
82 #include "gromacs/mdlib/coupling.h"
83 #include "gromacs/mdlib/ebin.h"
84 #include "gromacs/mdlib/enerdata_utils.h"
85 #include "gromacs/mdlib/energyoutput.h"
86 #include "gromacs/mdlib/expanded.h"
87 #include "gromacs/mdlib/force.h"
88 #include "gromacs/mdlib/force_flags.h"
89 #include "gromacs/mdlib/forcerec.h"
90 #include "gromacs/mdlib/freeenergyparameters.h"
91 #include "gromacs/mdlib/md_support.h"
92 #include "gromacs/mdlib/mdatoms.h"
93 #include "gromacs/mdlib/mdoutf.h"
94 #include "gromacs/mdlib/membed.h"
95 #include "gromacs/mdlib/resethandler.h"
96 #include "gromacs/mdlib/sighandler.h"
97 #include "gromacs/mdlib/simulationsignal.h"
98 #include "gromacs/mdlib/stat.h"
99 #include "gromacs/mdlib/stophandler.h"
100 #include "gromacs/mdlib/tgroup.h"
101 #include "gromacs/mdlib/trajectory_writing.h"
102 #include "gromacs/mdlib/update.h"
103 #include "gromacs/mdlib/update_constrain_gpu.h"
104 #include "gromacs/mdlib/vcm.h"
105 #include "gromacs/mdlib/vsite.h"
106 #include "gromacs/mdrunutility/handlerestart.h"
107 #include "gromacs/mdrunutility/multisim.h"
108 #include "gromacs/mdrunutility/printtime.h"
109 #include "gromacs/mdtypes/awh_history.h"
110 #include "gromacs/mdtypes/awh_params.h"
111 #include "gromacs/mdtypes/commrec.h"
112 #include "gromacs/mdtypes/df_history.h"
113 #include "gromacs/mdtypes/energyhistory.h"
114 #include "gromacs/mdtypes/fcdata.h"
115 #include "gromacs/mdtypes/forcebuffers.h"
116 #include "gromacs/mdtypes/forcerec.h"
117 #include "gromacs/mdtypes/group.h"
118 #include "gromacs/mdtypes/inputrec.h"
119 #include "gromacs/mdtypes/interaction_const.h"
120 #include "gromacs/mdtypes/md_enums.h"
121 #include "gromacs/mdtypes/mdatom.h"
122 #include "gromacs/mdtypes/mdrunoptions.h"
123 #include "gromacs/mdtypes/multipletimestepping.h"
124 #include "gromacs/mdtypes/observableshistory.h"
125 #include "gromacs/mdtypes/pullhistory.h"
126 #include "gromacs/mdtypes/simulation_workload.h"
127 #include "gromacs/mdtypes/state.h"
128 #include "gromacs/mdtypes/state_propagator_data_gpu.h"
129 #include "gromacs/modularsimulator/energydata.h"
130 #include "gromacs/nbnxm/gpu_data_mgmt.h"
131 #include "gromacs/nbnxm/nbnxm.h"
132 #include "gromacs/pbcutil/pbc.h"
133 #include "gromacs/pulling/output.h"
134 #include "gromacs/pulling/pull.h"
135 #include "gromacs/swap/swapcoords.h"
136 #include "gromacs/timing/wallcycle.h"
137 #include "gromacs/timing/walltime_accounting.h"
138 #include "gromacs/topology/atoms.h"
139 #include "gromacs/topology/idef.h"
140 #include "gromacs/topology/mtop_util.h"
141 #include "gromacs/topology/topology.h"
142 #include "gromacs/trajectory/trajectoryframe.h"
143 #include "gromacs/utility/basedefinitions.h"
144 #include "gromacs/utility/cstringutil.h"
145 #include "gromacs/utility/fatalerror.h"
146 #include "gromacs/utility/logger.h"
147 #include "gromacs/utility/real.h"
148 #include "gromacs/utility/smalloc.h"
150 #include "legacysimulator.h"
151 #include "replicaexchange.h"
154 using gmx::SimulationSignaller;
156 void gmx::LegacySimulator::do_md()
158 // TODO Historically, the EM and MD "integrators" used different
159 // names for the t_inputrec *parameter, but these must have the
160 // same name, now that it's a member of a struct. We use this ir
161 // alias to avoid a large ripple of nearly useless changes.
162 // t_inputrec is being replaced by IMdpOptionsProvider, so this
163 // will go away eventually.
164 t_inputrec* ir = inputrec;
165 int64_t step, step_rel;
166 double t, t0 = ir->init_t;
167 gmx_bool bGStatEveryStep, bGStat, bCalcVir, bCalcEnerStep, bCalcEner;
168 gmx_bool bNS = FALSE, bNStList, bStopCM, bFirstStep, bInitStep, bLastStep = FALSE;
169 gmx_bool bDoDHDL = FALSE, bDoFEP = FALSE, bDoExpanded = FALSE;
170 gmx_bool do_ene, do_log, do_verbose;
171 gmx_bool bMasterState;
172 unsigned int force_flags;
173 tensor force_vir = { { 0 } }, shake_vir = { { 0 } }, total_vir = { { 0 } }, pres = { { 0 } };
176 matrix pressureCouplingMu, M;
177 gmx_repl_ex_t repl_ex = nullptr;
178 gmx_global_stat_t gstat;
179 gmx_shellfc_t* shellfc;
180 gmx_bool bSumEkinhOld, bDoReplEx, bExchanged, bNeedRepartition;
181 gmx_bool bTemp, bPres, bTrotter;
183 std::vector<RVec> cbuf;
189 real saved_conserved_quantity = 0;
192 char sbuf[STEPSTRSIZE], sbuf2[STEPSTRSIZE];
194 /* PME load balancing data for GPU kernels */
195 gmx_bool bPMETune = FALSE;
196 gmx_bool bPMETunePrinting = FALSE;
198 bool bInteractiveMDstep = false;
200 /* Domain decomposition could incorrectly miss a bonded
201 interaction, but checking for that requires a global
202 communication stage, which does not otherwise happen in DD
203 code. So we do that alongside the first global energy reduction
204 after a new DD is made. These variables handle whether the
205 check happens, and the result it returns. */
206 bool shouldCheckNumberOfBondedInteractions = false;
207 int totalNumberOfBondedInteractions = -1;
209 SimulationSignals signals;
210 // Most global communnication stages don't propagate mdrun
211 // signals, and will use this object to achieve that.
212 SimulationSignaller nullSignaller(nullptr, nullptr, nullptr, false, false);
214 if (!mdrunOptions.writeConfout)
216 // This is on by default, and the main known use case for
217 // turning it off is for convenience in benchmarking, which is
218 // something that should not show up in the general user
223 "The -noconfout functionality is deprecated, and may be removed in a "
227 /* md-vv uses averaged full step velocities for T-control
228 md-vv-avek uses averaged half step velocities for T-control (but full step ekin for P control)
229 md uses averaged half step kinetic energies to determine temperature unless defined otherwise by GMX_EKIN_AVE_VEL; */
230 bTrotter = (EI_VV(ir->eI)
231 && (inputrecNptTrotter(ir) || inputrecNphTrotter(ir) || inputrecNvtTrotter(ir)));
233 const bool bRerunMD = false;
235 int nstglobalcomm = computeGlobalCommunicationPeriod(mdlog, ir, cr);
236 bGStatEveryStep = (nstglobalcomm == 1);
238 const SimulationGroups* groups = &top_global->groups;
240 std::unique_ptr<EssentialDynamics> ed = nullptr;
241 if (opt2bSet("-ei", nfile, fnm))
243 /* Initialize essential dynamics sampling */
244 ed = init_edsam(mdlog, opt2fn_null("-ei", nfile, fnm), opt2fn("-eo", nfile, fnm), top_global,
245 ir, cr, constr, state_global, observablesHistory, oenv, startingBehavior);
247 else if (observablesHistory->edsamHistory)
250 "The checkpoint is from a run with essential dynamics sampling, "
251 "but the current run did not specify the -ei option. "
252 "Either specify the -ei option to mdrun, or do not use this checkpoint file.");
255 int* fep_state = MASTER(cr) ? &state_global->fep_state : nullptr;
256 gmx::ArrayRef<real> lambda = MASTER(cr) ? state_global->lambda : gmx::ArrayRef<real>();
257 initialize_lambdas(fplog, *ir, MASTER(cr), fep_state, lambda);
258 Update upd(*ir, deform);
259 const bool doSimulatedAnnealing = initSimulatedAnnealing(ir, &upd);
260 const bool useReplicaExchange = (replExParams.exchangeInterval > 0);
262 const t_fcdata& fcdata = *fr->fcdata;
264 bool simulationsShareState = false;
265 int nstSignalComm = nstglobalcomm;
267 // TODO This implementation of ensemble orientation restraints is nasty because
268 // a user can't just do multi-sim with single-sim orientation restraints.
269 bool usingEnsembleRestraints =
270 (fcdata.disres->nsystems > 1) || ((ms != nullptr) && (fcdata.orires->nr != 0));
271 bool awhUsesMultiSim = (ir->bDoAwh && ir->awhParams->shareBiasMultisim && (ms != nullptr));
273 // Replica exchange, ensemble restraints and AWH need all
274 // simulations to remain synchronized, so they need
275 // checkpoints and stop conditions to act on the same step, so
276 // the propagation of such signals must take place between
277 // simulations, not just within simulations.
278 // TODO: Make algorithm initializers set these flags.
279 simulationsShareState = useReplicaExchange || usingEnsembleRestraints || awhUsesMultiSim;
281 if (simulationsShareState)
283 // Inter-simulation signal communication does not need to happen
284 // often, so we use a minimum of 200 steps to reduce overhead.
285 const int c_minimumInterSimulationSignallingInterval = 200;
286 nstSignalComm = ((c_minimumInterSimulationSignallingInterval + nstglobalcomm - 1) / nstglobalcomm)
291 if (startingBehavior != StartingBehavior::RestartWithAppending)
293 pleaseCiteCouplingAlgorithms(fplog, *ir);
296 init_mdoutf(fplog, nfile, fnm, mdrunOptions, cr, outputProvider, mdModulesNotifier, ir,
297 top_global, oenv, wcycle, startingBehavior, simulationsShareState, ms);
298 gmx::EnergyOutput energyOutput(mdoutf_get_fp_ene(outf), top_global, ir, pull_work,
299 mdoutf_get_fp_dhdl(outf), false, startingBehavior,
300 simulationsShareState, mdModulesNotifier);
302 gstat = global_stat_init(ir);
304 const auto& simulationWork = runScheduleWork->simulationWork;
305 const bool useGpuForPme = simulationWork.useGpuPme;
306 const bool useGpuForNonbonded = simulationWork.useGpuNonbonded;
307 const bool useGpuForBufferOps = simulationWork.useGpuBufferOps;
308 const bool useGpuForUpdate = simulationWork.useGpuUpdate;
310 /* Check for polarizable models and flexible constraints */
311 shellfc = init_shell_flexcon(fplog, top_global, constr ? constr->numFlexibleConstraints() : 0,
312 ir->nstcalcenergy, DOMAINDECOMP(cr), useGpuForPme);
315 double io = compute_io(ir, top_global->natoms, *groups, energyOutput.numEnergyTerms(), 1);
316 if ((io > 2000) && MASTER(cr))
318 fprintf(stderr, "\nWARNING: This run will generate roughly %.0f Mb of data\n\n", io);
322 // Local state only becomes valid now.
323 std::unique_ptr<t_state> stateInstance;
326 gmx_localtop_t top(top_global->ffparams);
328 auto mdatoms = mdAtoms->mdatoms();
330 ForceBuffers f(fr->useMts, ((useGpuForNonbonded && useGpuForBufferOps) || useGpuForUpdate)
331 ? PinningPolicy::PinnedIfSupported
332 : PinningPolicy::CannotBePinned);
333 if (DOMAINDECOMP(cr))
335 stateInstance = std::make_unique<t_state>();
336 state = stateInstance.get();
337 dd_init_local_state(cr->dd, state_global, state);
339 /* Distribute the charge groups over the nodes from the master node */
340 dd_partition_system(fplog, mdlog, ir->init_step, cr, TRUE, 1, state_global, *top_global, ir,
341 imdSession, pull_work, state, &f, mdAtoms, &top, fr, vsite, constr,
342 nrnb, nullptr, FALSE);
343 shouldCheckNumberOfBondedInteractions = true;
344 upd.setNumAtoms(state->natoms);
348 state_change_natoms(state_global, state_global->natoms);
349 /* Copy the pointer to the global state */
350 state = state_global;
352 /* Generate and initialize new topology */
353 mdAlgorithmsSetupAtomData(cr, ir, *top_global, &top, fr, &f, mdAtoms, constr, vsite, shellfc);
355 upd.setNumAtoms(state->natoms);
358 std::unique_ptr<UpdateConstrainGpu> integrator;
360 StatePropagatorDataGpu* stateGpu = fr->stateGpu;
362 // TODO: the assertions below should be handled by UpdateConstraintsBuilder.
365 GMX_RELEASE_ASSERT(!DOMAINDECOMP(cr) || ddUsesUpdateGroups(*cr->dd) || constr == nullptr
366 || constr->numConstraintsTotal() == 0,
367 "Constraints in domain decomposition are only supported with update "
368 "groups if using GPU update.\n");
369 GMX_RELEASE_ASSERT(ir->eConstrAlg != econtSHAKE || constr == nullptr
370 || constr->numConstraintsTotal() == 0,
371 "SHAKE is not supported with GPU update.");
372 GMX_RELEASE_ASSERT(useGpuForPme || (useGpuForNonbonded && simulationWork.useGpuBufferOps),
373 "Either PME or short-ranged non-bonded interaction tasks must run on "
374 "the GPU to use GPU update.\n");
375 GMX_RELEASE_ASSERT(ir->eI == eiMD,
376 "Only the md integrator is supported with the GPU update.\n");
378 ir->etc != etcNOSEHOOVER,
379 "Nose-Hoover temperature coupling is not supported with the GPU update.\n");
381 ir->epc == epcNO || ir->epc == epcPARRINELLORAHMAN || ir->epc == epcBERENDSEN
382 || ir->epc == epcCRESCALE,
383 "Only Parrinello-Rahman, Berendsen, and C-rescale pressure coupling are supported "
384 "with the GPU update.\n");
385 GMX_RELEASE_ASSERT(!mdatoms->haveVsites,
386 "Virtual sites are not supported with the GPU update.\n");
387 GMX_RELEASE_ASSERT(ed == nullptr,
388 "Essential dynamics is not supported with the GPU update.\n");
389 GMX_RELEASE_ASSERT(!ir->bPull || !pull_have_constraint(*ir->pull),
390 "Constraints pulling is not supported with the GPU update.\n");
391 GMX_RELEASE_ASSERT(fcdata.orires->nr == 0,
392 "Orientation restraints are not supported with the GPU update.\n");
395 || (!haveFepPerturbedMasses(*top_global) && !havePerturbedConstraints(*top_global)),
396 "Free energy perturbation of masses and constraints are not supported with the GPU "
399 if (constr != nullptr && constr->numConstraintsTotal() > 0)
403 .appendText("Updating coordinates and applying constraints on the GPU.");
407 GMX_LOG(mdlog.info).asParagraph().appendText("Updating coordinates on the GPU.");
409 GMX_RELEASE_ASSERT(fr->deviceStreamManager != nullptr,
410 "Device stream manager should be initialized in order to use GPU "
411 "update-constraints.");
413 fr->deviceStreamManager->streamIsValid(gmx::DeviceStreamType::UpdateAndConstraints),
414 "Update stream should be initialized in order to use GPU "
415 "update-constraints.");
416 integrator = std::make_unique<UpdateConstrainGpu>(
417 *ir, *top_global, fr->deviceStreamManager->context(),
418 fr->deviceStreamManager->stream(gmx::DeviceStreamType::UpdateAndConstraints),
419 stateGpu->xUpdatedOnDevice(), wcycle);
421 integrator->setPbc(PbcType::Xyz, state->box);
424 if (useGpuForPme || (useGpuForNonbonded && useGpuForBufferOps) || useGpuForUpdate)
426 changePinningPolicy(&state->x, PinningPolicy::PinnedIfSupported);
430 changePinningPolicy(&state->v, PinningPolicy::PinnedIfSupported);
433 // NOTE: The global state is no longer used at this point.
434 // But state_global is still used as temporary storage space for writing
435 // the global state to file and potentially for replica exchange.
436 // (Global topology should persist.)
438 update_mdatoms(mdatoms, state->lambda[efptMASS]);
442 /* Check nstexpanded here, because the grompp check was broken */
443 if (ir->expandedvals->nstexpanded % ir->nstcalcenergy != 0)
446 "With expanded ensemble, nstexpanded should be a multiple of nstcalcenergy");
448 init_expanded_ensemble(startingBehavior != StartingBehavior::NewSimulation, ir, state->dfhist);
453 EnergyData::initializeEnergyHistory(startingBehavior, observablesHistory, &energyOutput);
456 preparePrevStepPullCom(ir, pull_work, mdatoms->massT, state, state_global, cr,
457 startingBehavior != StartingBehavior::NewSimulation);
459 // TODO: Remove this by converting AWH into a ForceProvider
460 auto awh = prepareAwhModule(fplog, *ir, state_global, cr, ms,
461 startingBehavior != StartingBehavior::NewSimulation,
462 shellfc != nullptr, opt2fn("-awh", nfile, fnm), pull_work);
464 if (useReplicaExchange && MASTER(cr))
466 repl_ex = init_replica_exchange(fplog, ms, top_global->natoms, ir, replExParams);
468 /* PME tuning is only supported in the Verlet scheme, with PME for
469 * Coulomb. It is not supported with only LJ PME. */
470 bPMETune = (mdrunOptions.tunePme && EEL_PME(fr->ic->eeltype) && !mdrunOptions.reproducible
471 && ir->cutoff_scheme != ecutsGROUP);
473 pme_load_balancing_t* pme_loadbal = nullptr;
476 pme_loadbal_init(&pme_loadbal, cr, mdlog, *ir, state->box, *fr->ic, *fr->nbv, fr->pmedata,
480 if (!ir->bContinuation)
482 if (state->flags & (1U << estV))
484 auto v = makeArrayRef(state->v);
485 /* Set the velocities of vsites, shells and frozen atoms to zero */
486 for (i = 0; i < mdatoms->homenr; i++)
488 if (mdatoms->ptype[i] == eptVSite || mdatoms->ptype[i] == eptShell)
492 else if (mdatoms->cFREEZE)
494 for (m = 0; m < DIM; m++)
496 if (ir->opts.nFreeze[mdatoms->cFREEZE[i]][m])
507 /* Constrain the initial coordinates and velocities */
508 do_constrain_first(fplog, constr, ir, mdatoms->nr, mdatoms->homenr,
509 state->x.arrayRefWithPadding(), state->v.arrayRefWithPadding(),
510 state->box, state->lambda[efptBONDED]);
514 /* Construct the virtual sites for the initial configuration */
515 vsite->construct(state->x, ir->delta_t, {}, state->box);
519 if (ir->efep != efepNO)
521 /* Set free energy calculation frequency as the greatest common
522 * denominator of nstdhdl and repl_ex_nst. */
523 nstfep = ir->fepvals->nstdhdl;
526 nstfep = std::gcd(ir->expandedvals->nstexpanded, nstfep);
528 if (useReplicaExchange)
530 nstfep = std::gcd(replExParams.exchangeInterval, nstfep);
534 nstfep = std::gcd(ir->awhParams->nstSampleCoord, nstfep);
538 /* Be REALLY careful about what flags you set here. You CANNOT assume
539 * this is the first step, since we might be restarting from a checkpoint,
540 * and in that case we should not do any modifications to the state.
542 bStopCM = (ir->comm_mode != ecmNO && !ir->bContinuation);
544 // When restarting from a checkpoint, it can be appropriate to
545 // initialize ekind from quantities in the checkpoint. Otherwise,
546 // compute_globals must initialize ekind before the simulation
547 // starts/restarts. However, only the master rank knows what was
548 // found in the checkpoint file, so we have to communicate in
549 // order to coordinate the restart.
551 // TODO Consider removing this communication if/when checkpoint
552 // reading directly follows .tpr reading, because all ranks can
553 // agree on hasReadEkinState at that time.
554 bool hasReadEkinState = MASTER(cr) ? state_global->ekinstate.hasReadEkinState : false;
557 gmx_bcast(sizeof(hasReadEkinState), &hasReadEkinState, cr->mpi_comm_mygroup);
559 if (hasReadEkinState)
561 restore_ekinstate_from_state(cr, ekind, &state_global->ekinstate);
564 unsigned int cglo_flags =
565 (CGLO_TEMPERATURE | CGLO_GSTAT | (EI_VV(ir->eI) ? CGLO_PRESSURE : 0)
566 | (EI_VV(ir->eI) ? CGLO_CONSTRAINT : 0) | (hasReadEkinState ? CGLO_READEKIN : 0));
568 bSumEkinhOld = FALSE;
570 t_vcm vcm(top_global->groups, *ir);
571 reportComRemovalInfo(fplog, vcm);
573 /* To minimize communication, compute_globals computes the COM velocity
574 * and the kinetic energy for the velocities without COM motion removed.
575 * Thus to get the kinetic energy without the COM contribution, we need
576 * to call compute_globals twice.
578 for (int cgloIteration = 0; cgloIteration < (bStopCM ? 2 : 1); cgloIteration++)
580 unsigned int cglo_flags_iteration = cglo_flags;
581 if (bStopCM && cgloIteration == 0)
583 cglo_flags_iteration |= CGLO_STOPCM;
584 cglo_flags_iteration &= ~CGLO_TEMPERATURE;
586 compute_globals(gstat, cr, ir, fr, ekind, makeConstArrayRef(state->x),
587 makeConstArrayRef(state->v), state->box, mdatoms, nrnb, &vcm, nullptr,
588 enerd, force_vir, shake_vir, total_vir, pres, constr, &nullSignaller,
589 state->box, &totalNumberOfBondedInteractions, &bSumEkinhOld,
591 | (shouldCheckNumberOfBondedInteractions ? CGLO_CHECK_NUMBER_OF_BONDED_INTERACTIONS
593 if (cglo_flags_iteration & CGLO_STOPCM)
595 /* At initialization, do not pass x with acceleration-correction mode
596 * to avoid (incorrect) correction of the initial coordinates.
598 auto x = (vcm.mode == ecmLINEAR_ACCELERATION_CORRECTION) ? ArrayRef<RVec>()
599 : makeArrayRef(state->x);
600 process_and_stopcm_grp(fplog, &vcm, *mdatoms, x, makeArrayRef(state->v));
601 inc_nrnb(nrnb, eNR_STOPCM, mdatoms->homenr);
604 checkNumberOfBondedInteractions(mdlog, cr, totalNumberOfBondedInteractions, top_global, &top,
605 makeConstArrayRef(state->x), state->box,
606 &shouldCheckNumberOfBondedInteractions);
607 if (ir->eI == eiVVAK)
609 /* a second call to get the half step temperature initialized as well */
610 /* we do the same call as above, but turn the pressure off -- internally to
611 compute_globals, this is recognized as a velocity verlet half-step
612 kinetic energy calculation. This minimized excess variables, but
613 perhaps loses some logic?*/
615 compute_globals(gstat, cr, ir, fr, ekind, makeConstArrayRef(state->x),
616 makeConstArrayRef(state->v), state->box, mdatoms, nrnb, &vcm, nullptr,
617 enerd, force_vir, shake_vir, total_vir, pres, constr, &nullSignaller,
618 state->box, nullptr, &bSumEkinhOld, cglo_flags & ~CGLO_PRESSURE);
621 /* Calculate the initial half step temperature, and save the ekinh_old */
622 if (startingBehavior == StartingBehavior::NewSimulation)
624 for (i = 0; (i < ir->opts.ngtc); i++)
626 copy_mat(ekind->tcstat[i].ekinh, ekind->tcstat[i].ekinh_old);
630 /* need to make an initiation call to get the Trotter variables set, as well as other constants
631 for non-trotter temperature control */
632 auto trotter_seq = init_npt_vars(ir, state, &MassQ, bTrotter);
636 if (!ir->bContinuation)
638 if (constr && ir->eConstrAlg == econtLINCS)
640 fprintf(fplog, "RMS relative constraint deviation after constraining: %.2e\n",
643 if (EI_STATE_VELOCITY(ir->eI))
645 real temp = enerd->term[F_TEMP];
648 /* Result of Ekin averaged over velocities of -half
649 * and +half step, while we only have -half step here.
653 fprintf(fplog, "Initial temperature: %g K\n", temp);
658 fprintf(stderr, "starting mdrun '%s'\n", *(top_global->name));
661 sprintf(tbuf, "%8.1f", (ir->init_step + ir->nsteps) * ir->delta_t);
665 sprintf(tbuf, "%s", "infinite");
667 if (ir->init_step > 0)
669 fprintf(stderr, "%s steps, %s ps (continuing from step %s, %8.1f ps).\n",
670 gmx_step_str(ir->init_step + ir->nsteps, sbuf), tbuf,
671 gmx_step_str(ir->init_step, sbuf2), ir->init_step * ir->delta_t);
675 fprintf(stderr, "%s steps, %s ps.\n", gmx_step_str(ir->nsteps, sbuf), tbuf);
677 fprintf(fplog, "\n");
680 walltime_accounting_start_time(walltime_accounting);
681 wallcycle_start(wcycle, ewcRUN);
682 print_start(fplog, cr, walltime_accounting, "mdrun");
684 /***********************************************************
688 ************************************************************/
691 /* Skip the first Nose-Hoover integration when we get the state from tpx */
692 bInitStep = startingBehavior == StartingBehavior::NewSimulation || EI_VV(ir->eI);
693 bSumEkinhOld = FALSE;
695 bNeedRepartition = FALSE;
697 step = ir->init_step;
700 auto stopHandler = stopHandlerBuilder->getStopHandlerMD(
701 compat::not_null<SimulationSignal*>(&signals[eglsSTOPCOND]), simulationsShareState,
702 MASTER(cr), ir->nstlist, mdrunOptions.reproducible, nstSignalComm,
703 mdrunOptions.maximumHoursToRun, ir->nstlist == 0, fplog, step, bNS, walltime_accounting);
705 auto checkpointHandler = std::make_unique<CheckpointHandler>(
706 compat::make_not_null<SimulationSignal*>(&signals[eglsCHKPT]), simulationsShareState,
707 ir->nstlist == 0, MASTER(cr), mdrunOptions.writeConfout,
708 mdrunOptions.checkpointOptions.period);
710 const bool resetCountersIsLocal = true;
711 auto resetHandler = std::make_unique<ResetHandler>(
712 compat::make_not_null<SimulationSignal*>(&signals[eglsRESETCOUNTERS]),
713 !resetCountersIsLocal, ir->nsteps, MASTER(cr), mdrunOptions.timingOptions.resetHalfway,
714 mdrunOptions.maximumHoursToRun, mdlog, wcycle, walltime_accounting);
716 const DDBalanceRegionHandler ddBalanceRegionHandler(cr);
718 if (MASTER(cr) && isMultiSim(ms) && !useReplicaExchange)
720 logInitialMultisimStatus(ms, cr, mdlog, simulationsShareState, ir->nsteps, ir->init_step);
723 /* and stop now if we should */
724 bLastStep = (bLastStep || (ir->nsteps >= 0 && step_rel > ir->nsteps));
728 /* Determine if this is a neighbor search step */
729 bNStList = (ir->nstlist > 0 && step % ir->nstlist == 0);
731 if (bPMETune && bNStList)
733 // This has to be here because PME load balancing is called so early.
734 // TODO: Move to after all booleans are defined.
735 if (useGpuForUpdate && !bFirstStep)
737 stateGpu->copyCoordinatesFromGpu(state->x, AtomLocality::Local);
738 stateGpu->waitCoordinatesReadyOnHost(AtomLocality::Local);
740 /* PME grid + cut-off optimization with GPUs or PME nodes */
741 pme_loadbal_do(pme_loadbal, cr, (mdrunOptions.verbose && MASTER(cr)) ? stderr : nullptr,
742 fplog, mdlog, *ir, fr, state->box, state->x, wcycle, step, step_rel,
743 &bPMETunePrinting, simulationWork.useGpuPmePpCommunication);
746 wallcycle_start(wcycle, ewcSTEP);
748 bLastStep = (step_rel == ir->nsteps);
749 t = t0 + step * ir->delta_t;
751 // TODO Refactor this, so that nstfep does not need a default value of zero
752 if (ir->efep != efepNO || ir->bSimTemp)
754 /* find and set the current lambdas */
755 state->lambda = currentLambdas(step, *(ir->fepvals), state->fep_state);
757 bDoDHDL = do_per_step(step, ir->fepvals->nstdhdl);
758 bDoFEP = ((ir->efep != efepNO) && do_per_step(step, nstfep));
759 bDoExpanded = (do_per_step(step, ir->expandedvals->nstexpanded) && (ir->bExpanded)
763 bDoReplEx = (useReplicaExchange && (step > 0) && !bLastStep
764 && do_per_step(step, replExParams.exchangeInterval));
766 if (doSimulatedAnnealing)
768 update_annealing_target_temp(ir, t, &upd);
771 /* Stop Center of Mass motion */
772 bStopCM = (ir->comm_mode != ecmNO && do_per_step(step, ir->nstcomm));
774 /* Determine whether or not to do Neighbour Searching */
775 bNS = (bFirstStep || bNStList || bExchanged || bNeedRepartition);
777 /* Note that the stopHandler will cause termination at nstglobalcomm
778 * steps. Since this concides with nstcalcenergy, nsttcouple and/or
779 * nstpcouple steps, we have computed the half-step kinetic energy
780 * of the previous step and can always output energies at the last step.
782 bLastStep = bLastStep || stopHandler->stoppingAfterCurrentStep(bNS);
784 /* do_log triggers energy and virial calculation. Because this leads
785 * to different code paths, forces can be different. Thus for exact
786 * continuation we should avoid extra log output.
787 * Note that the || bLastStep can result in non-exact continuation
788 * beyond the last step. But we don't consider that to be an issue.
790 do_log = (do_per_step(step, ir->nstlog)
791 || (bFirstStep && startingBehavior == StartingBehavior::NewSimulation) || bLastStep);
792 do_verbose = mdrunOptions.verbose
793 && (step % mdrunOptions.verboseStepPrintInterval == 0 || bFirstStep || bLastStep);
795 if (useGpuForUpdate && !bFirstStep && bNS)
797 // Copy velocities from the GPU on search steps to keep a copy on host (device buffers are reinitialized).
798 stateGpu->copyVelocitiesFromGpu(state->v, AtomLocality::Local);
799 stateGpu->waitVelocitiesReadyOnHost(AtomLocality::Local);
800 // Copy coordinate from the GPU when needed at the search step.
801 // NOTE: The cases when coordinates needed on CPU for force evaluation are handled in sim_utils.
802 // NOTE: If the coordinates are to be written into output file they are also copied separately before the output.
803 stateGpu->copyCoordinatesFromGpu(state->x, AtomLocality::Local);
804 stateGpu->waitCoordinatesReadyOnHost(AtomLocality::Local);
807 if (bNS && !(bFirstStep && ir->bContinuation))
809 bMasterState = FALSE;
810 /* Correct the new box if it is too skewed */
811 if (inputrecDynamicBox(ir))
813 if (correct_box(fplog, step, state->box))
816 // If update is offloaded, it should be informed about the box size change
819 integrator->setPbc(PbcType::Xyz, state->box);
823 if (DOMAINDECOMP(cr) && bMasterState)
825 dd_collect_state(cr->dd, state, state_global);
828 if (DOMAINDECOMP(cr))
830 /* Repartition the domain decomposition */
831 dd_partition_system(fplog, mdlog, step, cr, bMasterState, nstglobalcomm, state_global,
832 *top_global, ir, imdSession, pull_work, state, &f, mdAtoms, &top,
833 fr, vsite, constr, nrnb, wcycle, do_verbose && !bPMETunePrinting);
834 shouldCheckNumberOfBondedInteractions = true;
835 upd.setNumAtoms(state->natoms);
839 // Allocate or re-size GPU halo exchange object, if necessary
840 if (bNS && havePPDomainDecomposition(cr) && simulationWork.useGpuHaloExchange)
842 GMX_RELEASE_ASSERT(fr->deviceStreamManager != nullptr,
843 "GPU device manager has to be initialized to use GPU "
844 "version of halo exchange.");
845 constructGpuHaloExchange(mdlog, *cr, *fr->deviceStreamManager, wcycle);
848 if (MASTER(cr) && do_log)
850 gmx::EnergyOutput::printHeader(fplog, step,
851 t); /* can we improve the information printed here? */
854 if (ir->efep != efepNO)
856 update_mdatoms(mdatoms, state->lambda[efptMASS]);
862 /* We need the kinetic energy at minus the half step for determining
863 * the full step kinetic energy and possibly for T-coupling.*/
864 /* This may not be quite working correctly yet . . . . */
865 compute_globals(gstat, cr, ir, fr, ekind, makeConstArrayRef(state->x),
866 makeConstArrayRef(state->v), state->box, mdatoms, nrnb, &vcm, wcycle,
867 enerd, nullptr, nullptr, nullptr, nullptr, constr, &nullSignaller,
868 state->box, &totalNumberOfBondedInteractions, &bSumEkinhOld,
869 CGLO_GSTAT | CGLO_TEMPERATURE | CGLO_CHECK_NUMBER_OF_BONDED_INTERACTIONS);
870 checkNumberOfBondedInteractions(mdlog, cr, totalNumberOfBondedInteractions, top_global,
871 &top, makeConstArrayRef(state->x), state->box,
872 &shouldCheckNumberOfBondedInteractions);
874 clear_mat(force_vir);
876 checkpointHandler->decideIfCheckpointingThisStep(bNS, bFirstStep, bLastStep);
878 /* Determine the energy and pressure:
879 * at nstcalcenergy steps and at energy output steps (set below).
881 if (EI_VV(ir->eI) && (!bInitStep))
883 bCalcEnerStep = do_per_step(step, ir->nstcalcenergy);
884 bCalcVir = bCalcEnerStep
886 && (do_per_step(step, ir->nstpcouple) || do_per_step(step - 1, ir->nstpcouple)));
890 bCalcEnerStep = do_per_step(step, ir->nstcalcenergy);
891 bCalcVir = bCalcEnerStep || (ir->epc != epcNO && do_per_step(step, ir->nstpcouple));
893 bCalcEner = bCalcEnerStep;
895 do_ene = (do_per_step(step, ir->nstenergy) || bLastStep);
897 if (do_ene || do_log || bDoReplEx)
903 /* Do we need global communication ? */
904 bGStat = (bCalcVir || bCalcEner || bStopCM || do_per_step(step, nstglobalcomm)
905 || (EI_VV(ir->eI) && inputrecNvtTrotter(ir) && do_per_step(step - 1, nstglobalcomm)));
907 force_flags = (GMX_FORCE_STATECHANGED | ((inputrecDynamicBox(ir)) ? GMX_FORCE_DYNAMICBOX : 0)
908 | GMX_FORCE_ALLFORCES | (bCalcVir ? GMX_FORCE_VIRIAL : 0)
909 | (bCalcEner ? GMX_FORCE_ENERGY : 0) | (bDoFEP ? GMX_FORCE_DHDL : 0));
910 if (fr->useMts && !do_per_step(step, ir->nstfout))
912 force_flags |= GMX_FORCE_DO_NOT_NEED_NORMAL_FORCE;
917 /* Now is the time to relax the shells */
918 relax_shell_flexcon(fplog, cr, ms, mdrunOptions.verbose, enforcedRotation, step, ir,
919 imdSession, pull_work, bNS, force_flags, &top, constr, enerd,
920 state->natoms, state->x.arrayRefWithPadding(),
921 state->v.arrayRefWithPadding(), state->box, state->lambda,
922 &state->hist, &f.view(), force_vir, mdatoms, nrnb, wcycle, shellfc,
923 fr, runScheduleWork, t, mu_tot, vsite, ddBalanceRegionHandler);
927 /* The AWH history need to be saved _before_ doing force calculations where the AWH bias
928 is updated (or the AWH update will be performed twice for one step when continuing).
929 It would be best to call this update function from do_md_trajectory_writing but that
930 would occur after do_force. One would have to divide the update_awh function into one
931 function applying the AWH force and one doing the AWH bias update. The update AWH
932 bias function could then be called after do_md_trajectory_writing (then containing
933 update_awh_history). The checkpointing will in the future probably moved to the start
934 of the md loop which will rid of this issue. */
935 if (awh && checkpointHandler->isCheckpointingStep() && MASTER(cr))
937 awh->updateHistory(state_global->awhHistory.get());
940 /* The coordinates (x) are shifted (to get whole molecules)
942 * This is parallellized as well, and does communication too.
943 * Check comments in sim_util.c
945 do_force(fplog, cr, ms, ir, awh.get(), enforcedRotation, imdSession, pull_work, step,
946 nrnb, wcycle, &top, state->box, state->x.arrayRefWithPadding(), &state->hist,
947 &f.view(), force_vir, mdatoms, enerd, state->lambda, fr, runScheduleWork,
948 vsite, mu_tot, t, ed ? ed->getLegacyED() : nullptr,
949 (bNS ? GMX_FORCE_NS : 0) | force_flags, ddBalanceRegionHandler);
952 // VV integrators do not need the following velocity half step
953 // if it is the first step after starting from a checkpoint.
954 // That is, the half step is needed on all other steps, and
955 // also the first step when starting from a .tpr file.
956 if (EI_VV(ir->eI) && (!bFirstStep || startingBehavior == StartingBehavior::NewSimulation))
957 /* ############### START FIRST UPDATE HALF-STEP FOR VV METHODS############### */
959 rvec* vbuf = nullptr;
961 wallcycle_start(wcycle, ewcUPDATE);
962 if (ir->eI == eiVV && bInitStep)
964 /* if using velocity verlet with full time step Ekin,
965 * take the first half step only to compute the
966 * virial for the first step. From there,
967 * revert back to the initial coordinates
968 * so that the input is actually the initial step.
970 snew(vbuf, state->natoms);
971 copy_rvecn(state->v.rvec_array(), vbuf, 0,
972 state->natoms); /* should make this better for parallelizing? */
976 /* this is for NHC in the Ekin(t+dt/2) version of vv */
977 trotter_update(ir, step, ekind, enerd, state, total_vir, mdatoms, &MassQ,
978 trotter_seq, ettTSEQ1);
981 upd.update_coords(*ir, step, mdatoms, state, f.view().forceWithPadding(), fcdata, ekind,
982 M, etrtVELOCITY1, cr, constr != nullptr);
984 wallcycle_stop(wcycle, ewcUPDATE);
985 constrain_velocities(constr, do_log, do_ene, step, state, nullptr, bCalcVir, shake_vir);
986 wallcycle_start(wcycle, ewcUPDATE);
987 /* if VV, compute the pressure and constraints */
988 /* For VV2, we strictly only need this if using pressure
989 * control, but we really would like to have accurate pressures
991 * Think about ways around this in the future?
992 * For now, keep this choice in comments.
994 /*bPres = (ir->eI==eiVV || inputrecNptTrotter(ir)); */
995 /*bTemp = ((ir->eI==eiVV &&(!bInitStep)) || (ir->eI==eiVVAK && inputrecNptTrotter(ir)));*/
997 bTemp = ((ir->eI == eiVV && (!bInitStep)) || (ir->eI == eiVVAK));
998 if (bCalcEner && ir->eI == eiVVAK)
1000 bSumEkinhOld = TRUE;
1002 /* for vv, the first half of the integration actually corresponds to the previous step.
1003 So we need information from the last step in the first half of the integration */
1004 if (bGStat || do_per_step(step - 1, nstglobalcomm))
1006 wallcycle_stop(wcycle, ewcUPDATE);
1007 compute_globals(gstat, cr, ir, fr, ekind, makeConstArrayRef(state->x),
1008 makeConstArrayRef(state->v), state->box, mdatoms, nrnb, &vcm, wcycle,
1009 enerd, force_vir, shake_vir, total_vir, pres, constr, &nullSignaller,
1010 state->box, &totalNumberOfBondedInteractions, &bSumEkinhOld,
1011 (bGStat ? CGLO_GSTAT : 0) | (bCalcEner ? CGLO_ENERGY : 0)
1012 | (bTemp ? CGLO_TEMPERATURE : 0) | (bPres ? CGLO_PRESSURE : 0)
1013 | (bPres ? CGLO_CONSTRAINT : 0) | (bStopCM ? CGLO_STOPCM : 0)
1014 | (shouldCheckNumberOfBondedInteractions ? CGLO_CHECK_NUMBER_OF_BONDED_INTERACTIONS
1017 /* explanation of above:
1018 a) We compute Ekin at the full time step
1019 if 1) we are using the AveVel Ekin, and it's not the
1020 initial step, or 2) if we are using AveEkin, but need the full
1021 time step kinetic energy for the pressure (always true now, since we want accurate statistics).
1022 b) If we are using EkinAveEkin for the kinetic energy for the temperature control, we still feed in
1023 EkinAveVel because it's needed for the pressure */
1024 checkNumberOfBondedInteractions(mdlog, cr, totalNumberOfBondedInteractions,
1025 top_global, &top, makeConstArrayRef(state->x),
1026 state->box, &shouldCheckNumberOfBondedInteractions);
1029 process_and_stopcm_grp(fplog, &vcm, *mdatoms, makeArrayRef(state->x),
1030 makeArrayRef(state->v));
1031 inc_nrnb(nrnb, eNR_STOPCM, mdatoms->homenr);
1033 wallcycle_start(wcycle, ewcUPDATE);
1035 /* temperature scaling and pressure scaling to produce the extended variables at t+dt */
1040 m_add(force_vir, shake_vir,
1041 total_vir); /* we need the un-dispersion corrected total vir here */
1042 trotter_update(ir, step, ekind, enerd, state, total_vir, mdatoms, &MassQ,
1043 trotter_seq, ettTSEQ2);
1045 /* TODO This is only needed when we're about to write
1046 * a checkpoint, because we use it after the restart
1047 * (in a kludge?). But what should we be doing if
1048 * the startingBehavior is NewSimulation or bInitStep are true? */
1049 if (inputrecNptTrotter(ir) || inputrecNphTrotter(ir))
1051 copy_mat(shake_vir, state->svir_prev);
1052 copy_mat(force_vir, state->fvir_prev);
1054 if ((inputrecNptTrotter(ir) || inputrecNvtTrotter(ir)) && ir->eI == eiVV)
1056 /* update temperature and kinetic energy now that step is over - this is the v(t+dt) point */
1057 enerd->term[F_TEMP] =
1058 sum_ekin(&(ir->opts), ekind, nullptr, (ir->eI == eiVV), FALSE);
1059 enerd->term[F_EKIN] = trace(ekind->ekin);
1062 else if (bExchanged)
1064 wallcycle_stop(wcycle, ewcUPDATE);
1065 /* We need the kinetic energy at minus the half step for determining
1066 * the full step kinetic energy and possibly for T-coupling.*/
1067 /* This may not be quite working correctly yet . . . . */
1068 compute_globals(gstat, cr, ir, fr, ekind, makeConstArrayRef(state->x),
1069 makeConstArrayRef(state->v), state->box, mdatoms, nrnb, &vcm, wcycle,
1070 enerd, nullptr, nullptr, nullptr, nullptr, constr, &nullSignaller,
1071 state->box, nullptr, &bSumEkinhOld, CGLO_GSTAT | CGLO_TEMPERATURE);
1072 wallcycle_start(wcycle, ewcUPDATE);
1075 /* if it's the initial step, we performed this first step just to get the constraint virial */
1076 if (ir->eI == eiVV && bInitStep)
1078 copy_rvecn(vbuf, state->v.rvec_array(), 0, state->natoms);
1081 wallcycle_stop(wcycle, ewcUPDATE);
1084 /* compute the conserved quantity */
1087 saved_conserved_quantity = NPT_energy(ir, state, &MassQ);
1090 last_ekin = enerd->term[F_EKIN];
1092 if ((ir->eDispCorr != edispcEnerPres) && (ir->eDispCorr != edispcAllEnerPres))
1094 saved_conserved_quantity -= enerd->term[F_DISPCORR];
1096 /* sum up the foreign kinetic energy and dK/dl terms for vv. currently done every step so that dhdl is correct in the .edr */
1097 if (ir->efep != efepNO)
1099 accumulateKineticLambdaComponents(enerd, state->lambda, *ir->fepvals);
1103 /* ######## END FIRST UPDATE STEP ############## */
1104 /* ######## If doing VV, we now have v(dt) ###### */
1107 /* perform extended ensemble sampling in lambda - we don't
1108 actually move to the new state before outputting
1109 statistics, but if performing simulated tempering, we
1110 do update the velocities and the tau_t. */
1112 lamnew = ExpandedEnsembleDynamics(fplog, ir, enerd, state, &MassQ, state->fep_state,
1113 state->dfhist, step, state->v.rvec_array(), mdatoms);
1114 /* history is maintained in state->dfhist, but state_global is what is sent to trajectory and log output */
1117 copy_df_history(state_global->dfhist, state->dfhist);
1121 // Copy coordinate from the GPU for the output/checkpointing if the update is offloaded and
1122 // coordinates have not already been copied for i) search or ii) CPU force tasks.
1123 if (useGpuForUpdate && !bNS && !runScheduleWork->domainWork.haveCpuLocalForceWork
1124 && (do_per_step(step, ir->nstxout) || do_per_step(step, ir->nstxout_compressed)
1125 || checkpointHandler->isCheckpointingStep()))
1127 stateGpu->copyCoordinatesFromGpu(state->x, AtomLocality::Local);
1128 stateGpu->waitCoordinatesReadyOnHost(AtomLocality::Local);
1130 // Copy velocities if needed for the output/checkpointing.
1131 // NOTE: Copy on the search steps is done at the beginning of the step.
1132 if (useGpuForUpdate && !bNS
1133 && (do_per_step(step, ir->nstvout) || checkpointHandler->isCheckpointingStep()))
1135 stateGpu->copyVelocitiesFromGpu(state->v, AtomLocality::Local);
1136 stateGpu->waitVelocitiesReadyOnHost(AtomLocality::Local);
1138 // Copy forces for the output if the forces were reduced on the GPU (not the case on virial steps)
1139 // and update is offloaded hence forces are kept on the GPU for update and have not been
1140 // already transferred in do_force().
1141 // TODO: There should be an improved, explicit mechanism that ensures this copy is only executed
1142 // when the forces are ready on the GPU -- the same synchronizer should be used as the one
1143 // prior to GPU update.
1144 // TODO: When the output flags will be included in step workload, this copy can be combined with the
1145 // copy call in do_force(...).
1146 // NOTE: The forces should not be copied here if the vsites are present, since they were modified
1147 // on host after the D2H copy in do_force(...).
1148 if (runScheduleWork->stepWork.useGpuFBufferOps && (simulationWork.useGpuUpdate && !vsite)
1149 && do_per_step(step, ir->nstfout))
1151 stateGpu->copyForcesFromGpu(f.view().force(), AtomLocality::Local);
1152 stateGpu->waitForcesReadyOnHost(AtomLocality::Local);
1154 /* Now we have the energies and forces corresponding to the
1155 * coordinates at time t. We must output all of this before
1158 do_md_trajectory_writing(fplog, cr, nfile, fnm, step, step_rel, t, ir, state, state_global,
1159 observablesHistory, top_global, fr, outf, energyOutput, ekind,
1160 f.view().force(), checkpointHandler->isCheckpointingStep(),
1161 bRerunMD, bLastStep, mdrunOptions.writeConfout, bSumEkinhOld);
1162 /* Check if IMD step and do IMD communication, if bIMD is TRUE. */
1163 bInteractiveMDstep = imdSession->run(step, bNS, state->box, state->x.rvec_array(), t);
1165 /* kludge -- virial is lost with restart for MTTK NPT control. Must reload (saved earlier). */
1166 if (startingBehavior != StartingBehavior::NewSimulation && bFirstStep
1167 && (inputrecNptTrotter(ir) || inputrecNphTrotter(ir)))
1169 copy_mat(state->svir_prev, shake_vir);
1170 copy_mat(state->fvir_prev, force_vir);
1173 stopHandler->setSignal();
1174 resetHandler->setSignal(walltime_accounting);
1176 if (bGStat || !PAR(cr))
1178 /* In parallel we only have to check for checkpointing in steps
1179 * where we do global communication,
1180 * otherwise the other nodes don't know.
1182 checkpointHandler->setSignal(walltime_accounting);
1185 /* ######### START SECOND UPDATE STEP ################# */
1187 /* at the start of step, randomize or scale the velocities ((if vv. Restriction of Andersen
1188 controlled in preprocessing */
1190 if (ETC_ANDERSEN(ir->etc)) /* keep this outside of update_tcouple because of the extra info required to pass */
1192 gmx_bool bIfRandomize;
1193 bIfRandomize = update_randomize_velocities(ir, step, cr, mdatoms, state->v, &upd, constr);
1194 /* if we have constraints, we have to remove the kinetic energy parallel to the bonds */
1195 if (constr && bIfRandomize)
1197 constrain_velocities(constr, do_log, do_ene, step, state, nullptr, false, nullptr);
1200 /* Box is changed in update() when we do pressure coupling,
1201 * but we should still use the old box for energy corrections and when
1202 * writing it to the energy file, so it matches the trajectory files for
1203 * the same timestep above. Make a copy in a separate array.
1205 copy_mat(state->box, lastbox);
1209 if (!useGpuForUpdate)
1211 wallcycle_start(wcycle, ewcUPDATE);
1213 /* UPDATE PRESSURE VARIABLES IN TROTTER FORMULATION WITH CONSTRAINTS */
1216 trotter_update(ir, step, ekind, enerd, state, total_vir, mdatoms, &MassQ, trotter_seq, ettTSEQ3);
1217 /* We can only do Berendsen coupling after we have summed
1218 * the kinetic energy or virial. Since the happens
1219 * in global_state after update, we should only do it at
1220 * step % nstlist = 1 with bGStatEveryStep=FALSE.
1225 update_tcouple(step, ir, state, ekind, &MassQ, mdatoms);
1226 update_pcouple_before_coordinates(fplog, step, ir, state, pressureCouplingMu, M, bInitStep);
1231 /* velocity half-step update */
1232 upd.update_coords(*ir, step, mdatoms, state, f.view().forceWithPadding(), fcdata, ekind,
1233 M, etrtVELOCITY2, cr, constr != nullptr);
1236 /* Above, initialize just copies ekinh into ekin,
1237 * it doesn't copy position (for VV),
1238 * and entire integrator for MD.
1241 if (ir->eI == eiVVAK)
1243 cbuf.resize(state->x.size());
1244 std::copy(state->x.begin(), state->x.end(), cbuf.begin());
1247 /* With leap-frog type integrators we compute the kinetic energy
1248 * at a whole time step as the average of the half-time step kinetic
1249 * energies of two subsequent steps. Therefore we need to compute the
1250 * half step kinetic energy also if we need energies at the next step.
1252 const bool needHalfStepKineticEnergy =
1253 (!EI_VV(ir->eI) && (do_per_step(step + 1, nstglobalcomm) || step_rel + 1 == ir->nsteps));
1255 // Parrinello-Rahman requires the pressure to be availible before the update to compute
1256 // the velocity scaling matrix. Hence, it runs one step after the nstpcouple step.
1257 const bool doParrinelloRahman = (ir->epc == epcPARRINELLORAHMAN
1258 && do_per_step(step + ir->nstpcouple - 1, ir->nstpcouple));
1260 if (useGpuForUpdate)
1262 if (bNS && (bFirstStep || DOMAINDECOMP(cr)))
1264 integrator->set(stateGpu->getCoordinates(), stateGpu->getVelocities(),
1265 stateGpu->getForces(), top.idef, *mdatoms, ekind->ngtc);
1267 // Copy data to the GPU after buffers might have being reinitialized
1268 stateGpu->copyVelocitiesToGpu(state->v, AtomLocality::Local);
1269 stateGpu->copyCoordinatesToGpu(state->x, AtomLocality::Local);
1272 if (simulationWork.useGpuPme && !runScheduleWork->simulationWork.useGpuPmePpCommunication
1273 && !thisRankHasDuty(cr, DUTY_PME))
1275 // The PME forces were recieved to the host, so have to be copied
1276 stateGpu->copyForcesToGpu(f.view().force(), AtomLocality::All);
1278 else if (!runScheduleWork->stepWork.useGpuFBufferOps)
1280 // The buffer ops were not offloaded this step, so the forces are on the
1281 // host and have to be copied
1282 stateGpu->copyForcesToGpu(f.view().force(), AtomLocality::Local);
1285 const bool doTemperatureScaling =
1286 (ir->etc != etcNO && do_per_step(step + ir->nsttcouple - 1, ir->nsttcouple));
1288 // This applies Leap-Frog, LINCS and SETTLE in succession
1289 integrator->integrate(stateGpu->getForcesReadyOnDeviceEvent(
1290 AtomLocality::Local, runScheduleWork->stepWork.useGpuFBufferOps),
1291 ir->delta_t, true, bCalcVir, shake_vir, doTemperatureScaling,
1292 ekind->tcstat, doParrinelloRahman, ir->nstpcouple * ir->delta_t, M);
1294 // Copy velocities D2H after update if:
1295 // - Globals are computed this step (includes the energy output steps).
1296 // - Temperature is needed for the next step.
1297 if (bGStat || needHalfStepKineticEnergy)
1299 stateGpu->copyVelocitiesFromGpu(state->v, AtomLocality::Local);
1300 stateGpu->waitVelocitiesReadyOnHost(AtomLocality::Local);
1305 /* With multiple time stepping we need to do an additional normal
1306 * update step to obtain the virial, as the actual MTS integration
1307 * using an acceleration where the slow forces are multiplied by mtsFactor.
1308 * Using that acceleration would result in a virial with the slow
1309 * force contribution would be a factor mtsFactor too large.
1311 if (fr->useMts && bCalcVir && constr != nullptr)
1313 upd.update_for_constraint_virial(*ir, *mdatoms, *state, f.view().forceWithPadding(), *ekind);
1315 constrain_coordinates(constr, do_log, do_ene, step, state,
1316 upd.xp()->arrayRefWithPadding(), &dvdl_constr, bCalcVir, shake_vir);
1319 ArrayRefWithPadding<const RVec> forceCombined =
1320 (fr->useMts && step % ir->mtsLevels[1].stepFactor == 0)
1321 ? f.view().forceMtsCombinedWithPadding()
1322 : f.view().forceWithPadding();
1323 upd.update_coords(*ir, step, mdatoms, state, forceCombined, fcdata, ekind, M,
1324 etrtPOSITION, cr, constr != nullptr);
1326 wallcycle_stop(wcycle, ewcUPDATE);
1328 constrain_coordinates(constr, do_log, do_ene, step, state, upd.xp()->arrayRefWithPadding(),
1329 &dvdl_constr, bCalcVir && !fr->useMts, shake_vir);
1331 upd.update_sd_second_half(*ir, step, &dvdl_constr, mdatoms, state, cr, nrnb, wcycle,
1332 constr, do_log, do_ene);
1333 upd.finish_update(*ir, mdatoms, state, wcycle, constr != nullptr);
1336 if (ir->bPull && ir->pull->bSetPbcRefToPrevStepCOM)
1338 updatePrevStepPullCom(pull_work, state);
1341 if (ir->eI == eiVVAK)
1343 /* erase F_EKIN and F_TEMP here? */
1344 /* just compute the kinetic energy at the half step to perform a trotter step */
1345 compute_globals(gstat, cr, ir, fr, ekind, makeConstArrayRef(state->x),
1346 makeConstArrayRef(state->v), state->box, mdatoms, nrnb, &vcm, wcycle, enerd,
1347 force_vir, shake_vir, total_vir, pres, constr, &nullSignaller, lastbox,
1348 nullptr, &bSumEkinhOld, (bGStat ? CGLO_GSTAT : 0) | CGLO_TEMPERATURE);
1349 wallcycle_start(wcycle, ewcUPDATE);
1350 trotter_update(ir, step, ekind, enerd, state, total_vir, mdatoms, &MassQ, trotter_seq, ettTSEQ4);
1351 /* now we know the scaling, we can compute the positions again */
1352 std::copy(cbuf.begin(), cbuf.end(), state->x.begin());
1354 upd.update_coords(*ir, step, mdatoms, state, f.view().forceWithPadding(), fcdata, ekind,
1355 M, etrtPOSITION, cr, constr != nullptr);
1356 wallcycle_stop(wcycle, ewcUPDATE);
1358 /* do we need an extra constraint here? just need to copy out of as_rvec_array(state->v.data()) to upd->xp? */
1359 /* are the small terms in the shake_vir here due
1360 * to numerical errors, or are they important
1361 * physically? I'm thinking they are just errors, but not completely sure.
1362 * For now, will call without actually constraining, constr=NULL*/
1363 upd.finish_update(*ir, mdatoms, state, wcycle, false);
1367 /* this factor or 2 correction is necessary
1368 because half of the constraint force is removed
1369 in the vv step, so we have to double it. See
1370 the Issue #1255. It is not yet clear
1371 if the factor of 2 is exact, or just a very
1372 good approximation, and this will be
1373 investigated. The next step is to see if this
1374 can be done adding a dhdl contribution from the
1375 rattle step, but this is somewhat more
1376 complicated with the current code. Will be
1377 investigated, hopefully for 4.6.3. However,
1378 this current solution is much better than
1379 having it completely wrong.
1381 enerd->term[F_DVDL_CONSTR] += 2 * dvdl_constr;
1385 enerd->term[F_DVDL_CONSTR] += dvdl_constr;
1388 if (vsite != nullptr)
1390 wallcycle_start(wcycle, ewcVSITECONSTR);
1391 vsite->construct(state->x, ir->delta_t, state->v, state->box);
1392 wallcycle_stop(wcycle, ewcVSITECONSTR);
1395 /* ############## IF NOT VV, Calculate globals HERE ############ */
1396 /* With Leap-Frog we can skip compute_globals at
1397 * non-communication steps, but we need to calculate
1398 * the kinetic energy one step before communication.
1401 // Organize to do inter-simulation signalling on steps if
1402 // and when algorithms require it.
1403 const bool doInterSimSignal = (simulationsShareState && do_per_step(step, nstSignalComm));
1405 if (bGStat || needHalfStepKineticEnergy || doInterSimSignal)
1407 // Copy coordinates when needed to stop the CM motion.
1408 if (useGpuForUpdate && !EI_VV(ir->eI) && bStopCM)
1410 stateGpu->copyCoordinatesFromGpu(state->x, AtomLocality::Local);
1411 stateGpu->waitCoordinatesReadyOnHost(AtomLocality::Local);
1413 // Since we're already communicating at this step, we
1414 // can propagate intra-simulation signals. Note that
1415 // check_nstglobalcomm has the responsibility for
1416 // choosing the value of nstglobalcomm that is one way
1417 // bGStat becomes true, so we can't get into a
1418 // situation where e.g. checkpointing can't be
1420 bool doIntraSimSignal = true;
1421 SimulationSignaller signaller(&signals, cr, ms, doInterSimSignal, doIntraSimSignal);
1423 compute_globals(gstat, cr, ir, fr, ekind, makeConstArrayRef(state->x),
1424 makeConstArrayRef(state->v), state->box, mdatoms, nrnb, &vcm,
1425 wcycle, enerd, force_vir, shake_vir, total_vir, pres, constr,
1426 &signaller, lastbox, &totalNumberOfBondedInteractions, &bSumEkinhOld,
1427 (bGStat ? CGLO_GSTAT : 0) | (!EI_VV(ir->eI) && bCalcEner ? CGLO_ENERGY : 0)
1428 | (!EI_VV(ir->eI) && bStopCM ? CGLO_STOPCM : 0)
1429 | (!EI_VV(ir->eI) ? CGLO_TEMPERATURE : 0)
1430 | (!EI_VV(ir->eI) ? CGLO_PRESSURE : 0) | CGLO_CONSTRAINT
1431 | (shouldCheckNumberOfBondedInteractions ? CGLO_CHECK_NUMBER_OF_BONDED_INTERACTIONS
1433 checkNumberOfBondedInteractions(mdlog, cr, totalNumberOfBondedInteractions,
1434 top_global, &top, makeConstArrayRef(state->x),
1435 state->box, &shouldCheckNumberOfBondedInteractions);
1436 if (!EI_VV(ir->eI) && bStopCM)
1438 process_and_stopcm_grp(fplog, &vcm, *mdatoms, makeArrayRef(state->x),
1439 makeArrayRef(state->v));
1440 inc_nrnb(nrnb, eNR_STOPCM, mdatoms->homenr);
1442 // TODO: The special case of removing CM motion should be dealt more gracefully
1443 if (useGpuForUpdate)
1445 stateGpu->copyCoordinatesToGpu(state->x, AtomLocality::Local);
1446 // Here we block until the H2D copy completes because event sync with the
1447 // force kernels that use the coordinates on the next steps is not implemented
1448 // (not because of a race on state->x being modified on the CPU while H2D is in progress).
1449 stateGpu->waitCoordinatesCopiedToDevice(AtomLocality::Local);
1450 // If the COM removal changed the velocities on the CPU, this has to be accounted for.
1451 if (vcm.mode != ecmNO)
1453 stateGpu->copyVelocitiesToGpu(state->v, AtomLocality::Local);
1460 /* ############# END CALC EKIN AND PRESSURE ################# */
1462 /* Note: this is OK, but there are some numerical precision issues with using the convergence of
1463 the virial that should probably be addressed eventually. state->veta has better properies,
1464 but what we actually need entering the new cycle is the new shake_vir value. Ideally, we could
1465 generate the new shake_vir, but test the veta value for convergence. This will take some thought. */
1467 if (ir->efep != efepNO && !EI_VV(ir->eI))
1469 /* Sum up the foreign energy and dK/dl terms for md and sd.
1470 Currently done every step so that dH/dl is correct in the .edr */
1471 accumulateKineticLambdaComponents(enerd, state->lambda, *ir->fepvals);
1474 update_pcouple_after_coordinates(fplog, step, ir, mdatoms, pres, force_vir, shake_vir,
1475 pressureCouplingMu, state, nrnb, upd.deform(), !useGpuForUpdate);
1477 const bool doBerendsenPressureCoupling =
1478 (inputrec->epc == epcBERENDSEN && do_per_step(step, inputrec->nstpcouple));
1479 const bool doCRescalePressureCoupling =
1480 (inputrec->epc == epcCRESCALE && do_per_step(step, inputrec->nstpcouple));
1482 && (doBerendsenPressureCoupling || doCRescalePressureCoupling || doParrinelloRahman))
1484 integrator->scaleCoordinates(pressureCouplingMu);
1485 if (doCRescalePressureCoupling)
1487 matrix pressureCouplingInvMu;
1488 gmx::invertBoxMatrix(pressureCouplingMu, pressureCouplingInvMu);
1489 integrator->scaleVelocities(pressureCouplingInvMu);
1491 integrator->setPbc(PbcType::Xyz, state->box);
1494 /* ################# END UPDATE STEP 2 ################# */
1495 /* #### We now have r(t+dt) and v(t+dt/2) ############# */
1497 /* The coordinates (x) were unshifted in update */
1500 /* We will not sum ekinh_old,
1501 * so signal that we still have to do it.
1503 bSumEkinhOld = TRUE;
1508 /* ######### BEGIN PREPARING EDR OUTPUT ########### */
1510 /* use the directly determined last velocity, not actually the averaged half steps */
1511 if (bTrotter && ir->eI == eiVV)
1513 enerd->term[F_EKIN] = last_ekin;
1515 enerd->term[F_ETOT] = enerd->term[F_EPOT] + enerd->term[F_EKIN];
1517 if (integratorHasConservedEnergyQuantity(ir))
1521 enerd->term[F_ECONSERVED] = enerd->term[F_ETOT] + saved_conserved_quantity;
1525 enerd->term[F_ECONSERVED] = enerd->term[F_ETOT] + NPT_energy(ir, state, &MassQ);
1528 /* ######### END PREPARING EDR OUTPUT ########### */
1534 if (fplog && do_log && bDoExpanded)
1536 /* only needed if doing expanded ensemble */
1537 PrintFreeEnergyInfoToFile(fplog, ir->fepvals, ir->expandedvals,
1538 ir->bSimTemp ? ir->simtempvals : nullptr,
1539 state_global->dfhist, state->fep_state, ir->nstlog, step);
1543 energyOutput.addDataAtEnergyStep(
1544 bDoDHDL, bCalcEnerStep, t, mdatoms->tmass, enerd, ir->fepvals,
1545 ir->expandedvals, lastbox,
1546 PTCouplingArrays{ state->boxv, state->nosehoover_xi, state->nosehoover_vxi,
1547 state->nhpres_xi, state->nhpres_vxi },
1548 state->fep_state, shake_vir, force_vir, total_vir, pres, ekind, mu_tot, constr);
1552 energyOutput.recordNonEnergyStep();
1555 gmx_bool do_dr = do_per_step(step, ir->nstdisreout);
1556 gmx_bool do_or = do_per_step(step, ir->nstorireout);
1558 if (doSimulatedAnnealing)
1560 gmx::EnergyOutput::printAnnealingTemperatures(do_log ? fplog : nullptr, groups,
1563 if (do_log || do_ene || do_dr || do_or)
1565 energyOutput.printStepToEnergyFile(mdoutf_get_fp_ene(outf), do_ene, do_dr, do_or,
1566 do_log ? fplog : nullptr, step, t,
1567 fr->fcdata.get(), awh.get());
1569 if (do_log && ir->bDoAwh && awh->hasFepLambdaDimension())
1571 const bool isInitialOutput = false;
1572 printLambdaStateToLog(fplog, state->lambda, isInitialOutput);
1577 pull_print_output(pull_work, step, t);
1580 if (do_per_step(step, ir->nstlog))
1582 if (fflush(fplog) != 0)
1584 gmx_fatal(FARGS, "Cannot flush logfile - maybe you are out of disk space?");
1590 /* Have to do this part _after_ outputting the logfile and the edr file */
1591 /* Gets written into the state at the beginning of next loop*/
1592 state->fep_state = lamnew;
1594 else if (ir->bDoAwh && awh->needForeignEnergyDifferences(step))
1596 state->fep_state = awh->fepLambdaState();
1598 /* Print the remaining wall clock time for the run */
1599 if (isMasterSimMasterRank(ms, MASTER(cr)) && (do_verbose || gmx_got_usr_signal()) && !bPMETunePrinting)
1603 fprintf(stderr, "\n");
1605 print_time(stderr, walltime_accounting, step, ir, cr);
1608 /* Ion/water position swapping.
1609 * Not done in last step since trajectory writing happens before this call
1610 * in the MD loop and exchanges would be lost anyway. */
1611 bNeedRepartition = FALSE;
1612 if ((ir->eSwapCoords != eswapNO) && (step > 0) && !bLastStep && do_per_step(step, ir->swap->nstswap))
1615 do_swapcoords(cr, step, t, ir, swap, wcycle, as_rvec_array(state->x.data()),
1616 state->box, MASTER(cr) && mdrunOptions.verbose, bRerunMD);
1618 if (bNeedRepartition && DOMAINDECOMP(cr))
1620 dd_collect_state(cr->dd, state, state_global);
1624 /* Replica exchange */
1628 bExchanged = replica_exchange(fplog, cr, ms, repl_ex, state_global, enerd, state, step, t);
1631 if ((bExchanged || bNeedRepartition) && DOMAINDECOMP(cr))
1633 dd_partition_system(fplog, mdlog, step, cr, TRUE, 1, state_global, *top_global, ir,
1634 imdSession, pull_work, state, &f, mdAtoms, &top, fr, vsite, constr,
1635 nrnb, wcycle, FALSE);
1636 shouldCheckNumberOfBondedInteractions = true;
1637 upd.setNumAtoms(state->natoms);
1643 /* ####### SET VARIABLES FOR NEXT ITERATION IF THEY STILL NEED IT ###### */
1644 /* With all integrators, except VV, we need to retain the pressure
1645 * at the current step for coupling at the next step.
1647 if ((state->flags & (1U << estPRES_PREV))
1648 && (bGStatEveryStep || (ir->nstpcouple > 0 && step % ir->nstpcouple == 0)))
1650 /* Store the pressure in t_state for pressure coupling
1651 * at the next MD step.
1653 copy_mat(pres, state->pres_prev);
1656 /* ####### END SET VARIABLES FOR NEXT ITERATION ###### */
1658 if ((membed != nullptr) && (!bLastStep))
1660 rescale_membed(step_rel, membed, as_rvec_array(state_global->x.data()));
1663 cycles = wallcycle_stop(wcycle, ewcSTEP);
1664 if (DOMAINDECOMP(cr) && wcycle)
1666 dd_cycles_add(cr->dd, cycles, ddCyclStep);
1669 /* increase the MD step number */
1676 fcReportProgress(ir->nsteps + ir->init_step, step);
1680 resetHandler->resetCounters(step, step_rel, mdlog, fplog, cr, fr->nbv.get(), nrnb,
1681 fr->pmedata, pme_loadbal, wcycle, walltime_accounting);
1683 /* If bIMD is TRUE, the master updates the IMD energy record and sends positions to VMD client */
1684 imdSession->updateEnergyRecordAndSendPositionsAndEnergies(bInteractiveMDstep, step, bCalcEner);
1686 /* End of main MD loop */
1688 /* Closing TNG files can include compressing data. Therefore it is good to do that
1689 * before stopping the time measurements. */
1690 mdoutf_tng_close(outf);
1692 /* Stop measuring walltime */
1693 walltime_accounting_end_time(walltime_accounting);
1695 if (!thisRankHasDuty(cr, DUTY_PME))
1697 /* Tell the PME only node to finish */
1698 gmx_pme_send_finish(cr);
1703 if (ir->nstcalcenergy > 0)
1705 energyOutput.printEnergyConservation(fplog, ir->simulation_part, EI_MD(ir->eI));
1707 gmx::EnergyOutput::printAnnealingTemperatures(fplog, groups, &(ir->opts));
1708 energyOutput.printAverages(fplog, groups);
1715 pme_loadbal_done(pme_loadbal, fplog, mdlog, fr->nbv->useGpu());
1718 done_shellfc(fplog, shellfc, step_rel);
1720 if (useReplicaExchange && MASTER(cr))
1722 print_replica_exchange_statistics(fplog, repl_ex);
1725 walltime_accounting_set_nsteps_done(walltime_accounting, step_rel);
1727 global_stat_destroy(gstat);