src/gromacs/mdlib/stophandler.cpp

   1 /*
   2  * This file is part of the GROMACS molecular simulation package.
   3  *
   4  * Copyright (c) 2018, by the GROMACS development team, led by
   5  * Mark Abraham, David van der Spoel, Berk Hess, and Erik Lindahl,
   6  * and including many others, as listed in the AUTHORS file in the
   7  * top-level source directory and at http://www.gromacs.org.
   8  *
   9  * GROMACS is free software; you can redistribute it and/or
  10  * modify it under the terms of the GNU Lesser General Public License
  11  * as published by the Free Software Foundation; either version 2.1
  12  * of the License, or (at your option) any later version.
  13  *
  14  * GROMACS is distributed in the hope that it will be useful,
  15  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  16  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  17  * Lesser General Public License for more details.
  18  *
  19  * You should have received a copy of the GNU Lesser General Public
  20  * License along with GROMACS; if not, see
  21  * http://www.gnu.org/licenses, or write to the Free Software Foundation,
  22  * Inc., 51 Franklin Street, Fifth Floor, Boston, MA  02110-1301  USA.
  23  *
  24  * If you want to redistribute modifications to GROMACS, please
  25  * consider that scientific software is very special. Version
  26  * control is crucial - bugs must be traceable. We will be happy to
  27  * consider code for inclusion in the official distribution, but
  28  * derived work must not be called official GROMACS. Details are found
  29  * in the README & COPYING files - if they are missing, get the
  30  * official version at http://www.gromacs.org.
  31  *
  32  * To help us fund GROMACS development, we humbly ask that you cite
  33  * the research papers on the package. Check out http://www.gromacs.org.
  34  */
  35 /*! \internal \file
  36  * \brief Defines StopHandler, a helper class and two stop conditions.
  37  *
  38  * \author Pascal Merz <pascal.merz@colorado.edu>
  39  * \ingroup module_mdlib
  40  */
  41 #include "gmxpre.h"
  42
  43 #include "stophandler.h"
  44
  45 #include "config.h"
  46
  47 #include "gromacs/compat/make_unique.h"
  48 #include "gromacs/timing/walltime_accounting.h"
  49 #include "gromacs/utility/cstringutil.h"
  50
  51 namespace gmx
  52 {
  53
  54 StopHandler::StopHandler(
  55         compat::not_null<SimulationSignal*>         signal,
  56         bool                                        simulationShareState,
  57         std::vector < std::function<StopSignal()> > stopConditions,
  58         bool                                        neverUpdateNeighborList) :
  59     signal_(*signal),
  60     stopConditions_(std::move(stopConditions)),
  61     neverUpdateNeighborlist_(neverUpdateNeighborList)
  62 {
  63     if (simulationShareState)
  64     {
  65         signal_.isLocal = false;
  66     }
  67 }
  68
  69 StopConditionSignal::StopConditionSignal(
  70         int  nstList,
  71         bool makeBinaryReproducibleSimulation,
  72         int  nstSignalComm) :
  73     handledStopCondition_(gmx_stop_cond_none),
  74     makeBinaryReproducibleSimulation_(makeBinaryReproducibleSimulation),
  75     nstSignalComm_(nstSignalComm),
  76     nstList_(nstList)
  77 {}
  78
  79 StopSignal StopConditionSignal::getSignal(FILE *fplog)
  80 {
  81     StopSignal signal = StopSignal::noSignal;
  82
  83     /* Check whether everything is still alright */
  84     if (static_cast<int>(gmx_get_stop_condition()) > handledStopCondition_)
  85     {
  86         int nsteps_stop   = -1;
  87
  88         /* this just makes signals[].sig compatible with the hack
  89            of sending signals around by MPI_Reduce together with
  90            other floats */
  91         if ((gmx_get_stop_condition() == gmx_stop_cond_next_ns) ||
  92             (makeBinaryReproducibleSimulation_ && gmx_get_stop_condition() == gmx_stop_cond_next))
  93         {
  94             /* We need at least two global communication steps to pass
  95              * around the signal. We stop at a pair-list creation step
  96              * to allow for exact continuation, when possible.
  97              */
  98             signal      = StopSignal::stopAtNextNSStep;
  99             nsteps_stop = std::max(nstList_, 2 * nstSignalComm_);
 100         }
 101         else if (gmx_get_stop_condition() == gmx_stop_cond_next)
 102         {
 103             /* Stop directly after the next global communication step.
 104              * This breaks exact continuation.
 105              */
 106             signal      = StopSignal::stopImmediately;
 107             nsteps_stop = nstSignalComm_ + 1;
 108         }
 109         if (fplog)
 110         {
 111             fprintf(fplog,
 112                     "\n\nReceived the %s signal, stopping within %d steps\n\n",
 113                     gmx_get_signal_name(), nsteps_stop);
 114             fflush(fplog);
 115         }
 116         fprintf(stderr,
 117                 "\n\nReceived the %s signal, stopping within %d steps\n\n",
 118                 gmx_get_signal_name(), nsteps_stop);
 119         fflush(stderr);
 120         handledStopCondition_ = static_cast<int>(gmx_get_stop_condition());
 121     }
 122
 123     return signal;
 124 }
 125
 126 StopConditionTime::StopConditionTime(
 127         int  nstList,
 128         real maximumHoursToRun,
 129         int  nstSignalComm) :
 130     signalSent_(false),
 131     maximumHoursToRun_(maximumHoursToRun),
 132     nstList_(nstList),
 133     nstSignalComm_(nstSignalComm),
 134     neverUpdateNeighborlist_(nstList <= 0)
 135 {}
 136
 137 StopSignal StopConditionTime::getSignal(
 138         bool                      bNS,
 139         int64_t                   step,
 140         FILE                     *fplog,
 141         gmx_walltime_accounting_t walltime_accounting)
 142 {
 143     if (signalSent_)
 144     {
 145         // We only want to send it once, but might be called again before run is terminated
 146         return StopSignal::noSignal;
 147     }
 148     if ((bNS || neverUpdateNeighborlist_) &&
 149         walltime_accounting_get_time_since_start(walltime_accounting) > maximumHoursToRun_ * 60.0 * 60.0 * 0.99)
 150     {
 151         /* Signal to terminate the run */
 152         char sbuf[STEPSTRSIZE];
 153         int  nsteps_stop = std::max(nstList_, 2 * nstSignalComm_);
 154         if (fplog)
 155         {
 156             fprintf(fplog, "\nStep %s: Run time exceeded %.3f hours, "
 157                     "will terminate the run within %d steps\n",
 158                     gmx_step_str(step, sbuf), maximumHoursToRun_ * 0.99, nsteps_stop);
 159         }
 160         fprintf(stderr, "\nStep %s: Run time exceeded %.3f hours, "
 161                 "will terminate the run within %d steps\n",
 162                 gmx_step_str(step, sbuf), maximumHoursToRun_ * 0.99, nsteps_stop);
 163         signalSent_ = true;
 164         return StopSignal::stopAtNextNSStep;
 165     }
 166     return StopSignal::noSignal;
 167 }
 168
 169 void StopHandlerBuilder::registerStopCondition(std::function<StopSignal()> stopCondition)
 170 {
 171     stopConditions_.emplace_back(std::move(stopCondition));
 172 };
 173
 174 std::unique_ptr<StopHandler> StopHandlerBuilder::getStopHandlerMD (
 175         compat::not_null<SimulationSignal*> signal,
 176         bool                                simulationShareState,
 177         bool                                isMaster,
 178         int                                 nstList,
 179         bool                                makeBinaryReproducibleSimulation,
 180         int                                 nstSignalComm,
 181         real                                maximumHoursToRun,
 182         bool                                neverUpdateNeighborList,
 183         FILE                               *fplog,
 184         const int64_t                      &step,
 185         const gmx_bool                     &bNS,
 186         gmx_walltime_accounting_t           walltime_accounting)
 187 {
 188     if (!GMX_THREAD_MPI || isMaster)
 189     {
 190         // TODO: Use unique_ptr once we switch to C++14 (unique_ptr can not easily be
 191         //       captured in lambda functions in C++11)
 192         auto stopConditionSignal = std::make_shared<StopConditionSignal>(
 193                     nstList, makeBinaryReproducibleSimulation, nstSignalComm);
 194         registerStopCondition(
 195                 [stopConditionSignal, fplog]()
 196                 {return stopConditionSignal->getSignal(fplog); });
 197     }
 198
 199     if (isMaster && maximumHoursToRun > 0)
 200     {
 201         // TODO: Use unique_ptr once we switch to C++14 (unique_ptr can not easily be
 202         //       captured in lambda functions in C++11)
 203         auto stopConditionTime = std::make_shared<StopConditionTime>(
 204                     nstList, maximumHoursToRun, nstSignalComm);
 205         registerStopCondition(
 206                 [stopConditionTime, &bNS, &step, fplog, walltime_accounting]()
 207                 {return stopConditionTime->getSignal(bNS, step, fplog, walltime_accounting); });
 208     }
 209
 210     return compat::make_unique<StopHandler>(
 211             signal, simulationShareState, stopConditions_, neverUpdateNeighborList);
 212 }
 213
 214 } // namespace gmx