| 1 |
tim |
1695 |
#include <math.h>
|
| 2 |
|
|
|
| 3 |
|
|
#include "primitives/Atom.hpp"
|
| 4 |
|
|
#include "primitives/SRI.hpp"
|
| 5 |
|
|
#include "primitives/AbstractClasses.hpp"
|
| 6 |
|
|
#include "brains/SimInfo.hpp"
|
| 7 |
|
|
#include "UseTheForce/ForceFields.hpp"
|
| 8 |
|
|
#include "brains/Thermo.hpp"
|
| 9 |
|
|
#include "io/ReadWrite.hpp"
|
| 10 |
|
|
#include "integrators/Integrator.hpp"
|
| 11 |
|
|
#include "utils/simError.h"
|
| 12 |
|
|
|
| 13 |
|
|
#ifdef IS_MPI
|
| 14 |
|
|
#include "brains/mpiSimulation.hpp"
|
| 15 |
|
|
#endif
|
| 16 |
|
|
|
| 17 |
|
|
|
| 18 |
|
|
// Basic isotropic thermostating and barostating via the Melchionna
|
| 19 |
|
|
// modification of the Hoover algorithm:
|
| 20 |
|
|
//
|
| 21 |
|
|
// Melchionna, S., Ciccotti, G., and Holian, B. L., 1993,
|
| 22 |
|
|
// Molec. Phys., 78, 533.
|
| 23 |
|
|
//
|
| 24 |
|
|
// and
|
| 25 |
|
|
//
|
| 26 |
|
|
// Hoover, W. G., 1986, Phys. Rev. A, 34, 2499.
|
| 27 |
|
|
|
| 28 |
|
|
template<typename T> NPT<T>::NPT ( SimInfo *theInfo, ForceFields* the_ff):
|
| 29 |
|
|
T( theInfo, the_ff )
|
| 30 |
|
|
{
|
| 31 |
|
|
GenericData* data;
|
| 32 |
|
|
DoubleGenericData * chiValue;
|
| 33 |
|
|
DoubleGenericData * integralOfChidtValue;
|
| 34 |
|
|
|
| 35 |
|
|
chiValue = NULL;
|
| 36 |
|
|
integralOfChidtValue = NULL;
|
| 37 |
|
|
|
| 38 |
|
|
chi = 0.0;
|
| 39 |
|
|
integralOfChidt = 0.0;
|
| 40 |
|
|
have_tau_thermostat = 0;
|
| 41 |
|
|
have_tau_barostat = 0;
|
| 42 |
|
|
have_target_temp = 0;
|
| 43 |
|
|
have_target_pressure = 0;
|
| 44 |
|
|
have_chi_tolerance = 0;
|
| 45 |
|
|
have_eta_tolerance = 0;
|
| 46 |
|
|
have_pos_iter_tolerance = 0;
|
| 47 |
|
|
|
| 48 |
|
|
// retrieve chi and integralOfChidt from simInfo
|
| 49 |
|
|
data = info->getProperty(CHIVALUE_ID);
|
| 50 |
|
|
if(data){
|
| 51 |
|
|
chiValue = dynamic_cast<DoubleGenericData*>(data);
|
| 52 |
|
|
}
|
| 53 |
|
|
|
| 54 |
|
|
data = info->getProperty(INTEGRALOFCHIDT_ID);
|
| 55 |
|
|
if(data){
|
| 56 |
|
|
integralOfChidtValue = dynamic_cast<DoubleGenericData*>(data);
|
| 57 |
|
|
}
|
| 58 |
|
|
|
| 59 |
|
|
// chi and integralOfChidt should appear by pair
|
| 60 |
|
|
if(chiValue && integralOfChidtValue){
|
| 61 |
|
|
chi = chiValue->getData();
|
| 62 |
|
|
integralOfChidt = integralOfChidtValue->getData();
|
| 63 |
|
|
}
|
| 64 |
|
|
|
| 65 |
|
|
oldPos = new double[3*integrableObjects.size()];
|
| 66 |
|
|
oldVel = new double[3*integrableObjects.size()];
|
| 67 |
|
|
oldJi = new double[3*integrableObjects.size()];
|
| 68 |
|
|
|
| 69 |
|
|
}
|
| 70 |
|
|
|
| 71 |
|
|
template<typename T> NPT<T>::~NPT() {
|
| 72 |
|
|
delete[] oldPos;
|
| 73 |
|
|
delete[] oldVel;
|
| 74 |
|
|
delete[] oldJi;
|
| 75 |
|
|
}
|
| 76 |
|
|
|
| 77 |
|
|
template<typename T> void NPT<T>::moveA() {
|
| 78 |
|
|
|
| 79 |
|
|
//new version of NPT
|
| 80 |
|
|
int i, j, k;
|
| 81 |
|
|
Vector3d Tb, ji;
|
| 82 |
|
|
double mass;
|
| 83 |
|
|
Vector3d vel;
|
| 84 |
|
|
Vector3d pos;
|
| 85 |
|
|
Vector3d frc;
|
| 86 |
|
|
Vector3d sc;
|
| 87 |
|
|
Vector3d COM;
|
| 88 |
|
|
|
| 89 |
|
|
instaTemp = tStats->getTemperature();
|
| 90 |
|
|
tStats->getPressureTensor( press );
|
| 91 |
|
|
instaPress = p_convert * (press[0][0] + press[1][1] + press[2][2]) / 3.0;
|
| 92 |
|
|
instaVol = tStats->getVolume();
|
| 93 |
|
|
|
| 94 |
|
|
tStats->getCOM(COM);
|
| 95 |
|
|
|
| 96 |
|
|
//evolve velocity half step
|
| 97 |
|
|
|
| 98 |
|
|
calcVelScale();
|
| 99 |
|
|
|
| 100 |
|
|
for( i=0; i<integrableObjects.size(); i++ ){
|
| 101 |
|
|
|
| 102 |
|
|
vel = integrableObjects[i]->getVel();
|
| 103 |
|
|
integrableObjects[i]->getFrc( frc );
|
| 104 |
|
|
|
| 105 |
|
|
mass = integrableObjects[i]->getMass();
|
| 106 |
|
|
|
| 107 |
|
|
getVelScaleA( sc, vel );
|
| 108 |
|
|
|
| 109 |
|
|
for (j=0; j < 3; j++) {
|
| 110 |
|
|
|
| 111 |
|
|
// velocity half step (use chi from previous step here):
|
| 112 |
|
|
vel[j] += dt2 * ((frc[j] / mass ) * eConvert - sc[j]);
|
| 113 |
|
|
|
| 114 |
|
|
}
|
| 115 |
|
|
|
| 116 |
|
|
integrableObjects[i]->setVel( vel );
|
| 117 |
|
|
|
| 118 |
|
|
if( integrableObjects[i]->isDirectional() ){
|
| 119 |
|
|
|
| 120 |
|
|
// get and convert the torque to body frame
|
| 121 |
|
|
|
| 122 |
|
|
Tb = integrableObjects[i]->getTrq();
|
| 123 |
|
|
integrableObjects[i]->lab2Body( Tb );
|
| 124 |
|
|
|
| 125 |
|
|
// get the angular momentum, and propagate a half step
|
| 126 |
|
|
|
| 127 |
|
|
ji = integrableObjects[i]->getJ();
|
| 128 |
|
|
|
| 129 |
|
|
for (j=0; j < 3; j++)
|
| 130 |
|
|
ji[j] += dt2 * (Tb[j] * eConvert - ji[j]*chi);
|
| 131 |
|
|
|
| 132 |
|
|
this->rotationPropagation( integrableObjects[i], ji );
|
| 133 |
|
|
|
| 134 |
|
|
integrableObjects[i]->setJ( ji );
|
| 135 |
|
|
}
|
| 136 |
|
|
}
|
| 137 |
|
|
|
| 138 |
|
|
// evolve chi and eta half step
|
| 139 |
|
|
|
| 140 |
|
|
evolveChiA();
|
| 141 |
|
|
evolveEtaA();
|
| 142 |
|
|
|
| 143 |
|
|
//calculate the integral of chidt
|
| 144 |
|
|
integralOfChidt += dt2*chi;
|
| 145 |
|
|
|
| 146 |
|
|
//save the old positions
|
| 147 |
|
|
for(i = 0; i < integrableObjects.size(); i++){
|
| 148 |
|
|
pos = integrableObjects[i]->getPos();
|
| 149 |
|
|
for(j = 0; j < 3; j++)
|
| 150 |
|
|
oldPos[i*3 + j] = pos[j];
|
| 151 |
|
|
}
|
| 152 |
|
|
|
| 153 |
|
|
//the first estimation of r(t+dt) is equal to r(t)
|
| 154 |
|
|
|
| 155 |
|
|
for(k = 0; k < 5; k ++){
|
| 156 |
|
|
|
| 157 |
|
|
for(i =0 ; i < integrableObjects.size(); i++){
|
| 158 |
|
|
|
| 159 |
|
|
vel = integrableObjects[i]->getVel();
|
| 160 |
|
|
pos = integrableObjects[i]->getPos();
|
| 161 |
|
|
|
| 162 |
|
|
this->getPosScale( pos, COM, i, sc );
|
| 163 |
|
|
|
| 164 |
|
|
for(j = 0; j < 3; j++)
|
| 165 |
|
|
pos[j] = oldPos[i*3 + j] + dt*(vel[j] + sc[j]);
|
| 166 |
|
|
|
| 167 |
|
|
integrableObjects[i]->setPos( pos );
|
| 168 |
|
|
}
|
| 169 |
|
|
|
| 170 |
|
|
if(nConstrained)
|
| 171 |
|
|
constrainA();
|
| 172 |
|
|
}
|
| 173 |
|
|
|
| 174 |
|
|
|
| 175 |
|
|
// Scale the box after all the positions have been moved:
|
| 176 |
|
|
|
| 177 |
|
|
this->scaleSimBox();
|
| 178 |
|
|
}
|
| 179 |
|
|
|
| 180 |
|
|
template<typename T> void NPT<T>::moveB( void ){
|
| 181 |
|
|
|
| 182 |
|
|
//new version of NPT
|
| 183 |
|
|
int i, j, k;
|
| 184 |
|
|
Vector3d Tb;
|
| 185 |
|
|
Vector3d ji;
|
| 186 |
|
|
Vector3d sc;
|
| 187 |
|
|
Vector3d vel;
|
| 188 |
|
|
Vector3d frc;
|
| 189 |
|
|
double mass;
|
| 190 |
|
|
|
| 191 |
|
|
// Set things up for the iteration:
|
| 192 |
|
|
|
| 193 |
|
|
for( i=0; i<integrableObjects.size(); i++ ){
|
| 194 |
|
|
|
| 195 |
|
|
vel = integrableObjects[i]->getVel();
|
| 196 |
|
|
|
| 197 |
|
|
for (j=0; j < 3; j++)
|
| 198 |
|
|
oldVel[3*i + j] = vel[j];
|
| 199 |
|
|
|
| 200 |
|
|
if( integrableObjects[i]->isDirectional() ){
|
| 201 |
|
|
|
| 202 |
|
|
ji = integrableObjects[i]->getJ();
|
| 203 |
|
|
|
| 204 |
|
|
for (j=0; j < 3; j++)
|
| 205 |
|
|
oldJi[3*i + j] = ji[j];
|
| 206 |
|
|
|
| 207 |
|
|
}
|
| 208 |
|
|
}
|
| 209 |
|
|
|
| 210 |
|
|
// do the iteration:
|
| 211 |
|
|
|
| 212 |
|
|
instaVol = tStats->getVolume();
|
| 213 |
|
|
|
| 214 |
|
|
for (k=0; k < 4; k++) {
|
| 215 |
|
|
|
| 216 |
|
|
instaTemp = tStats->getTemperature();
|
| 217 |
|
|
instaPress = tStats->getPressure();
|
| 218 |
|
|
|
| 219 |
|
|
// evolve chi another half step using the temperature at t + dt/2
|
| 220 |
|
|
|
| 221 |
|
|
this->evolveChiB();
|
| 222 |
|
|
this->evolveEtaB();
|
| 223 |
|
|
this->calcVelScale();
|
| 224 |
|
|
|
| 225 |
|
|
for( i=0; i<integrableObjects.size(); i++ ){
|
| 226 |
|
|
|
| 227 |
|
|
integrableObjects[i]->getFrc( frc );
|
| 228 |
|
|
vel = integrableObjects[i]->getVel();
|
| 229 |
|
|
|
| 230 |
|
|
mass = integrableObjects[i]->getMass();
|
| 231 |
|
|
|
| 232 |
|
|
getVelScaleB( sc, i );
|
| 233 |
|
|
|
| 234 |
|
|
// velocity half step
|
| 235 |
|
|
for (j=0; j < 3; j++)
|
| 236 |
|
|
vel[j] = oldVel[3*i+j] + dt2 * ((frc[j] / mass ) * eConvert - sc[j]);
|
| 237 |
|
|
|
| 238 |
|
|
integrableObjects[i]->setVel( vel );
|
| 239 |
|
|
|
| 240 |
|
|
if( integrableObjects[i]->isDirectional() ){
|
| 241 |
|
|
|
| 242 |
|
|
// get and convert the torque to body frame
|
| 243 |
|
|
|
| 244 |
|
|
Tb = integrableObjects[i]->getTrq();
|
| 245 |
|
|
integrableObjects[i]->lab2Body( Tb );
|
| 246 |
|
|
|
| 247 |
|
|
for (j=0; j < 3; j++)
|
| 248 |
|
|
ji[j] = oldJi[3*i + j] + dt2 * (Tb[j] * eConvert - oldJi[3*i+j]*chi);
|
| 249 |
|
|
|
| 250 |
|
|
integrableObjects[i]->setJ( ji );
|
| 251 |
|
|
}
|
| 252 |
|
|
}
|
| 253 |
|
|
|
| 254 |
|
|
if(nConstrained)
|
| 255 |
|
|
constrainB();
|
| 256 |
|
|
|
| 257 |
|
|
if ( this->chiConverged() && this->etaConverged() ) break;
|
| 258 |
|
|
}
|
| 259 |
|
|
|
| 260 |
|
|
//calculate integral of chida
|
| 261 |
|
|
integralOfChidt += dt2*chi;
|
| 262 |
|
|
|
| 263 |
|
|
|
| 264 |
|
|
}
|
| 265 |
|
|
|
| 266 |
|
|
template<typename T> void NPT<T>::resetIntegrator() {
|
| 267 |
|
|
chi = 0.0;
|
| 268 |
|
|
T::resetIntegrator();
|
| 269 |
|
|
}
|
| 270 |
|
|
|
| 271 |
|
|
template<typename T> void NPT<T>::evolveChiA() {
|
| 272 |
|
|
chi += dt2 * ( instaTemp / targetTemp - 1.0) / tt2;
|
| 273 |
|
|
oldChi = chi;
|
| 274 |
|
|
}
|
| 275 |
|
|
|
| 276 |
|
|
template<typename T> void NPT<T>::evolveChiB() {
|
| 277 |
|
|
|
| 278 |
|
|
prevChi = chi;
|
| 279 |
|
|
chi = oldChi + dt2 * ( instaTemp / targetTemp - 1.0) / tt2;
|
| 280 |
|
|
}
|
| 281 |
|
|
|
| 282 |
|
|
template<typename T> bool NPT<T>::chiConverged() {
|
| 283 |
|
|
|
| 284 |
|
|
return ( fabs( prevChi - chi ) <= chiTolerance );
|
| 285 |
|
|
}
|
| 286 |
|
|
|
| 287 |
|
|
template<typename T> int NPT<T>::readyCheck() {
|
| 288 |
|
|
|
| 289 |
|
|
//check parent's readyCheck() first
|
| 290 |
|
|
if (T::readyCheck() == -1)
|
| 291 |
|
|
return -1;
|
| 292 |
|
|
|
| 293 |
|
|
// First check to see if we have a target temperature.
|
| 294 |
|
|
// Not having one is fatal.
|
| 295 |
|
|
|
| 296 |
|
|
if (!have_target_temp) {
|
| 297 |
|
|
sprintf( painCave.errMsg,
|
| 298 |
|
|
"NPT error: You can't use the NPT integrator\n"
|
| 299 |
|
|
" without a targetTemp!\n"
|
| 300 |
|
|
);
|
| 301 |
|
|
painCave.isFatal = 1;
|
| 302 |
|
|
simError();
|
| 303 |
|
|
return -1;
|
| 304 |
|
|
}
|
| 305 |
|
|
|
| 306 |
|
|
if (!have_target_pressure) {
|
| 307 |
|
|
sprintf( painCave.errMsg,
|
| 308 |
|
|
"NPT error: You can't use the NPT integrator\n"
|
| 309 |
|
|
" without a targetPressure!\n"
|
| 310 |
|
|
);
|
| 311 |
|
|
painCave.isFatal = 1;
|
| 312 |
|
|
simError();
|
| 313 |
|
|
return -1;
|
| 314 |
|
|
}
|
| 315 |
|
|
|
| 316 |
|
|
// We must set tauThermostat.
|
| 317 |
|
|
|
| 318 |
|
|
if (!have_tau_thermostat) {
|
| 319 |
|
|
sprintf( painCave.errMsg,
|
| 320 |
|
|
"NPT error: If you use the NPT\n"
|
| 321 |
|
|
" integrator, you must set tauThermostat.\n");
|
| 322 |
|
|
painCave.isFatal = 1;
|
| 323 |
|
|
simError();
|
| 324 |
|
|
return -1;
|
| 325 |
|
|
}
|
| 326 |
|
|
|
| 327 |
|
|
// We must set tauBarostat.
|
| 328 |
|
|
|
| 329 |
|
|
if (!have_tau_barostat) {
|
| 330 |
|
|
sprintf( painCave.errMsg,
|
| 331 |
|
|
"If you use the NPT integrator, you must set tauBarostat.\n");
|
| 332 |
|
|
painCave.severity = OOPSE_ERROR;
|
| 333 |
|
|
painCave.isFatal = 1;
|
| 334 |
|
|
simError();
|
| 335 |
|
|
return -1;
|
| 336 |
|
|
}
|
| 337 |
|
|
|
| 338 |
|
|
if (!have_chi_tolerance) {
|
| 339 |
|
|
sprintf( painCave.errMsg,
|
| 340 |
|
|
"Setting chi tolerance to 1e-6 in NPT integrator\n");
|
| 341 |
|
|
chiTolerance = 1e-6;
|
| 342 |
|
|
have_chi_tolerance = 1;
|
| 343 |
|
|
painCave.severity = OOPSE_INFO;
|
| 344 |
|
|
painCave.isFatal = 0;
|
| 345 |
|
|
simError();
|
| 346 |
|
|
}
|
| 347 |
|
|
|
| 348 |
|
|
if (!have_eta_tolerance) {
|
| 349 |
|
|
sprintf( painCave.errMsg,
|
| 350 |
|
|
"Setting eta tolerance to 1e-6 in NPT integrator");
|
| 351 |
|
|
etaTolerance = 1e-6;
|
| 352 |
|
|
have_eta_tolerance = 1;
|
| 353 |
|
|
painCave.severity = OOPSE_INFO;
|
| 354 |
|
|
painCave.isFatal = 0;
|
| 355 |
|
|
simError();
|
| 356 |
|
|
}
|
| 357 |
|
|
|
| 358 |
|
|
// We need NkBT a lot, so just set it here: This is the RAW number
|
| 359 |
|
|
// of integrableObjects, so no subtraction or addition of constraints or
|
| 360 |
|
|
// orientational degrees of freedom:
|
| 361 |
|
|
|
| 362 |
|
|
NkBT = (double)(info->getTotIntegrableObjects()) * kB * targetTemp;
|
| 363 |
|
|
|
| 364 |
|
|
// fkBT is used because the thermostat operates on more degrees of freedom
|
| 365 |
|
|
// than the barostat (when there are particles with orientational degrees
|
| 366 |
|
|
// of freedom).
|
| 367 |
|
|
|
| 368 |
|
|
fkBT = (double)(info->getNDF()) * kB * targetTemp;
|
| 369 |
|
|
|
| 370 |
|
|
tt2 = tauThermostat * tauThermostat;
|
| 371 |
|
|
tb2 = tauBarostat * tauBarostat;
|
| 372 |
|
|
|
| 373 |
|
|
return 1;
|
| 374 |
|
|
}
|