ViewVC Help

View File | Revision Log | Show Annotations | View Changeset | Root Listing

root/OpenMD/trunk/src/rnemd/RNEMD.cpp

Comparing:
trunk/src/integrators/RNEMD.cpp (file contents), Revision 1350 by gezelter, Thu May 21 18:56:45 2009 UTC vs.
trunk/src/rnemd/RNEMD.cpp (file contents), Revision 2068 by gezelter, Thu Mar 5 15:40:58 2015 UTC

#	Line 6 | Line 6
6		* redistribute this software in source and binary code form, provided
7		* that the following conditions are met:
8		*
9	<	* 1. Acknowledgement of the program authors must be made in any
10	<	*    publication of scientific results based in part on use of the
11	<	*    program.  An acceptable form of acknowledgement is citation of
12	<	*    the article in which the program was described (Matthew
13	<	*    A. Meineke, Charles F. Vardeman II, Teng Lin, Christopher
14	<	*    J. Fennell and J. Daniel Gezelter, "OOPSE: An Object-Oriented
15	<	*    Parallel Simulation Engine for Molecular Dynamics,"
16	<	*    J. Comput. Chem. 26, pp. 252-271 (2005))
17	<	*
18	<	* 2. Redistributions of source code must retain the above copyright
9	>	* 1. Redistributions of source code must retain the above copyright
10		*    notice, this list of conditions and the following disclaimer.
11		*
12	<	* 3. Redistributions in binary form must reproduce the above copyright
12	>	* 2. Redistributions in binary form must reproduce the above copyright
13		*    notice, this list of conditions and the following disclaimer in the
14		*    documentation and/or other materials provided with the
15		*    distribution.
#	Line 37 | Line 28
28		* arising out of the use of or inability to use software, even if the
29		* University of Notre Dame has been advised of the possibility of
30		* such damages.
31	+	*
32	+	* SUPPORT OPEN SCIENCE!  If you use OpenMD or its source code in your
33	+	* research, please cite the appropriate papers when you publish your
34	+	* work.  Good starting points are:
35	+	*
36	+	* [1]  Meineke, et al., J. Comp. Chem. 26, 252-271 (2005).
37	+	* [2]  Fennell & Gezelter, J. Chem. Phys. 124, 234104 (2006).
38	+	* [3]  Sun, Lin & Gezelter, J. Chem. Phys. 128, 234107 (2008).
39	+	* [4]  Vardeman & Gezelter, in progress (2009).
40		*/
41	+	#ifdef IS_MPI
42	+	#include <mpi.h>
43	+	#endif
44
45	<	#include "integrators/RNEMD.hpp"
45	>	#include <cmath>
46	>	#include <sstream>
47	>	#include <string>
48	>
49	>	#include "rnemd/RNEMD.hpp"
50		#include "math/Vector3.hpp"
51	+	#include "math/Vector.hpp"
52		#include "math/SquareMatrix3.hpp"
53	+	#include "math/Polynomial.hpp"
54		#include "primitives/Molecule.hpp"
55		#include "primitives/StuntDouble.hpp"
56	<	#include "utils/OOPSEConstant.hpp"
56	>	#include "utils/PhysicalConstants.hpp"
57		#include "utils/Tuple.hpp"
58	+	#include "brains/Thermo.hpp"
59	+	#include "math/ConvexHull.hpp"
60
61	<	#ifndef IS_MPI
62	<	#include "math/SeqRandNumGen.hpp"
63	<	#else
53	<	#include "math/ParallelRandNumGen.hpp"
61	>	#ifdef _MSC_VER
62	>	#define isnan(x) _isnan((x))
63	>	#define isinf(x) (!_finite(x) && !_isnan(x))
64		#endif
65
66		#define HONKING_LARGE_VALUE 1.0e10
67
68	<	namespace oopse {
68	>	using namespace std;
69	>	namespace OpenMD {
70
71	<	RNEMD::RNEMD(SimInfo* info) : info_(info), evaluator_(info), seleMan_(info), usePeriodicBoundaryConditions_(info->getSimParams()->getUsePeriodicBoundaryConditions()) {
71	>	RNEMD::RNEMD(SimInfo* info) : info_(info), evaluator_(info), seleMan_(info),
72	>	evaluatorA_(info), seleManA_(info),
73	>	commonA_(info), evaluatorB_(info),
74	>	seleManB_(info), commonB_(info),
75	>	hasData_(false), hasDividingArea_(false),
76	>	usePeriodicBoundaryConditions_(info->getSimParams()->getUsePeriodicBoundaryConditions()) {
77	>
78	>	trialCount_ = 0;
79	>	failTrialCount_ = 0;
80	>	failRootCount_ = 0;
81	>
82	>	Globals* simParams = info->getSimParams();
83	>	RNEMDParameters* rnemdParams = simParams->getRNEMDParameters();
84	>
85	>	doRNEMD_ = rnemdParams->getUseRNEMD();
86	>	if (!doRNEMD_) return;
87	>
88	>	stringToMethod_["Swap"]  = rnemdSwap;
89	>	stringToMethod_["NIVS"]  = rnemdNIVS;
90	>	stringToMethod_["VSS"]   = rnemdVSS;
91	>
92	>	stringToFluxType_["KE"]  = rnemdKE;
93	>	stringToFluxType_["Px"]  = rnemdPx;
94	>	stringToFluxType_["Py"]  = rnemdPy;
95	>	stringToFluxType_["Pz"]  = rnemdPz;
96	>	stringToFluxType_["Pvector"]  = rnemdPvector;
97	>	stringToFluxType_["Lx"]  = rnemdLx;
98	>	stringToFluxType_["Ly"]  = rnemdLy;
99	>	stringToFluxType_["Lz"]  = rnemdLz;
100	>	stringToFluxType_["Lvector"]  = rnemdLvector;
101	>	stringToFluxType_["KE+Px"]  = rnemdKePx;
102	>	stringToFluxType_["KE+Py"]  = rnemdKePy;
103	>	stringToFluxType_["KE+Pvector"]  = rnemdKePvector;
104	>	stringToFluxType_["KE+Lx"]  = rnemdKeLx;
105	>	stringToFluxType_["KE+Ly"]  = rnemdKeLy;
106	>	stringToFluxType_["KE+Lz"]  = rnemdKeLz;
107	>	stringToFluxType_["KE+Lvector"]  = rnemdKeLvector;
108	>
109	>	runTime_ = simParams->getRunTime();
110	>	statusTime_ = simParams->getStatusTime();
111	>
112	>	const string methStr = rnemdParams->getMethod();
113	>	bool hasFluxType = rnemdParams->haveFluxType();
114	>
115	>	rnemdObjectSelection_ = rnemdParams->getObjectSelection();
116	>
117	>	string fluxStr;
118	>	if (hasFluxType) {
119	>	fluxStr = rnemdParams->getFluxType();
120	>	} else {
121	>	sprintf(painCave.errMsg,
122	>	"RNEMD: No fluxType was set in the md file.  This parameter,\n"
123	>	"\twhich must be one of the following values:\n"
124	>	"\tKE, Px, Py, Pz, Pvector, Lx, Ly, Lz, Lvector,\n"
125	>	"\tKE+Px, KE+Py, KE+Pvector, KE+Lx, KE+Ly, KE+Lz, KE+Lvector\n"
126	>	"\tmust be set to use RNEMD\n");
127	>	painCave.isFatal = 1;
128	>	painCave.severity = OPENMD_ERROR;
129	>	simError();
130	>	}
131	>
132	>	bool hasKineticFlux = rnemdParams->haveKineticFlux();
133	>	bool hasMomentumFlux = rnemdParams->haveMomentumFlux();
134	>	bool hasMomentumFluxVector = rnemdParams->haveMomentumFluxVector();
135	>	bool hasAngularMomentumFlux = rnemdParams->haveAngularMomentumFlux();
136	>	bool hasAngularMomentumFluxVector = rnemdParams->haveAngularMomentumFluxVector();
137	>	hasSelectionA_ = rnemdParams->haveSelectionA();
138	>	hasSelectionB_ = rnemdParams->haveSelectionB();
139	>	bool hasSlabWidth = rnemdParams->haveSlabWidth();
140	>	bool hasSlabACenter = rnemdParams->haveSlabACenter();
141	>	bool hasSlabBCenter = rnemdParams->haveSlabBCenter();
142	>	bool hasSphereARadius = rnemdParams->haveSphereARadius();
143	>	hasSphereBRadius_ = rnemdParams->haveSphereBRadius();
144	>	bool hasCoordinateOrigin = rnemdParams->haveCoordinateOrigin();
145	>	bool hasOutputFileName = rnemdParams->haveOutputFileName();
146	>	bool hasOutputFields = rnemdParams->haveOutputFields();
147
148	<	int seedValue;
149	<	Globals * simParams = info->getSimParams();
148	>	map<string, RNEMDMethod>::iterator i;
149	>	i = stringToMethod_.find(methStr);
150	>	if (i != stringToMethod_.end())
151	>	rnemdMethod_ = i->second;
152	>	else {
153	>	sprintf(painCave.errMsg,
154	>	"RNEMD: The current method,\n"
155	>	"\t\t%s is not one of the recognized\n"
156	>	"\texchange methods: Swap, NIVS, or VSS\n",
157	>	methStr.c_str());
158	>	painCave.isFatal = 1;
159	>	painCave.severity = OPENMD_ERROR;
160	>	simError();
161	>	}
162
163	<	stringToEnumMap_["Kinetic"] = rnemdKinetic;
164	<	stringToEnumMap_["Px"] = rnemdPx;
165	<	stringToEnumMap_["Py"] = rnemdPy;
166	<	stringToEnumMap_["Pz"] = rnemdPz;
167	<	stringToEnumMap_["Unknown"] = rnemdUnknown;
163	>	map<string, RNEMDFluxType>::iterator j;
164	>	j = stringToFluxType_.find(fluxStr);
165	>	if (j != stringToFluxType_.end())
166	>	rnemdFluxType_ = j->second;
167	>	else {
168	>	sprintf(painCave.errMsg,
169	>	"RNEMD: The current fluxType,\n"
170	>	"\t\t%s\n"
171	>	"\tis not one of the recognized flux types.\n",
172	>	fluxStr.c_str());
173	>	painCave.isFatal = 1;
174	>	painCave.severity = OPENMD_ERROR;
175	>	simError();
176	>	}
177
178	<	rnemdObjectSelection_ = simParams->getRNEMD_objectSelection();
179	<	evaluator_.loadScriptString(rnemdObjectSelection_);
180	<	seleMan_.setSelectionSet(evaluator_.evaluate());
178	>	bool methodFluxMismatch = false;
179	>	bool hasCorrectFlux = false;
180	>	switch(rnemdMethod_) {
181	>	case rnemdSwap:
182	>	switch (rnemdFluxType_) {
183	>	case rnemdKE:
184	>	hasCorrectFlux = hasKineticFlux;
185	>	break;
186	>	case rnemdPx:
187	>	case rnemdPy:
188	>	case rnemdPz:
189	>	hasCorrectFlux = hasMomentumFlux;
190	>	break;
191	>	default :
192	>	methodFluxMismatch = true;
193	>	break;
194	>	}
195	>	break;
196	>	case rnemdNIVS:
197	>	switch (rnemdFluxType_) {
198	>	case rnemdKE:
199	>	case rnemdRotKE:
200	>	case rnemdFullKE:
201	>	hasCorrectFlux = hasKineticFlux;
202	>	break;
203	>	case rnemdPx:
204	>	case rnemdPy:
205	>	case rnemdPz:
206	>	hasCorrectFlux = hasMomentumFlux;
207	>	break;
208	>	case rnemdKePx:
209	>	case rnemdKePy:
210	>	hasCorrectFlux = hasMomentumFlux && hasKineticFlux;
211	>	break;
212	>	default:
213	>	methodFluxMismatch = true;
214	>	break;
215	>	}
216	>	break;
217	>	case rnemdVSS:
218	>	switch (rnemdFluxType_) {
219	>	case rnemdKE:
220	>	case rnemdRotKE:
221	>	case rnemdFullKE:
222	>	hasCorrectFlux = hasKineticFlux;
223	>	break;
224	>	case rnemdPx:
225	>	case rnemdPy:
226	>	case rnemdPz:
227	>	hasCorrectFlux = hasMomentumFlux;
228	>	break;
229	>	case rnemdLx:
230	>	case rnemdLy:
231	>	case rnemdLz:
232	>	hasCorrectFlux = hasAngularMomentumFlux;
233	>	break;
234	>	case rnemdPvector:
235	>	hasCorrectFlux = hasMomentumFluxVector;
236	>	break;
237	>	case rnemdLvector:
238	>	hasCorrectFlux = hasAngularMomentumFluxVector;
239	>	break;
240	>	case rnemdKePx:
241	>	case rnemdKePy:
242	>	hasCorrectFlux = hasMomentumFlux && hasKineticFlux;
243	>	break;
244	>	case rnemdKeLx:
245	>	case rnemdKeLy:
246	>	case rnemdKeLz:
247	>	hasCorrectFlux = hasAngularMomentumFlux && hasKineticFlux;
248	>	break;
249	>	case rnemdKePvector:
250	>	hasCorrectFlux = hasMomentumFluxVector && hasKineticFlux;
251	>	break;
252	>	case rnemdKeLvector:
253	>	hasCorrectFlux = hasAngularMomentumFluxVector && hasKineticFlux;
254	>	break;
255	>	default:
256	>	methodFluxMismatch = true;
257	>	break;
258	>	}
259	>	default:
260	>	break;
261	>	}
262
263	+	if (methodFluxMismatch) {
264	+	sprintf(painCave.errMsg,
265	+	"RNEMD: The current method,\n"
266	+	"\t\t%s\n"
267	+	"\tcannot be used with the current flux type, %s\n",
268	+	methStr.c_str(), fluxStr.c_str());
269	+	painCave.isFatal = 1;
270	+	painCave.severity = OPENMD_ERROR;
271	+	simError();
272	+	}
273	+	if (!hasCorrectFlux) {
274	+	sprintf(painCave.errMsg,
275	+	"RNEMD: The current method, %s, and flux type, %s,\n"
276	+	"\tdid not have the correct flux value specified. Options\n"
277	+	"\tinclude: kineticFlux, momentumFlux, angularMomentumFlux,\n"
278	+	"\tmomentumFluxVector, and angularMomentumFluxVector.\n",
279	+	methStr.c_str(), fluxStr.c_str());
280	+	painCave.isFatal = 1;
281	+	painCave.severity = OPENMD_ERROR;
282	+	simError();
283	+	}
284
285	<	// do some sanity checking
285	>	if (hasKineticFlux) {
286	>	// convert the kcal / mol / Angstroms^2 / fs values in the md file
287	>	// into  amu / fs^3:
288	>	kineticFlux_ = rnemdParams->getKineticFlux()
289	>	* PhysicalConstants::energyConvert;
290	>	} else {
291	>	kineticFlux_ = 0.0;
292	>	}
293	>	if (hasMomentumFluxVector) {
294	>	std::vector<RealType> mf = rnemdParams->getMomentumFluxVector();
295	>	if (mf.size() != 3) {
296	>	sprintf(painCave.errMsg,
297	>	"RNEMD: Incorrect number of parameters specified for momentumFluxVector.\n"
298	>	"\tthere should be 3 parameters, but %lu were specified.\n",
299	>	mf.size());
300	>	painCave.isFatal = 1;
301	>	simError();
302	>	}
303	>	momentumFluxVector_.x() = mf[0];
304	>	momentumFluxVector_.y() = mf[1];
305	>	momentumFluxVector_.z() = mf[2];
306	>	} else {
307	>	momentumFluxVector_ = V3Zero;
308	>	if (hasMomentumFlux) {
309	>	RealType momentumFlux = rnemdParams->getMomentumFlux();
310	>	switch (rnemdFluxType_) {
311	>	case rnemdPx:
312	>	momentumFluxVector_.x() = momentumFlux;
313	>	break;
314	>	case rnemdPy:
315	>	momentumFluxVector_.y() = momentumFlux;
316	>	break;
317	>	case rnemdPz:
318	>	momentumFluxVector_.z() = momentumFlux;
319	>	break;
320	>	case rnemdKePx:
321	>	momentumFluxVector_.x() = momentumFlux;
322	>	break;
323	>	case rnemdKePy:
324	>	momentumFluxVector_.y() = momentumFlux;
325	>	break;
326	>	default:
327	>	break;
328	>	}
329	>	}
330	>	}
331	>	if (hasAngularMomentumFluxVector) {
332	>	std::vector<RealType> amf = rnemdParams->getAngularMomentumFluxVector();
333	>	if (amf.size() != 3) {
334	>	sprintf(painCave.errMsg,
335	>	"RNEMD: Incorrect number of parameters specified for angularMomentumFluxVector.\n"
336	>	"\tthere should be 3 parameters, but %lu were specified.\n",
337	>	amf.size());
338	>	painCave.isFatal = 1;
339	>	simError();
340	>	}
341	>	angularMomentumFluxVector_.x() = amf[0];
342	>	angularMomentumFluxVector_.y() = amf[1];
343	>	angularMomentumFluxVector_.z() = amf[2];
344	>	} else {
345	>	angularMomentumFluxVector_ = V3Zero;
346	>	if (hasAngularMomentumFlux) {
347	>	RealType angularMomentumFlux = rnemdParams->getAngularMomentumFlux();
348	>	switch (rnemdFluxType_) {
349	>	case rnemdLx:
350	>	angularMomentumFluxVector_.x() = angularMomentumFlux;
351	>	break;
352	>	case rnemdLy:
353	>	angularMomentumFluxVector_.y() = angularMomentumFlux;
354	>	break;
355	>	case rnemdLz:
356	>	angularMomentumFluxVector_.z() = angularMomentumFlux;
357	>	break;
358	>	case rnemdKeLx:
359	>	angularMomentumFluxVector_.x() = angularMomentumFlux;
360	>	break;
361	>	case rnemdKeLy:
362	>	angularMomentumFluxVector_.y() = angularMomentumFlux;
363	>	break;
364	>	case rnemdKeLz:
365	>	angularMomentumFluxVector_.z() = angularMomentumFlux;
366	>	break;
367	>	default:
368	>	break;
369	>	}
370	>	}
371	>	}
372
373	<	int selectionCount = seleMan_.getSelectionCount();
373	>	if (hasCoordinateOrigin) {
374	>	std::vector<RealType> co = rnemdParams->getCoordinateOrigin();
375	>	if (co.size() != 3) {
376	>	sprintf(painCave.errMsg,
377	>	"RNEMD: Incorrect number of parameters specified for coordinateOrigin.\n"
378	>	"\tthere should be 3 parameters, but %lu were specified.\n",
379	>	co.size());
380	>	painCave.isFatal = 1;
381	>	simError();
382	>	}
383	>	coordinateOrigin_.x() = co[0];
384	>	coordinateOrigin_.y() = co[1];
385	>	coordinateOrigin_.z() = co[2];
386	>	} else {
387	>	coordinateOrigin_ = V3Zero;
388	>	}
389	>
390	>	// do some sanity checking
391	>
392	>	int selectionCount = seleMan_.getSelectionCount();
393		int nIntegrable = info->getNGlobalIntegrableObjects();
80	–
394		if (selectionCount > nIntegrable) {
395		sprintf(painCave.errMsg,
396	<	"RNEMD warning: The current RNEMD_objectSelection,\n"
396	>	"RNEMD: The current objectSelection,\n"
397		"\t\t%s\n"
398		"\thas resulted in %d selected objects.  However,\n"
399		"\tthe total number of integrable objects in the system\n"
#	Line 90 | Line 403 | namespace oopse {
403		rnemdObjectSelection_.c_str(),
404		selectionCount, nIntegrable);
405		painCave.isFatal = 0;
406	+	painCave.severity = OPENMD_WARNING;
407		simError();
94	–
408		}
409
410	<	const std::string st = simParams->getRNEMD_swapType();
411	<
412	<	std::map<std::string, RNEMDTypeEnum>::iterator i;
413	<	i = stringToEnumMap_.find(st);
414	<	rnemdType_  = (i == stringToEnumMap_.end()) ? RNEMD::rnemdUnknown : i->second;
415	<
416	<	set_RNEMD_swapTime(simParams->getRNEMD_swapTime());
417	<	set_RNEMD_nBins(simParams->getRNEMD_nBins());
418	<	exchangeSum_ = 0.0;
419	<
420	<	#ifndef IS_MPI
421	<	if (simParams->haveSeed()) {
422	<	seedValue = simParams->getSeed();
423	<	randNumGen_ = new SeqRandNumGen(seedValue);
424	<	}else {
425	<	randNumGen_ = new SeqRandNumGen();
426	<	}
427	<	#else
428	<	if (simParams->haveSeed()) {
429	<	seedValue = simParams->getSeed();
430	<	randNumGen_ = new ParallelRandNumGen(seedValue);
431	<	}else {
432	<	randNumGen_ = new ParallelRandNumGen();
433	<	}
434	<	#endif
410	>	areaAccumulator_ = new Accumulator();
411	>
412	>	nBins_ = rnemdParams->getOutputBins();
413	>	binWidth_ = rnemdParams->getOutputBinWidth();
414	>
415	>	data_.resize(RNEMD::ENDINDEX);
416	>	OutputData z;
417	>	z.units =  "Angstroms";
418	>	z.title =  "Z";
419	>	z.dataType = "RealType";
420	>	z.accumulator.reserve(nBins_);
421	>	for (int i = 0; i < nBins_; i++)
422	>	z.accumulator.push_back( new Accumulator() );
423	>	data_[Z] = z;
424	>	outputMap_["Z"] =  Z;
425	>
426	>	OutputData r;
427	>	r.units =  "Angstroms";
428	>	r.title =  "R";
429	>	r.dataType = "RealType";
430	>	r.accumulator.reserve(nBins_);
431	>	for (int i = 0; i < nBins_; i++)
432	>	r.accumulator.push_back( new Accumulator() );
433	>	data_[R] = r;
434	>	outputMap_["R"] =  R;
435	>
436	>	OutputData temperature;
437	>	temperature.units =  "K";
438	>	temperature.title =  "Temperature";
439	>	temperature.dataType = "RealType";
440	>	temperature.accumulator.reserve(nBins_);
441	>	for (int i = 0; i < nBins_; i++)
442	>	temperature.accumulator.push_back( new Accumulator() );
443	>	data_[TEMPERATURE] = temperature;
444	>	outputMap_["TEMPERATURE"] =  TEMPERATURE;
445	>
446	>	OutputData velocity;
447	>	velocity.units = "angstroms/fs";
448	>	velocity.title =  "Velocity";
449	>	velocity.dataType = "Vector3d";
450	>	velocity.accumulator.reserve(nBins_);
451	>	for (int i = 0; i < nBins_; i++)
452	>	velocity.accumulator.push_back( new VectorAccumulator() );
453	>	data_[VELOCITY] = velocity;
454	>	outputMap_["VELOCITY"] = VELOCITY;
455	>
456	>	OutputData angularVelocity;
457	>	angularVelocity.units = "angstroms^2/fs";
458	>	angularVelocity.title =  "AngularVelocity";
459	>	angularVelocity.dataType = "Vector3d";
460	>	angularVelocity.accumulator.reserve(nBins_);
461	>	for (int i = 0; i < nBins_; i++)
462	>	angularVelocity.accumulator.push_back( new VectorAccumulator() );
463	>	data_[ANGULARVELOCITY] = angularVelocity;
464	>	outputMap_["ANGULARVELOCITY"] = ANGULARVELOCITY;
465	>
466	>	OutputData density;
467	>	density.units =  "g cm^-3";
468	>	density.title =  "Density";
469	>	density.dataType = "RealType";
470	>	density.accumulator.reserve(nBins_);
471	>	for (int i = 0; i < nBins_; i++)
472	>	density.accumulator.push_back( new Accumulator() );
473	>	data_[DENSITY] = density;
474	>	outputMap_["DENSITY"] =  DENSITY;
475	>
476	>	if (hasOutputFields) {
477	>	parseOutputFileFormat(rnemdParams->getOutputFields());
478	>	} else {
479	>	if (usePeriodicBoundaryConditions_)
480	>	outputMask_.set(Z);
481	>	else
482	>	outputMask_.set(R);
483	>	switch (rnemdFluxType_) {
484	>	case rnemdKE:
485	>	case rnemdRotKE:
486	>	case rnemdFullKE:
487	>	outputMask_.set(TEMPERATURE);
488	>	break;
489	>	case rnemdPx:
490	>	case rnemdPy:
491	>	outputMask_.set(VELOCITY);
492	>	break;
493	>	case rnemdPz:
494	>	case rnemdPvector:
495	>	outputMask_.set(VELOCITY);
496	>	outputMask_.set(DENSITY);
497	>	break;
498	>	case rnemdLx:
499	>	case rnemdLy:
500	>	case rnemdLz:
501	>	case rnemdLvector:
502	>	outputMask_.set(ANGULARVELOCITY);
503	>	break;
504	>	case rnemdKeLx:
505	>	case rnemdKeLy:
506	>	case rnemdKeLz:
507	>	case rnemdKeLvector:
508	>	outputMask_.set(TEMPERATURE);
509	>	outputMask_.set(ANGULARVELOCITY);
510	>	break;
511	>	case rnemdKePx:
512	>	case rnemdKePy:
513	>	outputMask_.set(TEMPERATURE);
514	>	outputMask_.set(VELOCITY);
515	>	break;
516	>	case rnemdKePvector:
517	>	outputMask_.set(TEMPERATURE);
518	>	outputMask_.set(VELOCITY);
519	>	outputMask_.set(DENSITY);
520	>	break;
521	>	default:
522	>	break;
523	>	}
524	>	}
525	>
526	>	if (hasOutputFileName) {
527	>	rnemdFileName_ = rnemdParams->getOutputFileName();
528	>	} else {
529	>	rnemdFileName_ = getPrefix(info->getFinalConfigFileName()) + ".rnemd";
530	>	}
531	>
532	>	exchangeTime_ = rnemdParams->getExchangeTime();
533	>
534	>	Snapshot* currentSnap_ = info->getSnapshotManager()->getCurrentSnapshot();
535	>	// total exchange sums are zeroed out at the beginning:
536	>
537	>	kineticExchange_ = 0.0;
538	>	momentumExchange_ = V3Zero;
539	>	angularMomentumExchange_ = V3Zero;
540	>
541	>	std::ostringstream selectionAstream;
542	>	std::ostringstream selectionBstream;
543	>
544	>	if (hasSelectionA_) {
545	>	selectionA_ = rnemdParams->getSelectionA();
546	>	} else {
547	>	if (usePeriodicBoundaryConditions_) {
548	>	Mat3x3d hmat = currentSnap_->getHmat();
549	>
550	>	if (hasSlabWidth)
551	>	slabWidth_ = rnemdParams->getSlabWidth();
552	>	else
553	>	slabWidth_ = hmat(2,2) / 10.0;
554	>
555	>	if (hasSlabACenter)
556	>	slabACenter_ = rnemdParams->getSlabACenter();
557	>	else
558	>	slabACenter_ = 0.0;
559	>
560	>	selectionAstream << "select wrappedz > "
561	>	<< slabACenter_ - 0.5*slabWidth_
562	>	<<  " && wrappedz < "
563	>	<< slabACenter_ + 0.5*slabWidth_;
564	>	selectionA_ = selectionAstream.str();
565	>	} else {
566	>	if (hasSphereARadius)
567	>	sphereARadius_ = rnemdParams->getSphereARadius();
568	>	else {
569	>	// use an initial guess to the size of the inner slab to be 1/10 the
570	>	// radius of an approximately spherical hull:
571	>	Thermo thermo(info);
572	>	RealType hVol = thermo.getHullVolume();
573	>	sphereARadius_ = 0.1 * pow((3.0 * hVol / (4.0 * M_PI)), 1.0/3.0);
574	>	}
575	>	selectionAstream << "select r < " << sphereARadius_;
576	>	selectionA_ = selectionAstream.str();
577	>	}
578	>	}
579	>
580	>	if (hasSelectionB_) {
581	>	selectionB_ = rnemdParams->getSelectionB();
582	>
583	>	} else {
584	>	if (usePeriodicBoundaryConditions_) {
585	>	Mat3x3d hmat = currentSnap_->getHmat();
586	>
587	>	if (hasSlabWidth)
588	>	slabWidth_ = rnemdParams->getSlabWidth();
589	>	else
590	>	slabWidth_ = hmat(2,2) / 10.0;
591	>
592	>	if (hasSlabBCenter)
593	>	slabBCenter_ = rnemdParams->getSlabBCenter();
594	>	else
595	>	slabBCenter_ = hmat(2,2) / 2.0;
596	>
597	>	selectionBstream << "select wrappedz > "
598	>	<< slabBCenter_ - 0.5*slabWidth_
599	>	<<  " && wrappedz < "
600	>	<< slabBCenter_ + 0.5*slabWidth_;
601	>	selectionB_ = selectionBstream.str();
602	>	} else {
603	>	if (hasSphereBRadius_) {
604	>	sphereBRadius_ = rnemdParams->getSphereBRadius();
605	>	selectionBstream << "select r > " << sphereBRadius_;
606	>	selectionB_ = selectionBstream.str();
607	>	} else {
608	>	selectionB_ = "select hull";
609	>	BisHull_ = true;
610	>	hasSelectionB_ = true;
611	>	}
612	>	}
613	>	}
614	>
615	>
616	>	// object evaluator:
617	>	evaluator_.loadScriptString(rnemdObjectSelection_);
618	>	seleMan_.setSelectionSet(evaluator_.evaluate());
619	>	evaluatorA_.loadScriptString(selectionA_);
620	>	evaluatorB_.loadScriptString(selectionB_);
621	>	seleManA_.setSelectionSet(evaluatorA_.evaluate());
622	>	seleManB_.setSelectionSet(evaluatorB_.evaluate());
623	>	commonA_ = seleManA_ & seleMan_;
624	>	commonB_ = seleManB_ & seleMan_;
625		}
626
627	+
628		RNEMD::~RNEMD() {
629	<	delete randNumGen_;
630	<	}
629	>	if (!doRNEMD_) return;
630	>	#ifdef IS_MPI
631	>	if (worldRank == 0) {
632	>	#endif
633
634	<	void RNEMD::doSwap() {
129	<	int midBin = nBins_ / 2;
634	>	writeOutputFile();
635
636	+	rnemdFile_.close();
637	+
638	+	#ifdef IS_MPI
639	+	}
640	+	#endif
641	+
642	+	// delete all of the objects we created:
643	+	delete areaAccumulator_;
644	+	data_.clear();
645	+	}
646	+
647	+	void RNEMD::doSwap(SelectionManager& smanA, SelectionManager& smanB) {
648	+	if (!doRNEMD_) return;
649	+	int selei;
650	+	int selej;
651	+
652		Snapshot* currentSnap_ = info_->getSnapshotManager()->getCurrentSnapshot();
653		Mat3x3d hmat = currentSnap_->getHmat();
654
134	–	seleMan_.setSelectionSet(evaluator_.evaluate());
135	–
136	–	int selei;
655		StuntDouble* sd;
138	–	int idx;
656
657	<	RealType min_val;
658	<	bool min_found = false;
659	<	StuntDouble* min_sd;
657	>	RealType min_val(0.0);
658	>	int min_found = 0;
659	>	StuntDouble* min_sd = NULL;
660
661	<	RealType max_val;
662	<	bool max_found = false;
663	<	StuntDouble* max_sd;
661	>	RealType max_val(0.0);
662	>	int max_found = 0;
663	>	StuntDouble* max_sd = NULL;
664
665	<	for (sd = seleMan_.beginSelected(selei); sd != NULL;
666	<	sd = seleMan_.nextSelected(selei)) {
665	>	for (sd = seleManA_.beginSelected(selei); sd != NULL;
666	>	sd = seleManA_.nextSelected(selei)) {
667
151	–	idx = sd->getLocalIndex();
152	–
668		Vector3d pos = sd->getPos();
669	<
669	>
670		// wrap the stuntdouble's position back into the box:
671	<
671	>
672		if (usePeriodicBoundaryConditions_)
673		currentSnap_->wrapVector(pos);
674	<
675	<	// which bin is this stuntdouble in?
676	<	// wrapped positions are in the range [-0.5*hmat(2,2), +0.5*hmat(2,2)]
677	<
678	<	int binNo = int(nBins_ * (pos.z() / hmat(2,2) + 0.5)) % nBins_;
679	<
680	<
166	<	// if we're in bin 0 or the middleBin
167	<	if (binNo == 0 || binNo == midBin) {
674	>
675	>	RealType mass = sd->getMass();
676	>	Vector3d vel = sd->getVel();
677	>	RealType value;
678	>
679	>	switch(rnemdFluxType_) {
680	>	case rnemdKE :
681
682	<	RealType mass = sd->getMass();
683	<	Vector3d vel = sd->getVel();
684	<	RealType value;
685	<
686	<	switch(rnemdType_) {
174	<	case rnemdKinetic :
682	>	value = mass * vel.lengthSquare();
683	>
684	>	if (sd->isDirectional()) {
685	>	Vector3d angMom = sd->getJ();
686	>	Mat3x3d I = sd->getI();
687
688	<	value = mass * (vel[0]*vel[0] + vel[1]*vel[1] +
689	<	vel[2]*vel[2]);
690	<	if (sd->isDirectional()) {
691	<	Vector3d angMom = sd->getJ();
692	<	Mat3x3d I = sd->getI();
693	<
694	<	if (sd->isLinear()) {
695	<	int i = sd->linearAxis();
696	<	int j = (i + 1) % 3;
697	<	int k = (i + 2) % 3;
186	<	value += angMom[j] * angMom[j] / I(j, j) +
187	<	angMom[k] * angMom[k] / I(k, k);
188	<	} else {
189	<	value += angMom[0]*angMom[0]/I(0, 0)
190	<	+ angMom[1]*angMom[1]/I(1, 1)
191	<	+ angMom[2]*angMom[2]/I(2, 2);
192	<	}
688	>	if (sd->isLinear()) {
689	>	int i = sd->linearAxis();
690	>	int j = (i + 1) % 3;
691	>	int k = (i + 2) % 3;
692	>	value += angMom[j] * angMom[j] / I(j, j) +
693	>	angMom[k] * angMom[k] / I(k, k);
694	>	} else {
695	>	value += angMom[0]*angMom[0]/I(0, 0)
696	>	+ angMom[1]*angMom[1]/I(1, 1)
697	>	+ angMom[2]*angMom[2]/I(2, 2);
698		}
699	<	value = value * 0.5 / OOPSEConstant::energyConvert;
700	<	break;
701	<	case rnemdPx :
702	<	value = mass * vel[0];
703	<	break;
704	<	case rnemdPy :
705	<	value = mass * vel[1];
706	<	break;
707	<	case rnemdPz :
708	<	value = mass * vel[2];
709	<	break;
710	<	case rnemdUnknown :
711	<	default :
712	<	break;
699	>	} //angular momenta exchange enabled
700	>	value *= 0.5;
701	>	break;
702	>	case rnemdPx :
703	>	value = mass * vel[0];
704	>	break;
705	>	case rnemdPy :
706	>	value = mass * vel[1];
707	>	break;
708	>	case rnemdPz :
709	>	value = mass * vel[2];
710	>	break;
711	>	default :
712	>	break;
713	>	}
714	>	if (!max_found) {
715	>	max_val = value;
716	>	max_sd = sd;
717	>	max_found = 1;
718	>	} else {
719	>	if (max_val < value) {
720	>	max_val = value;
721	>	max_sd = sd;
722		}
723	+	}
724	+	}
725
726	<	if (binNo == 0) {
727	<	if (!min_found) {
728	<	min_val = value;
729	<	min_sd = sd;
730	<	min_found = true;
731	<	} else {
732	<	if (min_val > value) {
733	<	min_val = value;
734	<	min_sd = sd;
735	<	}
736	<	}
737	<	} else {
738	<	if (!max_found) {
739	<	max_val = value;
740	<	max_sd = sd;
741	<	max_found = true;
742	<	} else {
743	<	if (max_val < value) {
744	<	max_val = value;
745	<	max_sd = sd;
746	<	}
747	<	}
748	<	}
726	>	for (sd = seleManB_.beginSelected(selej); sd != NULL;
727	>	sd = seleManB_.nextSelected(selej)) {
728	>
729	>	Vector3d pos = sd->getPos();
730	>
731	>	// wrap the stuntdouble's position back into the box:
732	>
733	>	if (usePeriodicBoundaryConditions_)
734	>	currentSnap_->wrapVector(pos);
735	>
736	>	RealType mass = sd->getMass();
737	>	Vector3d vel = sd->getVel();
738	>	RealType value;
739	>
740	>	switch(rnemdFluxType_) {
741	>	case rnemdKE :
742	>
743	>	value = mass * vel.lengthSquare();
744	>
745	>	if (sd->isDirectional()) {
746	>	Vector3d angMom = sd->getJ();
747	>	Mat3x3d I = sd->getI();
748	>
749	>	if (sd->isLinear()) {
750	>	int i = sd->linearAxis();
751	>	int j = (i + 1) % 3;
752	>	int k = (i + 2) % 3;
753	>	value += angMom[j] * angMom[j] / I(j, j) +
754	>	angMom[k] * angMom[k] / I(k, k);
755	>	} else {
756	>	value += angMom[0]*angMom[0]/I(0, 0)
757	>	+ angMom[1]*angMom[1]/I(1, 1)
758	>	+ angMom[2]*angMom[2]/I(2, 2);
759	>	}
760	>	} //angular momenta exchange enabled
761	>	value *= 0.5;
762	>	break;
763	>	case rnemdPx :
764	>	value = mass * vel[0];
765	>	break;
766	>	case rnemdPy :
767	>	value = mass * vel[1];
768	>	break;
769	>	case rnemdPz :
770	>	value = mass * vel[2];
771	>	break;
772	>	default :
773	>	break;
774		}
775	+
776	+	if (!min_found) {
777	+	min_val = value;
778	+	min_sd = sd;
779	+	min_found = 1;
780	+	} else {
781	+	if (min_val > value) {
782	+	min_val = value;
783	+	min_sd = sd;
784	+	}
785	+	}
786		}
787	+
788	+	#ifdef IS_MPI
789	+	int worldRank;
790	+	MPI_Comm_rank( MPI_COMM_WORLD, &worldRank);
791	+
792	+	int my_min_found = min_found;
793	+	int my_max_found = max_found;
794
236	–	#ifdef IS_MPI
237	–	int nProc, worldRank;
238	–
239	–	nProc = MPI::COMM_WORLD.Get_size();
240	–	worldRank = MPI::COMM_WORLD.Get_rank();
241	–
242	–	bool my_min_found = min_found;
243	–	bool my_max_found = max_found;
244	–
795		// Even if we didn't find a minimum, did someone else?
796	<	MPI::COMM_WORLD.Allreduce(&my_min_found, &min_found,
797	<	1, MPI::BOOL, MPI::LAND);
248	<
796	>	MPI_Allreduce(&my_min_found, &min_found, 1, MPI_INT, MPI_LOR,
797	>	MPI_COMM_WORLD);
798		// Even if we didn't find a maximum, did someone else?
799	<	MPI::COMM_WORLD.Allreduce(&my_max_found, &max_found,
800	<	1, MPI::BOOL, MPI::LAND);
801	<
802	<	struct {
803	<	RealType val;
804	<	int rank;
805	<	} max_vals, min_vals;
806	<
807	<	if (min_found) {
808	<	if (my_min_found)
799	>	MPI_Allreduce(&my_max_found, &max_found, 1, MPI_INT, MPI_LOR,
800	>	MPI_COMM_WORLD);
801	>	#endif
802	>
803	>	if (max_found && min_found) {
804	>
805	>	#ifdef IS_MPI
806	>	struct {
807	>	RealType val;
808	>	int rank;
809	>	} max_vals, min_vals;
810	>
811	>	if (my_min_found) {
812		min_vals.val = min_val;
813	<	else
813	>	} else {
814		min_vals.val = HONKING_LARGE_VALUE;
815	<
815	>	}
816		min_vals.rank = worldRank;
817
818		// Who had the minimum?
819	<	MPI::COMM_WORLD.Allreduce(&min_vals, &min_vals,
820	<	1, MPI::REALTYPE_INT, MPI::MINLOC);
819	>	MPI_Allreduce(&min_vals, &min_vals,
820	>	1, MPI_REALTYPE_INT, MPI_MINLOC, MPI_COMM_WORLD);
821		min_val = min_vals.val;
270	–	}
822
823	<	if (max_found) {
273	<	if (my_max_found)
823	>	if (my_max_found) {
824		max_vals.val = max_val;
825	<	else
825	>	} else {
826		max_vals.val = -HONKING_LARGE_VALUE;
827	<
827	>	}
828		max_vals.rank = worldRank;
829
830		// Who had the maximum?
831	<	MPI::COMM_WORLD.Allreduce(&max_vals, &max_vals,
832	<	1, MPI::REALTYPE_INT, MPI::MAXLOC);
831	>	MPI_Allreduce(&max_vals, &max_vals,
832	>	1, MPI_REALTYPE_INT, MPI_MAXLOC, MPI_COMM_WORLD);
833		max_val = max_vals.val;
284	–	}
834		#endif
835	<
836	<	if (max_found && min_found) {
837	<	if (min_val< max_val) {
289	<
835	>
836	>	if (min_val < max_val) {
837	>
838		#ifdef IS_MPI
839		if (max_vals.rank == worldRank && min_vals.rank == worldRank) {
840		// I have both maximum and minimum, so proceed like a single
841		// processor version:
842		#endif
843	<	// objects to be swapped: velocity & angular velocity
843	>
844		Vector3d min_vel = min_sd->getVel();
845		Vector3d max_vel = max_sd->getVel();
846		RealType temp_vel;
847
848	<	switch(rnemdType_) {
849	<	case rnemdKinetic :
848	>	switch(rnemdFluxType_) {
849	>	case rnemdKE :
850		min_sd->setVel(max_vel);
851		max_sd->setVel(min_vel);
852	<	if (min_sd->isDirectional() && max_sd->isDirectional()) {
852	>	if (min_sd->isDirectional() && max_sd->isDirectional()) {
853		Vector3d min_angMom = min_sd->getJ();
854		Vector3d max_angMom = max_sd->getJ();
855		min_sd->setJ(max_angMom);
856		max_sd->setJ(min_angMom);
857	<	}
857	>	}//angular momenta exchange enabled
858	>	//assumes same rigid body identity
859		break;
860		case rnemdPx :
861		temp_vel = min_vel.x();
#	Line 329 | Line 878 | namespace oopse {
878		min_sd->setVel(min_vel);
879		max_sd->setVel(max_vel);
880		break;
332	–	case rnemdUnknown :
881		default :
882		break;
883		}
884	+
885		#ifdef IS_MPI
886		// the rest of the cases only apply in parallel simulations:
887		} else if (max_vals.rank == worldRank) {
#	Line 340 | Line 889 | namespace oopse {
889
890		Vector3d min_vel;
891		Vector3d max_vel = max_sd->getVel();
892	<	MPI::Status status;
892	>	MPI_Status status;
893
894		// point-to-point swap of the velocity vector
895	<	MPI::COMM_WORLD.Sendrecv(max_vel.getArrayPointer(), 3, MPI::REALTYPE,
896	<	min_vals.rank, 0,
897	<	min_vel.getArrayPointer(), 3, MPI::REALTYPE,
898	<	min_vals.rank, 0, status);
895	>	MPI_Sendrecv(max_vel.getArrayPointer(), 3, MPI_REALTYPE,
896	>	min_vals.rank, 0,
897	>	min_vel.getArrayPointer(), 3, MPI_REALTYPE,
898	>	min_vals.rank, 0, MPI_COMM_WORLD, &status);
899
900	<	switch(rnemdType_) {
901	<	case rnemdKinetic :
900	>	switch(rnemdFluxType_) {
901	>	case rnemdKE :
902		max_sd->setVel(min_vel);
903	<
903	>	//angular momenta exchange enabled
904		if (max_sd->isDirectional()) {
905		Vector3d min_angMom;
906		Vector3d max_angMom = max_sd->getJ();
907	<
907	>
908		// point-to-point swap of the angular momentum vector
909	<	MPI::COMM_WORLD.Sendrecv(max_angMom.getArrayPointer(), 3,
910	<	MPI::REALTYPE, min_vals.rank, 1,
911	<	min_angMom.getArrayPointer(), 3,
912	<	MPI::REALTYPE, min_vals.rank, 1,
913	<	status);
914	<
909	>	MPI_Sendrecv(max_angMom.getArrayPointer(), 3,
910	>	MPI_REALTYPE, min_vals.rank, 1,
911	>	min_angMom.getArrayPointer(), 3,
912	>	MPI_REALTYPE, min_vals.rank, 1,
913	>	MPI_COMM_WORLD, &status);
914	>
915		max_sd->setJ(min_angMom);
916	<	}
916	>	}
917		break;
918		case rnemdPx :
919		max_vel.x() = min_vel.x();
#	Line 378 | Line 927 | namespace oopse {
927		max_vel.z() = min_vel.z();
928		max_sd->setVel(max_vel);
929		break;
381	–	case rnemdUnknown :
930		default :
931		break;
932		}
#	Line 387 | Line 935 | namespace oopse {
935
936		Vector3d max_vel;
937		Vector3d min_vel = min_sd->getVel();
938	<	MPI::Status status;
938	>	MPI_Status status;
939
940		// point-to-point swap of the velocity vector
941	<	MPI::COMM_WORLD.Sendrecv(min_vel.getArrayPointer(), 3, MPI::REALTYPE,
942	<	max_vals.rank, 0,
943	<	max_vel.getArrayPointer(), 3, MPI::REALTYPE,
944	<	max_vals.rank, 0, status);
941	>	MPI_Sendrecv(min_vel.getArrayPointer(), 3, MPI_REALTYPE,
942	>	max_vals.rank, 0,
943	>	max_vel.getArrayPointer(), 3, MPI_REALTYPE,
944	>	max_vals.rank, 0, MPI_COMM_WORLD, &status);
945
946	<	switch(rnemdType_) {
947	<	case rnemdKinetic :
946	>	switch(rnemdFluxType_) {
947	>	case rnemdKE :
948		min_sd->setVel(max_vel);
949	<
949	>	//angular momenta exchange enabled
950		if (min_sd->isDirectional()) {
951		Vector3d min_angMom = min_sd->getJ();
952		Vector3d max_angMom;
953	<
953	>
954		// point-to-point swap of the angular momentum vector
955	<	MPI::COMM_WORLD.Sendrecv(min_angMom.getArrayPointer(), 3,
956	<	MPI::REALTYPE, max_vals.rank, 1,
957	<	max_angMom.getArrayPointer(), 3,
958	<	MPI::REALTYPE, max_vals.rank, 1,
959	<	status);
960	<
955	>	MPI_Sendrecv(min_angMom.getArrayPointer(), 3,
956	>	MPI_REALTYPE, max_vals.rank, 1,
957	>	max_angMom.getArrayPointer(), 3,
958	>	MPI_REALTYPE, max_vals.rank, 1,
959	>	MPI_COMM_WORLD, &status);
960	>
961		min_sd->setJ(max_angMom);
962		}
963		break;
#	Line 425 | Line 973 | namespace oopse {
973		min_vel.z() = max_vel.z();
974		min_sd->setVel(min_vel);
975		break;
428	–	case rnemdUnknown :
976		default :
977		break;
978		}
979		}
980		#endif
981	<	exchangeSum_ += max_val - min_val;
982	<	} else {
983	<	std::cerr << "exchange NOT performed.\nmin_val > max_val.\n";
981	>
982	>	switch(rnemdFluxType_) {
983	>	case rnemdKE:
984	>	kineticExchange_ += max_val - min_val;
985	>	break;
986	>	case rnemdPx:
987	>	momentumExchange_.x() += max_val - min_val;
988	>	break;
989	>	case rnemdPy:
990	>	momentumExchange_.y() += max_val - min_val;
991	>	break;
992	>	case rnemdPz:
993	>	momentumExchange_.z() += max_val - min_val;
994	>	break;
995	>	default:
996	>	break;
997	>	}
998	>	} else {
999	>	sprintf(painCave.errMsg,
1000	>	"RNEMD::doSwap exchange NOT performed because min_val > max_val\n");
1001	>	painCave.isFatal = 0;
1002	>	painCave.severity = OPENMD_INFO;
1003	>	simError();
1004	>	failTrialCount_++;
1005		}
1006		} else {
1007	<	std::cerr << "exchange NOT performed.\none of the two slabs empty.\n";
1008	<	}
1009	<
1007	>	sprintf(painCave.errMsg,
1008	>	"RNEMD::doSwap exchange NOT performed because selected object\n"
1009	>	"\twas not present in at least one of the two slabs.\n");
1010	>	painCave.isFatal = 0;
1011	>	painCave.severity = OPENMD_INFO;
1012	>	simError();
1013	>	failTrialCount_++;
1014	>	}
1015		}
1016
1017	<	void RNEMD::getStatus() {
1017	>	void RNEMD::doNIVS(SelectionManager& smanA, SelectionManager& smanB) {
1018	>	if (!doRNEMD_) return;
1019	>	int selei;
1020	>	int selej;
1021
1022		Snapshot* currentSnap_ = info_->getSnapshotManager()->getCurrentSnapshot();
1023		Mat3x3d hmat = currentSnap_->getHmat();
448	–	Stats& stat = currentSnap_->statData;
449	–	RealType time = currentSnap_->getTime();
1024
1025	<	stat[Stats::RNEMD_SWAP_TOTAL] = exchangeSum_;
1025	>	StuntDouble* sd;
1026
1027	<	seleMan_.setSelectionSet(evaluator_.evaluate());
1027	>	vector<StuntDouble*> hotBin, coldBin;
1028
1029	<	int selei;
1030	<	StuntDouble* sd;
1031	<	int idx;
1029	>	RealType Phx = 0.0;
1030	>	RealType Phy = 0.0;
1031	>	RealType Phz = 0.0;
1032	>	RealType Khx = 0.0;
1033	>	RealType Khy = 0.0;
1034	>	RealType Khz = 0.0;
1035	>	RealType Khw = 0.0;
1036	>	RealType Pcx = 0.0;
1037	>	RealType Pcy = 0.0;
1038	>	RealType Pcz = 0.0;
1039	>	RealType Kcx = 0.0;
1040	>	RealType Kcy = 0.0;
1041	>	RealType Kcz = 0.0;
1042	>	RealType Kcw = 0.0;
1043
1044	<	std::vector<RealType> valueHist(nBins_, 0.0); // keeps track of what's
1045	<	// being averaged
461	<	std::vector<int> valueCount(nBins_, 0);       // keeps track of the
462	<	// number of degrees of
463	<	// freedom being averaged
1044	>	for (sd = smanA.beginSelected(selei); sd != NULL;
1045	>	sd = smanA.nextSelected(selei)) {
1046
465	–	for (sd = seleMan_.beginSelected(selei); sd != NULL;
466	–	sd = seleMan_.nextSelected(selei)) {
467	–
468	–	idx = sd->getLocalIndex();
469	–
1047		Vector3d pos = sd->getPos();
1048	<
1048	>
1049		// wrap the stuntdouble's position back into the box:
1050
1051		if (usePeriodicBoundaryConditions_)
1052		currentSnap_->wrapVector(pos);
1053
477	–	// which bin is this stuntdouble in?
478	–	// wrapped positions are in the range [-0.5*hmat(2,2), +0.5*hmat(2,2)]
1054
1055	<	int binNo = int(nBins_ * (pos.z() / hmat(2,2) + 0.5)) % nBins_;
1055	>	RealType mass = sd->getMass();
1056	>	Vector3d vel = sd->getVel();
1057
1058	+	hotBin.push_back(sd);
1059	+	Phx += mass * vel.x();
1060	+	Phy += mass * vel.y();
1061	+	Phz += mass * vel.z();
1062	+	Khx += mass * vel.x() * vel.x();
1063	+	Khy += mass * vel.y() * vel.y();
1064	+	Khz += mass * vel.z() * vel.z();
1065	+	if (sd->isDirectional()) {
1066	+	Vector3d angMom = sd->getJ();
1067	+	Mat3x3d I = sd->getI();
1068	+	if (sd->isLinear()) {
1069	+	int i = sd->linearAxis();
1070	+	int j = (i + 1) % 3;
1071	+	int k = (i + 2) % 3;
1072	+	Khw += angMom[j] * angMom[j] / I(j, j) +
1073	+	angMom[k] * angMom[k] / I(k, k);
1074	+	} else {
1075	+	Khw += angMom[0]*angMom[0]/I(0, 0)
1076	+	+ angMom[1]*angMom[1]/I(1, 1)
1077	+	+ angMom[2]*angMom[2]/I(2, 2);
1078	+	}
1079	+	}
1080	+	}
1081	+	for (sd = smanB.beginSelected(selej); sd != NULL;
1082	+	sd = smanB.nextSelected(selej)) {
1083	+	Vector3d pos = sd->getPos();
1084	+
1085	+	// wrap the stuntdouble's position back into the box:
1086	+
1087	+	if (usePeriodicBoundaryConditions_)
1088	+	currentSnap_->wrapVector(pos);
1089	+
1090		RealType mass = sd->getMass();
1091		Vector3d vel = sd->getVel();
484	–	RealType value;
1092
1093	<	switch(rnemdType_) {
1094	<	case rnemdKinetic :
1095	<
1096	<	value = mass * (vel[0]*vel[0] + vel[1]*vel[1] +
1097	<	vel[2]*vel[2]);
1098	<
1099	<	valueCount[binNo] += 3;
1100	<	if (sd->isDirectional()) {
1101	<	Vector3d angMom = sd->getJ();
1102	<	Mat3x3d I = sd->getI();
1103	<
1104	<	if (sd->isLinear()) {
1105	<	int i = sd->linearAxis();
1106	<	int j = (i + 1) % 3;
1107	<	int k = (i + 2) % 3;
1108	<	value += angMom[j] * angMom[j] / I(j, j) +
1109	<	angMom[k] * angMom[k] / I(k, k);
1093	>	coldBin.push_back(sd);
1094	>	Pcx += mass * vel.x();
1095	>	Pcy += mass * vel.y();
1096	>	Pcz += mass * vel.z();
1097	>	Kcx += mass * vel.x() * vel.x();
1098	>	Kcy += mass * vel.y() * vel.y();
1099	>	Kcz += mass * vel.z() * vel.z();
1100	>	if (sd->isDirectional()) {
1101	>	Vector3d angMom = sd->getJ();
1102	>	Mat3x3d I = sd->getI();
1103	>	if (sd->isLinear()) {
1104	>	int i = sd->linearAxis();
1105	>	int j = (i + 1) % 3;
1106	>	int k = (i + 2) % 3;
1107	>	Kcw += angMom[j] * angMom[j] / I(j, j) +
1108	>	angMom[k] * angMom[k] / I(k, k);
1109	>	} else {
1110	>	Kcw += angMom[0]*angMom[0]/I(0, 0)
1111	>	+ angMom[1]*angMom[1]/I(1, 1)
1112	>	+ angMom[2]*angMom[2]/I(2, 2);
1113	>	}
1114	>	}
1115	>	}
1116	>
1117	>	Khx *= 0.5;
1118	>	Khy *= 0.5;
1119	>	Khz *= 0.5;
1120	>	Khw *= 0.5;
1121	>	Kcx *= 0.5;
1122	>	Kcy *= 0.5;
1123	>	Kcz *= 0.5;
1124	>	Kcw *= 0.5;
1125
1126	<	valueCount[binNo] +=2;
1126	>	#ifdef IS_MPI
1127	>	MPI_Allreduce(MPI_IN_PLACE, &Phx, 1, MPI_REALTYPE, MPI_SUM, MPI_COMM_WORLD);
1128	>	MPI_Allreduce(MPI_IN_PLACE, &Phy, 1, MPI_REALTYPE, MPI_SUM, MPI_COMM_WORLD);
1129	>	MPI_Allreduce(MPI_IN_PLACE, &Phz, 1, MPI_REALTYPE, MPI_SUM, MPI_COMM_WORLD);
1130	>	MPI_Allreduce(MPI_IN_PLACE, &Pcx, 1, MPI_REALTYPE, MPI_SUM, MPI_COMM_WORLD);
1131	>	MPI_Allreduce(MPI_IN_PLACE, &Pcy, 1, MPI_REALTYPE, MPI_SUM, MPI_COMM_WORLD);
1132	>	MPI_Allreduce(MPI_IN_PLACE, &Pcz, 1, MPI_REALTYPE, MPI_SUM, MPI_COMM_WORLD);
1133
1134	<	} else {
1135	<	value += angMom[0]*angMom[0]/I(0, 0)
1136	<	+ angMom[1]*angMom[1]/I(1, 1)
1137	<	+ angMom[2]*angMom[2]/I(2, 2);
1138	<	valueCount[binNo] +=3;
1134	>	MPI_Allreduce(MPI_IN_PLACE, &Khx, 1, MPI_REALTYPE, MPI_SUM, MPI_COMM_WORLD);
1135	>	MPI_Allreduce(MPI_IN_PLACE, &Khy, 1, MPI_REALTYPE, MPI_SUM, MPI_COMM_WORLD);
1136	>	MPI_Allreduce(MPI_IN_PLACE, &Khz, 1, MPI_REALTYPE, MPI_SUM, MPI_COMM_WORLD);
1137	>	MPI_Allreduce(MPI_IN_PLACE, &Khw, 1, MPI_REALTYPE, MPI_SUM, MPI_COMM_WORLD);
1138	>
1139	>	MPI_Allreduce(MPI_IN_PLACE, &Kcx, 1, MPI_REALTYPE, MPI_SUM, MPI_COMM_WORLD);
1140	>	MPI_Allreduce(MPI_IN_PLACE, &Kcy, 1, MPI_REALTYPE, MPI_SUM, MPI_COMM_WORLD);
1141	>	MPI_Allreduce(MPI_IN_PLACE, &Kcz, 1, MPI_REALTYPE, MPI_SUM, MPI_COMM_WORLD);
1142	>	MPI_Allreduce(MPI_IN_PLACE, &Kcw, 1, MPI_REALTYPE, MPI_SUM, MPI_COMM_WORLD);
1143	>	#endif
1144	>
1145	>	//solve coldBin coeff's first
1146	>	RealType px = Pcx / Phx;
1147	>	RealType py = Pcy / Phy;
1148	>	RealType pz = Pcz / Phz;
1149	>	RealType c, x, y, z;
1150	>	bool successfulScale = false;
1151	>	if ((rnemdFluxType_ == rnemdFullKE) ||
1152	>	(rnemdFluxType_ == rnemdRotKE)) {
1153	>	//may need sanity check Khw & Kcw > 0
1154	>
1155	>	if (rnemdFluxType_ == rnemdFullKE) {
1156	>	c = 1.0 - kineticTarget_ / (Kcx + Kcy + Kcz + Kcw);
1157	>	} else {
1158	>	c = 1.0 - kineticTarget_ / Kcw;
1159	>	}
1160	>
1161	>	if ((c > 0.81) && (c < 1.21)) {//restrict scaling coefficients
1162	>	c = sqrt(c);
1163	>
1164	>	RealType w = 0.0;
1165	>	if (rnemdFluxType_ ==  rnemdFullKE) {
1166	>	x = 1.0 + px * (1.0 - c);
1167	>	y = 1.0 + py * (1.0 - c);
1168	>	z = 1.0 + pz * (1.0 - c);
1169	>	/* more complicated way
1170	>	w = 1.0 + (Kcw - Kcw * c * c - (c * c * (Kcx + Kcy + Kcz
1171	>	+ Khx * px * px + Khy * py * py + Khz * pz * pz)
1172	>	- 2.0 * c * (Khx * px * (1.0 + px) + Khy * py * (1.0 + py)
1173	>	+ Khz * pz * (1.0 + pz)) + Khx * px * (2.0 + px)
1174	>	+ Khy * py * (2.0 + py) + Khz * pz * (2.0 + pz)
1175	>	- Kcx - Kcy - Kcz)) / Khw; the following is simpler
1176	>	*/
1177	>	if ((fabs(x - 1.0) < 0.1) && (fabs(y - 1.0) < 0.1) &&
1178	>	(fabs(z - 1.0) < 0.1)) {
1179	>	w = 1.0 + (kineticTarget_
1180	>	+ Khx * (1.0 - x * x) + Khy * (1.0 - y * y)
1181	>	+ Khz * (1.0 - z * z)) / Khw;
1182	>	}//no need to calculate w if x, y or z is out of range
1183	>	} else {
1184	>	w = 1.0 + kineticTarget_ / Khw;
1185	>	}
1186	>	if ((w > 0.81) && (w < 1.21)) {//restrict scaling coefficients
1187	>	//if w is in the right range, so should be x, y, z.
1188	>	vector<StuntDouble*>::iterator sdi;
1189	>	Vector3d vel;
1190	>	for (sdi = coldBin.begin(); sdi != coldBin.end(); ++sdi) {
1191	>	if (rnemdFluxType_ == rnemdFullKE) {
1192	>	vel = (*sdi)->getVel() * c;
1193	>	(*sdi)->setVel(vel);
1194	>	}
1195	>	if ((*sdi)->isDirectional()) {
1196	>	Vector3d angMom = (*sdi)->getJ() * c;
1197	>	(*sdi)->setJ(angMom);
1198	>	}
1199		}
1200	+	w = sqrt(w);
1201	+	for (sdi = hotBin.begin(); sdi != hotBin.end(); ++sdi) {
1202	+	if (rnemdFluxType_ == rnemdFullKE) {
1203	+	vel = (*sdi)->getVel();
1204	+	vel.x() *= x;
1205	+	vel.y() *= y;
1206	+	vel.z() *= z;
1207	+	(*sdi)->setVel(vel);
1208	+	}
1209	+	if ((*sdi)->isDirectional()) {
1210	+	Vector3d angMom = (*sdi)->getJ() * w;
1211	+	(*sdi)->setJ(angMom);
1212	+	}
1213	+	}
1214	+	successfulScale = true;
1215	+	kineticExchange_ += kineticTarget_;
1216		}
1217	<	value = value / OOPSEConstant::energyConvert / OOPSEConstant::kb;
1218	<
1217	>	}
1218	>	} else {
1219	>	RealType a000, a110, c0, a001, a111, b01, b11, c1;
1220	>	switch(rnemdFluxType_) {
1221	>	case rnemdKE :
1222	>	/* used hotBin coeff's & only scale x & y dimensions
1223	>	RealType px = Phx / Pcx;
1224	>	RealType py = Phy / Pcy;
1225	>	a110 = Khy;
1226	>	c0 = - Khx - Khy - kineticTarget_;
1227	>	a000 = Khx;
1228	>	a111 = Kcy * py * py;
1229	>	b11 = -2.0 * Kcy * py * (1.0 + py);
1230	>	c1 = Kcy * py * (2.0 + py) + Kcx * px * ( 2.0 + px) + kineticTarget_;
1231	>	b01 = -2.0 * Kcx * px * (1.0 + px);
1232	>	a001 = Kcx * px * px;
1233	>	*/
1234	>	//scale all three dimensions, let c_x = c_y
1235	>	a000 = Kcx + Kcy;
1236	>	a110 = Kcz;
1237	>	c0 = kineticTarget_ - Kcx - Kcy - Kcz;
1238	>	a001 = Khx * px * px + Khy * py * py;
1239	>	a111 = Khz * pz * pz;
1240	>	b01 = -2.0 * (Khx * px * (1.0 + px) + Khy * py * (1.0 + py));
1241	>	b11 = -2.0 * Khz * pz * (1.0 + pz);
1242	>	c1 = Khx * px * (2.0 + px) + Khy * py * (2.0 + py)
1243	>	+ Khz * pz * (2.0 + pz) - kineticTarget_;
1244		break;
1245		case rnemdPx :
1246	<	value = mass * vel[0];
1247	<	valueCount[binNo]++;
1246	>	c = 1 - momentumTarget_.x() / Pcx;
1247	>	a000 = Kcy;
1248	>	a110 = Kcz;
1249	>	c0 = Kcx * c * c - Kcx - Kcy - Kcz;
1250	>	a001 = py * py * Khy;
1251	>	a111 = pz * pz * Khz;
1252	>	b01 = -2.0 * Khy * py * (1.0 + py);
1253	>	b11 = -2.0 * Khz * pz * (1.0 + pz);
1254	>	c1 = Khy * py * (2.0 + py) + Khz * pz * (2.0 + pz)
1255	>	+ Khx * (fastpow(c * px - px - 1.0, 2) - 1.0);
1256		break;
1257		case rnemdPy :
1258	<	value = mass * vel[1];
1259	<	valueCount[binNo]++;
1258	>	c = 1 - momentumTarget_.y() / Pcy;
1259	>	a000 = Kcx;
1260	>	a110 = Kcz;
1261	>	c0 = Kcy * c * c - Kcx - Kcy - Kcz;
1262	>	a001 = px * px * Khx;
1263	>	a111 = pz * pz * Khz;
1264	>	b01 = -2.0 * Khx * px * (1.0 + px);
1265	>	b11 = -2.0 * Khz * pz * (1.0 + pz);
1266	>	c1 = Khx * px * (2.0 + px) + Khz * pz * (2.0 + pz)
1267	>	+ Khy * (fastpow(c * py - py - 1.0, 2) - 1.0);
1268		break;
1269	<	case rnemdPz :
1270	<	value = mass * vel[2];
1271	<	valueCount[binNo]++;
1269	>	case rnemdPz ://we don't really do this, do we?
1270	>	c = 1 - momentumTarget_.z() / Pcz;
1271	>	a000 = Kcx;
1272	>	a110 = Kcy;
1273	>	c0 = Kcz * c * c - Kcx - Kcy - Kcz;
1274	>	a001 = px * px * Khx;
1275	>	a111 = py * py * Khy;
1276	>	b01 = -2.0 * Khx * px * (1.0 + px);
1277	>	b11 = -2.0 * Khy * py * (1.0 + py);
1278	>	c1 = Khx * px * (2.0 + px) + Khy * py * (2.0 + py)
1279	>	+ Khz * (fastpow(c * pz - pz - 1.0, 2) - 1.0);
1280		break;
528	–	case rnemdUnknown :
1281		default :
1282		break;
1283		}
1284	<	valueHist[binNo] += value;
1284	>
1285	>	RealType v1 = a000 * a111 - a001 * a110;
1286	>	RealType v2 = a000 * b01;
1287	>	RealType v3 = a000 * b11;
1288	>	RealType v4 = a000 * c1 - a001 * c0;
1289	>	RealType v8 = a110 * b01;
1290	>	RealType v10 = - b01 * c0;
1291	>
1292	>	RealType u0 = v2 * v10 - v4 * v4;
1293	>	RealType u1 = -2.0 * v3 * v4;
1294	>	RealType u2 = -v2 * v8 - v3 * v3 - 2.0 * v1 * v4;
1295	>	RealType u3 = -2.0 * v1 * v3;
1296	>	RealType u4 = - v1 * v1;
1297	>	//rescale coefficients
1298	>	RealType maxAbs = fabs(u0);
1299	>	if (maxAbs < fabs(u1)) maxAbs = fabs(u1);
1300	>	if (maxAbs < fabs(u2)) maxAbs = fabs(u2);
1301	>	if (maxAbs < fabs(u3)) maxAbs = fabs(u3);
1302	>	if (maxAbs < fabs(u4)) maxAbs = fabs(u4);
1303	>	u0 /= maxAbs;
1304	>	u1 /= maxAbs;
1305	>	u2 /= maxAbs;
1306	>	u3 /= maxAbs;
1307	>	u4 /= maxAbs;
1308	>	//max_element(start, end) is also available.
1309	>	Polynomial<RealType> poly; //same as DoublePolynomial poly;
1310	>	poly.setCoefficient(4, u4);
1311	>	poly.setCoefficient(3, u3);
1312	>	poly.setCoefficient(2, u2);
1313	>	poly.setCoefficient(1, u1);
1314	>	poly.setCoefficient(0, u0);
1315	>	vector<RealType> realRoots = poly.FindRealRoots();
1316	>
1317	>	vector<RealType>::iterator ri;
1318	>	RealType r1, r2, alpha0;
1319	>	vector<pair<RealType,RealType> > rps;
1320	>	for (ri = realRoots.begin(); ri !=realRoots.end(); ++ri) {
1321	>	r2 = *ri;
1322	>	//check if FindRealRoots() give the right answer
1323	>	if ( fabs(u0 + r2 * (u1 + r2 * (u2 + r2 * (u3 + r2 * u4)))) > 1e-6 ) {
1324	>	sprintf(painCave.errMsg,
1325	>	"RNEMD Warning: polynomial solve seems to have an error!");
1326	>	painCave.isFatal = 0;
1327	>	simError();
1328	>	failRootCount_++;
1329	>	}
1330	>	//might not be useful w/o rescaling coefficients
1331	>	alpha0 = -c0 - a110 * r2 * r2;
1332	>	if (alpha0 >= 0.0) {
1333	>	r1 = sqrt(alpha0 / a000);
1334	>	if (fabs(c1 + r1 * (b01 + r1 * a001) + r2 * (b11 + r2 * a111))
1335	>	< 1e-6)
1336	>	{ rps.push_back(make_pair(r1, r2)); }
1337	>	if (r1 > 1e-6) { //r1 non-negative
1338	>	r1 = -r1;
1339	>	if (fabs(c1 + r1 * (b01 + r1 * a001) + r2 * (b11 + r2 * a111))
1340	>	< 1e-6)
1341	>	{ rps.push_back(make_pair(r1, r2)); }
1342	>	}
1343	>	}
1344	>	}
1345	>	// Consider combining together the solving pair part w/ the searching
1346	>	// best solution part so that we don't need the pairs vector
1347	>	if (!rps.empty()) {
1348	>	RealType smallestDiff = HONKING_LARGE_VALUE;
1349	>	RealType diff;
1350	>	pair<RealType,RealType> bestPair = make_pair(1.0, 1.0);
1351	>	vector<pair<RealType,RealType> >::iterator rpi;
1352	>	for (rpi = rps.begin(); rpi != rps.end(); ++rpi) {
1353	>	r1 = (*rpi).first;
1354	>	r2 = (*rpi).second;
1355	>	switch(rnemdFluxType_) {
1356	>	case rnemdKE :
1357	>	diff = fastpow(1.0 - r1, 2) + fastpow(1.0 - r2, 2)
1358	>	+ fastpow(r1 * r1 / r2 / r2 - Kcz/Kcx, 2)
1359	>	+ fastpow(r1 * r1 / r2 / r2 - Kcz/Kcy, 2);
1360	>	break;
1361	>	case rnemdPx :
1362	>	diff = fastpow(1.0 - r1, 2) + fastpow(1.0 - r2, 2)
1363	>	+ fastpow(r1 * r1 / r2 / r2 - Kcz/Kcy, 2);
1364	>	break;
1365	>	case rnemdPy :
1366	>	diff = fastpow(1.0 - r1, 2) + fastpow(1.0 - r2, 2)
1367	>	+ fastpow(r1 * r1 / r2 / r2 - Kcz/Kcx, 2);
1368	>	break;
1369	>	case rnemdPz :
1370	>	diff = fastpow(1.0 - r1, 2) + fastpow(1.0 - r2, 2)
1371	>	+ fastpow(r1 * r1 / r2 / r2 - Kcy/Kcx, 2);
1372	>	default :
1373	>	break;
1374	>	}
1375	>	if (diff < smallestDiff) {
1376	>	smallestDiff = diff;
1377	>	bestPair = *rpi;
1378	>	}
1379	>	}
1380	>	#ifdef IS_MPI
1381	>	if (worldRank == 0) {
1382	>	#endif
1383	>	// sprintf(painCave.errMsg,
1384	>	//         "RNEMD: roots r1= %lf\tr2 = %lf\n",
1385	>	//         bestPair.first, bestPair.second);
1386	>	// painCave.isFatal = 0;
1387	>	// painCave.severity = OPENMD_INFO;
1388	>	// simError();
1389	>	#ifdef IS_MPI
1390	>	}
1391	>	#endif
1392	>
1393	>	switch(rnemdFluxType_) {
1394	>	case rnemdKE :
1395	>	x = bestPair.first;
1396	>	y = bestPair.first;
1397	>	z = bestPair.second;
1398	>	break;
1399	>	case rnemdPx :
1400	>	x = c;
1401	>	y = bestPair.first;
1402	>	z = bestPair.second;
1403	>	break;
1404	>	case rnemdPy :
1405	>	x = bestPair.first;
1406	>	y = c;
1407	>	z = bestPair.second;
1408	>	break;
1409	>	case rnemdPz :
1410	>	x = bestPair.first;
1411	>	y = bestPair.second;
1412	>	z = c;
1413	>	break;
1414	>	default :
1415	>	break;
1416	>	}
1417	>	vector<StuntDouble*>::iterator sdi;
1418	>	Vector3d vel;
1419	>	for (sdi = coldBin.begin(); sdi != coldBin.end(); ++sdi) {
1420	>	vel = (*sdi)->getVel();
1421	>	vel.x() *= x;
1422	>	vel.y() *= y;
1423	>	vel.z() *= z;
1424	>	(*sdi)->setVel(vel);
1425	>	}
1426	>	//convert to hotBin coefficient
1427	>	x = 1.0 + px * (1.0 - x);
1428	>	y = 1.0 + py * (1.0 - y);
1429	>	z = 1.0 + pz * (1.0 - z);
1430	>	for (sdi = hotBin.begin(); sdi != hotBin.end(); ++sdi) {
1431	>	vel = (*sdi)->getVel();
1432	>	vel.x() *= x;
1433	>	vel.y() *= y;
1434	>	vel.z() *= z;
1435	>	(*sdi)->setVel(vel);
1436	>	}
1437	>	successfulScale = true;
1438	>	switch(rnemdFluxType_) {
1439	>	case rnemdKE :
1440	>	kineticExchange_ += kineticTarget_;
1441	>	break;
1442	>	case rnemdPx :
1443	>	case rnemdPy :
1444	>	case rnemdPz :
1445	>	momentumExchange_ += momentumTarget_;
1446	>	break;
1447	>	default :
1448	>	break;
1449	>	}
1450	>	}
1451		}
1452	+	if (successfulScale != true) {
1453	+	sprintf(painCave.errMsg,
1454	+	"RNEMD::doNIVS exchange NOT performed - roots that solve\n"
1455	+	"\tthe constraint equations may not exist or there may be\n"
1456	+	"\tno selected objects in one or both slabs.\n");
1457	+	painCave.isFatal = 0;
1458	+	painCave.severity = OPENMD_INFO;
1459	+	simError();
1460	+	failTrialCount_++;
1461	+	}
1462	+	}
1463	+
1464	+	void RNEMD::doVSS(SelectionManager& smanA, SelectionManager& smanB) {
1465	+	if (!doRNEMD_) return;
1466	+	int selei;
1467	+	int selej;
1468
1469	+	Snapshot* currentSnap_ = info_->getSnapshotManager()->getCurrentSnapshot();
1470	+	Mat3x3d hmat = currentSnap_->getHmat();
1471	+
1472	+	StuntDouble* sd;
1473	+
1474	+	vector<StuntDouble*> hotBin, coldBin;
1475	+
1476	+	Vector3d Ph(V3Zero);
1477	+	Vector3d Lh(V3Zero);
1478	+	RealType Mh = 0.0;
1479	+	Mat3x3d Ih(0.0);
1480	+	RealType Kh = 0.0;
1481	+	Vector3d Pc(V3Zero);
1482	+	Vector3d Lc(V3Zero);
1483	+	RealType Mc = 0.0;
1484	+	Mat3x3d Ic(0.0);
1485	+	RealType Kc = 0.0;
1486	+
1487	+	// Constraints can be on only the linear or angular momentum, but
1488	+	// not both.  Usually, the user will specify which they want, but
1489	+	// in case they don't, the use of periodic boundaries should make
1490	+	// the choice for us.
1491	+	bool doLinearPart = false;
1492	+	bool doAngularPart = false;
1493	+
1494	+	switch (rnemdFluxType_) {
1495	+	case rnemdPx:
1496	+	case rnemdPy:
1497	+	case rnemdPz:
1498	+	case rnemdPvector:
1499	+	case rnemdKePx:
1500	+	case rnemdKePy:
1501	+	case rnemdKePvector:
1502	+	doLinearPart = true;
1503	+	break;
1504	+	case rnemdLx:
1505	+	case rnemdLy:
1506	+	case rnemdLz:
1507	+	case rnemdLvector:
1508	+	case rnemdKeLx:
1509	+	case rnemdKeLy:
1510	+	case rnemdKeLz:
1511	+	case rnemdKeLvector:
1512	+	doAngularPart = true;
1513	+	break;
1514	+	case rnemdKE:
1515	+	case rnemdRotKE:
1516	+	case rnemdFullKE:
1517	+	default:
1518	+	if (usePeriodicBoundaryConditions_)
1519	+	doLinearPart = true;
1520	+	else
1521	+	doAngularPart = true;
1522	+	break;
1523	+	}
1524	+
1525	+	for (sd = smanA.beginSelected(selei); sd != NULL;
1526	+	sd = smanA.nextSelected(selei)) {
1527	+
1528	+	Vector3d pos = sd->getPos();
1529	+
1530	+	// wrap the stuntdouble's position back into the box:
1531	+
1532	+	if (usePeriodicBoundaryConditions_)
1533	+	currentSnap_->wrapVector(pos);
1534	+
1535	+	RealType mass = sd->getMass();
1536	+	Vector3d vel = sd->getVel();
1537	+	Vector3d rPos = sd->getPos() - coordinateOrigin_;
1538	+	RealType r2;
1539	+
1540	+	hotBin.push_back(sd);
1541	+	Ph += mass * vel;
1542	+	Mh += mass;
1543	+	Kh += mass * vel.lengthSquare();
1544	+	Lh += mass * cross(rPos, vel);
1545	+	Ih -= outProduct(rPos, rPos) * mass;
1546	+	r2 = rPos.lengthSquare();
1547	+	Ih(0, 0) += mass * r2;
1548	+	Ih(1, 1) += mass * r2;
1549	+	Ih(2, 2) += mass * r2;
1550	+
1551	+	if (rnemdFluxType_ == rnemdFullKE) {
1552	+	if (sd->isDirectional()) {
1553	+	Vector3d angMom = sd->getJ();
1554	+	Mat3x3d I = sd->getI();
1555	+	if (sd->isLinear()) {
1556	+	int i = sd->linearAxis();
1557	+	int j = (i + 1) % 3;
1558	+	int k = (i + 2) % 3;
1559	+	Kh += angMom[j] * angMom[j] / I(j, j) +
1560	+	angMom[k] * angMom[k] / I(k, k);
1561	+	} else {
1562	+	Kh += angMom[0] * angMom[0] / I(0, 0) +
1563	+	angMom[1] * angMom[1] / I(1, 1) +
1564	+	angMom[2] * angMom[2] / I(2, 2);
1565	+	}
1566	+	}
1567	+	}
1568	+	}
1569	+	for (sd = smanB.beginSelected(selej); sd != NULL;
1570	+	sd = smanB.nextSelected(selej)) {
1571	+
1572	+	Vector3d pos = sd->getPos();
1573	+
1574	+	// wrap the stuntdouble's position back into the box:
1575	+
1576	+	if (usePeriodicBoundaryConditions_)
1577	+	currentSnap_->wrapVector(pos);
1578	+
1579	+	RealType mass = sd->getMass();
1580	+	Vector3d vel = sd->getVel();
1581	+	Vector3d rPos = sd->getPos() - coordinateOrigin_;
1582	+	RealType r2;
1583	+
1584	+	coldBin.push_back(sd);
1585	+	Pc += mass * vel;
1586	+	Mc += mass;
1587	+	Kc += mass * vel.lengthSquare();
1588	+	Lc += mass * cross(rPos, vel);
1589	+	Ic -= outProduct(rPos, rPos) * mass;
1590	+	r2 = rPos.lengthSquare();
1591	+	Ic(0, 0) += mass * r2;
1592	+	Ic(1, 1) += mass * r2;
1593	+	Ic(2, 2) += mass * r2;
1594	+
1595	+	if (rnemdFluxType_ == rnemdFullKE) {
1596	+	if (sd->isDirectional()) {
1597	+	Vector3d angMom = sd->getJ();
1598	+	Mat3x3d I = sd->getI();
1599	+	if (sd->isLinear()) {
1600	+	int i = sd->linearAxis();
1601	+	int j = (i + 1) % 3;
1602	+	int k = (i + 2) % 3;
1603	+	Kc += angMom[j] * angMom[j] / I(j, j) +
1604	+	angMom[k] * angMom[k] / I(k, k);
1605	+	} else {
1606	+	Kc += angMom[0] * angMom[0] / I(0, 0) +
1607	+	angMom[1] * angMom[1] / I(1, 1) +
1608	+	angMom[2] * angMom[2] / I(2, 2);
1609	+	}
1610	+	}
1611	+	}
1612	+	}
1613	+
1614	+	Kh *= 0.5;
1615	+	Kc *= 0.5;
1616	+
1617		#ifdef IS_MPI
1618	+	MPI_Allreduce(MPI_IN_PLACE, &Ph[0], 3, MPI_REALTYPE, MPI_SUM,
1619	+	MPI_COMM_WORLD);
1620	+	MPI_Allreduce(MPI_IN_PLACE, &Pc[0], 3, MPI_REALTYPE, MPI_SUM,
1621	+	MPI_COMM_WORLD);
1622	+	MPI_Allreduce(MPI_IN_PLACE, &Lh[0], 3, MPI_REALTYPE, MPI_SUM,
1623	+	MPI_COMM_WORLD);
1624	+	MPI_Allreduce(MPI_IN_PLACE, &Lc[0], 3, MPI_REALTYPE, MPI_SUM,
1625	+	MPI_COMM_WORLD);
1626	+	MPI_Allreduce(MPI_IN_PLACE, &Mh, 1, MPI_REALTYPE, MPI_SUM, MPI_COMM_WORLD);
1627	+	MPI_Allreduce(MPI_IN_PLACE, &Kh, 1, MPI_REALTYPE, MPI_SUM, MPI_COMM_WORLD);
1628	+	MPI_Allreduce(MPI_IN_PLACE, &Mc, 1, MPI_REALTYPE, MPI_SUM, MPI_COMM_WORLD);
1629	+	MPI_Allreduce(MPI_IN_PLACE, &Kc, 1, MPI_REALTYPE, MPI_SUM, MPI_COMM_WORLD);
1630	+	MPI_Allreduce(MPI_IN_PLACE, Ih.getArrayPointer(), 9,
1631	+	MPI_REALTYPE, MPI_SUM, MPI_COMM_WORLD);
1632	+	MPI_Allreduce(MPI_IN_PLACE, Ic.getArrayPointer(), 9,
1633	+	MPI_REALTYPE, MPI_SUM, MPI_COMM_WORLD);
1634	+	#endif
1635	+
1636
1637	<	// all processors have the same number of bins, and STL vectors pack their
1638	<	// arrays, so in theory, this should be safe:
1637	>	Vector3d ac, acrec, bc, bcrec;
1638	>	Vector3d ah, ahrec, bh, bhrec;
1639
1640	<	MPI::COMM_WORLD.Allreduce(MPI::IN_PLACE, &valueHist[0],
1641	<	nBins_, MPI::REALTYPE, MPI::SUM);
1642	<	MPI::COMM_WORLD.Allreduce(MPI::IN_PLACE, &valueCount[0],
1643	<	nBins_, MPI::INT, MPI::SUM);
1640	>	bool successfulExchange = false;
1641	>	if ((Mh > 0.0) && (Mc > 0.0)) {//both slabs are not empty
1642	>	Vector3d vc = Pc / Mc;
1643	>	ac = -momentumTarget_ / Mc + vc;
1644	>	acrec = -momentumTarget_ / Mc;
1645	>
1646	>	// We now need the inverse of the inertia tensor to calculate the
1647	>	// angular velocity of the cold slab;
1648	>	Mat3x3d Ici = Ic.inverse();
1649	>	Vector3d omegac = Ici * Lc;
1650	>	bc  = -(Ici * angularMomentumTarget_) + omegac;
1651	>	bcrec = bc - omegac;
1652	>
1653	>	RealType cNumerator = Kc - kineticTarget_;
1654	>	if (doLinearPart)
1655	>	cNumerator -= 0.5 * Mc * ac.lengthSquare();
1656	>
1657	>	if (doAngularPart)
1658	>	cNumerator -= 0.5 * ( dot(bc, Ic * bc));
1659
1660	+	if (cNumerator > 0.0) {
1661	+
1662	+	RealType cDenominator = Kc;
1663	+
1664	+	if (doLinearPart)
1665	+	cDenominator -= 0.5 * Mc * vc.lengthSquare();
1666	+
1667	+	if (doAngularPart)
1668	+	cDenominator -= 0.5*(dot(omegac, Ic * omegac));
1669	+
1670	+	if (cDenominator > 0.0) {
1671	+	RealType c = sqrt(cNumerator / cDenominator);
1672	+	if ((c > 0.9) && (c < 1.1)) {//restrict scaling coefficients
1673	+
1674	+	Vector3d vh = Ph / Mh;
1675	+	ah = momentumTarget_ / Mh + vh;
1676	+	ahrec = momentumTarget_ / Mh;
1677	+
1678	+	// We now need the inverse of the inertia tensor to
1679	+	// calculate the angular velocity of the hot slab;
1680	+	Mat3x3d Ihi = Ih.inverse();
1681	+	Vector3d omegah = Ihi * Lh;
1682	+	bh  = (Ihi * angularMomentumTarget_) + omegah;
1683	+	bhrec = bh - omegah;
1684	+
1685	+	RealType hNumerator = Kh + kineticTarget_;
1686	+	if (doLinearPart)
1687	+	hNumerator -= 0.5 * Mh * ah.lengthSquare();
1688	+
1689	+	if (doAngularPart)
1690	+	hNumerator -= 0.5 * ( dot(bh, Ih * bh));
1691	+
1692	+	if (hNumerator > 0.0) {
1693	+
1694	+	RealType hDenominator = Kh;
1695	+	if (doLinearPart)
1696	+	hDenominator -= 0.5 * Mh * vh.lengthSquare();
1697	+	if (doAngularPart)
1698	+	hDenominator -= 0.5*(dot(omegah, Ih * omegah));
1699	+
1700	+	if (hDenominator > 0.0) {
1701	+	RealType h = sqrt(hNumerator / hDenominator);
1702	+	if ((h > 0.9) && (h < 1.1)) {
1703	+
1704	+	vector<StuntDouble*>::iterator sdi;
1705	+	Vector3d vel;
1706	+	Vector3d rPos;
1707	+
1708	+	for (sdi = coldBin.begin(); sdi != coldBin.end(); ++sdi) {
1709	+	//vel = (*sdi)->getVel();
1710	+	rPos = (*sdi)->getPos() - coordinateOrigin_;
1711	+	if (doLinearPart)
1712	+	vel = ((*sdi)->getVel() - vc) * c + ac;
1713	+	if (doAngularPart)
1714	+	vel = ((*sdi)->getVel() - cross(omegac, rPos)) * c + cross(bc, rPos);
1715	+
1716	+	(*sdi)->setVel(vel);
1717	+	if (rnemdFluxType_ == rnemdFullKE) {
1718	+	if ((*sdi)->isDirectional()) {
1719	+	Vector3d angMom = (*sdi)->getJ() * c;
1720	+	(*sdi)->setJ(angMom);
1721	+	}
1722	+	}
1723	+	}
1724	+	for (sdi = hotBin.begin(); sdi != hotBin.end(); ++sdi) {
1725	+	//vel = (*sdi)->getVel();
1726	+	rPos = (*sdi)->getPos() - coordinateOrigin_;
1727	+	if (doLinearPart)
1728	+	vel = ((*sdi)->getVel() - vh) * h + ah;
1729	+	if (doAngularPart)
1730	+	vel = ((*sdi)->getVel() - cross(omegah, rPos)) * h + cross(bh, rPos);
1731	+
1732	+	(*sdi)->setVel(vel);
1733	+	if (rnemdFluxType_ == rnemdFullKE) {
1734	+	if ((*sdi)->isDirectional()) {
1735	+	Vector3d angMom = (*sdi)->getJ() * h;
1736	+	(*sdi)->setJ(angMom);
1737	+	}
1738	+	}
1739	+	}
1740	+	successfulExchange = true;
1741	+	kineticExchange_ += kineticTarget_;
1742	+	momentumExchange_ += momentumTarget_;
1743	+	angularMomentumExchange_ += angularMomentumTarget_;
1744	+	}
1745	+	}
1746	+	}
1747	+	}
1748	+	}
1749	+	}
1750	+	}
1751	+	if (successfulExchange != true) {
1752	+	sprintf(painCave.errMsg,
1753	+	"RNEMD::doVSS exchange NOT performed - roots that solve\n"
1754	+	"\tthe constraint equations may not exist or there may be\n"
1755	+	"\tno selected objects in one or both slabs.\n");
1756	+	painCave.isFatal = 0;
1757	+	painCave.severity = OPENMD_INFO;
1758	+	simError();
1759	+	failTrialCount_++;
1760	+	}
1761	+	}
1762	+
1763	+	RealType RNEMD::getDividingArea() {
1764	+
1765	+	if (hasDividingArea_) return dividingArea_;
1766	+
1767	+	RealType areaA, areaB;
1768	+	Snapshot* snap = info_->getSnapshotManager()->getCurrentSnapshot();
1769	+
1770	+	if (hasSelectionA_) {
1771	+
1772	+	if (evaluatorA_.hasSurfaceArea())
1773	+	areaA = evaluatorA_.getSurfaceArea();
1774	+	else {
1775	+
1776	+	int isd;
1777	+	StuntDouble* sd;
1778	+	vector<StuntDouble*> aSites;
1779	+	seleManA_.setSelectionSet(evaluatorA_.evaluate());
1780	+	for (sd = seleManA_.beginSelected(isd); sd != NULL;
1781	+	sd = seleManA_.nextSelected(isd)) {
1782	+	aSites.push_back(sd);
1783	+	}
1784	+	#if defined(HAVE_QHULL)
1785	+	ConvexHull* surfaceMeshA = new ConvexHull();
1786	+	surfaceMeshA->computeHull(aSites);
1787	+	areaA = surfaceMeshA->getArea();
1788	+	delete surfaceMeshA;
1789	+	#else
1790	+	sprintf( painCave.errMsg,
1791	+	"RNEMD::getDividingArea : Hull calculation is not possible\n"
1792	+	"\twithout libqhull. Please rebuild OpenMD with qhull enabled.");
1793	+	painCave.severity = OPENMD_ERROR;
1794	+	painCave.isFatal = 1;
1795	+	simError();
1796	+	#endif
1797	+	}
1798	+
1799	+	} else {
1800	+	if (usePeriodicBoundaryConditions_) {
1801	+	// in periodic boundaries, the surface area is twice the x-y
1802	+	// area of the current box:
1803	+	areaA = 2.0 * snap->getXYarea();
1804	+	} else {
1805	+	// in non-periodic simulations, without explicitly setting
1806	+	// selections, the sphere radius sets the surface area of the
1807	+	// dividing surface:
1808	+	areaA = 4.0 * M_PI * pow(sphereARadius_, 2);
1809	+	}
1810	+	}
1811	+
1812	+	if (hasSelectionB_) {
1813	+	if (evaluatorB_.hasSurfaceArea()) {
1814	+	areaB = evaluatorB_.getSurfaceArea();
1815	+	} else {
1816	+
1817	+	int isd;
1818	+	StuntDouble* sd;
1819	+	vector<StuntDouble*> bSites;
1820	+	seleManB_.setSelectionSet(evaluatorB_.evaluate());
1821	+	for (sd = seleManB_.beginSelected(isd); sd != NULL;
1822	+	sd = seleManB_.nextSelected(isd)) {
1823	+	bSites.push_back(sd);
1824	+	}
1825	+
1826	+	#if defined(HAVE_QHULL)
1827	+	ConvexHull* surfaceMeshB = new ConvexHull();
1828	+	surfaceMeshB->computeHull(bSites);
1829	+	areaB = surfaceMeshB->getArea();
1830	+	delete surfaceMeshB;
1831	+	#else
1832	+	sprintf( painCave.errMsg,
1833	+	"RNEMD::getDividingArea : Hull calculation is not possible\n"
1834	+	"\twithout libqhull. Please rebuild OpenMD with qhull enabled.");
1835	+	painCave.severity = OPENMD_ERROR;
1836	+	painCave.isFatal = 1;
1837	+	simError();
1838	+	#endif
1839	+	}
1840	+
1841	+	} else {
1842	+	if (usePeriodicBoundaryConditions_) {
1843	+	// in periodic boundaries, the surface area is twice the x-y
1844	+	// area of the current box:
1845	+	areaB = 2.0 * snap->getXYarea();
1846	+	} else {
1847	+	// in non-periodic simulations, without explicitly setting
1848	+	// selections, but if a sphereBradius has been set, just use that:
1849	+	areaB = 4.0 * M_PI * pow(sphereBRadius_, 2);
1850	+	}
1851	+	}
1852	+
1853	+	dividingArea_ = min(areaA, areaB);
1854	+	hasDividingArea_ = true;
1855	+	return dividingArea_;
1856	+	}
1857	+
1858	+	void RNEMD::doRNEMD() {
1859	+	if (!doRNEMD_) return;
1860	+	trialCount_++;
1861	+
1862	+	// object evaluator:
1863	+	evaluator_.loadScriptString(rnemdObjectSelection_);
1864	+	seleMan_.setSelectionSet(evaluator_.evaluate());
1865	+
1866	+	evaluatorA_.loadScriptString(selectionA_);
1867	+	evaluatorB_.loadScriptString(selectionB_);
1868	+
1869	+	seleManA_.setSelectionSet(evaluatorA_.evaluate());
1870	+	seleManB_.setSelectionSet(evaluatorB_.evaluate());
1871	+
1872	+	commonA_ = seleManA_ & seleMan_;
1873	+	commonB_ = seleManB_ & seleMan_;
1874	+
1875	+	// Target exchange quantities (in each exchange) = dividingArea * dt * flux
1876	+	// dt = exchange time interval
1877	+	// flux = target flux
1878	+	// dividingArea = smallest dividing surface between the two regions
1879	+
1880	+	hasDividingArea_ = false;
1881	+	RealType area = getDividingArea();
1882	+
1883	+	kineticTarget_ = kineticFlux_ * exchangeTime_ * area;
1884	+	momentumTarget_ = momentumFluxVector_ * exchangeTime_ * area;
1885	+	angularMomentumTarget_ = angularMomentumFluxVector_ * exchangeTime_ * area;
1886	+
1887	+	switch(rnemdMethod_) {
1888	+	case rnemdSwap:
1889	+	doSwap(commonA_, commonB_);
1890	+	break;
1891	+	case rnemdNIVS:
1892	+	doNIVS(commonA_, commonB_);
1893	+	break;
1894	+	case rnemdVSS:
1895	+	doVSS(commonA_, commonB_);
1896	+	break;
1897	+	case rnemdUnkownMethod:
1898	+	default :
1899	+	break;
1900	+	}
1901	+	}
1902	+
1903	+	void RNEMD::collectData() {
1904	+	if (!doRNEMD_) return;
1905	+	Snapshot* currentSnap_ = info_->getSnapshotManager()->getCurrentSnapshot();
1906	+
1907	+	// collectData can be called more frequently than the doRNEMD, so use the
1908	+	// computed area from the last exchange time:
1909	+	RealType area = getDividingArea();
1910	+	areaAccumulator_->add(area);
1911	+	Mat3x3d hmat = currentSnap_->getHmat();
1912	+	Vector3d u = angularMomentumFluxVector_;
1913	+	u.normalize();
1914	+
1915	+	seleMan_.setSelectionSet(evaluator_.evaluate());
1916	+
1917	+	int selei(0);
1918	+	StuntDouble* sd;
1919	+	int binNo;
1920	+	RealType mass;
1921	+	Vector3d vel;
1922	+	Vector3d rPos;
1923	+	RealType KE;
1924	+	Vector3d L;
1925	+	Mat3x3d I;
1926	+	RealType r2;
1927	+
1928	+	vector<RealType> binMass(nBins_, 0.0);
1929	+	vector<Vector3d> binP(nBins_, V3Zero);
1930	+	vector<RealType> binOmega(nBins_, 0.0);
1931	+	vector<Vector3d> binL(nBins_, V3Zero);
1932	+	vector<Mat3x3d>  binI(nBins_);
1933	+	vector<RealType> binKE(nBins_, 0.0);
1934	+	vector<int> binDOF(nBins_, 0);
1935	+	vector<int> binCount(nBins_, 0);
1936	+
1937	+	// alternative approach, track all molecules instead of only those
1938	+	// selected for scaling/swapping:
1939	+	/*
1940	+	SimInfo::MoleculeIterator miter;
1941	+	vector<StuntDouble*>::iterator iiter;
1942	+	Molecule* mol;
1943	+	StuntDouble* sd;
1944	+	for (mol = info_->beginMolecule(miter); mol != NULL;
1945	+	mol = info_->nextMolecule(miter))
1946	+	sd is essentially sd
1947	+	for (sd = mol->beginIntegrableObject(iiter);
1948	+	sd != NULL;
1949	+	sd = mol->nextIntegrableObject(iiter))
1950	+	*/
1951	+
1952	+	for (sd = seleMan_.beginSelected(selei); sd != NULL;
1953	+	sd = seleMan_.nextSelected(selei)) {
1954	+
1955	+	Vector3d pos = sd->getPos();
1956	+
1957	+	// wrap the stuntdouble's position back into the box:
1958	+
1959	+	if (usePeriodicBoundaryConditions_) {
1960	+	currentSnap_->wrapVector(pos);
1961	+	// which bin is this stuntdouble in?
1962	+	// wrapped positions are in the range [-0.5*hmat(2,2), +0.5*hmat(2,2)]
1963	+	// Shift molecules by half a box to have bins start at 0
1964	+	// The modulo operator is used to wrap the case when we are
1965	+	// beyond the end of the bins back to the beginning.
1966	+	binNo = int(nBins_ * (pos.z() / hmat(2,2) + 0.5)) % nBins_;
1967	+	} else {
1968	+	Vector3d rPos = pos - coordinateOrigin_;
1969	+	binNo = int(rPos.length() / binWidth_);
1970	+	}
1971	+
1972	+	mass = sd->getMass();
1973	+	vel = sd->getVel();
1974	+	rPos = sd->getPos() - coordinateOrigin_;
1975	+	KE = 0.5 * mass * vel.lengthSquare();
1976	+	L = mass * cross(rPos, vel);
1977	+	I = outProduct(rPos, rPos) * mass;
1978	+	r2 = rPos.lengthSquare();
1979	+	I(0, 0) += mass * r2;
1980	+	I(1, 1) += mass * r2;
1981	+	I(2, 2) += mass * r2;
1982	+
1983	+	// Project the relative position onto a plane perpendicular to
1984	+	// the angularMomentumFluxVector:
1985	+	// Vector3d rProj = rPos - dot(rPos, u) * u;
1986	+	// Project the velocity onto a plane perpendicular to the
1987	+	// angularMomentumFluxVector:
1988	+	// Vector3d vProj = vel  - dot(vel, u) * u;
1989	+	// Compute angular velocity vector (should be nearly parallel to
1990	+	// angularMomentumFluxVector
1991	+	// Vector3d aVel = cross(rProj, vProj);
1992	+
1993	+	if (binNo >= 0 && binNo < nBins_)  {
1994	+	binCount[binNo]++;
1995	+	binMass[binNo] += mass;
1996	+	binP[binNo] += mass*vel;
1997	+	binKE[binNo] += KE;
1998	+	binI[binNo] += I;
1999	+	binL[binNo] += L;
2000	+	binDOF[binNo] += 3;
2001	+
2002	+	if (sd->isDirectional()) {
2003	+	Vector3d angMom = sd->getJ();
2004	+	Mat3x3d Ia = sd->getI();
2005	+	if (sd->isLinear()) {
2006	+	int i = sd->linearAxis();
2007	+	int j = (i + 1) % 3;
2008	+	int k = (i + 2) % 3;
2009	+	binKE[binNo] += 0.5 * (angMom[j] * angMom[j] / Ia(j, j) +
2010	+	angMom[k] * angMom[k] / Ia(k, k));
2011	+	binDOF[binNo] += 2;
2012	+	} else {
2013	+	binKE[binNo] += 0.5 * (angMom[0] * angMom[0] / Ia(0, 0) +
2014	+	angMom[1] * angMom[1] / Ia(1, 1) +
2015	+	angMom[2] * angMom[2] / Ia(2, 2));
2016	+	binDOF[binNo] += 3;
2017	+	}
2018	+	}
2019	+	}
2020	+	}
2021	+
2022	+	#ifdef IS_MPI
2023	+
2024	+	for (int i = 0; i < nBins_; i++) {
2025	+
2026	+	MPI_Allreduce(MPI_IN_PLACE, &binCount[i],
2027	+	1, MPI_INT, MPI_SUM, MPI_COMM_WORLD);
2028	+	MPI_Allreduce(MPI_IN_PLACE, &binMass[i],
2029	+	1, MPI_REALTYPE, MPI_SUM, MPI_COMM_WORLD);
2030	+	MPI_Allreduce(MPI_IN_PLACE, binP[i].getArrayPointer(),
2031	+	3, MPI_REALTYPE, MPI_SUM, MPI_COMM_WORLD);
2032	+	MPI_Allreduce(MPI_IN_PLACE, binL[i].getArrayPointer(),
2033	+	3, MPI_REALTYPE, MPI_SUM, MPI_COMM_WORLD);
2034	+	MPI_Allreduce(MPI_IN_PLACE, binI[i].getArrayPointer(),
2035	+	9, MPI_REALTYPE, MPI_SUM, MPI_COMM_WORLD);
2036	+	MPI_Allreduce(MPI_IN_PLACE, &binKE[i],
2037	+	1, MPI_REALTYPE, MPI_SUM, MPI_COMM_WORLD);
2038	+	MPI_Allreduce(MPI_IN_PLACE, &binDOF[i],
2039	+	1, MPI_INT, MPI_SUM, MPI_COMM_WORLD);
2040	+	//MPI_Allreduce(MPI_IN_PLACE, &binOmega[i],
2041	+	//                          1, MPI_REALTYPE, MPI_SUM, MPI_COMM_WORLD);
2042	+	}
2043	+
2044	+	#endif
2045	+
2046	+	Vector3d omega;
2047	+	RealType den;
2048	+	RealType temp;
2049	+	RealType z;
2050	+	RealType r;
2051	+	for (int i = 0; i < nBins_; i++) {
2052	+	if (usePeriodicBoundaryConditions_) {
2053	+	z = (((RealType)i + 0.5) / (RealType)nBins_) * hmat(2,2);
2054	+	den = binMass[i] * nBins_ * PhysicalConstants::densityConvert
2055	+	/ currentSnap_->getVolume() ;
2056	+	} else {
2057	+	r = (((RealType)i + 0.5) * binWidth_);
2058	+	RealType rinner = (RealType)i * binWidth_;
2059	+	RealType router = (RealType)(i+1) * binWidth_;
2060	+	den = binMass[i] * 3.0 * PhysicalConstants::densityConvert
2061	+	/ (4.0 * M_PI * (pow(router,3) - pow(rinner,3)));
2062	+	}
2063	+	vel = binP[i] / binMass[i];
2064	+
2065	+	omega = binI[i].inverse() * binL[i];
2066	+
2067	+	// omega = binOmega[i] / binCount[i];
2068	+
2069	+	if (binCount[i] > 0) {
2070	+	// only add values if there are things to add
2071	+	temp = 2.0 * binKE[i] / (binDOF[i] * PhysicalConstants::kb *
2072	+	PhysicalConstants::energyConvert);
2073	+
2074	+	for (unsigned int j = 0; j < outputMask_.size(); ++j) {
2075	+	if(outputMask_[j]) {
2076	+	switch(j) {
2077	+	case Z:
2078	+	dynamic_cast<Accumulator *>(data_[j].accumulator[i])->add(z);
2079	+	break;
2080	+	case R:
2081	+	dynamic_cast<Accumulator *>(data_[j].accumulator[i])->add(r);
2082	+	break;
2083	+	case TEMPERATURE:
2084	+	dynamic_cast<Accumulator *>(data_[j].accumulator[i])->add(temp);
2085	+	break;
2086	+	case VELOCITY:
2087	+	dynamic_cast<VectorAccumulator *>(data_[j].accumulator[i])->add(vel);
2088	+	break;
2089	+	case ANGULARVELOCITY:
2090	+	dynamic_cast<VectorAccumulator *>(data_[j].accumulator[i])->add(omega);
2091	+	break;
2092	+	case DENSITY:
2093	+	dynamic_cast<Accumulator *>(data_[j].accumulator[i])->add(den);
2094	+	break;
2095	+	}
2096	+	}
2097	+	}
2098	+	}
2099	+	}
2100	+	hasData_ = true;
2101	+	}
2102	+
2103	+	void RNEMD::getStarted() {
2104	+	if (!doRNEMD_) return;
2105	+	hasDividingArea_ = false;
2106	+	collectData();
2107	+	writeOutputFile();
2108	+	}
2109	+
2110	+	void RNEMD::parseOutputFileFormat(const std::string& format) {
2111	+	if (!doRNEMD_) return;
2112	+	StringTokenizer tokenizer(format, " ,;|\t\n\r");
2113	+
2114	+	while(tokenizer.hasMoreTokens()) {
2115	+	std::string token(tokenizer.nextToken());
2116	+	toUpper(token);
2117	+	OutputMapType::iterator i = outputMap_.find(token);
2118	+	if (i != outputMap_.end()) {
2119	+	outputMask_.set(i->second);
2120	+	} else {
2121	+	sprintf( painCave.errMsg,
2122	+	"RNEMD::parseOutputFileFormat: %s is not a recognized\n"
2123	+	"\toutputFileFormat keyword.\n", token.c_str() );
2124	+	painCave.isFatal = 0;
2125	+	painCave.severity = OPENMD_ERROR;
2126	+	simError();
2127	+	}
2128	+	}
2129	+	}
2130	+
2131	+	void RNEMD::writeOutputFile() {
2132	+	if (!doRNEMD_) return;
2133	+	if (!hasData_) return;
2134	+
2135	+	#ifdef IS_MPI
2136		// If we're the root node, should we print out the results
2137	<	int worldRank = MPI::COMM_WORLD.Get_rank();
2137	>	int worldRank;
2138	>	MPI_Comm_rank( MPI_COMM_WORLD, &worldRank);
2139	>
2140		if (worldRank == 0) {
2141		#endif
2142	<
550	<	std::cout << time;
551	<	for (int j = 0; j < nBins_; j++)
552	<	std::cout << "\t" << valueHist[j] / (RealType)valueCount[j];
553	<	std::cout << "\n";
2142	>	rnemdFile_.open(rnemdFileName_.c_str(), std::ios::out | std::ios::trunc );
2143
2144	+	if( !rnemdFile_ ){
2145	+	sprintf( painCave.errMsg,
2146	+	"Could not open \"%s\" for RNEMD output.\n",
2147	+	rnemdFileName_.c_str());
2148	+	painCave.isFatal = 1;
2149	+	simError();
2150	+	}
2151	+
2152	+	Snapshot* currentSnap_ = info_->getSnapshotManager()->getCurrentSnapshot();
2153	+
2154	+	RealType time = currentSnap_->getTime();
2155	+	RealType avgArea;
2156	+	areaAccumulator_->getAverage(avgArea);
2157	+
2158	+	RealType Jz(0.0);
2159	+	Vector3d JzP(V3Zero);
2160	+	Vector3d JzL(V3Zero);
2161	+	if (time >= info_->getSimParams()->getDt()) {
2162	+	Jz = kineticExchange_ / (time * avgArea)
2163	+	/ PhysicalConstants::energyConvert;
2164	+	JzP = momentumExchange_ / (time * avgArea);
2165	+	JzL = angularMomentumExchange_ / (time * avgArea);
2166	+	}
2167	+
2168	+	rnemdFile_ << "#######################################################\n";
2169	+	rnemdFile_ << "# RNEMD {\n";
2170	+
2171	+	map<string, RNEMDMethod>::iterator mi;
2172	+	for(mi = stringToMethod_.begin(); mi != stringToMethod_.end(); ++mi) {
2173	+	if ( (*mi).second == rnemdMethod_)
2174	+	rnemdFile_ << "#    exchangeMethod  = \"" << (*mi).first << "\";\n";
2175	+	}
2176	+	map<string, RNEMDFluxType>::iterator fi;
2177	+	for(fi = stringToFluxType_.begin(); fi != stringToFluxType_.end(); ++fi) {
2178	+	if ( (*fi).second == rnemdFluxType_)
2179	+	rnemdFile_ << "#    fluxType  = \"" << (*fi).first << "\";\n";
2180	+	}
2181	+
2182	+	rnemdFile_ << "#    exchangeTime = " << exchangeTime_ << ";\n";
2183	+
2184	+	rnemdFile_ << "#    objectSelection = \""
2185	+	<< rnemdObjectSelection_ << "\";\n";
2186	+	rnemdFile_ << "#    selectionA = \"" << selectionA_ << "\";\n";
2187	+	rnemdFile_ << "#    selectionB = \"" << selectionB_ << "\";\n";
2188	+	rnemdFile_ << "# }\n";
2189	+	rnemdFile_ << "#######################################################\n";
2190	+	rnemdFile_ << "# RNEMD report:\n";
2191	+	rnemdFile_ << "#      running time = " << time << " fs\n";
2192	+	rnemdFile_ << "# Target flux:\n";
2193	+	rnemdFile_ << "#           kinetic = "
2194	+	<< kineticFlux_ / PhysicalConstants::energyConvert
2195	+	<< " (kcal/mol/A^2/fs)\n";
2196	+	rnemdFile_ << "#          momentum = " << momentumFluxVector_
2197	+	<< " (amu/A/fs^2)\n";
2198	+	rnemdFile_ << "#  angular momentum = " << angularMomentumFluxVector_
2199	+	<< " (amu/A^2/fs^2)\n";
2200	+	rnemdFile_ << "# Target one-time exchanges:\n";
2201	+	rnemdFile_ << "#          kinetic = "
2202	+	<< kineticTarget_ / PhysicalConstants::energyConvert
2203	+	<< " (kcal/mol)\n";
2204	+	rnemdFile_ << "#          momentum = " << momentumTarget_
2205	+	<< " (amu*A/fs)\n";
2206	+	rnemdFile_ << "#  angular momentum = " << angularMomentumTarget_
2207	+	<< " (amu*A^2/fs)\n";
2208	+	rnemdFile_ << "# Actual exchange totals:\n";
2209	+	rnemdFile_ << "#          kinetic = "
2210	+	<< kineticExchange_ / PhysicalConstants::energyConvert
2211	+	<< " (kcal/mol)\n";
2212	+	rnemdFile_ << "#          momentum = " << momentumExchange_
2213	+	<< " (amu*A/fs)\n";
2214	+	rnemdFile_ << "#  angular momentum = " << angularMomentumExchange_
2215	+	<< " (amu*A^2/fs)\n";
2216	+	rnemdFile_ << "# Actual flux:\n";
2217	+	rnemdFile_ << "#          kinetic = " << Jz
2218	+	<< " (kcal/mol/A^2/fs)\n";
2219	+	rnemdFile_ << "#          momentum = " << JzP
2220	+	<< " (amu/A/fs^2)\n";
2221	+	rnemdFile_ << "#  angular momentum = " << JzL
2222	+	<< " (amu/A^2/fs^2)\n";
2223	+	rnemdFile_ << "# Exchange statistics:\n";
2224	+	rnemdFile_ << "#               attempted = " << trialCount_ << "\n";
2225	+	rnemdFile_ << "#                  failed = " << failTrialCount_ << "\n";
2226	+	if (rnemdMethod_ == rnemdNIVS) {
2227	+	rnemdFile_ << "#  NIVS root-check errors = "
2228	+	<< failRootCount_ << "\n";
2229	+	}
2230	+	rnemdFile_ << "#######################################################\n";
2231	+
2232	+
2233	+
2234	+	//write title
2235	+	rnemdFile_ << "#";
2236	+	for (unsigned int i = 0; i < outputMask_.size(); ++i) {
2237	+	if (outputMask_[i]) {
2238	+	rnemdFile_ << "\t" << data_[i].title <<
2239	+	"(" << data_[i].units << ")";
2240	+	// add some extra tabs for column alignment
2241	+	if (data_[i].dataType == "Vector3d") rnemdFile_ << "\t\t";
2242	+	}
2243	+	}
2244	+	rnemdFile_ << std::endl;
2245	+
2246	+	rnemdFile_.precision(8);
2247	+
2248	+	for (int j = 0; j < nBins_; j++) {
2249	+
2250	+	for (unsigned int i = 0; i < outputMask_.size(); ++i) {
2251	+	if (outputMask_[i]) {
2252	+	if (data_[i].dataType == "RealType")
2253	+	writeReal(i,j);
2254	+	else if (data_[i].dataType == "Vector3d")
2255	+	writeVector(i,j);
2256	+	else {
2257	+	sprintf( painCave.errMsg,
2258	+	"RNEMD found an unknown data type for: %s ",
2259	+	data_[i].title.c_str());
2260	+	painCave.isFatal = 1;
2261	+	simError();
2262	+	}
2263	+	}
2264	+	}
2265	+	rnemdFile_ << std::endl;
2266	+
2267	+	}
2268	+
2269	+	rnemdFile_ << "#######################################################\n";
2270	+	rnemdFile_ << "# 95% confidence intervals in those quantities follow:\n";
2271	+	rnemdFile_ << "#######################################################\n";
2272	+
2273	+
2274	+	for (int j = 0; j < nBins_; j++) {
2275	+	rnemdFile_ << "#";
2276	+	for (unsigned int i = 0; i < outputMask_.size(); ++i) {
2277	+	if (outputMask_[i]) {
2278	+	if (data_[i].dataType == "RealType")
2279	+	writeRealErrorBars(i,j);
2280	+	else if (data_[i].dataType == "Vector3d")
2281	+	writeVectorErrorBars(i,j);
2282	+	else {
2283	+	sprintf( painCave.errMsg,
2284	+	"RNEMD found an unknown data type for: %s ",
2285	+	data_[i].title.c_str());
2286	+	painCave.isFatal = 1;
2287	+	simError();
2288	+	}
2289	+	}
2290	+	}
2291	+	rnemdFile_ << std::endl;
2292	+
2293	+	}
2294	+
2295	+	rnemdFile_.flush();
2296	+	rnemdFile_.close();
2297	+
2298		#ifdef IS_MPI
2299		}
2300		#endif
2301	+
2302		}
2303	+
2304	+	void RNEMD::writeReal(int index, unsigned int bin) {
2305	+	if (!doRNEMD_) return;
2306	+	assert(index >=0 && index < ENDINDEX);
2307	+	assert(int(bin) < nBins_);
2308	+	RealType s;
2309	+	int count;
2310	+
2311	+	count = data_[index].accumulator[bin]->count();
2312	+	if (count == 0) return;
2313	+
2314	+	dynamic_cast<Accumulator *>(data_[index].accumulator[bin])->getAverage(s);
2315	+
2316	+	if (! isinf(s) && ! isnan(s)) {
2317	+	rnemdFile_ << "\t" << s;
2318	+	} else{
2319	+	sprintf( painCave.errMsg,
2320	+	"RNEMD detected a numerical error writing: %s for bin %u",
2321	+	data_[index].title.c_str(), bin);
2322	+	painCave.isFatal = 1;
2323	+	simError();
2324	+	}
2325	+	}
2326	+
2327	+	void RNEMD::writeVector(int index, unsigned int bin) {
2328	+	if (!doRNEMD_) return;
2329	+	assert(index >=0 && index < ENDINDEX);
2330	+	assert(int(bin) < nBins_);
2331	+	Vector3d s;
2332	+	int count;
2333	+
2334	+	count = data_[index].accumulator[bin]->count();
2335	+
2336	+	if (count == 0) return;
2337	+
2338	+	dynamic_cast<VectorAccumulator*>(data_[index].accumulator[bin])->getAverage(s);
2339	+	if (isinf(s[0]) || isnan(s[0]) ||
2340	+	isinf(s[1]) || isnan(s[1]) ||
2341	+	isinf(s[2]) || isnan(s[2]) ) {
2342	+	sprintf( painCave.errMsg,
2343	+	"RNEMD detected a numerical error writing: %s for bin %u",
2344	+	data_[index].title.c_str(), bin);
2345	+	painCave.isFatal = 1;
2346	+	simError();
2347	+	} else {
2348	+	rnemdFile_ << "\t" << s[0] << "\t" << s[1] << "\t" << s[2];
2349	+	}
2350	+	}
2351	+
2352	+	void RNEMD::writeRealErrorBars(int index, unsigned int bin) {
2353	+	if (!doRNEMD_) return;
2354	+	assert(index >=0 && index < ENDINDEX);
2355	+	assert(int(bin) < nBins_);
2356	+	RealType s;
2357	+	int count;
2358	+
2359	+	count = data_[index].accumulator[bin]->count();
2360	+	if (count == 0) return;
2361	+
2362	+	dynamic_cast<Accumulator *>(data_[index].accumulator[bin])->get95percentConfidenceInterval(s);
2363	+
2364	+	if (! isinf(s) && ! isnan(s)) {
2365	+	rnemdFile_ << "\t" << s;
2366	+	} else{
2367	+	sprintf( painCave.errMsg,
2368	+	"RNEMD detected a numerical error writing: %s std. dev. for bin %u",
2369	+	data_[index].title.c_str(), bin);
2370	+	painCave.isFatal = 1;
2371	+	simError();
2372	+	}
2373	+	}
2374	+
2375	+	void RNEMD::writeVectorErrorBars(int index, unsigned int bin) {
2376	+	if (!doRNEMD_) return;
2377	+	assert(index >=0 && index < ENDINDEX);
2378	+	assert(int(bin) < nBins_);
2379	+	Vector3d s;
2380	+	int count;
2381	+
2382	+	count = data_[index].accumulator[bin]->count();
2383	+	if (count == 0) return;
2384	+
2385	+	dynamic_cast<VectorAccumulator*>(data_[index].accumulator[bin])->get95percentConfidenceInterval(s);
2386	+	if (isinf(s[0]) || isnan(s[0]) ||
2387	+	isinf(s[1]) || isnan(s[1]) ||
2388	+	isinf(s[2]) || isnan(s[2]) ) {
2389	+	sprintf( painCave.errMsg,
2390	+	"RNEMD detected a numerical error writing: %s std. dev. for bin %u",
2391	+	data_[index].title.c_str(), bin);
2392	+	painCave.isFatal = 1;
2393	+	simError();
2394	+	} else {
2395	+	rnemdFile_ << "\t" << s[0] << "\t" << s[1] << "\t" << s[2];
2396	+	}
2397	+	}
2398		}
2399	+

Comparing:
trunk/src/integrators/RNEMD.cpp (property svn:keywords), Revision 1350 by gezelter, Thu May 21 18:56:45 2009 UTC vs.
trunk/src/rnemd/RNEMD.cpp (property svn:keywords), Revision 2068 by gezelter, Thu Mar 5 15:40:58 2015 UTC

#	Line 0 | Line 1
1	+	Author Id Revision Date

Diff Legend

–	Removed lines
+	Added lines
<	Changed lines
>	Changed lines