src/rnemd/RNEMD.cpp

/*
 * Copyright (c) 2005 The University of Notre Dame. All Rights Reserved.
 *
 * The University of Notre Dame grants you ("Licensee") a
 * non-exclusive, royalty free, license to use, modify and
 * redistribute this software in source and binary code form, provided
 * that the following conditions are met:
 *
 * 1. Redistributions of source code must retain the above copyright
 *    notice, this list of conditions and the following disclaimer.
 *
 * 2. Redistributions in binary form must reproduce the above copyright
 *    notice, this list of conditions and the following disclaimer in the
 *    documentation and/or other materials provided with the
 *    distribution.
 *
 * This software is provided "AS IS," without a warranty of any
 * kind. All express or implied conditions, representations and
 * warranties, including any implied warranty of merchantability,
 * fitness for a particular purpose or non-infringement, are hereby
 * excluded.  The University of Notre Dame and its licensors shall not
 * be liable for any damages suffered by licensee as a result of
 * using, modifying or distributing the software or its
 * derivatives. In no event will the University of Notre Dame or its
 * licensors be liable for any lost revenue, profit or data, or for
 * direct, indirect, special, consequential, incidental or punitive
 * damages, however caused and regardless of the theory of liability,
 * arising out of the use of or inability to use software, even if the
 * University of Notre Dame has been advised of the possibility of
 * such damages.
 *
 * SUPPORT OPEN SCIENCE!  If you use OpenMD or its source code in your
 * research, please cite the appropriate papers when you publish your
 * work.  Good starting points are:
 *                                                                      
 * [1]  Meineke, et al., J. Comp. Chem. 26, 252-271 (2005).             
 * [2]  Fennell & Gezelter, J. Chem. Phys. 124, 234104 (2006).          
 * [3]  Sun, Lin & Gezelter, J. Chem. Phys. 128, 234107 (2008).          
 * [4]  Vardeman & Gezelter, in progress (2009).                        
 */
#ifdef IS_MPI
#include <mpi.h>
#endif

#include <cmath>
#include <sstream>
#include <string>

#include "rnemd/RNEMD.hpp"
#include "math/Vector3.hpp"
#include "math/Vector.hpp"
#include "math/SquareMatrix3.hpp"
#include "math/Polynomial.hpp"
#include "primitives/Molecule.hpp"
#include "primitives/StuntDouble.hpp"
#include "utils/PhysicalConstants.hpp"
#include "utils/Tuple.hpp"
#include "brains/Thermo.hpp"
#include "math/ConvexHull.hpp"

#ifdef _MSC_VER
#define isnan(x) _isnan((x))
#define isinf(x) (!_finite(x) && !_isnan(x))
#endif

#define HONKING_LARGE_VALUE 1.0e10

using namespace std;
namespace OpenMD {
  
  RNEMD::RNEMD(SimInfo* info) : info_(info), evaluator_(info), seleMan_(info), 
                                evaluatorA_(info), seleManA_(info), 
                                commonA_(info), evaluatorB_(info), 
                                seleManB_(info), commonB_(info), 
                                hasData_(false), hasDividingArea_(false),
                                usePeriodicBoundaryConditions_(info->getSimParams()->getUsePeriodicBoundaryConditions()) {

    trialCount_ = 0;
    failTrialCount_ = 0;
    failRootCount_ = 0;

    Globals* simParams = info->getSimParams();
    RNEMDParameters* rnemdParams = simParams->getRNEMDParameters();

    doRNEMD_ = rnemdParams->getUseRNEMD();
    if (!doRNEMD_) return;

    stringToMethod_["Swap"]  = rnemdSwap;
    stringToMethod_["NIVS"]  = rnemdNIVS;
    stringToMethod_["VSS"]   = rnemdVSS;

    stringToFluxType_["KE"]  = rnemdKE;
    stringToFluxType_["Px"]  = rnemdPx;
    stringToFluxType_["Py"]  = rnemdPy;
    stringToFluxType_["Pz"]  = rnemdPz;
    stringToFluxType_["Pvector"]  = rnemdPvector;
    stringToFluxType_["Lx"]  = rnemdLx;
    stringToFluxType_["Ly"]  = rnemdLy;
    stringToFluxType_["Lz"]  = rnemdLz;
    stringToFluxType_["Lvector"]  = rnemdLvector;
    stringToFluxType_["KE+Px"]  = rnemdKePx;
    stringToFluxType_["KE+Py"]  = rnemdKePy;
    stringToFluxType_["KE+Pvector"]  = rnemdKePvector;
    stringToFluxType_["KE+Lx"]  = rnemdKeLx;
    stringToFluxType_["KE+Ly"]  = rnemdKeLy;
    stringToFluxType_["KE+Lz"]  = rnemdKeLz;
    stringToFluxType_["KE+Lvector"]  = rnemdKeLvector;

    runTime_ = simParams->getRunTime();
    statusTime_ = simParams->getStatusTime();

    const string methStr = rnemdParams->getMethod();
    bool hasFluxType = rnemdParams->haveFluxType();

    rnemdObjectSelection_ = rnemdParams->getObjectSelection();

    string fluxStr;
    if (hasFluxType) {
      fluxStr = rnemdParams->getFluxType();
    } else {
      sprintf(painCave.errMsg, 
              "RNEMD: No fluxType was set in the md file.  This parameter,\n"
              "\twhich must be one of the following values:\n"
              "\tKE, Px, Py, Pz, Pvector, Lx, Ly, Lz, Lvector,\n"
              "\tKE+Px, KE+Py, KE+Pvector, KE+Lx, KE+Ly, KE+Lz, KE+Lvector\n"
              "\tmust be set to use RNEMD\n");
      painCave.isFatal = 1;
      painCave.severity = OPENMD_ERROR;
      simError();
    }

    bool hasKineticFlux = rnemdParams->haveKineticFlux();
    bool hasMomentumFlux = rnemdParams->haveMomentumFlux();
    bool hasMomentumFluxVector = rnemdParams->haveMomentumFluxVector();
    bool hasAngularMomentumFlux = rnemdParams->haveAngularMomentumFlux();
    bool hasAngularMomentumFluxVector = rnemdParams->haveAngularMomentumFluxVector();
    hasSelectionA_ = rnemdParams->haveSelectionA();
    hasSelectionB_ = rnemdParams->haveSelectionB();
    bool hasSlabWidth = rnemdParams->haveSlabWidth();
    bool hasSlabACenter = rnemdParams->haveSlabACenter();
    bool hasSlabBCenter = rnemdParams->haveSlabBCenter();
    bool hasSphereARadius = rnemdParams->haveSphereARadius();
    hasSphereBRadius_ = rnemdParams->haveSphereBRadius();
    bool hasCoordinateOrigin = rnemdParams->haveCoordinateOrigin();
    bool hasOutputFileName = rnemdParams->haveOutputFileName();
    bool hasOutputFields = rnemdParams->haveOutputFields();
    
    map<string, RNEMDMethod>::iterator i;
    i = stringToMethod_.find(methStr);
    if (i != stringToMethod_.end()) 
      rnemdMethod_ = i->second;
    else {
      sprintf(painCave.errMsg, 
              "RNEMD: The current method,\n"
              "\t\t%s is not one of the recognized\n"
              "\texchange methods: Swap, NIVS, or VSS\n",
              methStr.c_str());
      painCave.isFatal = 1;
      painCave.severity = OPENMD_ERROR;
      simError();
    }

    map<string, RNEMDFluxType>::iterator j;
    j = stringToFluxType_.find(fluxStr);
    if (j != stringToFluxType_.end()) 
      rnemdFluxType_ = j->second;
    else {
      sprintf(painCave.errMsg, 
              "RNEMD: The current fluxType,\n"
              "\t\t%s\n"
              "\tis not one of the recognized flux types.\n",
              fluxStr.c_str());
      painCave.isFatal = 1;
      painCave.severity = OPENMD_ERROR;
      simError();
    }

    bool methodFluxMismatch = false;
    bool hasCorrectFlux = false;
    switch(rnemdMethod_) {
    case rnemdSwap:
      switch (rnemdFluxType_) {
      case rnemdKE:
        hasCorrectFlux = hasKineticFlux;
        break;
      case rnemdPx:
      case rnemdPy:
      case rnemdPz:
        hasCorrectFlux = hasMomentumFlux;
        break;
      default :
        methodFluxMismatch = true;
        break;
      }
      break;
    case rnemdNIVS:
      switch (rnemdFluxType_) {
      case rnemdKE:
      case rnemdRotKE:
      case rnemdFullKE:
        hasCorrectFlux = hasKineticFlux;
        break;
      case rnemdPx:
      case rnemdPy:
      case rnemdPz:
        hasCorrectFlux = hasMomentumFlux;
        break;
      case rnemdKePx:
      case rnemdKePy:
        hasCorrectFlux = hasMomentumFlux && hasKineticFlux;
        break;
      default:
        methodFluxMismatch = true;
        break;
      }
      break;
    case rnemdVSS:
      switch (rnemdFluxType_) {
      case rnemdKE:
      case rnemdRotKE:
      case rnemdFullKE:
        hasCorrectFlux = hasKineticFlux;
        break;
      case rnemdPx:
      case rnemdPy:
      case rnemdPz:
        hasCorrectFlux = hasMomentumFlux;
        break;
      case rnemdLx:
      case rnemdLy:
      case rnemdLz:
        hasCorrectFlux = hasAngularMomentumFlux;
        break;
      case rnemdPvector:
        hasCorrectFlux = hasMomentumFluxVector;
        break;
      case rnemdLvector:
        hasCorrectFlux = hasAngularMomentumFluxVector;
        break;
      case rnemdKePx:
      case rnemdKePy:
        hasCorrectFlux = hasMomentumFlux && hasKineticFlux;
        break;
      case rnemdKeLx:
      case rnemdKeLy:
      case rnemdKeLz:
        hasCorrectFlux = hasAngularMomentumFlux && hasKineticFlux;
        break;
      case rnemdKePvector:
        hasCorrectFlux = hasMomentumFluxVector && hasKineticFlux;
        break;
      case rnemdKeLvector:
        hasCorrectFlux = hasAngularMomentumFluxVector && hasKineticFlux;
        break;
      default:
        methodFluxMismatch = true;
        break;
      }
    default:
      break;
    }

    if (methodFluxMismatch) {
      sprintf(painCave.errMsg, 
              "RNEMD: The current method,\n"
              "\t\t%s\n"
              "\tcannot be used with the current flux type, %s\n",
              methStr.c_str(), fluxStr.c_str());
      painCave.isFatal = 1;
      painCave.severity = OPENMD_ERROR;
      simError();        
    } 
    if (!hasCorrectFlux) {
      sprintf(painCave.errMsg, 
              "RNEMD: The current method, %s, and flux type, %s,\n"
              "\tdid not have the correct flux value specified. Options\n"
              "\tinclude: kineticFlux, momentumFlux, angularMomentumFlux,\n"
              "\tmomentumFluxVector, and angularMomentumFluxVector.\n",
              methStr.c_str(), fluxStr.c_str());
      painCave.isFatal = 1;
      painCave.severity = OPENMD_ERROR;
      simError();        
    } 

    if (hasKineticFlux) {
      // convert the kcal / mol / Angstroms^2 / fs values in the md file 
      // into  amu / fs^3:
      kineticFlux_ = rnemdParams->getKineticFlux() 
        * PhysicalConstants::energyConvert;
    } else {
      kineticFlux_ = 0.0;
    }
    if (hasMomentumFluxVector) {
      std::vector<RealType> mf = rnemdParams->getMomentumFluxVector();
      if (mf.size() != 3) {
          sprintf(painCave.errMsg,
                  "RNEMD: Incorrect number of parameters specified for momentumFluxVector.\n"
                  "\tthere should be 3 parameters, but %lu were specified.\n", 
                  mf.size());
          painCave.isFatal = 1;
          simError();      
      }
      momentumFluxVector_.x() = mf[0];
      momentumFluxVector_.y() = mf[1];
      momentumFluxVector_.z() = mf[2];
    } else {
      momentumFluxVector_ = V3Zero;
      if (hasMomentumFlux) {
        RealType momentumFlux = rnemdParams->getMomentumFlux();
        switch (rnemdFluxType_) {
        case rnemdPx:
          momentumFluxVector_.x() = momentumFlux;
          break;
        case rnemdPy:
          momentumFluxVector_.y() = momentumFlux;
          break;
        case rnemdPz:
          momentumFluxVector_.z() = momentumFlux;
          break;
        case rnemdKePx:
          momentumFluxVector_.x() = momentumFlux;
          break;
        case rnemdKePy:
          momentumFluxVector_.y() = momentumFlux;
          break;
        default:
          break;
        }
      }
      if (hasAngularMomentumFluxVector) {
        std::vector<RealType> amf = rnemdParams->getAngularMomentumFluxVector();
        if (amf.size() != 3) {
          sprintf(painCave.errMsg,
                  "RNEMD: Incorrect number of parameters specified for angularMomentumFluxVector.\n"
                  "\tthere should be 3 parameters, but %lu were specified.\n", 
                  amf.size());
          painCave.isFatal = 1;
          simError();      
        }
        angularMomentumFluxVector_.x() = amf[0];
        angularMomentumFluxVector_.y() = amf[1];
        angularMomentumFluxVector_.z() = amf[2];
      } else {
        angularMomentumFluxVector_ = V3Zero;
        if (hasAngularMomentumFlux) {
          RealType angularMomentumFlux = rnemdParams->getAngularMomentumFlux();
          switch (rnemdFluxType_) {
          case rnemdLx:
            angularMomentumFluxVector_.x() = angularMomentumFlux;
            break;
          case rnemdLy:
            angularMomentumFluxVector_.y() = angularMomentumFlux;
            break;
          case rnemdLz:
            angularMomentumFluxVector_.z() = angularMomentumFlux;
            break;
          case rnemdKeLx:
            angularMomentumFluxVector_.x() = angularMomentumFlux;
            break;
          case rnemdKeLy:
            angularMomentumFluxVector_.y() = angularMomentumFlux;
            break;
          case rnemdKeLz:
            angularMomentumFluxVector_.z() = angularMomentumFlux;
            break;
          default:
            break;
          }
        }        
      }

      if (hasCoordinateOrigin) {
        std::vector<RealType> co = rnemdParams->getCoordinateOrigin();
        if (co.size() != 3) {
          sprintf(painCave.errMsg,
                  "RNEMD: Incorrect number of parameters specified for coordinateOrigin.\n"
                  "\tthere should be 3 parameters, but %lu were specified.\n", 
                  co.size());
          painCave.isFatal = 1;
          simError();      
        }
        coordinateOrigin_.x() = co[0];
        coordinateOrigin_.y() = co[1];
        coordinateOrigin_.z() = co[2];
      } else {
        coordinateOrigin_ = V3Zero;
      }

      // do some sanity checking

      int selectionCount = seleMan_.getSelectionCount();

      int nIntegrable = info->getNGlobalIntegrableObjects();

      if (selectionCount > nIntegrable) {
        sprintf(painCave.errMsg, 
                "RNEMD: The current objectSelection,\n"
                "\t\t%s\n"
                "\thas resulted in %d selected objects.  However,\n"
                "\tthe total number of integrable objects in the system\n"
                "\tis only %d.  This is almost certainly not what you want\n"
                "\tto do.  A likely cause of this is forgetting the _RB_0\n"
                "\tselector in the selection script!\n", 
                rnemdObjectSelection_.c_str(), 
                selectionCount, nIntegrable);
        painCave.isFatal = 0;
        painCave.severity = OPENMD_WARNING;
        simError();
      }

      areaAccumulator_ = new Accumulator();

      nBins_ = rnemdParams->getOutputBins();
      binWidth_ = rnemdParams->getOutputBinWidth();

      data_.resize(RNEMD::ENDINDEX);
      OutputData z;
      z.units =  "Angstroms";
      z.title =  "Z";
      z.dataType = "RealType";
      z.accumulator.reserve(nBins_);
      for (int i = 0; i < nBins_; i++) 
        z.accumulator.push_back( new Accumulator() );
      data_[Z] = z;
      outputMap_["Z"] =  Z;

      OutputData r;
      r.units =  "Angstroms";
      r.title =  "R";
      r.dataType = "RealType";
      r.accumulator.reserve(nBins_);
      for (int i = 0; i < nBins_; i++) 
        r.accumulator.push_back( new Accumulator() );
      data_[R] = r;
      outputMap_["R"] =  R;

      OutputData temperature;
      temperature.units =  "K";
      temperature.title =  "Temperature";
      temperature.dataType = "RealType";
      temperature.accumulator.reserve(nBins_);
      for (int i = 0; i < nBins_; i++) 
        temperature.accumulator.push_back( new Accumulator() );
      data_[TEMPERATURE] = temperature;
      outputMap_["TEMPERATURE"] =  TEMPERATURE;

      OutputData velocity;
      velocity.units = "angstroms/fs";
      velocity.title =  "Velocity";  
      velocity.dataType = "Vector3d";
      velocity.accumulator.reserve(nBins_);
      for (int i = 0; i < nBins_; i++) 
        velocity.accumulator.push_back( new VectorAccumulator() );
      data_[VELOCITY] = velocity;
      outputMap_["VELOCITY"] = VELOCITY;

      OutputData angularVelocity;
      angularVelocity.units = "angstroms^2/fs";
      angularVelocity.title =  "AngularVelocity";  
      angularVelocity.dataType = "Vector3d";
      angularVelocity.accumulator.reserve(nBins_);
      for (int i = 0; i < nBins_; i++) 
        angularVelocity.accumulator.push_back( new VectorAccumulator() );
      data_[ANGULARVELOCITY] = angularVelocity;
      outputMap_["ANGULARVELOCITY"] = ANGULARVELOCITY;

      OutputData density;
      density.units =  "g cm^-3";
      density.title =  "Density";
      density.dataType = "RealType";
      density.accumulator.reserve(nBins_);
      for (int i = 0; i < nBins_; i++) 
        density.accumulator.push_back( new Accumulator() );
      data_[DENSITY] = density;
      outputMap_["DENSITY"] =  DENSITY;

      if (hasOutputFields) {
        parseOutputFileFormat(rnemdParams->getOutputFields());
      } else {
        if (usePeriodicBoundaryConditions_)
          outputMask_.set(Z);
        else
          outputMask_.set(R);
        switch (rnemdFluxType_) {
        case rnemdKE:
        case rnemdRotKE:
        case rnemdFullKE:
          outputMask_.set(TEMPERATURE);
          break;
        case rnemdPx:
        case rnemdPy:
          outputMask_.set(VELOCITY);
          break;
        case rnemdPz:        
        case rnemdPvector:
          outputMask_.set(VELOCITY);
          outputMask_.set(DENSITY);
          break;
        case rnemdLx:
        case rnemdLy:
        case rnemdLz:
        case rnemdLvector:
          outputMask_.set(ANGULARVELOCITY);
          break;
        case rnemdKeLx:
        case rnemdKeLy:
        case rnemdKeLz:
        case rnemdKeLvector:
          outputMask_.set(TEMPERATURE);
          outputMask_.set(ANGULARVELOCITY);
          break;
        case rnemdKePx:
        case rnemdKePy:
          outputMask_.set(TEMPERATURE);
          outputMask_.set(VELOCITY);
          break;
        case rnemdKePvector:
          outputMask_.set(TEMPERATURE);
          outputMask_.set(VELOCITY);
          outputMask_.set(DENSITY);        
          break;
        default:
          break;
        }
      }
      
      if (hasOutputFileName) {
        rnemdFileName_ = rnemdParams->getOutputFileName();
      } else {
        rnemdFileName_ = getPrefix(info->getFinalConfigFileName()) + ".rnemd";
      }          

      exchangeTime_ = rnemdParams->getExchangeTime();

      Snapshot* currentSnap_ = info->getSnapshotManager()->getCurrentSnapshot();
      // total exchange sums are zeroed out at the beginning:

      kineticExchange_ = 0.0;
      momentumExchange_ = V3Zero;
      angularMomentumExchange_ = V3Zero;

      std::ostringstream selectionAstream;
      std::ostringstream selectionBstream;
    
      if (hasSelectionA_) {
        selectionA_ = rnemdParams->getSelectionA();
      } else {
        if (usePeriodicBoundaryConditions_) {     
          Mat3x3d hmat = currentSnap_->getHmat();
        
          if (hasSlabWidth) 
            slabWidth_ = rnemdParams->getSlabWidth();
          else
            slabWidth_ = hmat(2,2) / 10.0;
        
          if (hasSlabACenter) 
            slabACenter_ = rnemdParams->getSlabACenter();
          else 
            slabACenter_ = 0.0;
        
          selectionAstream << "select wrappedz > " 
                           << slabACenter_ - 0.5*slabWidth_ 
                           <<  " && wrappedz < "
                           << slabACenter_ + 0.5*slabWidth_;
          selectionA_ = selectionAstream.str();
        } else {
          if (hasSphereARadius) 
            sphereARadius_ = rnemdParams->getSphereARadius();
          else {
            // use an initial guess to the size of the inner slab to be 1/10 the
            // radius of an approximately spherical hull:
            Thermo thermo(info);
            RealType hVol = thermo.getHullVolume();
            sphereARadius_ = 0.1 * pow((3.0 * hVol / (4.0 * M_PI)), 1.0/3.0);
          }
          selectionAstream << "select r < " << sphereARadius_;
          selectionA_ = selectionAstream.str();
        }
      }
    
      if (hasSelectionB_) {
        selectionB_ = rnemdParams->getSelectionB();

      } else {
        if (usePeriodicBoundaryConditions_) {     
          Mat3x3d hmat = currentSnap_->getHmat();
        
          if (hasSlabWidth) 
            slabWidth_ = rnemdParams->getSlabWidth();
          else
            slabWidth_ = hmat(2,2) / 10.0;
        
          if (hasSlabBCenter) 
            slabBCenter_ = rnemdParams->getSlabBCenter();
          else 
            slabBCenter_ = hmat(2,2) / 2.0;
        
          selectionBstream << "select wrappedz > " 
                           << slabBCenter_ - 0.5*slabWidth_ 
                           <<  " && wrappedz < "
                           << slabBCenter_ + 0.5*slabWidth_;
          selectionB_ = selectionBstream.str();
        } else {
          if (hasSphereBRadius_) {
            sphereBRadius_ = rnemdParams->getSphereBRadius();
            selectionBstream << "select r > " << sphereBRadius_;
            selectionB_ = selectionBstream.str();
          } else {
            selectionB_ = "select hull";
            BisHull_ = true;
            hasSelectionB_ = true;
          }
        }
      }
    }

    // object evaluator:
    evaluator_.loadScriptString(rnemdObjectSelection_);
    seleMan_.setSelectionSet(evaluator_.evaluate());
    evaluatorA_.loadScriptString(selectionA_);
    evaluatorB_.loadScriptString(selectionB_);
    seleManA_.setSelectionSet(evaluatorA_.evaluate());
    seleManB_.setSelectionSet(evaluatorB_.evaluate());
    commonA_ = seleManA_ & seleMan_;
    commonB_ = seleManB_ & seleMan_;  
  }
  
    
  RNEMD::~RNEMD() {
    if (!doRNEMD_) return;
#ifdef IS_MPI
    if (worldRank == 0) {
#endif

      writeOutputFile();

      rnemdFile_.close();
      
#ifdef IS_MPI
    }
#endif

    // delete all of the objects we created:
    delete areaAccumulator_;    
    data_.clear();
  }
  
  void RNEMD::doSwap(SelectionManager& smanA, SelectionManager& smanB) {
    if (!doRNEMD_) return;
    int selei;
    int selej;

    Snapshot* currentSnap_ = info_->getSnapshotManager()->getCurrentSnapshot();
    Mat3x3d hmat = currentSnap_->getHmat();

    StuntDouble* sd;

    RealType min_val;
    int min_found = 0;   
    StuntDouble* min_sd;

    RealType max_val;
    int max_found = 0;
    StuntDouble* max_sd;

    for (sd = seleManA_.beginSelected(selei); sd != NULL; 
         sd = seleManA_.nextSelected(selei)) {

      Vector3d pos = sd->getPos();
      
      // wrap the stuntdouble's position back into the box:
      
      if (usePeriodicBoundaryConditions_)
        currentSnap_->wrapVector(pos);
      
      RealType mass = sd->getMass();
      Vector3d vel = sd->getVel();
      RealType value;
      
      switch(rnemdFluxType_) {
      case rnemdKE :
        
        value = mass * vel.lengthSquare();
        
        if (sd->isDirectional()) {
          Vector3d angMom = sd->getJ();
          Mat3x3d I = sd->getI();
          
          if (sd->isLinear()) {
            int i = sd->linearAxis();
            int j = (i + 1) % 3;
            int k = (i + 2) % 3;
            value += angMom[j] * angMom[j] / I(j, j) + 
              angMom[k] * angMom[k] / I(k, k);
          } else {                        
            value += angMom[0]*angMom[0]/I(0, 0) 
              + angMom[1]*angMom[1]/I(1, 1) 
              + angMom[2]*angMom[2]/I(2, 2);
          }
        } //angular momenta exchange enabled
        value *= 0.5;
        break;
      case rnemdPx :
        value = mass * vel[0];
        break;
      case rnemdPy :
        value = mass * vel[1];
        break;
      case rnemdPz :
        value = mass * vel[2];
        break;
      default :
        break;
      }
      if (!max_found) {
        max_val = value;
        max_sd = sd;
        max_found = 1;
      } else {
        if (max_val < value) {
          max_val = value;
          max_sd = sd;
        }
      }   
    }
        
    for (sd = seleManB_.beginSelected(selej); sd != NULL; 
         sd = seleManB_.nextSelected(selej)) {

      Vector3d pos = sd->getPos();
      
      // wrap the stuntdouble's position back into the box:
      
      if (usePeriodicBoundaryConditions_)
        currentSnap_->wrapVector(pos);
      
      RealType mass = sd->getMass();
      Vector3d vel = sd->getVel();
      RealType value;
      
      switch(rnemdFluxType_) {
      case rnemdKE :
        
        value = mass * vel.lengthSquare();
        
        if (sd->isDirectional()) {
          Vector3d angMom = sd->getJ();
          Mat3x3d I = sd->getI();
          
          if (sd->isLinear()) {
            int i = sd->linearAxis();
            int j = (i + 1) % 3;
            int k = (i + 2) % 3;
            value += angMom[j] * angMom[j] / I(j, j) + 
              angMom[k] * angMom[k] / I(k, k);
          } else {                        
            value += angMom[0]*angMom[0]/I(0, 0) 
              + angMom[1]*angMom[1]/I(1, 1) 
              + angMom[2]*angMom[2]/I(2, 2);
          }
        } //angular momenta exchange enabled
        value *= 0.5;
        break;
      case rnemdPx :
        value = mass * vel[0];
        break;
      case rnemdPy :
        value = mass * vel[1];
        break;
      case rnemdPz :
        value = mass * vel[2];
        break;
      default :
        break;
      }
      
      if (!min_found) {
        min_val = value;
        min_sd = sd;
        min_found = 1;
      } else {
        if (min_val > value) {
          min_val = value;
          min_sd = sd;
        }
      }
    }
    
#ifdef IS_MPI    
    int worldRank;
    MPI_Comm_rank( MPI_COMM_WORLD, &worldRank);
         
    int my_min_found = min_found;
    int my_max_found = max_found;

    // Even if we didn't find a minimum, did someone else?
    MPI_Allreduce(&my_min_found, &min_found, 1, MPI_INT, MPI_LOR, 
                  MPI_COMM_WORLD);
    // Even if we didn't find a maximum, did someone else?
    MPI_Allreduce(&my_max_found, &max_found, 1, MPI_INT, MPI_LOR, 
                  MPI_COMM_WORLD);
#endif

    if (max_found && min_found) {

#ifdef IS_MPI
      struct {
        RealType val;
        int rank;
      } max_vals, min_vals;
      
      if (my_min_found) {
        min_vals.val = min_val;
      } else {
        min_vals.val = HONKING_LARGE_VALUE;
      }
      min_vals.rank = worldRank;    
      
      // Who had the minimum?
      MPI_Allreduce(&min_vals, &min_vals, 
                    1, MPI_REALTYPE_INT, MPI_MINLOC, MPI_COMM_WORLD);
      min_val = min_vals.val;
      
      if (my_max_found) {
        max_vals.val = max_val;
      } else {
        max_vals.val = -HONKING_LARGE_VALUE;
      }
      max_vals.rank = worldRank;    
      
      // Who had the maximum?
      MPI_Allreduce(&max_vals, &max_vals, 
                    1, MPI_REALTYPE_INT, MPI_MAXLOC, MPI_COMM_WORLD);
      max_val = max_vals.val;
#endif
      
      if (min_val < max_val) {
        
#ifdef IS_MPI       
        if (max_vals.rank == worldRank && min_vals.rank == worldRank) {
          // I have both maximum and minimum, so proceed like a single
          // processor version:
#endif

          Vector3d min_vel = min_sd->getVel();
          Vector3d max_vel = max_sd->getVel();
          RealType temp_vel;
          
          switch(rnemdFluxType_) {
          case rnemdKE :
            min_sd->setVel(max_vel);
            max_sd->setVel(min_vel);
            if (min_sd->isDirectional() && max_sd->isDirectional()) {
              Vector3d min_angMom = min_sd->getJ();
              Vector3d max_angMom = max_sd->getJ();
              min_sd->setJ(max_angMom);
              max_sd->setJ(min_angMom);
            }//angular momenta exchange enabled
            //assumes same rigid body identity
            break;
          case rnemdPx :
            temp_vel = min_vel.x();
            min_vel.x() = max_vel.x();
            max_vel.x() = temp_vel;
            min_sd->setVel(min_vel);
            max_sd->setVel(max_vel);
            break;
          case rnemdPy :
            temp_vel = min_vel.y();
            min_vel.y() = max_vel.y();
            max_vel.y() = temp_vel;
            min_sd->setVel(min_vel);
            max_sd->setVel(max_vel);
            break;
          case rnemdPz :
            temp_vel = min_vel.z();
            min_vel.z() = max_vel.z();
            max_vel.z() = temp_vel;
            min_sd->setVel(min_vel);
            max_sd->setVel(max_vel);
            break;
          default :
            break;
          }

#ifdef IS_MPI
          // the rest of the cases only apply in parallel simulations:
        } else if (max_vals.rank == worldRank) {
          // I had the max, but not the minimum
          
          Vector3d min_vel;
          Vector3d max_vel = max_sd->getVel();
          MPI_Status status;

          // point-to-point swap of the velocity vector
          MPI_Sendrecv(max_vel.getArrayPointer(), 3, MPI_REALTYPE,
                       min_vals.rank, 0, 
                       min_vel.getArrayPointer(), 3, MPI_REALTYPE,
                       min_vals.rank, 0, MPI_COMM_WORLD, &status);
          
          switch(rnemdFluxType_) {
          case rnemdKE :
            max_sd->setVel(min_vel);
            //angular momenta exchange enabled
            if (max_sd->isDirectional()) {
              Vector3d min_angMom;
              Vector3d max_angMom = max_sd->getJ();
              
              // point-to-point swap of the angular momentum vector
              MPI_Sendrecv(max_angMom.getArrayPointer(), 3, 
                           MPI_REALTYPE, min_vals.rank, 1, 
                           min_angMom.getArrayPointer(), 3, 
                           MPI_REALTYPE, min_vals.rank, 1, 
                           MPI_COMM_WORLD, &status);
              
              max_sd->setJ(min_angMom);
            }
            break;
          case rnemdPx :
            max_vel.x() = min_vel.x();
            max_sd->setVel(max_vel);
            break;
          case rnemdPy :
            max_vel.y() = min_vel.y();
            max_sd->setVel(max_vel);
            break;
          case rnemdPz :
            max_vel.z() = min_vel.z();
            max_sd->setVel(max_vel);
            break;
          default :
            break;
          }
        } else if (min_vals.rank == worldRank) {
          // I had the minimum but not the maximum:
          
          Vector3d max_vel;
          Vector3d min_vel = min_sd->getVel();
          MPI_Status status;
          
          // point-to-point swap of the velocity vector
          MPI_Sendrecv(min_vel.getArrayPointer(), 3, MPI_REALTYPE,
                       max_vals.rank, 0, 
                       max_vel.getArrayPointer(), 3, MPI_REALTYPE,
                       max_vals.rank, 0, MPI_COMM_WORLD, &status);
          
          switch(rnemdFluxType_) {
          case rnemdKE :
            min_sd->setVel(max_vel);
            //angular momenta exchange enabled
            if (min_sd->isDirectional()) {
              Vector3d min_angMom = min_sd->getJ();
              Vector3d max_angMom;
              
              // point-to-point swap of the angular momentum vector
              MPI_Sendrecv(min_angMom.getArrayPointer(), 3, 
                           MPI_REALTYPE, max_vals.rank, 1, 
                           max_angMom.getArrayPointer(), 3, 
                           MPI_REALTYPE, max_vals.rank, 1, 
                           MPI_COMM_WORLD, &status);
              
              min_sd->setJ(max_angMom);
            }
            break;
          case rnemdPx :
            min_vel.x() = max_vel.x();
            min_sd->setVel(min_vel);
            break;
          case rnemdPy :
            min_vel.y() = max_vel.y();
            min_sd->setVel(min_vel);
            break;
          case rnemdPz :
            min_vel.z() = max_vel.z();
            min_sd->setVel(min_vel);
            break;
          default :
            break;
          }
        }
#endif
        
        switch(rnemdFluxType_) {
        case rnemdKE:
          kineticExchange_ += max_val - min_val;
          break;
        case rnemdPx:
          momentumExchange_.x() += max_val - min_val;
          break;
        case rnemdPy:
          momentumExchange_.y() += max_val - min_val;
          break;
        case rnemdPz:
          momentumExchange_.z() += max_val - min_val;
          break;
        default:
          break;
        }
      } else {        
        sprintf(painCave.errMsg, 
                "RNEMD::doSwap exchange NOT performed because min_val > max_val\n");
        painCave.isFatal = 0;
        painCave.severity = OPENMD_INFO;
        simError();        
        failTrialCount_++;
      }
    } else {
      sprintf(painCave.errMsg, 
              "RNEMD::doSwap exchange NOT performed because selected object\n"
              "\twas not present in at least one of the two slabs.\n");
      painCave.isFatal = 0;
      painCave.severity = OPENMD_INFO;
      simError();        
      failTrialCount_++;
    }    
  }
  
  void RNEMD::doNIVS(SelectionManager& smanA, SelectionManager& smanB) {
    if (!doRNEMD_) return;
    int selei;
    int selej;

    Snapshot* currentSnap_ = info_->getSnapshotManager()->getCurrentSnapshot();
    RealType time = currentSnap_->getTime();     
    Mat3x3d hmat = currentSnap_->getHmat();

    StuntDouble* sd;

    vector<StuntDouble*> hotBin, coldBin;

    RealType Phx = 0.0;
    RealType Phy = 0.0;
    RealType Phz = 0.0;
    RealType Khx = 0.0;
    RealType Khy = 0.0;
    RealType Khz = 0.0;
    RealType Khw = 0.0;
    RealType Pcx = 0.0;
    RealType Pcy = 0.0;
    RealType Pcz = 0.0;
    RealType Kcx = 0.0;
    RealType Kcy = 0.0;
    RealType Kcz = 0.0;
    RealType Kcw = 0.0;

    for (sd = smanA.beginSelected(selei); sd != NULL; 
         sd = smanA.nextSelected(selei)) {

      Vector3d pos = sd->getPos();
      
      // wrap the stuntdouble's position back into the box:
      
      if (usePeriodicBoundaryConditions_)
        currentSnap_->wrapVector(pos);
      
      
      RealType mass = sd->getMass();
      Vector3d vel = sd->getVel();
      
      hotBin.push_back(sd);
      Phx += mass * vel.x();
      Phy += mass * vel.y();
      Phz += mass * vel.z();
      Khx += mass * vel.x() * vel.x();
      Khy += mass * vel.y() * vel.y();
      Khz += mass * vel.z() * vel.z();
      if (sd->isDirectional()) {
        Vector3d angMom = sd->getJ();
        Mat3x3d I = sd->getI();
        if (sd->isLinear()) {
          int i = sd->linearAxis();
          int j = (i + 1) % 3;
          int k = (i + 2) % 3;
          Khw += angMom[j] * angMom[j] / I(j, j) +
            angMom[k] * angMom[k] / I(k, k);
        } else {
          Khw += angMom[0]*angMom[0]/I(0, 0)
            + angMom[1]*angMom[1]/I(1, 1)
            + angMom[2]*angMom[2]/I(2, 2);
        }
      }
    }
    for (sd = smanB.beginSelected(selej); sd != NULL; 
         sd = smanB.nextSelected(selej)) {
      Vector3d pos = sd->getPos();
      
      // wrap the stuntdouble's position back into the box:
      
      if (usePeriodicBoundaryConditions_)
        currentSnap_->wrapVector(pos);
            
      RealType mass = sd->getMass();
      Vector3d vel = sd->getVel();

      coldBin.push_back(sd);
      Pcx += mass * vel.x();
      Pcy += mass * vel.y();
      Pcz += mass * vel.z();
      Kcx += mass * vel.x() * vel.x();
      Kcy += mass * vel.y() * vel.y();
      Kcz += mass * vel.z() * vel.z();
      if (sd->isDirectional()) {
        Vector3d angMom = sd->getJ();
        Mat3x3d I = sd->getI();
        if (sd->isLinear()) {
          int i = sd->linearAxis();
          int j = (i + 1) % 3;
          int k = (i + 2) % 3;
          Kcw += angMom[j] * angMom[j] / I(j, j) +
            angMom[k] * angMom[k] / I(k, k);
        } else {
          Kcw += angMom[0]*angMom[0]/I(0, 0)
            + angMom[1]*angMom[1]/I(1, 1)
            + angMom[2]*angMom[2]/I(2, 2);
        }
      }
    }
    
    Khx *= 0.5;
    Khy *= 0.5;
    Khz *= 0.5;
    Khw *= 0.5;
    Kcx *= 0.5;
    Kcy *= 0.5;
    Kcz *= 0.5;
    Kcw *= 0.5;

#ifdef IS_MPI
    MPI_Allreduce(MPI_IN_PLACE, &Phx, 1, MPI_REALTYPE, MPI_SUM, MPI_COMM_WORLD);
    MPI_Allreduce(MPI_IN_PLACE, &Phy, 1, MPI_REALTYPE, MPI_SUM, MPI_COMM_WORLD);
    MPI_Allreduce(MPI_IN_PLACE, &Phz, 1, MPI_REALTYPE, MPI_SUM, MPI_COMM_WORLD);
    MPI_Allreduce(MPI_IN_PLACE, &Pcx, 1, MPI_REALTYPE, MPI_SUM, MPI_COMM_WORLD);
    MPI_Allreduce(MPI_IN_PLACE, &Pcy, 1, MPI_REALTYPE, MPI_SUM, MPI_COMM_WORLD);
    MPI_Allreduce(MPI_IN_PLACE, &Pcz, 1, MPI_REALTYPE, MPI_SUM, MPI_COMM_WORLD);

    MPI_Allreduce(MPI_IN_PLACE, &Khx, 1, MPI_REALTYPE, MPI_SUM, MPI_COMM_WORLD);
    MPI_Allreduce(MPI_IN_PLACE, &Khy, 1, MPI_REALTYPE, MPI_SUM, MPI_COMM_WORLD);
    MPI_Allreduce(MPI_IN_PLACE, &Khz, 1, MPI_REALTYPE, MPI_SUM, MPI_COMM_WORLD);
    MPI_Allreduce(MPI_IN_PLACE, &Khw, 1, MPI_REALTYPE, MPI_SUM, MPI_COMM_WORLD);

    MPI_Allreduce(MPI_IN_PLACE, &Kcx, 1, MPI_REALTYPE, MPI_SUM, MPI_COMM_WORLD);
    MPI_Allreduce(MPI_IN_PLACE, &Kcy, 1, MPI_REALTYPE, MPI_SUM, MPI_COMM_WORLD);
    MPI_Allreduce(MPI_IN_PLACE, &Kcz, 1, MPI_REALTYPE, MPI_SUM, MPI_COMM_WORLD);
    MPI_Allreduce(MPI_IN_PLACE, &Kcw, 1, MPI_REALTYPE, MPI_SUM, MPI_COMM_WORLD);
#endif

    //solve coldBin coeff's first
    RealType px = Pcx / Phx;
    RealType py = Pcy / Phy;
    RealType pz = Pcz / Phz;
    RealType c, x, y, z;
    bool successfulScale = false;
    if ((rnemdFluxType_ == rnemdFullKE) ||
        (rnemdFluxType_ == rnemdRotKE)) {
      //may need sanity check Khw & Kcw > 0

      if (rnemdFluxType_ == rnemdFullKE) {
        c = 1.0 - kineticTarget_ / (Kcx + Kcy + Kcz + Kcw);
      } else {
        c = 1.0 - kineticTarget_ / Kcw;
      }

      if ((c > 0.81) && (c < 1.21)) {//restrict scaling coefficients
        c = sqrt(c);

        RealType w = 0.0;
        if (rnemdFluxType_ ==  rnemdFullKE) {
          x = 1.0 + px * (1.0 - c);
          y = 1.0 + py * (1.0 - c);
          z = 1.0 + pz * (1.0 - c);
          /* more complicated way
             w = 1.0 + (Kcw - Kcw * c * c - (c * c * (Kcx + Kcy + Kcz
             + Khx * px * px + Khy * py * py + Khz * pz * pz)
             - 2.0 * c * (Khx * px * (1.0 + px) + Khy * py * (1.0 + py)
             + Khz * pz * (1.0 + pz)) + Khx * px * (2.0 + px)
             + Khy * py * (2.0 + py) + Khz * pz * (2.0 + pz)
             - Kcx - Kcy - Kcz)) / Khw; the following is simpler
          */
          if ((fabs(x - 1.0) < 0.1) && (fabs(y - 1.0) < 0.1) &&
              (fabs(z - 1.0) < 0.1)) {
            w = 1.0 + (kineticTarget_ 
                       + Khx * (1.0 - x * x) + Khy * (1.0 - y * y)
                       + Khz * (1.0 - z * z)) / Khw;
          }//no need to calculate w if x, y or z is out of range
        } else {
          w = 1.0 + kineticTarget_ / Khw;
        }
        if ((w > 0.81) && (w < 1.21)) {//restrict scaling coefficients
          //if w is in the right range, so should be x, y, z.
          vector<StuntDouble*>::iterator sdi;
          Vector3d vel;
          for (sdi = coldBin.begin(); sdi != coldBin.end(); ++sdi) {
            if (rnemdFluxType_ == rnemdFullKE) {
              vel = (*sdi)->getVel() * c;
              (*sdi)->setVel(vel);
            }
            if ((*sdi)->isDirectional()) {
              Vector3d angMom = (*sdi)->getJ() * c;
              (*sdi)->setJ(angMom);
            }
          }
          w = sqrt(w);
          for (sdi = hotBin.begin(); sdi != hotBin.end(); ++sdi) {
            if (rnemdFluxType_ == rnemdFullKE) {
              vel = (*sdi)->getVel();
              vel.x() *= x;
              vel.y() *= y;
              vel.z() *= z;
              (*sdi)->setVel(vel);
            }
            if ((*sdi)->isDirectional()) {
              Vector3d angMom = (*sdi)->getJ() * w;
              (*sdi)->setJ(angMom);
            }
          }
          successfulScale = true;
          kineticExchange_ += kineticTarget_;
        }
      }
    } else {
      RealType a000, a110, c0, a001, a111, b01, b11, c1;
      switch(rnemdFluxType_) {
      case rnemdKE :
        /* used hotBin coeff's & only scale x & y dimensions
           RealType px = Phx / Pcx;
           RealType py = Phy / Pcy;
           a110 = Khy;
           c0 = - Khx - Khy - kineticTarget_;
           a000 = Khx;
           a111 = Kcy * py * py;
           b11 = -2.0 * Kcy * py * (1.0 + py);
           c1 = Kcy * py * (2.0 + py) + Kcx * px * ( 2.0 + px) + kineticTarget_;
           b01 = -2.0 * Kcx * px * (1.0 + px);
           a001 = Kcx * px * px;
        */
        //scale all three dimensions, let c_x = c_y
        a000 = Kcx + Kcy;
        a110 = Kcz;
        c0 = kineticTarget_ - Kcx - Kcy - Kcz;
        a001 = Khx * px * px + Khy * py * py;
        a111 = Khz * pz * pz;
        b01 = -2.0 * (Khx * px * (1.0 + px) + Khy * py * (1.0 + py));
        b11 = -2.0 * Khz * pz * (1.0 + pz);
        c1 = Khx * px * (2.0 + px) + Khy * py * (2.0 + py)
          + Khz * pz * (2.0 + pz) - kineticTarget_;
        break;
      case rnemdPx :
        c = 1 - momentumTarget_.x() / Pcx;
        a000 = Kcy;
        a110 = Kcz;
        c0 = Kcx * c * c - Kcx - Kcy - Kcz;
        a001 = py * py * Khy;
        a111 = pz * pz * Khz;
        b01 = -2.0 * Khy * py * (1.0 + py);
        b11 = -2.0 * Khz * pz * (1.0 + pz);
        c1 = Khy * py * (2.0 + py) + Khz * pz * (2.0 + pz)
          + Khx * (fastpow(c * px - px - 1.0, 2) - 1.0);
        break;
      case rnemdPy :
        c = 1 - momentumTarget_.y() / Pcy;
        a000 = Kcx;
        a110 = Kcz;
        c0 = Kcy * c * c - Kcx - Kcy - Kcz;
        a001 = px * px * Khx;
        a111 = pz * pz * Khz;
        b01 = -2.0 * Khx * px * (1.0 + px);
        b11 = -2.0 * Khz * pz * (1.0 + pz);
        c1 = Khx * px * (2.0 + px) + Khz * pz * (2.0 + pz)
          + Khy * (fastpow(c * py - py - 1.0, 2) - 1.0);
        break;
      case rnemdPz ://we don't really do this, do we?
        c = 1 - momentumTarget_.z() / Pcz;
        a000 = Kcx;
        a110 = Kcy;
        c0 = Kcz * c * c - Kcx - Kcy - Kcz;
        a001 = px * px * Khx;
        a111 = py * py * Khy;
        b01 = -2.0 * Khx * px * (1.0 + px);
        b11 = -2.0 * Khy * py * (1.0 + py);
        c1 = Khx * px * (2.0 + px) + Khy * py * (2.0 + py)
          + Khz * (fastpow(c * pz - pz - 1.0, 2) - 1.0);
        break;
      default :
        break;
      }
      
      RealType v1 = a000 * a111 - a001 * a110;
      RealType v2 = a000 * b01;
      RealType v3 = a000 * b11;
      RealType v4 = a000 * c1 - a001 * c0;
      RealType v8 = a110 * b01;
      RealType v10 = - b01 * c0;
      
      RealType u0 = v2 * v10 - v4 * v4;
      RealType u1 = -2.0 * v3 * v4;
      RealType u2 = -v2 * v8 - v3 * v3 - 2.0 * v1 * v4;
      RealType u3 = -2.0 * v1 * v3;
      RealType u4 = - v1 * v1;
      //rescale coefficients
      RealType maxAbs = fabs(u0);
      if (maxAbs < fabs(u1)) maxAbs = fabs(u1);
      if (maxAbs < fabs(u2)) maxAbs = fabs(u2);
      if (maxAbs < fabs(u3)) maxAbs = fabs(u3);
      if (maxAbs < fabs(u4)) maxAbs = fabs(u4);
      u0 /= maxAbs;
      u1 /= maxAbs;
      u2 /= maxAbs;
      u3 /= maxAbs;
      u4 /= maxAbs;
      //max_element(start, end) is also available.
      Polynomial<RealType> poly; //same as DoublePolynomial poly;
      poly.setCoefficient(4, u4);
      poly.setCoefficient(3, u3);
      poly.setCoefficient(2, u2);
      poly.setCoefficient(1, u1);
      poly.setCoefficient(0, u0);
      vector<RealType> realRoots = poly.FindRealRoots();
      
      vector<RealType>::iterator ri;
      RealType r1, r2, alpha0;
      vector<pair<RealType,RealType> > rps;
      for (ri = realRoots.begin(); ri !=realRoots.end(); ++ri) {
        r2 = *ri;
        //check if FindRealRoots() give the right answer
        if ( fabs(u0 + r2 * (u1 + r2 * (u2 + r2 * (u3 + r2 * u4)))) > 1e-6 ) {
          sprintf(painCave.errMsg, 
                  "RNEMD Warning: polynomial solve seems to have an error!");
          painCave.isFatal = 0;
          simError();
          failRootCount_++;
        }
        //might not be useful w/o rescaling coefficients
        alpha0 = -c0 - a110 * r2 * r2;
        if (alpha0 >= 0.0) {
          r1 = sqrt(alpha0 / a000);
          if (fabs(c1 + r1 * (b01 + r1 * a001) + r2 * (b11 + r2 * a111))
              < 1e-6)
            { rps.push_back(make_pair(r1, r2)); }
          if (r1 > 1e-6) { //r1 non-negative
            r1 = -r1;
            if (fabs(c1 + r1 * (b01 + r1 * a001) + r2 * (b11 + r2 * a111))
                < 1e-6)
              { rps.push_back(make_pair(r1, r2)); }
          }
        }
      }
      // Consider combining together the solving pair part w/ the searching
      // best solution part so that we don't need the pairs vector
      if (!rps.empty()) {
        RealType smallestDiff = HONKING_LARGE_VALUE;
        RealType diff;
        pair<RealType,RealType> bestPair = make_pair(1.0, 1.0);
        vector<pair<RealType,RealType> >::iterator rpi;
        for (rpi = rps.begin(); rpi != rps.end(); ++rpi) {
          r1 = (*rpi).first;
          r2 = (*rpi).second;
          switch(rnemdFluxType_) {
          case rnemdKE :
            diff = fastpow(1.0 - r1, 2) + fastpow(1.0 - r2, 2)
              + fastpow(r1 * r1 / r2 / r2 - Kcz/Kcx, 2)
              + fastpow(r1 * r1 / r2 / r2 - Kcz/Kcy, 2);
            break;
          case rnemdPx :
            diff = fastpow(1.0 - r1, 2) + fastpow(1.0 - r2, 2)
              + fastpow(r1 * r1 / r2 / r2 - Kcz/Kcy, 2);
            break;
          case rnemdPy :
            diff = fastpow(1.0 - r1, 2) + fastpow(1.0 - r2, 2)
              + fastpow(r1 * r1 / r2 / r2 - Kcz/Kcx, 2);
            break;
          case rnemdPz :
            diff = fastpow(1.0 - r1, 2) + fastpow(1.0 - r2, 2)
              + fastpow(r1 * r1 / r2 / r2 - Kcy/Kcx, 2);
          default :
            break;
          }
          if (diff < smallestDiff) {
            smallestDiff = diff;
            bestPair = *rpi;
          }
        }
#ifdef IS_MPI
        if (worldRank == 0) {
#endif
          // sprintf(painCave.errMsg, 
          //         "RNEMD: roots r1= %lf\tr2 = %lf\n",
          //         bestPair.first, bestPair.second);
          // painCave.isFatal = 0;
          // painCave.severity = OPENMD_INFO;
          // simError();
#ifdef IS_MPI
        }
#endif
        
        switch(rnemdFluxType_) {
        case rnemdKE :
          x = bestPair.first;
          y = bestPair.first;
          z = bestPair.second;
          break;
        case rnemdPx :
          x = c;
          y = bestPair.first;
          z = bestPair.second;
          break;
        case rnemdPy :
          x = bestPair.first;
          y = c;
          z = bestPair.second;
          break;
        case rnemdPz :
          x = bestPair.first;
          y = bestPair.second;
          z = c;
          break;          
        default :
          break;
        }
        vector<StuntDouble*>::iterator sdi;
        Vector3d vel;
        for (sdi = coldBin.begin(); sdi != coldBin.end(); ++sdi) {
          vel = (*sdi)->getVel();
          vel.x() *= x;
          vel.y() *= y;
          vel.z() *= z;
          (*sdi)->setVel(vel);
        }
        //convert to hotBin coefficient
        x = 1.0 + px * (1.0 - x);
        y = 1.0 + py * (1.0 - y);
        z = 1.0 + pz * (1.0 - z);
        for (sdi = hotBin.begin(); sdi != hotBin.end(); ++sdi) {
          vel = (*sdi)->getVel();
          vel.x() *= x;
          vel.y() *= y;
          vel.z() *= z;
          (*sdi)->setVel(vel);
        }
        successfulScale = true;
        switch(rnemdFluxType_) {
        case rnemdKE :
          kineticExchange_ += kineticTarget_;
          break;
        case rnemdPx :
        case rnemdPy :
        case rnemdPz :
          momentumExchange_ += momentumTarget_;
          break;          
        default :
          break;
        }       
      }
    }
    if (successfulScale != true) {
      sprintf(painCave.errMsg, 
              "RNEMD::doNIVS exchange NOT performed - roots that solve\n"
              "\tthe constraint equations may not exist or there may be\n"
              "\tno selected objects in one or both slabs.\n");
      painCave.isFatal = 0;
      painCave.severity = OPENMD_INFO;
      simError();        
      failTrialCount_++;
    }
  }
  
  void RNEMD::doVSS(SelectionManager& smanA, SelectionManager& smanB) {
    if (!doRNEMD_) return;
    int selei;
    int selej;

    Snapshot* currentSnap_ = info_->getSnapshotManager()->getCurrentSnapshot();
    RealType time = currentSnap_->getTime();     
    Mat3x3d hmat = currentSnap_->getHmat();

    StuntDouble* sd;

    vector<StuntDouble*> hotBin, coldBin;

    Vector3d Ph(V3Zero);
    Vector3d Lh(V3Zero);
    RealType Mh = 0.0;
    Mat3x3d Ih(0.0);
    RealType Kh = 0.0;
    Vector3d Pc(V3Zero);
    Vector3d Lc(V3Zero);
    RealType Mc = 0.0;
    Mat3x3d Ic(0.0);
    RealType Kc = 0.0;

    // Constraints can be on only the linear or angular momentum, but
    // not both.  Usually, the user will specify which they want, but
    // in case they don't, the use of periodic boundaries should make
    // the choice for us.
    bool doLinearPart = false;
    bool doAngularPart = false;

    switch (rnemdFluxType_) {
    case rnemdPx:
    case rnemdPy:
    case rnemdPz:
    case rnemdPvector:
    case rnemdKePx:
    case rnemdKePy:
    case rnemdKePvector:
      doLinearPart = true;
      break;
    case rnemdLx:
    case rnemdLy:
    case rnemdLz:
    case rnemdLvector:
    case rnemdKeLx:
    case rnemdKeLy:
    case rnemdKeLz:
    case rnemdKeLvector:
      doAngularPart = true;
      break;
    case rnemdKE:
    case rnemdRotKE:
    case rnemdFullKE:
    default:
      if (usePeriodicBoundaryConditions_) 
        doLinearPart = true;
      else
        doAngularPart = true;
      break;
    }
    
    for (sd = smanA.beginSelected(selei); sd != NULL; 
         sd = smanA.nextSelected(selei)) {

      Vector3d pos = sd->getPos();

      // wrap the stuntdouble's position back into the box:
      
      if (usePeriodicBoundaryConditions_)
        currentSnap_->wrapVector(pos);
      
      RealType mass = sd->getMass();
      Vector3d vel = sd->getVel();
      Vector3d rPos = sd->getPos() - coordinateOrigin_;
      RealType r2;
      
      hotBin.push_back(sd);
      Ph += mass * vel;
      Mh += mass;
      Kh += mass * vel.lengthSquare();
      Lh += mass * cross(rPos, vel);
      Ih -= outProduct(rPos, rPos) * mass;
      r2 = rPos.lengthSquare();
      Ih(0, 0) += mass * r2;
      Ih(1, 1) += mass * r2;
      Ih(2, 2) += mass * r2;
      
      if (rnemdFluxType_ == rnemdFullKE) {
        if (sd->isDirectional()) {
          Vector3d angMom = sd->getJ();
          Mat3x3d I = sd->getI();
          if (sd->isLinear()) {
            int i = sd->linearAxis();
            int j = (i + 1) % 3;
            int k = (i + 2) % 3;
            Kh += angMom[j] * angMom[j] / I(j, j) +
              angMom[k] * angMom[k] / I(k, k);
          } else {
            Kh += angMom[0] * angMom[0] / I(0, 0) +
              angMom[1] * angMom[1] / I(1, 1) +
              angMom[2] * angMom[2] / I(2, 2);
          }
        }
      }
    }
    for (sd = smanB.beginSelected(selej); sd != NULL; 
         sd = smanB.nextSelected(selej)) {

      Vector3d pos = sd->getPos();
      
      // wrap the stuntdouble's position back into the box:
      
      if (usePeriodicBoundaryConditions_)
        currentSnap_->wrapVector(pos);
      
      RealType mass = sd->getMass();
      Vector3d vel = sd->getVel();
      Vector3d rPos = sd->getPos() - coordinateOrigin_;
      RealType r2;

      coldBin.push_back(sd);
      Pc += mass * vel;
      Mc += mass;
      Kc += mass * vel.lengthSquare();
      Lc += mass * cross(rPos, vel);
      Ic -= outProduct(rPos, rPos) * mass;
      r2 = rPos.lengthSquare();
      Ic(0, 0) += mass * r2;
      Ic(1, 1) += mass * r2;
      Ic(2, 2) += mass * r2;
      
      if (rnemdFluxType_ == rnemdFullKE) {
        if (sd->isDirectional()) {
          Vector3d angMom = sd->getJ();
          Mat3x3d I = sd->getI();
          if (sd->isLinear()) {
            int i = sd->linearAxis();
            int j = (i + 1) % 3;
            int k = (i + 2) % 3;
            Kc += angMom[j] * angMom[j] / I(j, j) +
              angMom[k] * angMom[k] / I(k, k);
          } else {
            Kc += angMom[0] * angMom[0] / I(0, 0) +
              angMom[1] * angMom[1] / I(1, 1) +
              angMom[2] * angMom[2] / I(2, 2);
          }
        }
      }
    }
    
    Kh *= 0.5;
    Kc *= 0.5;
    
#ifdef IS_MPI
    MPI_Allreduce(MPI_IN_PLACE, &Ph[0], 3, MPI_REALTYPE, MPI_SUM, 
                  MPI_COMM_WORLD);
    MPI_Allreduce(MPI_IN_PLACE, &Pc[0], 3, MPI_REALTYPE, MPI_SUM, 
                  MPI_COMM_WORLD);
    MPI_Allreduce(MPI_IN_PLACE, &Lh[0], 3, MPI_REALTYPE, MPI_SUM, 
                  MPI_COMM_WORLD);
    MPI_Allreduce(MPI_IN_PLACE, &Lc[0], 3, MPI_REALTYPE, MPI_SUM, 
                  MPI_COMM_WORLD);
    MPI_Allreduce(MPI_IN_PLACE, &Mh, 1, MPI_REALTYPE, MPI_SUM, MPI_COMM_WORLD);
    MPI_Allreduce(MPI_IN_PLACE, &Kh, 1, MPI_REALTYPE, MPI_SUM, MPI_COMM_WORLD);
    MPI_Allreduce(MPI_IN_PLACE, &Mc, 1, MPI_REALTYPE, MPI_SUM, MPI_COMM_WORLD);
    MPI_Allreduce(MPI_IN_PLACE, &Kc, 1, MPI_REALTYPE, MPI_SUM, MPI_COMM_WORLD);
    MPI_Allreduce(MPI_IN_PLACE, Ih.getArrayPointer(), 9, 
                  MPI_REALTYPE, MPI_SUM, MPI_COMM_WORLD);
    MPI_Allreduce(MPI_IN_PLACE, Ic.getArrayPointer(), 9, 
                  MPI_REALTYPE, MPI_SUM, MPI_COMM_WORLD);
#endif
    

    Vector3d ac, acrec, bc, bcrec;
    Vector3d ah, ahrec, bh, bhrec;

    bool successfulExchange = false;
    if ((Mh > 0.0) && (Mc > 0.0)) {//both slabs are not empty
      Vector3d vc = Pc / Mc;
      ac = -momentumTarget_ / Mc + vc;
      acrec = -momentumTarget_ / Mc;
      
      // We now need the inverse of the inertia tensor to calculate the 
      // angular velocity of the cold slab;
      Mat3x3d Ici = Ic.inverse();
      Vector3d omegac = Ici * Lc;
      bc  = -(Ici * angularMomentumTarget_) + omegac;
      bcrec = bc - omegac;
      
      RealType cNumerator = Kc - kineticTarget_;
      if (doLinearPart) 
        cNumerator -= 0.5 * Mc * ac.lengthSquare();
      
      if (doAngularPart)
        cNumerator -= 0.5 * ( dot(bc, Ic * bc));

      if (cNumerator > 0.0) {
        
        RealType cDenominator = Kc;

        if (doLinearPart)
          cDenominator -= 0.5 * Mc * vc.lengthSquare();

        if (doAngularPart)
          cDenominator -= 0.5*(dot(omegac, Ic * omegac));
        
        if (cDenominator > 0.0) {
          RealType c = sqrt(cNumerator / cDenominator);
          if ((c > 0.9) && (c < 1.1)) {//restrict scaling coefficients
            
            Vector3d vh = Ph / Mh;
            ah = momentumTarget_ / Mh + vh;
            ahrec = momentumTarget_ / Mh;
            
            // We now need the inverse of the inertia tensor to
            // calculate the angular velocity of the hot slab;
            Mat3x3d Ihi = Ih.inverse();
            Vector3d omegah = Ihi * Lh;
            bh  = (Ihi * angularMomentumTarget_) + omegah;
            bhrec = bh - omegah;
            
            RealType hNumerator = Kh + kineticTarget_;
            if (doLinearPart) 
              hNumerator -= 0.5 * Mh * ah.lengthSquare();
            
            if (doAngularPart)
              hNumerator -= 0.5 * ( dot(bh, Ih * bh));
              
            if (hNumerator > 0.0) {
              
              RealType hDenominator = Kh;
              if (doLinearPart) 
                hDenominator -= 0.5 * Mh * vh.lengthSquare();
              if (doAngularPart)
                hDenominator -= 0.5*(dot(omegah, Ih * omegah));
              
              if (hDenominator > 0.0) {
                RealType h = sqrt(hNumerator / hDenominator);
                if ((h > 0.9) && (h < 1.1)) {
                  
                  vector<StuntDouble*>::iterator sdi;
                  Vector3d vel;
                  Vector3d rPos;
                  
                  for (sdi = coldBin.begin(); sdi != coldBin.end(); ++sdi) {
                    //vel = (*sdi)->getVel();
                    rPos = (*sdi)->getPos() - coordinateOrigin_;
                    if (doLinearPart)
                      vel = ((*sdi)->getVel() - vc) * c + ac;
                    if (doAngularPart)
                      vel = ((*sdi)->getVel() - cross(omegac, rPos)) * c + cross(bc, rPos);

                    (*sdi)->setVel(vel);
                    if (rnemdFluxType_ == rnemdFullKE) {
                      if ((*sdi)->isDirectional()) {
                        Vector3d angMom = (*sdi)->getJ() * c;
                        (*sdi)->setJ(angMom);
                      }
                    }
                  }
                  for (sdi = hotBin.begin(); sdi != hotBin.end(); ++sdi) {
                    //vel = (*sdi)->getVel();
                    rPos = (*sdi)->getPos() - coordinateOrigin_;
                    if (doLinearPart)
                      vel = ((*sdi)->getVel() - vh) * h + ah;     
                    if (doAngularPart)
                      vel = ((*sdi)->getVel() - cross(omegah, rPos)) * h + cross(bh, rPos);     

                    (*sdi)->setVel(vel);
                    if (rnemdFluxType_ == rnemdFullKE) {
                      if ((*sdi)->isDirectional()) {
                        Vector3d angMom = (*sdi)->getJ() * h;
                        (*sdi)->setJ(angMom);
                      }
                    }
                  }
                  successfulExchange = true;
                  kineticExchange_ += kineticTarget_;
                  momentumExchange_ += momentumTarget_;
                  angularMomentumExchange_ += angularMomentumTarget_;
                }
              }
            }
          }
        }
      }
    }
    if (successfulExchange != true) {
      sprintf(painCave.errMsg, 
              "RNEMD::doVSS exchange NOT performed - roots that solve\n"
              "\tthe constraint equations may not exist or there may be\n"
              "\tno selected objects in one or both slabs.\n");
      painCave.isFatal = 0;
      painCave.severity = OPENMD_INFO;
      simError();        
      failTrialCount_++;
    }
  }

  RealType RNEMD::getDividingArea() {

    if (hasDividingArea_) return dividingArea_;

    RealType areaA, areaB;
    Snapshot* snap = info_->getSnapshotManager()->getCurrentSnapshot();

    if (hasSelectionA_) {

      if (evaluatorA_.hasSurfaceArea()) 
        areaA = evaluatorA_.getSurfaceArea();
      else {
         
        cerr << "selection A did not have surface area, recomputing\n";
        int isd;
        StuntDouble* sd;
        vector<StuntDouble*> aSites;
        seleManA_.setSelectionSet(evaluatorA_.evaluate());
        for (sd = seleManA_.beginSelected(isd); sd != NULL; 
             sd = seleManA_.nextSelected(isd)) {
          aSites.push_back(sd);
        }
#if defined(HAVE_QHULL)
        ConvexHull* surfaceMeshA = new ConvexHull();
        surfaceMeshA->computeHull(aSites);
        cerr << "flag1\n";
        areaA = surfaceMeshA->getArea();
        cerr << "Flag2 " << areaA << "\n";
        delete surfaceMeshA;
#else
        sprintf( painCave.errMsg,
               "RNEMD::getDividingArea : Hull calculation is not possible\n"
                 "\twithout libqhull. Please rebuild OpenMD with qhull enabled.");
        painCave.severity = OPENMD_ERROR;
        painCave.isFatal = 1;
        simError();
#endif
      }

    } else {
      if (usePeriodicBoundaryConditions_) {
        // in periodic boundaries, the surface area is twice the x-y
        // area of the current box:
        areaA = 2.0 * snap->getXYarea();
      } else {
        // in non-periodic simulations, without explicitly setting
        // selections, the sphere radius sets the surface area of the
        // dividing surface:
        areaA = 4.0 * M_PI * pow(sphereARadius_, 2);
      }
    }

    if (hasSelectionB_) {
      if (evaluatorB_.hasSurfaceArea()) 
        areaB = evaluatorB_.getSurfaceArea();
      else {
        cerr << "selection B did not have surface area, recomputing\n";

        int isd;
        StuntDouble* sd;
        vector<StuntDouble*> bSites;
        seleManB_.setSelectionSet(evaluatorB_.evaluate());
        for (sd = seleManB_.beginSelected(isd); sd != NULL; 
             sd = seleManB_.nextSelected(isd)) {
          bSites.push_back(sd);
        }
        
#if defined(HAVE_QHULL)
        ConvexHull* surfaceMeshB = new ConvexHull();    
        surfaceMeshB->computeHull(bSites);
        areaB = surfaceMeshB->getArea();
        delete surfaceMeshB;
#else
        sprintf( painCave.errMsg,
                 "RNEMD::getDividingArea : Hull calculation is not possible\n"
                 "\twithout libqhull. Please rebuild OpenMD with qhull enabled.");
        painCave.severity = OPENMD_ERROR;
        painCave.isFatal = 1;
        simError();
#endif
      }
      
    } else {
      if (usePeriodicBoundaryConditions_) {
        // in periodic boundaries, the surface area is twice the x-y
        // area of the current box:
        areaB = 2.0 * snap->getXYarea();
      } else {
        // in non-periodic simulations, without explicitly setting
        // selections, but if a sphereBradius has been set, just use that:
        areaB = 4.0 * M_PI * pow(sphereBRadius_, 2);
      }
    }
      
    dividingArea_ = min(areaA, areaB);
    hasDividingArea_ = true;
    return dividingArea_;
  }
  
  void RNEMD::doRNEMD() {
    if (!doRNEMD_) return;
    trialCount_++;

    // object evaluator:
    evaluator_.loadScriptString(rnemdObjectSelection_);
    seleMan_.setSelectionSet(evaluator_.evaluate());

    evaluatorA_.loadScriptString(selectionA_);
    evaluatorB_.loadScriptString(selectionB_);

    seleManA_.setSelectionSet(evaluatorA_.evaluate());
    seleManB_.setSelectionSet(evaluatorB_.evaluate());

    commonA_ = seleManA_ & seleMan_;
    commonB_ = seleManB_ & seleMan_;

    // Target exchange quantities (in each exchange) = dividingArea * dt * flux
    // dt = exchange time interval
    // flux = target flux
    // dividingArea = smallest dividing surface between the two regions

    hasDividingArea_ = false;
    RealType area = getDividingArea();

    kineticTarget_ = kineticFlux_ * exchangeTime_ * area;
    momentumTarget_ = momentumFluxVector_ * exchangeTime_ * area;
    angularMomentumTarget_ = angularMomentumFluxVector_ * exchangeTime_ * area; 

    switch(rnemdMethod_) {
    case rnemdSwap: 
      doSwap(commonA_, commonB_);
      break;
    case rnemdNIVS:
      doNIVS(commonA_, commonB_);
      break;
    case rnemdVSS:
      doVSS(commonA_, commonB_);
      break;
    case rnemdUnkownMethod:
    default :
      break;
    }
  }

  void RNEMD::collectData() {
    if (!doRNEMD_) return;
    Snapshot* currentSnap_ = info_->getSnapshotManager()->getCurrentSnapshot();
    
    // collectData can be called more frequently than the doRNEMD, so use the 
    // computed area from the last exchange time:
    RealType area = getDividingArea();
    areaAccumulator_->add(area);
    Mat3x3d hmat = currentSnap_->getHmat();
    Vector3d u = angularMomentumFluxVector_;
    u.normalize();

    seleMan_.setSelectionSet(evaluator_.evaluate());

    int selei(0);
    StuntDouble* sd;
    int binNo;
    RealType mass;
    Vector3d vel; 
    Vector3d rPos;
    RealType KE;
    Vector3d L;
    Mat3x3d I;
    RealType r2;

    vector<RealType> binMass(nBins_, 0.0);
    vector<Vector3d> binP(nBins_, V3Zero);
    vector<RealType> binOmega(nBins_, 0.0);
    vector<Vector3d> binL(nBins_, V3Zero);
    vector<Mat3x3d>  binI(nBins_);
    vector<RealType> binKE(nBins_, 0.0);
    vector<int> binDOF(nBins_, 0);
    vector<int> binCount(nBins_, 0);

    // alternative approach, track all molecules instead of only those
    // selected for scaling/swapping:
    /*
      SimInfo::MoleculeIterator miter;
      vector<StuntDouble*>::iterator iiter;
      Molecule* mol;
      StuntDouble* sd;
      for (mol = info_->beginMolecule(miter); mol != NULL;
      mol = info_->nextMolecule(miter))
      sd is essentially sd
      for (sd = mol->beginIntegrableObject(iiter);
      sd != NULL;
      sd = mol->nextIntegrableObject(iiter))
    */

    for (sd = seleMan_.beginSelected(selei); sd != NULL; 
         sd = seleMan_.nextSelected(selei)) {     
    
      Vector3d pos = sd->getPos();

      // wrap the stuntdouble's position back into the box:
      
      if (usePeriodicBoundaryConditions_) {
        currentSnap_->wrapVector(pos);
        // which bin is this stuntdouble in?
        // wrapped positions are in the range [-0.5*hmat(2,2), +0.5*hmat(2,2)]
        // Shift molecules by half a box to have bins start at 0
        // The modulo operator is used to wrap the case when we are 
        // beyond the end of the bins back to the beginning.
        binNo = int(nBins_ * (pos.z() / hmat(2,2) + 0.5)) % nBins_;
      } else {
        Vector3d rPos = pos - coordinateOrigin_;
        binNo = int(rPos.length() / binWidth_);
      }

      mass = sd->getMass();
      vel = sd->getVel();
      rPos = sd->getPos() - coordinateOrigin_;
      KE = 0.5 * mass * vel.lengthSquare();
      L = mass * cross(rPos, vel);
      I = outProduct(rPos, rPos) * mass;
      r2 = rPos.lengthSquare();
      I(0, 0) += mass * r2;
      I(1, 1) += mass * r2;
      I(2, 2) += mass * r2;

      // Project the relative position onto a plane perpendicular to
      // the angularMomentumFluxVector:
      // Vector3d rProj = rPos - dot(rPos, u) * u;
      // Project the velocity onto a plane perpendicular to the
      // angularMomentumFluxVector:
      // Vector3d vProj = vel  - dot(vel, u) * u;
      // Compute angular velocity vector (should be nearly parallel to
      // angularMomentumFluxVector
      // Vector3d aVel = cross(rProj, vProj);

      if (binNo >= 0 && binNo < nBins_)  {
        binCount[binNo]++;
        binMass[binNo] += mass;
        binP[binNo] += mass*vel;
        binKE[binNo] += KE;
        binI[binNo] += I;
        binL[binNo] += L;
        binDOF[binNo] += 3;
        
        if (sd->isDirectional()) {
          Vector3d angMom = sd->getJ();
          Mat3x3d Ia = sd->getI();
          if (sd->isLinear()) {
            int i = sd->linearAxis();
            int j = (i + 1) % 3;
            int k = (i + 2) % 3;
            binKE[binNo] += 0.5 * (angMom[j] * angMom[j] / Ia(j, j) + 
                                   angMom[k] * angMom[k] / Ia(k, k));
            binDOF[binNo] += 2;
          } else {
            binKE[binNo] += 0.5 * (angMom[0] * angMom[0] / Ia(0, 0) +
                                   angMom[1] * angMom[1] / Ia(1, 1) +
                                   angMom[2] * angMom[2] / Ia(2, 2));
            binDOF[binNo] += 3;
          }
        }
      }
    }
    
#ifdef IS_MPI

    for (int i = 0; i < nBins_; i++) {
      
      MPI_Allreduce(MPI_IN_PLACE, &binCount[i],
                    1, MPI_INT, MPI_SUM, MPI_COMM_WORLD);
      MPI_Allreduce(MPI_IN_PLACE, &binMass[i],
                    1, MPI_REALTYPE, MPI_SUM, MPI_COMM_WORLD);
      MPI_Allreduce(MPI_IN_PLACE, binP[i].getArrayPointer(),
                    3, MPI_REALTYPE, MPI_SUM, MPI_COMM_WORLD);
      MPI_Allreduce(MPI_IN_PLACE, binL[i].getArrayPointer(),
                    3, MPI_REALTYPE, MPI_SUM, MPI_COMM_WORLD);
      MPI_Allreduce(MPI_IN_PLACE, binI[i].getArrayPointer(),
                    9, MPI_REALTYPE, MPI_SUM, MPI_COMM_WORLD);
      MPI_Allreduce(MPI_IN_PLACE, &binKE[i],
                    1, MPI_REALTYPE, MPI_SUM, MPI_COMM_WORLD);
      MPI_Allreduce(MPI_IN_PLACE, &binDOF[i],
                    1, MPI_INT, MPI_SUM, MPI_COMM_WORLD);
      //MPI_Allreduce(MPI_IN_PLACE, &binOmega[i],
      //                          1, MPI_REALTYPE, MPI_SUM, MPI_COMM_WORLD);
    }
    
#endif

    Vector3d omega;
    RealType den;
    RealType temp;
    RealType z;
    RealType r;
    for (int i = 0; i < nBins_; i++) {
      if (usePeriodicBoundaryConditions_) {
        z = (((RealType)i + 0.5) / (RealType)nBins_) * hmat(2,2);
        den = binMass[i] * nBins_ * PhysicalConstants::densityConvert 
          / currentSnap_->getVolume() ;
      } else {
        r = (((RealType)i + 0.5) * binWidth_);
        RealType rinner = (RealType)i * binWidth_;
        RealType router = (RealType)(i+1) * binWidth_;
        den = binMass[i] * 3.0 * PhysicalConstants::densityConvert
          / (4.0 * M_PI * (pow(router,3) - pow(rinner,3)));
      }
      vel = binP[i] / binMass[i];

      omega = binI[i].inverse() * binL[i];

      // omega = binOmega[i] / binCount[i];

      if (binCount[i] > 0) {
        // only add values if there are things to add
        temp = 2.0 * binKE[i] / (binDOF[i] * PhysicalConstants::kb *
                                 PhysicalConstants::energyConvert);
        
        for (unsigned int j = 0; j < outputMask_.size(); ++j) {
          if(outputMask_[j]) {
            switch(j) {
            case Z:
              dynamic_cast<Accumulator *>(data_[j].accumulator[i])->add(z);
              break;
            case R:
              dynamic_cast<Accumulator *>(data_[j].accumulator[i])->add(r);
              break;
            case TEMPERATURE:
              dynamic_cast<Accumulator *>(data_[j].accumulator[i])->add(temp);
              break;
            case VELOCITY:
              dynamic_cast<VectorAccumulator *>(data_[j].accumulator[i])->add(vel);
              break;
            case ANGULARVELOCITY:  
              dynamic_cast<VectorAccumulator *>(data_[j].accumulator[i])->add(omega);
              break;
            case DENSITY:
              dynamic_cast<Accumulator *>(data_[j].accumulator[i])->add(den);
              break;
            }
          }
        }
      }
    }
    hasData_ = true;
  }

  void RNEMD::getStarted() {
    if (!doRNEMD_) return;
    hasDividingArea_ = false;
    collectData();
    writeOutputFile();
  }

  void RNEMD::parseOutputFileFormat(const std::string& format) {
    if (!doRNEMD_) return;
    StringTokenizer tokenizer(format, " ,;|\t\n\r");
    
    while(tokenizer.hasMoreTokens()) {
      std::string token(tokenizer.nextToken());
      toUpper(token);
      OutputMapType::iterator i = outputMap_.find(token);
      if (i != outputMap_.end()) {
        outputMask_.set(i->second);
      } else {
        sprintf( painCave.errMsg,
                 "RNEMD::parseOutputFileFormat: %s is not a recognized\n"
                 "\toutputFileFormat keyword.\n", token.c_str() );
        painCave.isFatal = 0;
        painCave.severity = OPENMD_ERROR;
        simError();            
      }
    }   
  }
  
  void RNEMD::writeOutputFile() {
    if (!doRNEMD_) return;
    if (!hasData_) return;
    
#ifdef IS_MPI
    // If we're the root node, should we print out the results
    int worldRank;
    MPI_Comm_rank( MPI_COMM_WORLD, &worldRank);

    if (worldRank == 0) {
#endif
      rnemdFile_.open(rnemdFileName_.c_str(), std::ios::out | std::ios::trunc );
      
      if( !rnemdFile_ ){        
        sprintf( painCave.errMsg,
                 "Could not open \"%s\" for RNEMD output.\n",
                 rnemdFileName_.c_str());
        painCave.isFatal = 1;
        simError();
      }

      Snapshot* currentSnap_ = info_->getSnapshotManager()->getCurrentSnapshot();

      RealType time = currentSnap_->getTime();
      RealType avgArea;
      areaAccumulator_->getAverage(avgArea);

      RealType Jz(0.0);
      Vector3d JzP(V3Zero);
      Vector3d JzL(V3Zero);
      if (time >= info_->getSimParams()->getDt()) {
        Jz = kineticExchange_ / (time * avgArea)
          / PhysicalConstants::energyConvert;
        JzP = momentumExchange_ / (time * avgArea);
        JzL = angularMomentumExchange_ / (time * avgArea);
      }

      rnemdFile_ << "#######################################################\n";
      rnemdFile_ << "# RNEMD {\n";

      map<string, RNEMDMethod>::iterator mi;
      for(mi = stringToMethod_.begin(); mi != stringToMethod_.end(); ++mi) {
        if ( (*mi).second == rnemdMethod_) 
          rnemdFile_ << "#    exchangeMethod  = \"" << (*mi).first << "\";\n";
      }
      map<string, RNEMDFluxType>::iterator fi;
      for(fi = stringToFluxType_.begin(); fi != stringToFluxType_.end(); ++fi) {
        if ( (*fi).second == rnemdFluxType_) 
          rnemdFile_ << "#    fluxType  = \"" << (*fi).first << "\";\n";
      }
      
      rnemdFile_ << "#    exchangeTime = " << exchangeTime_ << ";\n";

      rnemdFile_ << "#    objectSelection = \"" 
                 << rnemdObjectSelection_ << "\";\n";
      rnemdFile_ << "#    selectionA = \"" << selectionA_ << "\";\n";
      rnemdFile_ << "#    selectionB = \"" << selectionB_ << "\";\n";
      rnemdFile_ << "# }\n";
      rnemdFile_ << "#######################################################\n";
      rnemdFile_ << "# RNEMD report:\n";      
      rnemdFile_ << "#      running time = " << time << " fs\n";
      rnemdFile_ << "# Target flux:\n";
      rnemdFile_ << "#           kinetic = " 
                 << kineticFlux_ / PhysicalConstants::energyConvert 
                 << " (kcal/mol/A^2/fs)\n";
      rnemdFile_ << "#          momentum = " << momentumFluxVector_ 
                 << " (amu/A/fs^2)\n";
      rnemdFile_ << "#  angular momentum = " << angularMomentumFluxVector_ 
                 << " (amu/A^2/fs^2)\n";
      rnemdFile_ << "# Target one-time exchanges:\n";
      rnemdFile_ << "#          kinetic = " 
                 << kineticTarget_ / PhysicalConstants::energyConvert 
                 << " (kcal/mol)\n";
      rnemdFile_ << "#          momentum = " << momentumTarget_ 
                 << " (amu*A/fs)\n";
      rnemdFile_ << "#  angular momentum = " << angularMomentumTarget_ 
                 << " (amu*A^2/fs)\n";
      rnemdFile_ << "# Actual exchange totals:\n";
      rnemdFile_ << "#          kinetic = " 
                 << kineticExchange_ / PhysicalConstants::energyConvert 
                 << " (kcal/mol)\n";
      rnemdFile_ << "#          momentum = " << momentumExchange_ 
                 << " (amu*A/fs)\n";      
      rnemdFile_ << "#  angular momentum = " << angularMomentumExchange_ 
                 << " (amu*A^2/fs)\n";      
      rnemdFile_ << "# Actual flux:\n";
      rnemdFile_ << "#          kinetic = " << Jz
                 << " (kcal/mol/A^2/fs)\n";
      rnemdFile_ << "#          momentum = " << JzP 
                 << " (amu/A/fs^2)\n";
      rnemdFile_ << "#  angular momentum = " << JzL
                 << " (amu/A^2/fs^2)\n";
      rnemdFile_ << "# Exchange statistics:\n";
      rnemdFile_ << "#               attempted = " << trialCount_ << "\n";
      rnemdFile_ << "#                  failed = " << failTrialCount_ << "\n";
      if (rnemdMethod_ == rnemdNIVS) {
        rnemdFile_ << "#  NIVS root-check errors = "
                   << failRootCount_ << "\n";
      }
      rnemdFile_ << "#######################################################\n";
      
      
      //write title
      rnemdFile_ << "#";
      for (unsigned int i = 0; i < outputMask_.size(); ++i) {
        if (outputMask_[i]) {
          rnemdFile_ << "\t" << data_[i].title << 
            "(" << data_[i].units << ")";
          // add some extra tabs for column alignment
          if (data_[i].dataType == "Vector3d") rnemdFile_ << "\t\t";
        }
      }
      rnemdFile_ << std::endl;
      
      rnemdFile_.precision(8);
      
      for (int j = 0; j < nBins_; j++) {        
        
        for (unsigned int i = 0; i < outputMask_.size(); ++i) {
          if (outputMask_[i]) {
            if (data_[i].dataType == "RealType")
              writeReal(i,j);
            else if (data_[i].dataType == "Vector3d") 
              writeVector(i,j);
            else {
              sprintf( painCave.errMsg,
                       "RNEMD found an unknown data type for: %s ",
                       data_[i].title.c_str());
              painCave.isFatal = 1;
              simError();
            }
          }
        }
        rnemdFile_ << std::endl;
        
      }        

      rnemdFile_ << "#######################################################\n";
      rnemdFile_ << "# 95% confidence intervals in those quantities follow:\n";
      rnemdFile_ << "#######################################################\n";


      for (int j = 0; j < nBins_; j++) {        
        rnemdFile_ << "#";
        for (unsigned int i = 0; i < outputMask_.size(); ++i) {
          if (outputMask_[i]) {
            if (data_[i].dataType == "RealType")
              writeRealErrorBars(i,j);
            else if (data_[i].dataType == "Vector3d")
              writeVectorErrorBars(i,j);
            else {
              sprintf( painCave.errMsg,
                       "RNEMD found an unknown data type for: %s ",
                       data_[i].title.c_str());
              painCave.isFatal = 1;
              simError();
            }
          }
        }
        rnemdFile_ << std::endl;
        
      }        
      
      rnemdFile_.flush();
      rnemdFile_.close();
      
#ifdef IS_MPI
    }
#endif
    
  }
  
  void RNEMD::writeReal(int index, unsigned int bin) {
    if (!doRNEMD_) return;
    assert(index >=0 && index < ENDINDEX);
    assert(int(bin) < nBins_);
    RealType s;
    int count;
    
    count = data_[index].accumulator[bin]->count();
    if (count == 0) return;
    
    dynamic_cast<Accumulator *>(data_[index].accumulator[bin])->getAverage(s);
    
    if (! isinf(s) && ! isnan(s)) {
      rnemdFile_ << "\t" << s;
    } else{
      sprintf( painCave.errMsg,
               "RNEMD detected a numerical error writing: %s for bin %u",
               data_[index].title.c_str(), bin);
      painCave.isFatal = 1;
      simError();
    }    
  }
  
  void RNEMD::writeVector(int index, unsigned int bin) {
    if (!doRNEMD_) return;
    assert(index >=0 && index < ENDINDEX);
    assert(int(bin) < nBins_);
    Vector3d s;
    int count;
    
    count = data_[index].accumulator[bin]->count();

    if (count == 0) return;

    dynamic_cast<VectorAccumulator*>(data_[index].accumulator[bin])->getAverage(s);
    if (isinf(s[0]) || isnan(s[0]) || 
        isinf(s[1]) || isnan(s[1]) || 
        isinf(s[2]) || isnan(s[2]) ) {      
      sprintf( painCave.errMsg,
               "RNEMD detected a numerical error writing: %s for bin %u",
               data_[index].title.c_str(), bin);
      painCave.isFatal = 1;
      simError();
    } else {
      rnemdFile_ << "\t" << s[0] << "\t" << s[1] << "\t" << s[2];
    }
  }  

  void RNEMD::writeRealErrorBars(int index, unsigned int bin) {
    if (!doRNEMD_) return;
    assert(index >=0 && index < ENDINDEX);
    assert(int(bin) < nBins_);
    RealType s;
    int count;
    
    count = data_[index].accumulator[bin]->count();
    if (count == 0) return;
    
    dynamic_cast<Accumulator *>(data_[index].accumulator[bin])->get95percentConfidenceInterval(s);
    
    if (! isinf(s) && ! isnan(s)) {
      rnemdFile_ << "\t" << s;
    } else{
      sprintf( painCave.errMsg,
               "RNEMD detected a numerical error writing: %s std. dev. for bin %u",
               data_[index].title.c_str(), bin);
      painCave.isFatal = 1;
      simError();
    }    
  }
  
  void RNEMD::writeVectorErrorBars(int index, unsigned int bin) {
    if (!doRNEMD_) return;
    assert(index >=0 && index < ENDINDEX);
    assert(int(bin) < nBins_);
    Vector3d s;
    int count;
    
    count = data_[index].accumulator[bin]->count();
    if (count == 0) return;

    dynamic_cast<VectorAccumulator*>(data_[index].accumulator[bin])->get95percentConfidenceInterval(s);
    if (isinf(s[0]) || isnan(s[0]) || 
        isinf(s[1]) || isnan(s[1]) || 
        isinf(s[2]) || isnan(s[2]) ) {      
      sprintf( painCave.errMsg,
               "RNEMD detected a numerical error writing: %s std. dev. for bin %u",
               data_[index].title.c_str(), bin);
      painCave.isFatal = 1;
      simError();
    } else {
      rnemdFile_ << "\t" << s[0] << "\t" << s[1] << "\t" << s[2];
    }
  }  
}

Revision:	2026
Committed:	Wed Oct 22 12:23:59 2014 UTC (10 years, 6 months ago) by gezelter
File size:	78086 byte(s)
Log Message:	Starting to add support for UniformGradient. Changed Vector3d input type to a more general std::vector<RealType> input. This change alters RNEMD and UniformField inputs.