ViewVC Help

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

root/OpenMD/branches/development/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.
branches/development/src/rnemd/RNEMD.cpp (file contents), Revision 1775 by gezelter, Wed Aug 8 18:45:52 2012 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, 24107 (2008).
39	+	* [4]  Vardeman & Gezelter, in progress (2009).
40		*/
41
42	<	#include "integrators/RNEMD.hpp"
42	>	#include <cmath>
43	>	#include "rnemd/RNEMD.hpp"
44		#include "math/Vector3.hpp"
45	+	#include "math/Vector.hpp"
46		#include "math/SquareMatrix3.hpp"
47	+	#include "math/Polynomial.hpp"
48		#include "primitives/Molecule.hpp"
49		#include "primitives/StuntDouble.hpp"
50	<	#include "utils/OOPSEConstant.hpp"
50	>	#include "utils/PhysicalConstants.hpp"
51		#include "utils/Tuple.hpp"
52	<
53	<	#ifndef IS_MPI
51	<	#include "math/SeqRandNumGen.hpp"
52	<	#else
53	<	#include "math/ParallelRandNumGen.hpp"
52	>	#ifdef IS_MPI
53	>	#include <mpi.h>
54		#endif
55
56		#define HONKING_LARGE_VALUE 1.0e10
57
58	<	namespace oopse {
58	>	using namespace std;
59	>	namespace OpenMD {
60
61	<	RNEMD::RNEMD(SimInfo* info) : info_(info), evaluator_(info), seleMan_(info), usePeriodicBoundaryConditions_(info->getSimParams()->getUsePeriodicBoundaryConditions()) {
62	<
61	>	RNEMD::RNEMD(SimInfo* info) : info_(info), evaluator_(info), seleMan_(info),
62	>	usePeriodicBoundaryConditions_(info->getSimParams()->getUsePeriodicBoundaryConditions()) {
63	>
64	>	trialCount_ = 0;
65	>	failTrialCount_ = 0;
66	>	failRootCount_ = 0;
67	>
68		int seedValue;
69		Globals * simParams = info->getSimParams();
70	+	RNEMDParameters* rnemdParams = simParams->getRNEMDParameters();
71
72	<	stringToEnumMap_["Kinetic"] = rnemdKinetic;
73	<	stringToEnumMap_["Px"] = rnemdPx;
74	<	stringToEnumMap_["Py"] = rnemdPy;
68	<	stringToEnumMap_["Pz"] = rnemdPz;
69	<	stringToEnumMap_["Unknown"] = rnemdUnknown;
72	>	stringToMethod_["Swap"]  = rnemdSwap;
73	>	stringToMethod_["NIVS"]  = rnemdNIVS;
74	>	stringToMethod_["VSS"]   = rnemdVSS;
75
76	<	rnemdObjectSelection_ = simParams->getRNEMD_objectSelection();
76	>	stringToFluxType_["KE"]  = rnemdKE;
77	>	stringToFluxType_["Px"]  = rnemdPx;
78	>	stringToFluxType_["Py"]  = rnemdPy;
79	>	stringToFluxType_["Pz"]  = rnemdPz;
80	>	stringToFluxType_["KE+Px"]  = rnemdKePx;
81	>	stringToFluxType_["KE+Py"]  = rnemdKePy;
82	>	stringToFluxType_["KE+Pvector"]  = rnemdKePvector;
83	>
84	>	runTime_ = simParams->getRunTime();
85	>	statusTime_ = simParams->getStatusTime();
86	>
87	>	rnemdObjectSelection_ = rnemdParams->getObjectSelection();
88		evaluator_.loadScriptString(rnemdObjectSelection_);
89		seleMan_.setSelectionSet(evaluator_.evaluate());
90
91	+	const string methStr = rnemdParams->getMethod();
92	+	bool hasFluxType = rnemdParams->haveFluxType();
93
94	+	string fluxStr;
95	+	if (hasFluxType) {
96	+	fluxStr = rnemdParams->getFluxType();
97	+	} else {
98	+	sprintf(painCave.errMsg,
99	+	"RNEMD: No fluxType was set in the md file.  This parameter,\n"
100	+	"\twhich must be one of the following values:\n"
101	+	"\tKE, Px, Py, Pz, KE+Px, KE+Py, KE+Pvector, must be set to\n"
102	+	"\tuse RNEMD\n");
103	+	painCave.isFatal = 1;
104	+	painCave.severity = OPENMD_ERROR;
105	+	simError();
106	+	}
107	+
108	+	bool hasKineticFlux = rnemdParams->haveKineticFlux();
109	+	bool hasMomentumFlux = rnemdParams->haveMomentumFlux();
110	+	bool hasMomentumFluxVector = rnemdParams->haveMomentumFluxVector();
111	+	bool hasSlabWidth = rnemdParams->haveSlabWidth();
112	+	bool hasSlabACenter = rnemdParams->haveSlabACenter();
113	+	bool hasSlabBCenter = rnemdParams->haveSlabBCenter();
114	+	bool hasOutputFileName = rnemdParams->haveOutputFileName();
115	+	bool hasOutputFields = rnemdParams->haveOutputFields();
116	+
117	+	map<string, RNEMDMethod>::iterator i;
118	+	i = stringToMethod_.find(methStr);
119	+	if (i != stringToMethod_.end())
120	+	rnemdMethod_ = i->second;
121	+	else {
122	+	sprintf(painCave.errMsg,
123	+	"RNEMD: The current method,\n"
124	+	"\t\t%s is not one of the recognized\n"
125	+	"\texchange methods: Swap, NIVS, or VSS\n",
126	+	methStr.c_str());
127	+	painCave.isFatal = 1;
128	+	painCave.severity = OPENMD_ERROR;
129	+	simError();
130	+	}
131	+
132	+	map<string, RNEMDFluxType>::iterator j;
133	+	j = stringToFluxType_.find(fluxStr);
134	+	if (j != stringToFluxType_.end())
135	+	rnemdFluxType_ = j->second;
136	+	else {
137	+	sprintf(painCave.errMsg,
138	+	"RNEMD: The current fluxType,\n"
139	+	"\t\t%s\n"
140	+	"\tis not one of the recognized flux types.\n",
141	+	fluxStr.c_str());
142	+	painCave.isFatal = 1;
143	+	painCave.severity = OPENMD_ERROR;
144	+	simError();
145	+	}
146	+
147	+	bool methodFluxMismatch = false;
148	+	bool hasCorrectFlux = false;
149	+	switch(rnemdMethod_) {
150	+	case rnemdSwap:
151	+	switch (rnemdFluxType_) {
152	+	case rnemdKE:
153	+	hasCorrectFlux = hasKineticFlux;
154	+	break;
155	+	case rnemdPx:
156	+	case rnemdPy:
157	+	case rnemdPz:
158	+	hasCorrectFlux = hasMomentumFlux;
159	+	break;
160	+	default :
161	+	methodFluxMismatch = true;
162	+	break;
163	+	}
164	+	break;
165	+	case rnemdNIVS:
166	+	switch (rnemdFluxType_) {
167	+	case rnemdKE:
168	+	case rnemdRotKE:
169	+	case rnemdFullKE:
170	+	hasCorrectFlux = hasKineticFlux;
171	+	break;
172	+	case rnemdPx:
173	+	case rnemdPy:
174	+	case rnemdPz:
175	+	hasCorrectFlux = hasMomentumFlux;
176	+	break;
177	+	case rnemdKePx:
178	+	case rnemdKePy:
179	+	hasCorrectFlux = hasMomentumFlux && hasKineticFlux;
180	+	break;
181	+	default:
182	+	methodFluxMismatch = true;
183	+	break;
184	+	}
185	+	break;
186	+	case rnemdVSS:
187	+	switch (rnemdFluxType_) {
188	+	case rnemdKE:
189	+	case rnemdRotKE:
190	+	case rnemdFullKE:
191	+	hasCorrectFlux = hasKineticFlux;
192	+	break;
193	+	case rnemdPx:
194	+	case rnemdPy:
195	+	case rnemdPz:
196	+	hasCorrectFlux = hasMomentumFlux;
197	+	break;
198	+	case rnemdPvector:
199	+	hasCorrectFlux = hasMomentumFluxVector;
200	+	case rnemdKePx:
201	+	case rnemdKePy:
202	+	hasCorrectFlux = hasMomentumFlux && hasKineticFlux;
203	+	break;
204	+	case rnemdKePvector:
205	+	hasCorrectFlux = hasMomentumFluxVector && hasKineticFlux;
206	+	break;
207	+	default:
208	+	methodFluxMismatch = true;
209	+	break;
210	+	}
211	+	default:
212	+	break;
213	+	}
214	+
215	+	if (methodFluxMismatch) {
216	+	sprintf(painCave.errMsg,
217	+	"RNEMD: The current method,\n"
218	+	"\t\t%s\n"
219	+	"\tcannot be used with the current flux type, %s\n",
220	+	methStr.c_str(), fluxStr.c_str());
221	+	painCave.isFatal = 1;
222	+	painCave.severity = OPENMD_ERROR;
223	+	simError();
224	+	}
225	+	if (!hasCorrectFlux) {
226	+	sprintf(painCave.errMsg,
227	+	"RNEMD: The current method,\n"
228	+	"\t%s, and flux type %s\n"
229	+	"\tdid not have the correct flux value specified. Options\n"
230	+	"\tinclude: kineticFlux, momentumFlux, and momentumFluxVector\n",
231	+	methStr.c_str(), fluxStr.c_str());
232	+	painCave.isFatal = 1;
233	+	painCave.severity = OPENMD_ERROR;
234	+	simError();
235	+	}
236	+
237	+	if (hasKineticFlux) {
238	+	kineticFlux_ = rnemdParams->getKineticFlux();
239	+	} else {
240	+	kineticFlux_ = 0.0;
241	+	}
242	+	if (hasMomentumFluxVector) {
243	+	momentumFluxVector_ = rnemdParams->getMomentumFluxVector();
244	+	} else {
245	+	momentumFluxVector_ = V3Zero;
246	+	if (hasMomentumFlux) {
247	+	RealType momentumFlux = rnemdParams->getMomentumFlux();
248	+	switch (rnemdFluxType_) {
249	+	case rnemdPx:
250	+	momentumFluxVector_.x() = momentumFlux;
251	+	break;
252	+	case rnemdPy:
253	+	momentumFluxVector_.y() = momentumFlux;
254	+	break;
255	+	case rnemdPz:
256	+	momentumFluxVector_.z() = momentumFlux;
257	+	break;
258	+	case rnemdKePx:
259	+	momentumFluxVector_.x() = momentumFlux;
260	+	break;
261	+	case rnemdKePy:
262	+	momentumFluxVector_.y() = momentumFlux;
263	+	break;
264	+	default:
265	+	break;
266	+	}
267	+	}
268	+	}
269	+
270		// do some sanity checking
271
272		int selectionCount = seleMan_.getSelectionCount();
#	Line 80 | Line 274 | namespace oopse {
274
275		if (selectionCount > nIntegrable) {
276		sprintf(painCave.errMsg,
277	<	"RNEMD warning: The current RNEMD_objectSelection,\n"
277	>	"RNEMD: The current objectSelection,\n"
278		"\t\t%s\n"
279		"\thas resulted in %d selected objects.  However,\n"
280		"\tthe total number of integrable objects in the system\n"
#	Line 90 | Line 284 | namespace oopse {
284		rnemdObjectSelection_.c_str(),
285		selectionCount, nIntegrable);
286		painCave.isFatal = 0;
287	+	painCave.severity = OPENMD_WARNING;
288		simError();
289	+	}
290
291	+	areaAccumulator_ = new Accumulator();
292	+
293	+	nBins_ = rnemdParams->getOutputBins();
294	+
295	+	data_.resize(RNEMD::ENDINDEX);
296	+	OutputData z;
297	+	z.units =  "Angstroms";
298	+	z.title =  "Z";
299	+	z.dataType = "RealType";
300	+	z.accumulator.reserve(nBins_);
301	+	for (unsigned int i = 0; i < nBins_; i++)
302	+	z.accumulator.push_back( new Accumulator() );
303	+	data_[Z] = z;
304	+	outputMap_["Z"] =  Z;
305	+
306	+	OutputData temperature;
307	+	temperature.units =  "K";
308	+	temperature.title =  "Temperature";
309	+	temperature.dataType = "RealType";
310	+	temperature.accumulator.reserve(nBins_);
311	+	for (unsigned int i = 0; i < nBins_; i++)
312	+	temperature.accumulator.push_back( new Accumulator() );
313	+	data_[TEMPERATURE] = temperature;
314	+	outputMap_["TEMPERATURE"] =  TEMPERATURE;
315	+
316	+	OutputData velocity;
317	+	velocity.units = "amu/fs";
318	+	velocity.title =  "Velocity";
319	+	velocity.dataType = "Vector3d";
320	+	velocity.accumulator.reserve(nBins_);
321	+	for (unsigned int i = 0; i < nBins_; i++)
322	+	velocity.accumulator.push_back( new VectorAccumulator() );
323	+	data_[VELOCITY] = velocity;
324	+	outputMap_["VELOCITY"] = VELOCITY;
325	+
326	+	OutputData density;
327	+	density.units =  "g cm^-3";
328	+	density.title =  "Density";
329	+	density.dataType = "RealType";
330	+	density.accumulator.reserve(nBins_);
331	+	for (unsigned int i = 0; i < nBins_; i++)
332	+	density.accumulator.push_back( new Accumulator() );
333	+	data_[DENSITY] = density;
334	+	outputMap_["DENSITY"] =  DENSITY;
335	+
336	+	if (hasOutputFields) {
337	+	parseOutputFileFormat(rnemdParams->getOutputFields());
338	+	} else {
339	+	outputMask_.set(Z);
340	+	switch (rnemdFluxType_) {
341	+	case rnemdKE:
342	+	case rnemdRotKE:
343	+	case rnemdFullKE:
344	+	outputMask_.set(TEMPERATURE);
345	+	break;
346	+	case rnemdPx:
347	+	case rnemdPy:
348	+	outputMask_.set(VELOCITY);
349	+	break;
350	+	case rnemdPz:
351	+	case rnemdPvector:
352	+	outputMask_.set(VELOCITY);
353	+	outputMask_.set(DENSITY);
354	+	break;
355	+	case rnemdKePx:
356	+	case rnemdKePy:
357	+	outputMask_.set(TEMPERATURE);
358	+	outputMask_.set(VELOCITY);
359	+	break;
360	+	case rnemdKePvector:
361	+	outputMask_.set(TEMPERATURE);
362	+	outputMask_.set(VELOCITY);
363	+	outputMask_.set(DENSITY);
364	+	break;
365	+	default:
366	+	break;
367	+	}
368		}
369	<
370	<	const std::string st = simParams->getRNEMD_swapType();
369	>
370	>	if (hasOutputFileName) {
371	>	rnemdFileName_ = rnemdParams->getOutputFileName();
372	>	} else {
373	>	rnemdFileName_ = getPrefix(info->getFinalConfigFileName()) + ".rnemd";
374	>	}
375
376	<	std::map<std::string, RNEMDTypeEnum>::iterator i;
100	<	i = stringToEnumMap_.find(st);
101	<	rnemdType_  = (i == stringToEnumMap_.end()) ? RNEMD::rnemdUnknown : i->second;
376	>	exchangeTime_ = rnemdParams->getExchangeTime();
377
378	<	set_RNEMD_swapTime(simParams->getRNEMD_swapTime());
379	<	set_RNEMD_nBins(simParams->getRNEMD_nBins());
380	<	exchangeSum_ = 0.0;
378	>	Snapshot* currentSnap_ = info->getSnapshotManager()->getCurrentSnapshot();
379	>	Mat3x3d hmat = currentSnap_->getHmat();
380	>
381	>	// Target exchange quantities (in each exchange) =  2 Lx Ly dt flux
382	>	// Lx, Ly = box dimensions in x & y
383	>	// dt = exchange time interval
384	>	// flux = target flux
385
386	<	#ifndef IS_MPI
387	<	if (simParams->haveSeed()) {
388	<	seedValue = simParams->getSeed();
389	<	randNumGen_ = new SeqRandNumGen(seedValue);
390	<	}else {
391	<	randNumGen_ = new SeqRandNumGen();
392	<	}
393	<	#else
394	<	if (simParams->haveSeed()) {
395	<	seedValue = simParams->getSeed();
396	<	randNumGen_ = new ParallelRandNumGen(seedValue);
397	<	}else {
398	<	randNumGen_ = new ParallelRandNumGen();
120	<	}
121	<	#endif
122	<	}
386	>	RealType area = currentSnap_->getXYarea();
387	>	kineticTarget_ = 2.0 * kineticFlux_ * exchangeTime_ * area;
388	>	momentumTarget_ = 2.0 * momentumFluxVector_ * exchangeTime_ * area;
389	>
390	>	// total exchange sums are zeroed out at the beginning:
391	>
392	>	kineticExchange_ = 0.0;
393	>	momentumExchange_ = V3Zero;
394	>
395	>	if (hasSlabWidth)
396	>	slabWidth_ = rnemdParams->getSlabWidth();
397	>	else
398	>	slabWidth_ = hmat(2,2) / 10.0;
399
400	+	if (hasSlabACenter)
401	+	slabACenter_ = rnemdParams->getSlabACenter();
402	+	else
403	+	slabACenter_ = 0.0;
404	+
405	+	if (hasSlabBCenter)
406	+	slabBCenter_ = rnemdParams->getSlabBCenter();
407	+	else
408	+	slabBCenter_ = hmat(2,2) / 2.0;
409	+
410	+	}
411	+
412		RNEMD::~RNEMD() {
413	<	delete randNumGen_;
413	>
414	>	#ifdef IS_MPI
415	>	if (worldRank == 0) {
416	>	#endif
417	>
418	>	writeOutputFile();
419	>
420	>	rnemdFile_.close();
421	>
422	>	#ifdef IS_MPI
423	>	}
424	>	#endif
425		}
426	+
427	+	bool RNEMD::inSlabA(Vector3d pos) {
428	+	return (abs(pos.z() - slabACenter_) < 0.5*slabWidth_);
429	+	}
430	+	bool RNEMD::inSlabB(Vector3d pos) {
431	+	return (abs(pos.z() - slabBCenter_) < 0.5*slabWidth_);
432	+	}
433
434		void RNEMD::doSwap() {
129	–	int midBin = nBins_ / 2;
435
436		Snapshot* currentSnap_ = info_->getSnapshotManager()->getCurrentSnapshot();
437		Mat3x3d hmat = currentSnap_->getHmat();
#	Line 156 | Line 461 | namespace oopse {
461
462		if (usePeriodicBoundaryConditions_)
463		currentSnap_->wrapVector(pos);
464	+	bool inA = inSlabA(pos);
465	+	bool inB = inSlabB(pos);
466
467	<	// which bin is this stuntdouble in?
161	<	// wrapped positions are in the range [-0.5*hmat(2,2), +0.5*hmat(2,2)]
162	<
163	<	int binNo = int(nBins_ * (pos.z() / hmat(2,2) + 0.5)) % nBins_;
164	<
165	<
166	<	// if we're in bin 0 or the middleBin
167	<	if (binNo == 0 || binNo == midBin) {
467	>	if (inA || inB) {
468
469		RealType mass = sd->getMass();
470		Vector3d vel = sd->getVel();
471		RealType value;
472	<
473	<	switch(rnemdType_) {
474	<	case rnemdKinetic :
472	>
473	>	switch(rnemdFluxType_) {
474	>	case rnemdKE :
475
476	<	value = mass * (vel[0]*vel[0] + vel[1]*vel[1] +
477	<	vel[2]*vel[2]);
478	<	if (sd->isDirectional()) {
476	>	value = mass * vel.lengthSquare();
477	>
478	>	if (sd->isDirectional()) {
479		Vector3d angMom = sd->getJ();
480		Mat3x3d I = sd->getI();
481
482		if (sd->isLinear()) {
483	<	int i = sd->linearAxis();
484	<	int j = (i + 1) % 3;
485	<	int k = (i + 2) % 3;
486	<	value += angMom[j] * angMom[j] / I(j, j) +
487	<	angMom[k] * angMom[k] / I(k, k);
483	>	int i = sd->linearAxis();
484	>	int j = (i + 1) % 3;
485	>	int k = (i + 2) % 3;
486	>	value += angMom[j] * angMom[j] / I(j, j) +
487	>	angMom[k] * angMom[k] / I(k, k);
488		} else {
489	<	value += angMom[0]*angMom[0]/I(0, 0)
490	<	+ angMom[1]*angMom[1]/I(1, 1)
491	<	+ angMom[2]*angMom[2]/I(2, 2);
489	>	value += angMom[0]*angMom[0]/I(0, 0)
490	>	+ angMom[1]*angMom[1]/I(1, 1)
491	>	+ angMom[2]*angMom[2]/I(2, 2);
492		}
493	<	}
494	<	value = value * 0.5 / OOPSEConstant::energyConvert;
493	>	} //angular momenta exchange enabled
494	>	//energyConvert temporarily disabled
495	>	//make kineticExchange_ comparable between swap & scale
496	>	//value = value * 0.5 / PhysicalConstants::energyConvert;
497	>	value *= 0.5;
498		break;
499		case rnemdPx :
500		value = mass * vel[0];
#	Line 202 | Line 505 | namespace oopse {
505		case rnemdPz :
506		value = mass * vel[2];
507		break;
205	–	case rnemdUnknown :
508		default :
509		break;
510		}
511
512	<	if (binNo == 0) {
512	>	if (inA == 0) {
513		if (!min_found) {
514		min_val = value;
515		min_sd = sd;
#	Line 218 | Line 520 | namespace oopse {
520		min_sd = sd;
521		}
522		}
523	<	} else {
523	>	} else {
524		if (!max_found) {
525		max_val = value;
526		max_sd = sd;
#	Line 232 | Line 534 | namespace oopse {
534		}
535		}
536		}
537	<
537	>
538		#ifdef IS_MPI
539		int nProc, worldRank;
540	<
540	>
541		nProc = MPI::COMM_WORLD.Get_size();
542		worldRank = MPI::COMM_WORLD.Get_rank();
543
#	Line 243 | Line 545 | namespace oopse {
545		bool my_max_found = max_found;
546
547		// Even if we didn't find a minimum, did someone else?
548	<	MPI::COMM_WORLD.Allreduce(&my_min_found, &min_found,
247	<	1, MPI::BOOL, MPI::LAND);
248	<
548	>	MPI::COMM_WORLD.Allreduce(&my_min_found, &min_found, 1, MPI::BOOL, MPI::LOR);
549		// Even if we didn't find a maximum, did someone else?
550	<	MPI::COMM_WORLD.Allreduce(&my_max_found, &max_found,
551	<	1, MPI::BOOL, MPI::LAND);
552	<
553	<	struct {
554	<	RealType val;
555	<	int rank;
556	<	} max_vals, min_vals;
557	<
558	<	if (min_found) {
559	<	if (my_min_found)
550	>	MPI::COMM_WORLD.Allreduce(&my_max_found, &max_found, 1, MPI::BOOL, MPI::LOR);
551	>	#endif
552	>
553	>	if (max_found && min_found) {
554	>
555	>	#ifdef IS_MPI
556	>	struct {
557	>	RealType val;
558	>	int rank;
559	>	} max_vals, min_vals;
560	>
561	>	if (my_min_found) {
562		min_vals.val = min_val;
563	<	else
563	>	} else {
564		min_vals.val = HONKING_LARGE_VALUE;
565	<
565	>	}
566		min_vals.rank = worldRank;
567
568		// Who had the minimum?
569		MPI::COMM_WORLD.Allreduce(&min_vals, &min_vals,
570		1, MPI::REALTYPE_INT, MPI::MINLOC);
571		min_val = min_vals.val;
270	–	}
572
573	<	if (max_found) {
273	<	if (my_max_found)
573	>	if (my_max_found) {
574		max_vals.val = max_val;
575	<	else
575	>	} else {
576		max_vals.val = -HONKING_LARGE_VALUE;
577	<
577	>	}
578		max_vals.rank = worldRank;
579
580		// Who had the maximum?
581		MPI::COMM_WORLD.Allreduce(&max_vals, &max_vals,
582		1, MPI::REALTYPE_INT, MPI::MAXLOC);
583		max_val = max_vals.val;
284	–	}
584		#endif
585	<
586	<	if (max_found && min_found) {
587	<	if (min_val< max_val) {
289	<
585	>
586	>	if (min_val < max_val) {
587	>
588		#ifdef IS_MPI
589		if (max_vals.rank == worldRank && min_vals.rank == worldRank) {
590		// I have both maximum and minimum, so proceed like a single
591		// processor version:
592		#endif
593	<	// objects to be swapped: velocity & angular velocity
593	>
594		Vector3d min_vel = min_sd->getVel();
595		Vector3d max_vel = max_sd->getVel();
596		RealType temp_vel;
597
598	<	switch(rnemdType_) {
599	<	case rnemdKinetic :
598	>	switch(rnemdFluxType_) {
599	>	case rnemdKE :
600		min_sd->setVel(max_vel);
601		max_sd->setVel(min_vel);
602	<	if (min_sd->isDirectional() && max_sd->isDirectional()) {
602	>	if (min_sd->isDirectional() && max_sd->isDirectional()) {
603		Vector3d min_angMom = min_sd->getJ();
604		Vector3d max_angMom = max_sd->getJ();
605		min_sd->setJ(max_angMom);
606		max_sd->setJ(min_angMom);
607	<	}
607	>	}//angular momenta exchange enabled
608	>	//assumes same rigid body identity
609		break;
610		case rnemdPx :
611		temp_vel = min_vel.x();
#	Line 329 | Line 628 | namespace oopse {
628		min_sd->setVel(min_vel);
629		max_sd->setVel(max_vel);
630		break;
332	–	case rnemdUnknown :
631		default :
632		break;
633		}
634	+
635		#ifdef IS_MPI
636		// the rest of the cases only apply in parallel simulations:
637		} else if (max_vals.rank == worldRank) {
#	Line 348 | Line 647 | namespace oopse {
647		min_vel.getArrayPointer(), 3, MPI::REALTYPE,
648		min_vals.rank, 0, status);
649
650	<	switch(rnemdType_) {
651	<	case rnemdKinetic :
650	>	switch(rnemdFluxType_) {
651	>	case rnemdKE :
652		max_sd->setVel(min_vel);
653	<
653	>	//angular momenta exchange enabled
654		if (max_sd->isDirectional()) {
655		Vector3d min_angMom;
656		Vector3d max_angMom = max_sd->getJ();
657	<
657	>
658		// point-to-point swap of the angular momentum vector
659		MPI::COMM_WORLD.Sendrecv(max_angMom.getArrayPointer(), 3,
660		MPI::REALTYPE, min_vals.rank, 1,
661		min_angMom.getArrayPointer(), 3,
662		MPI::REALTYPE, min_vals.rank, 1,
663		status);
664	<
664	>
665		max_sd->setJ(min_angMom);
666	<	}
666	>	}
667		break;
668		case rnemdPx :
669		max_vel.x() = min_vel.x();
#	Line 378 | Line 677 | namespace oopse {
677		max_vel.z() = min_vel.z();
678		max_sd->setVel(max_vel);
679		break;
381	–	case rnemdUnknown :
680		default :
681		break;
682		}
#	Line 395 | Line 693 | namespace oopse {
693		max_vel.getArrayPointer(), 3, MPI::REALTYPE,
694		max_vals.rank, 0, status);
695
696	<	switch(rnemdType_) {
697	<	case rnemdKinetic :
696	>	switch(rnemdFluxType_) {
697	>	case rnemdKE :
698		min_sd->setVel(max_vel);
699	<
699	>	//angular momenta exchange enabled
700		if (min_sd->isDirectional()) {
701		Vector3d min_angMom = min_sd->getJ();
702		Vector3d max_angMom;
703	<
703	>
704		// point-to-point swap of the angular momentum vector
705		MPI::COMM_WORLD.Sendrecv(min_angMom.getArrayPointer(), 3,
706		MPI::REALTYPE, max_vals.rank, 1,
707		max_angMom.getArrayPointer(), 3,
708		MPI::REALTYPE, max_vals.rank, 1,
709		status);
710	<
710	>
711		min_sd->setJ(max_angMom);
712		}
713		break;
#	Line 425 | Line 723 | namespace oopse {
723		min_vel.z() = max_vel.z();
724		min_sd->setVel(min_vel);
725		break;
428	–	case rnemdUnknown :
726		default :
727		break;
728		}
729		}
730		#endif
731	<	exchangeSum_ += max_val - min_val;
732	<	} else {
733	<	std::cerr << "exchange NOT performed.\nmin_val > max_val.\n";
731	>
732	>	switch(rnemdFluxType_) {
733	>	case rnemdKE:
734	>	cerr << "KE\n";
735	>	kineticExchange_ += max_val - min_val;
736	>	break;
737	>	case rnemdPx:
738	>	momentumExchange_.x() += max_val - min_val;
739	>	break;
740	>	case rnemdPy:
741	>	momentumExchange_.y() += max_val - min_val;
742	>	break;
743	>	case rnemdPz:
744	>	momentumExchange_.z() += max_val - min_val;
745	>	break;
746	>	default:
747	>	cerr << "default\n";
748	>	break;
749	>	}
750	>	} else {
751	>	sprintf(painCave.errMsg,
752	>	"RNEMD::doSwap exchange NOT performed because min_val > max_val\n");
753	>	painCave.isFatal = 0;
754	>	painCave.severity = OPENMD_INFO;
755	>	simError();
756	>	failTrialCount_++;
757		}
758		} else {
759	<	std::cerr << "exchange NOT performed.\none of the two slabs empty.\n";
760	<	}
761	<
759	>	sprintf(painCave.errMsg,
760	>	"RNEMD::doSwap exchange NOT performed because selected object\n"
761	>	"\twas not present in at least one of the two slabs.\n");
762	>	painCave.isFatal = 0;
763	>	painCave.severity = OPENMD_INFO;
764	>	simError();
765	>	failTrialCount_++;
766	>	}
767		}
768
769	<	void RNEMD::getStatus() {
769	>	void RNEMD::doNIVS() {
770
771		Snapshot* currentSnap_ = info_->getSnapshotManager()->getCurrentSnapshot();
772		Mat3x3d hmat = currentSnap_->getHmat();
448	–	Stats& stat = currentSnap_->statData;
449	–	RealType time = currentSnap_->getTime();
773
451	–	stat[Stats::RNEMD_SWAP_TOTAL] = exchangeSum_;
452	–
774		seleMan_.setSelectionSet(evaluator_.evaluate());
775
776		int selei;
777		StuntDouble* sd;
778		int idx;
779
780	<	std::vector<RealType> valueHist(nBins_, 0.0); // keeps track of what's
460	<	// being averaged
461	<	std::vector<int> valueCount(nBins_, 0);       // keeps track of the
462	<	// number of degrees of
463	<	// freedom being averaged
780	>	vector<StuntDouble*> hotBin, coldBin;
781
782	+	RealType Phx = 0.0;
783	+	RealType Phy = 0.0;
784	+	RealType Phz = 0.0;
785	+	RealType Khx = 0.0;
786	+	RealType Khy = 0.0;
787	+	RealType Khz = 0.0;
788	+	RealType Khw = 0.0;
789	+	RealType Pcx = 0.0;
790	+	RealType Pcy = 0.0;
791	+	RealType Pcz = 0.0;
792	+	RealType Kcx = 0.0;
793	+	RealType Kcy = 0.0;
794	+	RealType Kcz = 0.0;
795	+	RealType Kcw = 0.0;
796	+
797		for (sd = seleMan_.beginSelected(selei); sd != NULL;
798		sd = seleMan_.nextSelected(selei)) {
799	<
799	>
800		idx = sd->getLocalIndex();
801	<
801	>
802		Vector3d pos = sd->getPos();
803
804		// wrap the stuntdouble's position back into the box:
805	<
805	>
806		if (usePeriodicBoundaryConditions_)
807		currentSnap_->wrapVector(pos);
808	<
808	>
809		// which bin is this stuntdouble in?
810	<	// wrapped positions are in the range [-0.5*hmat(2,2), +0.5*hmat(2,2)]
811	<
480	<	int binNo = int(nBins_ * (pos.z() / hmat(2,2) + 0.5)) % nBins_;
481	<
482	<	RealType mass = sd->getMass();
483	<	Vector3d vel = sd->getVel();
484	<	RealType value;
810	>	bool inA = inSlabA(pos);
811	>	bool inB = inSlabB(pos);
812
813	<	switch(rnemdType_) {
814	<	case rnemdKinetic :
815	<
816	<	value = mass * (vel[0]*vel[0] + vel[1]*vel[1] +
817	<	vel[2]*vel[2]);
818	<
819	<	valueCount[binNo] += 3;
820	<	if (sd->isDirectional()) {
821	<	Vector3d angMom = sd->getJ();
822	<	Mat3x3d I = sd->getI();
823	<
824	<	if (sd->isLinear()) {
825	<	int i = sd->linearAxis();
826	<	int j = (i + 1) % 3;
827	<	int k = (i + 2) % 3;
828	<	value += angMom[j] * angMom[j] / I(j, j) +
829	<	angMom[k] * angMom[k] / I(k, k);
813	>	if (inA || inB) {
814	>
815	>	RealType mass = sd->getMass();
816	>	Vector3d vel = sd->getVel();
817	>
818	>	if (inA) {
819	>	hotBin.push_back(sd);
820	>	Phx += mass * vel.x();
821	>	Phy += mass * vel.y();
822	>	Phz += mass * vel.z();
823	>	Khx += mass * vel.x() * vel.x();
824	>	Khy += mass * vel.y() * vel.y();
825	>	Khz += mass * vel.z() * vel.z();
826	>	if (sd->isDirectional()) {
827	>	Vector3d angMom = sd->getJ();
828	>	Mat3x3d I = sd->getI();
829	>	if (sd->isLinear()) {
830	>	int i = sd->linearAxis();
831	>	int j = (i + 1) % 3;
832	>	int k = (i + 2) % 3;
833	>	Khw += angMom[j] * angMom[j] / I(j, j) +
834	>	angMom[k] * angMom[k] / I(k, k);
835	>	} else {
836	>	Khw += angMom[0]*angMom[0]/I(0, 0)
837	>	+ angMom[1]*angMom[1]/I(1, 1)
838	>	+ angMom[2]*angMom[2]/I(2, 2);
839	>	}
840	>	}
841	>	} else {
842	>	coldBin.push_back(sd);
843	>	Pcx += mass * vel.x();
844	>	Pcy += mass * vel.y();
845	>	Pcz += mass * vel.z();
846	>	Kcx += mass * vel.x() * vel.x();
847	>	Kcy += mass * vel.y() * vel.y();
848	>	Kcz += mass * vel.z() * vel.z();
849	>	if (sd->isDirectional()) {
850	>	Vector3d angMom = sd->getJ();
851	>	Mat3x3d I = sd->getI();
852	>	if (sd->isLinear()) {
853	>	int i = sd->linearAxis();
854	>	int j = (i + 1) % 3;
855	>	int k = (i + 2) % 3;
856	>	Kcw += angMom[j] * angMom[j] / I(j, j) +
857	>	angMom[k] * angMom[k] / I(k, k);
858	>	} else {
859	>	Kcw += angMom[0]*angMom[0]/I(0, 0)
860	>	+ angMom[1]*angMom[1]/I(1, 1)
861	>	+ angMom[2]*angMom[2]/I(2, 2);
862	>	}
863	>	}
864	>	}
865	>	}
866	>	}
867	>
868	>	Khx *= 0.5;
869	>	Khy *= 0.5;
870	>	Khz *= 0.5;
871	>	Khw *= 0.5;
872	>	Kcx *= 0.5;
873	>	Kcy *= 0.5;
874	>	Kcz *= 0.5;
875	>	Kcw *= 0.5;
876
877	<	valueCount[binNo] +=2;
877	>	#ifdef IS_MPI
878	>	MPI::COMM_WORLD.Allreduce(MPI::IN_PLACE, &Phx, 1, MPI::REALTYPE, MPI::SUM);
879	>	MPI::COMM_WORLD.Allreduce(MPI::IN_PLACE, &Phy, 1, MPI::REALTYPE, MPI::SUM);
880	>	MPI::COMM_WORLD.Allreduce(MPI::IN_PLACE, &Phz, 1, MPI::REALTYPE, MPI::SUM);
881	>	MPI::COMM_WORLD.Allreduce(MPI::IN_PLACE, &Pcx, 1, MPI::REALTYPE, MPI::SUM);
882	>	MPI::COMM_WORLD.Allreduce(MPI::IN_PLACE, &Pcy, 1, MPI::REALTYPE, MPI::SUM);
883	>	MPI::COMM_WORLD.Allreduce(MPI::IN_PLACE, &Pcz, 1, MPI::REALTYPE, MPI::SUM);
884
885	<	} else {
886	<	value += angMom[0]*angMom[0]/I(0, 0)
887	<	+ angMom[1]*angMom[1]/I(1, 1)
888	<	+ angMom[2]*angMom[2]/I(2, 2);
889	<	valueCount[binNo] +=3;
885	>	MPI::COMM_WORLD.Allreduce(MPI::IN_PLACE, &Khx, 1, MPI::REALTYPE, MPI::SUM);
886	>	MPI::COMM_WORLD.Allreduce(MPI::IN_PLACE, &Khy, 1, MPI::REALTYPE, MPI::SUM);
887	>	MPI::COMM_WORLD.Allreduce(MPI::IN_PLACE, &Khz, 1, MPI::REALTYPE, MPI::SUM);
888	>	MPI::COMM_WORLD.Allreduce(MPI::IN_PLACE, &Khw, 1, MPI::REALTYPE, MPI::SUM);
889	>
890	>	MPI::COMM_WORLD.Allreduce(MPI::IN_PLACE, &Kcx, 1, MPI::REALTYPE, MPI::SUM);
891	>	MPI::COMM_WORLD.Allreduce(MPI::IN_PLACE, &Kcy, 1, MPI::REALTYPE, MPI::SUM);
892	>	MPI::COMM_WORLD.Allreduce(MPI::IN_PLACE, &Kcz, 1, MPI::REALTYPE, MPI::SUM);
893	>	MPI::COMM_WORLD.Allreduce(MPI::IN_PLACE, &Kcw, 1, MPI::REALTYPE, MPI::SUM);
894	>	#endif
895	>
896	>	//solve coldBin coeff's first
897	>	RealType px = Pcx / Phx;
898	>	RealType py = Pcy / Phy;
899	>	RealType pz = Pcz / Phz;
900	>	RealType c, x, y, z;
901	>	bool successfulScale = false;
902	>	if ((rnemdFluxType_ == rnemdFullKE) ||
903	>	(rnemdFluxType_ == rnemdRotKE)) {
904	>	//may need sanity check Khw & Kcw > 0
905	>
906	>	if (rnemdFluxType_ == rnemdFullKE) {
907	>	c = 1.0 - kineticTarget_ / (Kcx + Kcy + Kcz + Kcw);
908	>	} else {
909	>	c = 1.0 - kineticTarget_ / Kcw;
910	>	}
911	>
912	>	if ((c > 0.81) && (c < 1.21)) {//restrict scaling coefficients
913	>	c = sqrt(c);
914	>	//std::cerr << "cold slab scaling coefficient: " << c << endl;
915	>	//now convert to hotBin coefficient
916	>	RealType w = 0.0;
917	>	if (rnemdFluxType_ ==  rnemdFullKE) {
918	>	x = 1.0 + px * (1.0 - c);
919	>	y = 1.0 + py * (1.0 - c);
920	>	z = 1.0 + pz * (1.0 - c);
921	>	/* more complicated way
922	>	w = 1.0 + (Kcw - Kcw * c * c - (c * c * (Kcx + Kcy + Kcz
923	>	+ Khx * px * px + Khy * py * py + Khz * pz * pz)
924	>	- 2.0 * c * (Khx * px * (1.0 + px) + Khy * py * (1.0 + py)
925	>	+ Khz * pz * (1.0 + pz)) + Khx * px * (2.0 + px)
926	>	+ Khy * py * (2.0 + py) + Khz * pz * (2.0 + pz)
927	>	- Kcx - Kcy - Kcz)) / Khw; the following is simpler
928	>	*/
929	>	if ((fabs(x - 1.0) < 0.1) && (fabs(y - 1.0) < 0.1) &&
930	>	(fabs(z - 1.0) < 0.1)) {
931	>	w = 1.0 + (kineticTarget_
932	>	+ Khx * (1.0 - x * x) + Khy * (1.0 - y * y)
933	>	+ Khz * (1.0 - z * z)) / Khw;
934	>	}//no need to calculate w if x, y or z is out of range
935	>	} else {
936	>	w = 1.0 + kineticTarget_ / Khw;
937	>	}
938	>	if ((w > 0.81) && (w < 1.21)) {//restrict scaling coefficients
939	>	//if w is in the right range, so should be x, y, z.
940	>	vector<StuntDouble*>::iterator sdi;
941	>	Vector3d vel;
942	>	for (sdi = coldBin.begin(); sdi != coldBin.end(); sdi++) {
943	>	if (rnemdFluxType_ == rnemdFullKE) {
944	>	vel = (*sdi)->getVel() * c;
945	>	(*sdi)->setVel(vel);
946	>	}
947	>	if ((*sdi)->isDirectional()) {
948	>	Vector3d angMom = (*sdi)->getJ() * c;
949	>	(*sdi)->setJ(angMom);
950	>	}
951		}
952	+	w = sqrt(w);
953	+	// std::cerr << "xh= " << x << "\tyh= " << y << "\tzh= " << z
954	+	//           << "\twh= " << w << endl;
955	+	for (sdi = hotBin.begin(); sdi != hotBin.end(); sdi++) {
956	+	if (rnemdFluxType_ == rnemdFullKE) {
957	+	vel = (*sdi)->getVel();
958	+	vel.x() *= x;
959	+	vel.y() *= y;
960	+	vel.z() *= z;
961	+	(*sdi)->setVel(vel);
962	+	}
963	+	if ((*sdi)->isDirectional()) {
964	+	Vector3d angMom = (*sdi)->getJ() * w;
965	+	(*sdi)->setJ(angMom);
966	+	}
967	+	}
968	+	successfulScale = true;
969	+	kineticExchange_ += kineticTarget_;
970		}
971	<	value = value / OOPSEConstant::energyConvert / OOPSEConstant::kb;
972	<
973	<	break;
974	<	case rnemdPx :
975	<	value = mass * vel[0];
976	<	valueCount[binNo]++;
971	>	}
972	>	} else {
973	>	RealType a000, a110, c0, a001, a111, b01, b11, c1;
974	>	switch(rnemdFluxType_) {
975	>	case rnemdKE :
976	>	/* used hotBin coeff's & only scale x & y dimensions
977	>	RealType px = Phx / Pcx;
978	>	RealType py = Phy / Pcy;
979	>	a110 = Khy;
980	>	c0 = - Khx - Khy - kineticTarget_;
981	>	a000 = Khx;
982	>	a111 = Kcy * py * py;
983	>	b11 = -2.0 * Kcy * py * (1.0 + py);
984	>	c1 = Kcy * py * (2.0 + py) + Kcx * px * ( 2.0 + px) + kineticTarget_;
985	>	b01 = -2.0 * Kcx * px * (1.0 + px);
986	>	a001 = Kcx * px * px;
987	>	*/
988	>	//scale all three dimensions, let c_x = c_y
989	>	a000 = Kcx + Kcy;
990	>	a110 = Kcz;
991	>	c0 = kineticTarget_ - Kcx - Kcy - Kcz;
992	>	a001 = Khx * px * px + Khy * py * py;
993	>	a111 = Khz * pz * pz;
994	>	b01 = -2.0 * (Khx * px * (1.0 + px) + Khy * py * (1.0 + py));
995	>	b11 = -2.0 * Khz * pz * (1.0 + pz);
996	>	c1 = Khx * px * (2.0 + px) + Khy * py * (2.0 + py)
997	>	+ Khz * pz * (2.0 + pz) - kineticTarget_;
998		break;
999	+	case rnemdPx :
1000	+	c = 1 - momentumTarget_.x() / Pcx;
1001	+	a000 = Kcy;
1002	+	a110 = Kcz;
1003	+	c0 = Kcx * c * c - Kcx - Kcy - Kcz;
1004	+	a001 = py * py * Khy;
1005	+	a111 = pz * pz * Khz;
1006	+	b01 = -2.0 * Khy * py * (1.0 + py);
1007	+	b11 = -2.0 * Khz * pz * (1.0 + pz);
1008	+	c1 = Khy * py * (2.0 + py) + Khz * pz * (2.0 + pz)
1009	+	+ Khx * (fastpow(c * px - px - 1.0, 2) - 1.0);
1010	+	break;
1011		case rnemdPy :
1012	<	value = mass * vel[1];
1013	<	valueCount[binNo]++;
1012	>	c = 1 - momentumTarget_.y() / Pcy;
1013	>	a000 = Kcx;
1014	>	a110 = Kcz;
1015	>	c0 = Kcy * c * c - Kcx - Kcy - Kcz;
1016	>	a001 = px * px * Khx;
1017	>	a111 = pz * pz * Khz;
1018	>	b01 = -2.0 * Khx * px * (1.0 + px);
1019	>	b11 = -2.0 * Khz * pz * (1.0 + pz);
1020	>	c1 = Khx * px * (2.0 + px) + Khz * pz * (2.0 + pz)
1021	>	+ Khy * (fastpow(c * py - py - 1.0, 2) - 1.0);
1022		break;
1023	<	case rnemdPz :
1024	<	value = mass * vel[2];
1025	<	valueCount[binNo]++;
1023	>	case rnemdPz ://we don't really do this, do we?
1024	>	c = 1 - momentumTarget_.z() / Pcz;
1025	>	a000 = Kcx;
1026	>	a110 = Kcy;
1027	>	c0 = Kcz * c * c - Kcx - Kcy - Kcz;
1028	>	a001 = px * px * Khx;
1029	>	a111 = py * py * Khy;
1030	>	b01 = -2.0 * Khx * px * (1.0 + px);
1031	>	b11 = -2.0 * Khy * py * (1.0 + py);
1032	>	c1 = Khx * px * (2.0 + px) + Khy * py * (2.0 + py)
1033	>	+ Khz * (fastpow(c * pz - pz - 1.0, 2) - 1.0);
1034		break;
528	–	case rnemdUnknown :
1035		default :
1036		break;
1037		}
1038	<	valueHist[binNo] += value;
1038	>
1039	>	RealType v1 = a000 * a111 - a001 * a110;
1040	>	RealType v2 = a000 * b01;
1041	>	RealType v3 = a000 * b11;
1042	>	RealType v4 = a000 * c1 - a001 * c0;
1043	>	RealType v8 = a110 * b01;
1044	>	RealType v10 = - b01 * c0;
1045	>
1046	>	RealType u0 = v2 * v10 - v4 * v4;
1047	>	RealType u1 = -2.0 * v3 * v4;
1048	>	RealType u2 = -v2 * v8 - v3 * v3 - 2.0 * v1 * v4;
1049	>	RealType u3 = -2.0 * v1 * v3;
1050	>	RealType u4 = - v1 * v1;
1051	>	//rescale coefficients
1052	>	RealType maxAbs = fabs(u0);
1053	>	if (maxAbs < fabs(u1)) maxAbs = fabs(u1);
1054	>	if (maxAbs < fabs(u2)) maxAbs = fabs(u2);
1055	>	if (maxAbs < fabs(u3)) maxAbs = fabs(u3);
1056	>	if (maxAbs < fabs(u4)) maxAbs = fabs(u4);
1057	>	u0 /= maxAbs;
1058	>	u1 /= maxAbs;
1059	>	u2 /= maxAbs;
1060	>	u3 /= maxAbs;
1061	>	u4 /= maxAbs;
1062	>	//max_element(start, end) is also available.
1063	>	Polynomial<RealType> poly; //same as DoublePolynomial poly;
1064	>	poly.setCoefficient(4, u4);
1065	>	poly.setCoefficient(3, u3);
1066	>	poly.setCoefficient(2, u2);
1067	>	poly.setCoefficient(1, u1);
1068	>	poly.setCoefficient(0, u0);
1069	>	vector<RealType> realRoots = poly.FindRealRoots();
1070	>
1071	>	vector<RealType>::iterator ri;
1072	>	RealType r1, r2, alpha0;
1073	>	vector<pair<RealType,RealType> > rps;
1074	>	for (ri = realRoots.begin(); ri !=realRoots.end(); ri++) {
1075	>	r2 = *ri;
1076	>	//check if FindRealRoots() give the right answer
1077	>	if ( fabs(u0 + r2 * (u1 + r2 * (u2 + r2 * (u3 + r2 * u4)))) > 1e-6 ) {
1078	>	sprintf(painCave.errMsg,
1079	>	"RNEMD Warning: polynomial solve seems to have an error!");
1080	>	painCave.isFatal = 0;
1081	>	simError();
1082	>	failRootCount_++;
1083	>	}
1084	>	//might not be useful w/o rescaling coefficients
1085	>	alpha0 = -c0 - a110 * r2 * r2;
1086	>	if (alpha0 >= 0.0) {
1087	>	r1 = sqrt(alpha0 / a000);
1088	>	if (fabs(c1 + r1 * (b01 + r1 * a001) + r2 * (b11 + r2 * a111))
1089	>	< 1e-6)
1090	>	{ rps.push_back(make_pair(r1, r2)); }
1091	>	if (r1 > 1e-6) { //r1 non-negative
1092	>	r1 = -r1;
1093	>	if (fabs(c1 + r1 * (b01 + r1 * a001) + r2 * (b11 + r2 * a111))
1094	>	< 1e-6)
1095	>	{ rps.push_back(make_pair(r1, r2)); }
1096	>	}
1097	>	}
1098	>	}
1099	>	// Consider combining together the solving pair part w/ the searching
1100	>	// best solution part so that we don't need the pairs vector
1101	>	if (!rps.empty()) {
1102	>	RealType smallestDiff = HONKING_LARGE_VALUE;
1103	>	RealType diff;
1104	>	pair<RealType,RealType> bestPair = make_pair(1.0, 1.0);
1105	>	vector<pair<RealType,RealType> >::iterator rpi;
1106	>	for (rpi = rps.begin(); rpi != rps.end(); rpi++) {
1107	>	r1 = (*rpi).first;
1108	>	r2 = (*rpi).second;
1109	>	switch(rnemdFluxType_) {
1110	>	case rnemdKE :
1111	>	diff = fastpow(1.0 - r1, 2) + fastpow(1.0 - r2, 2)
1112	>	+ fastpow(r1 * r1 / r2 / r2 - Kcz/Kcx, 2)
1113	>	+ fastpow(r1 * r1 / r2 / r2 - Kcz/Kcy, 2);
1114	>	break;
1115	>	case rnemdPx :
1116	>	diff = fastpow(1.0 - r1, 2) + fastpow(1.0 - r2, 2)
1117	>	+ fastpow(r1 * r1 / r2 / r2 - Kcz/Kcy, 2);
1118	>	break;
1119	>	case rnemdPy :
1120	>	diff = fastpow(1.0 - r1, 2) + fastpow(1.0 - r2, 2)
1121	>	+ fastpow(r1 * r1 / r2 / r2 - Kcz/Kcx, 2);
1122	>	break;
1123	>	case rnemdPz :
1124	>	diff = fastpow(1.0 - r1, 2) + fastpow(1.0 - r2, 2)
1125	>	+ fastpow(r1 * r1 / r2 / r2 - Kcy/Kcx, 2);
1126	>	default :
1127	>	break;
1128	>	}
1129	>	if (diff < smallestDiff) {
1130	>	smallestDiff = diff;
1131	>	bestPair = *rpi;
1132	>	}
1133	>	}
1134	>	#ifdef IS_MPI
1135	>	if (worldRank == 0) {
1136	>	#endif
1137	>	// sprintf(painCave.errMsg,
1138	>	//         "RNEMD: roots r1= %lf\tr2 = %lf\n",
1139	>	//         bestPair.first, bestPair.second);
1140	>	// painCave.isFatal = 0;
1141	>	// painCave.severity = OPENMD_INFO;
1142	>	// simError();
1143	>	#ifdef IS_MPI
1144	>	}
1145	>	#endif
1146	>
1147	>	switch(rnemdFluxType_) {
1148	>	case rnemdKE :
1149	>	x = bestPair.first;
1150	>	y = bestPair.first;
1151	>	z = bestPair.second;
1152	>	break;
1153	>	case rnemdPx :
1154	>	x = c;
1155	>	y = bestPair.first;
1156	>	z = bestPair.second;
1157	>	break;
1158	>	case rnemdPy :
1159	>	x = bestPair.first;
1160	>	y = c;
1161	>	z = bestPair.second;
1162	>	break;
1163	>	case rnemdPz :
1164	>	x = bestPair.first;
1165	>	y = bestPair.second;
1166	>	z = c;
1167	>	break;
1168	>	default :
1169	>	break;
1170	>	}
1171	>	vector<StuntDouble*>::iterator sdi;
1172	>	Vector3d vel;
1173	>	for (sdi = coldBin.begin(); sdi != coldBin.end(); sdi++) {
1174	>	vel = (*sdi)->getVel();
1175	>	vel.x() *= x;
1176	>	vel.y() *= y;
1177	>	vel.z() *= z;
1178	>	(*sdi)->setVel(vel);
1179	>	}
1180	>	//convert to hotBin coefficient
1181	>	x = 1.0 + px * (1.0 - x);
1182	>	y = 1.0 + py * (1.0 - y);
1183	>	z = 1.0 + pz * (1.0 - z);
1184	>	for (sdi = hotBin.begin(); sdi != hotBin.end(); sdi++) {
1185	>	vel = (*sdi)->getVel();
1186	>	vel.x() *= x;
1187	>	vel.y() *= y;
1188	>	vel.z() *= z;
1189	>	(*sdi)->setVel(vel);
1190	>	}
1191	>	successfulScale = true;
1192	>	switch(rnemdFluxType_) {
1193	>	case rnemdKE :
1194	>	kineticExchange_ += kineticTarget_;
1195	>	break;
1196	>	case rnemdPx :
1197	>	case rnemdPy :
1198	>	case rnemdPz :
1199	>	momentumExchange_ += momentumTarget_;
1200	>	break;
1201	>	default :
1202	>	break;
1203	>	}
1204	>	}
1205		}
1206	+	if (successfulScale != true) {
1207	+	sprintf(painCave.errMsg,
1208	+	"RNEMD::doNIVS exchange NOT performed - roots that solve\n"
1209	+	"\tthe constraint equations may not exist or there may be\n"
1210	+	"\tno selected objects in one or both slabs.\n");
1211	+	painCave.isFatal = 0;
1212	+	painCave.severity = OPENMD_INFO;
1213	+	simError();
1214	+	failTrialCount_++;
1215	+	}
1216	+	}
1217
1218	<	#ifdef IS_MPI
1218	>	void RNEMD::doVSS() {
1219
1220	<	// all processors have the same number of bins, and STL vectors pack their
1221	<	// arrays, so in theory, this should be safe:
1220	>	Snapshot* currentSnap_ = info_->getSnapshotManager()->getCurrentSnapshot();
1221	>	RealType time = currentSnap_->getTime();
1222	>	Mat3x3d hmat = currentSnap_->getHmat();
1223
1224	<	MPI::COMM_WORLD.Allreduce(MPI::IN_PLACE, &valueHist[0],
1225	<	nBins_, MPI::REALTYPE, MPI::SUM);
1226	<	MPI::COMM_WORLD.Allreduce(MPI::IN_PLACE, &valueCount[0],
1227	<	nBins_, MPI::INT, MPI::SUM);
1224	>	seleMan_.setSelectionSet(evaluator_.evaluate());
1225	>
1226	>	int selei;
1227	>	StuntDouble* sd;
1228	>	int idx;
1229	>
1230	>	vector<StuntDouble*> hotBin, coldBin;
1231	>
1232	>	Vector3d Ph(V3Zero);
1233	>	RealType Mh = 0.0;
1234	>	RealType Kh = 0.0;
1235	>	Vector3d Pc(V3Zero);
1236	>	RealType Mc = 0.0;
1237	>	RealType Kc = 0.0;
1238	>
1239	>
1240	>	for (sd = seleMan_.beginSelected(selei); sd != NULL;
1241	>	sd = seleMan_.nextSelected(selei)) {
1242	>
1243	>	idx = sd->getLocalIndex();
1244	>
1245	>	Vector3d pos = sd->getPos();
1246	>
1247	>	// wrap the stuntdouble's position back into the box:
1248	>
1249	>	if (usePeriodicBoundaryConditions_)
1250	>	currentSnap_->wrapVector(pos);
1251	>
1252	>	// which bin is this stuntdouble in?
1253	>	bool inA = inSlabA(pos);
1254	>	bool inB = inSlabB(pos);
1255	>
1256	>	if (inA || inB) {
1257	>
1258	>	RealType mass = sd->getMass();
1259	>	Vector3d vel = sd->getVel();
1260	>
1261	>	if (inA) {
1262	>	hotBin.push_back(sd);
1263	>	//std::cerr << "before, velocity = " << vel << endl;
1264	>	Ph += mass * vel;
1265	>	//std::cerr << "after, velocity = " << vel << endl;
1266	>	Mh += mass;
1267	>	Kh += mass * vel.lengthSquare();
1268	>	if (rnemdFluxType_ == rnemdFullKE) {
1269	>	if (sd->isDirectional()) {
1270	>	Vector3d angMom = sd->getJ();
1271	>	Mat3x3d I = sd->getI();
1272	>	if (sd->isLinear()) {
1273	>	int i = sd->linearAxis();
1274	>	int j = (i + 1) % 3;
1275	>	int k = (i + 2) % 3;
1276	>	Kh += angMom[j] * angMom[j] / I(j, j) +
1277	>	angMom[k] * angMom[k] / I(k, k);
1278	>	} else {
1279	>	Kh += angMom[0] * angMom[0] / I(0, 0) +
1280	>	angMom[1] * angMom[1] / I(1, 1) +
1281	>	angMom[2] * angMom[2] / I(2, 2);
1282	>	}
1283	>	}
1284	>	}
1285	>	} else { //midBin_
1286	>	coldBin.push_back(sd);
1287	>	Pc += mass * vel;
1288	>	Mc += mass;
1289	>	Kc += mass * vel.lengthSquare();
1290	>	if (rnemdFluxType_ == rnemdFullKE) {
1291	>	if (sd->isDirectional()) {
1292	>	Vector3d angMom = sd->getJ();
1293	>	Mat3x3d I = sd->getI();
1294	>	if (sd->isLinear()) {
1295	>	int i = sd->linearAxis();
1296	>	int j = (i + 1) % 3;
1297	>	int k = (i + 2) % 3;
1298	>	Kc += angMom[j] * angMom[j] / I(j, j) +
1299	>	angMom[k] * angMom[k] / I(k, k);
1300	>	} else {
1301	>	Kc += angMom[0] * angMom[0] / I(0, 0) +
1302	>	angMom[1] * angMom[1] / I(1, 1) +
1303	>	angMom[2] * angMom[2] / I(2, 2);
1304	>	}
1305	>	}
1306	>	}
1307	>	}
1308	>	}
1309	>	}
1310	>
1311	>	Kh *= 0.5;
1312	>	Kc *= 0.5;
1313	>
1314	>	// std::cerr << "Mh= " << Mh << "\tKh= " << Kh << "\tMc= " << Mc
1315	>	//        << "\tKc= " << Kc << endl;
1316	>	// std::cerr << "Ph= " << Ph << "\tPc= " << Pc << endl;
1317	>
1318	>	#ifdef IS_MPI
1319	>	MPI::COMM_WORLD.Allreduce(MPI::IN_PLACE, &Ph[0], 3, MPI::REALTYPE, MPI::SUM);
1320	>	MPI::COMM_WORLD.Allreduce(MPI::IN_PLACE, &Pc[0], 3, MPI::REALTYPE, MPI::SUM);
1321	>	MPI::COMM_WORLD.Allreduce(MPI::IN_PLACE, &Mh, 1, MPI::REALTYPE, MPI::SUM);
1322	>	MPI::COMM_WORLD.Allreduce(MPI::IN_PLACE, &Kh, 1, MPI::REALTYPE, MPI::SUM);
1323	>	MPI::COMM_WORLD.Allreduce(MPI::IN_PLACE, &Mc, 1, MPI::REALTYPE, MPI::SUM);
1324	>	MPI::COMM_WORLD.Allreduce(MPI::IN_PLACE, &Kc, 1, MPI::REALTYPE, MPI::SUM);
1325	>	#endif
1326	>
1327	>	bool successfulExchange = false;
1328	>	if ((Mh > 0.0) && (Mc > 0.0)) {//both slabs are not empty
1329	>	Vector3d vc = Pc / Mc;
1330	>	Vector3d ac = -momentumTarget_ / Mc + vc;
1331	>	Vector3d acrec = -momentumTarget_ / Mc;
1332	>	RealType cNumerator = Kc - kineticTarget_ - 0.5 * Mc * ac.lengthSquare();
1333	>	if (cNumerator > 0.0) {
1334	>	RealType cDenominator = Kc - 0.5 * Mc * vc.lengthSquare();
1335	>	if (cDenominator > 0.0) {
1336	>	RealType c = sqrt(cNumerator / cDenominator);
1337	>	if ((c > 0.9) && (c < 1.1)) {//restrict scaling coefficients
1338	>	Vector3d vh = Ph / Mh;
1339	>	Vector3d ah = momentumTarget_ / Mh + vh;
1340	>	Vector3d ahrec = momentumTarget_ / Mh;
1341	>	RealType hNumerator = Kh + kineticTarget_
1342	>	- 0.5 * Mh * ah.lengthSquare();
1343	>	if (hNumerator > 0.0) {
1344	>	RealType hDenominator = Kh - 0.5 * Mh * vh.lengthSquare();
1345	>	if (hDenominator > 0.0) {
1346	>	RealType h = sqrt(hNumerator / hDenominator);
1347	>	if ((h > 0.9) && (h < 1.1)) {
1348	>	// std::cerr << "cold slab scaling coefficient: " << c << "\n";
1349	>	// std::cerr << "hot slab scaling coefficient: " << h <<  "\n";
1350	>	vector<StuntDouble*>::iterator sdi;
1351	>	Vector3d vel;
1352	>	for (sdi = coldBin.begin(); sdi != coldBin.end(); sdi++) {
1353	>	//vel = (*sdi)->getVel();
1354	>	vel = ((*sdi)->getVel() - vc) * c + ac;
1355	>	(*sdi)->setVel(vel);
1356	>	if (rnemdFluxType_ == rnemdFullKE) {
1357	>	if ((*sdi)->isDirectional()) {
1358	>	Vector3d angMom = (*sdi)->getJ() * c;
1359	>	(*sdi)->setJ(angMom);
1360	>	}
1361	>	}
1362	>	}
1363	>	for (sdi = hotBin.begin(); sdi != hotBin.end(); sdi++) {
1364	>	//vel = (*sdi)->getVel();
1365	>	vel = ((*sdi)->getVel() - vh) * h + ah;
1366	>	(*sdi)->setVel(vel);
1367	>	if (rnemdFluxType_ == rnemdFullKE) {
1368	>	if ((*sdi)->isDirectional()) {
1369	>	Vector3d angMom = (*sdi)->getJ() * h;
1370	>	(*sdi)->setJ(angMom);
1371	>	}
1372	>	}
1373	>	}
1374	>	successfulExchange = true;
1375	>	kineticExchange_ += kineticTarget_;
1376	>	momentumExchange_ += momentumTarget_;
1377	>	}
1378	>	}
1379	>	}
1380	>	}
1381	>	}
1382	>	}
1383	>	}
1384	>	if (successfulExchange != true) {
1385	>	sprintf(painCave.errMsg,
1386	>	"RNEMD::doVSS exchange NOT performed - roots that solve\n"
1387	>	"\tthe constraint equations may not exist or there may be\n"
1388	>	"\tno selected objects in one or both slabs.\n");
1389	>	painCave.isFatal = 0;
1390	>	painCave.severity = OPENMD_INFO;
1391	>	simError();
1392	>	failTrialCount_++;
1393	>	}
1394	>	}
1395
1396	+	void RNEMD::doRNEMD() {
1397	+
1398	+	trialCount_++;
1399	+	switch(rnemdMethod_) {
1400	+	case rnemdSwap:
1401	+	doSwap();
1402	+	break;
1403	+	case rnemdNIVS:
1404	+	doNIVS();
1405	+	break;
1406	+	case rnemdVSS:
1407	+	doVSS();
1408	+	break;
1409	+	case rnemdUnkownMethod:
1410	+	default :
1411	+	break;
1412	+	}
1413	+	}
1414	+
1415	+	void RNEMD::collectData() {
1416	+
1417	+	Snapshot* currentSnap_ = info_->getSnapshotManager()->getCurrentSnapshot();
1418	+	Mat3x3d hmat = currentSnap_->getHmat();
1419	+
1420	+	areaAccumulator_->add(currentSnap_->getXYarea());
1421	+
1422	+	seleMan_.setSelectionSet(evaluator_.evaluate());
1423	+
1424	+	int selei;
1425	+	StuntDouble* sd;
1426	+	int idx;
1427	+
1428	+	vector<RealType> binMass(nBins_, 0.0);
1429	+	vector<RealType> binPx(nBins_, 0.0);
1430	+	vector<RealType> binPy(nBins_, 0.0);
1431	+	vector<RealType> binPz(nBins_, 0.0);
1432	+	vector<RealType> binKE(nBins_, 0.0);
1433	+	vector<int> binDOF(nBins_, 0);
1434	+	vector<int> binCount(nBins_, 0);
1435	+
1436	+	// alternative approach, track all molecules instead of only those
1437	+	// selected for scaling/swapping:
1438	+	/*
1439	+	SimInfo::MoleculeIterator miter;
1440	+	vector<StuntDouble*>::iterator iiter;
1441	+	Molecule* mol;
1442	+	StuntDouble* sd;
1443	+	for (mol = info_->beginMolecule(miter); mol != NULL;
1444	+	mol = info_->nextMolecule(miter))
1445	+	sd is essentially sd
1446	+	for (sd = mol->beginIntegrableObject(iiter);
1447	+	sd != NULL;
1448	+	sd = mol->nextIntegrableObject(iiter))
1449	+	*/
1450	+	for (sd = seleMan_.beginSelected(selei); sd != NULL;
1451	+	sd = seleMan_.nextSelected(selei)) {
1452	+
1453	+	idx = sd->getLocalIndex();
1454	+
1455	+	Vector3d pos = sd->getPos();
1456	+
1457	+	// wrap the stuntdouble's position back into the box:
1458	+
1459	+	if (usePeriodicBoundaryConditions_)
1460	+	currentSnap_->wrapVector(pos);
1461	+
1462	+
1463	+	// which bin is this stuntdouble in?
1464	+	// wrapped positions are in the range [-0.5*hmat(2,2), +0.5*hmat(2,2)]
1465	+	// Shift molecules by half a box to have bins start at 0
1466	+	// The modulo operator is used to wrap the case when we are
1467	+	// beyond the end of the bins back to the beginning.
1468	+	int binNo = int(nBins_ * (pos.z() / hmat(2,2) + 0.5)) % nBins_;
1469	+
1470	+	RealType mass = sd->getMass();
1471	+	Vector3d vel = sd->getVel();
1472	+
1473	+	binCount[binNo]++;
1474	+	binMass[binNo] += mass;
1475	+	binPx[binNo] += mass*vel.x();
1476	+	binPy[binNo] += mass*vel.y();
1477	+	binPz[binNo] += mass*vel.z();
1478	+	binKE[binNo] += 0.5 * (mass * vel.lengthSquare());
1479	+	binDOF[binNo] += 3;
1480	+
1481	+	if (sd->isDirectional()) {
1482	+	Vector3d angMom = sd->getJ();
1483	+	Mat3x3d I = sd->getI();
1484	+	if (sd->isLinear()) {
1485	+	int i = sd->linearAxis();
1486	+	int j = (i + 1) % 3;
1487	+	int k = (i + 2) % 3;
1488	+	binKE[binNo] += 0.5 * (angMom[j] * angMom[j] / I(j, j) +
1489	+	angMom[k] * angMom[k] / I(k, k));
1490	+	binDOF[binNo] += 2;
1491	+	} else {
1492	+	binKE[binNo] += 0.5 * (angMom[0] * angMom[0] / I(0, 0) +
1493	+	angMom[1] * angMom[1] / I(1, 1) +
1494	+	angMom[2] * angMom[2] / I(2, 2));
1495	+	binDOF[binNo] += 3;
1496	+	}
1497	+	}
1498	+	}
1499	+
1500	+
1501	+	#ifdef IS_MPI
1502	+	MPI::COMM_WORLD.Allreduce(MPI::IN_PLACE, &binCount[0],
1503	+	nBins_, MPI::INT, MPI::SUM);
1504	+	MPI::COMM_WORLD.Allreduce(MPI::IN_PLACE, &binMass[0],
1505	+	nBins_, MPI::REALTYPE, MPI::SUM);
1506	+	MPI::COMM_WORLD.Allreduce(MPI::IN_PLACE, &binPx[0],
1507	+	nBins_, MPI::REALTYPE, MPI::SUM);
1508	+	MPI::COMM_WORLD.Allreduce(MPI::IN_PLACE, &binPy[0],
1509	+	nBins_, MPI::REALTYPE, MPI::SUM);
1510	+	MPI::COMM_WORLD.Allreduce(MPI::IN_PLACE, &binPz[0],
1511	+	nBins_, MPI::REALTYPE, MPI::SUM);
1512	+	MPI::COMM_WORLD.Allreduce(MPI::IN_PLACE, &binKE[0],
1513	+	nBins_, MPI::REALTYPE, MPI::SUM);
1514	+	MPI::COMM_WORLD.Allreduce(MPI::IN_PLACE, &binDOF[0],
1515	+	nBins_, MPI::INT, MPI::SUM);
1516	+	#endif
1517	+
1518	+	Vector3d vel;
1519	+	RealType den;
1520	+	RealType temp;
1521	+	RealType z;
1522	+	for (int i = 0; i < nBins_; i++) {
1523	+	z = (((RealType)i + 0.5) / (RealType)nBins_) * hmat(2,2);
1524	+	vel.x() = binPx[i] / binMass[i];
1525	+	vel.y() = binPy[i] / binMass[i];
1526	+	vel.z() = binPz[i] / binMass[i];
1527	+	den = binCount[i] * nBins_ / currentSnap_->getVolume();
1528	+	temp = 2.0 * binKE[i] / (binDOF[i] * PhysicalConstants::kb *
1529	+	PhysicalConstants::energyConvert);
1530	+
1531	+	for (unsigned int j = 0; j < outputMask_.size(); ++j) {
1532	+	if(outputMask_[j]) {
1533	+	switch(j) {
1534	+	case Z:
1535	+	(data_[j].accumulator[i])->add(z);
1536	+	break;
1537	+	case TEMPERATURE:
1538	+	data_[j].accumulator[i]->add(temp);
1539	+	break;
1540	+	case VELOCITY:
1541	+	dynamic_cast<VectorAccumulator *>(data_[j].accumulator[i])->add(vel);
1542	+	break;
1543	+	case DENSITY:
1544	+	data_[j].accumulator[i]->add(den);
1545	+	break;
1546	+	}
1547	+	}
1548	+	}
1549	+	}
1550	+	}
1551	+
1552	+	void RNEMD::getStarted() {
1553	+	collectData();
1554	+	writeOutputFile();
1555	+	}
1556	+
1557	+	void RNEMD::parseOutputFileFormat(const std::string& format) {
1558	+	StringTokenizer tokenizer(format, " ,;|\t\n\r");
1559	+
1560	+	while(tokenizer.hasMoreTokens()) {
1561	+	std::string token(tokenizer.nextToken());
1562	+	toUpper(token);
1563	+	OutputMapType::iterator i = outputMap_.find(token);
1564	+	if (i != outputMap_.end()) {
1565	+	outputMask_.set(i->second);
1566	+	} else {
1567	+	sprintf( painCave.errMsg,
1568	+	"RNEMD::parseOutputFileFormat: %s is not a recognized\n"
1569	+	"\toutputFileFormat keyword.\n", token.c_str() );
1570	+	painCave.isFatal = 0;
1571	+	painCave.severity = OPENMD_ERROR;
1572	+	simError();
1573	+	}
1574	+	}
1575	+	}
1576	+
1577	+	void RNEMD::writeOutputFile() {
1578	+
1579	+	#ifdef IS_MPI
1580		// If we're the root node, should we print out the results
1581		int worldRank = MPI::COMM_WORLD.Get_rank();
1582		if (worldRank == 0) {
1583		#endif
1584	<
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";
1584	>	rnemdFile_.open(rnemdFileName_.c_str(), std::ios::out | std::ios::trunc );
1585
1586	+	if( !rnemdFile_ ){
1587	+	sprintf( painCave.errMsg,
1588	+	"Could not open \"%s\" for RNEMD output.\n",
1589	+	rnemdFileName_.c_str());
1590	+	painCave.isFatal = 1;
1591	+	simError();
1592	+	}
1593	+
1594	+	Snapshot* currentSnap_ = info_->getSnapshotManager()->getCurrentSnapshot();
1595	+
1596	+	RealType time = currentSnap_->getTime();
1597	+	RealType avgArea;
1598	+	areaAccumulator_->getAverage(avgArea);
1599	+	RealType Jz = kineticExchange_ / (2.0 * time * avgArea);
1600	+	Vector3d JzP = momentumExchange_ / (2.0 * time * avgArea);
1601	+
1602	+	rnemdFile_ << "#######################################################\n";
1603	+	rnemdFile_ << "# RNEMD {\n";
1604	+
1605	+	map<string, RNEMDMethod>::iterator mi;
1606	+	for(mi = stringToMethod_.begin(); mi != stringToMethod_.end(); ++mi) {
1607	+	if ( (*mi).second == rnemdMethod_)
1608	+	rnemdFile_ << "#    exchangeMethod  = \"" << (*mi).first << "\";\n";
1609	+	}
1610	+	map<string, RNEMDFluxType>::iterator fi;
1611	+	for(fi = stringToFluxType_.begin(); fi != stringToFluxType_.end(); ++fi) {
1612	+	if ( (*fi).second == rnemdFluxType_)
1613	+	rnemdFile_ << "#    fluxType  = \"" << (*fi).first << "\";\n";
1614	+	}
1615	+
1616	+	rnemdFile_ << "#    exchangeTime = " << exchangeTime_ << ";\n";
1617	+
1618	+	rnemdFile_ << "#    objectSelection = \""
1619	+	<< rnemdObjectSelection_ << "\";\n";
1620	+	rnemdFile_ << "#    slabWidth = " << slabWidth_ << ";\n";
1621	+	rnemdFile_ << "#    slabAcenter = " << slabACenter_ << ";\n";
1622	+	rnemdFile_ << "#    slabBcenter = " << slabBCenter_ << ";\n";
1623	+	rnemdFile_ << "# }\n";
1624	+	rnemdFile_ << "#######################################################\n";
1625	+	rnemdFile_ << "# RNEMD report:\n";
1626	+	rnemdFile_ << "#     running time = " << time << " fs\n";
1627	+	rnemdFile_ << "#     target flux:\n";
1628	+	rnemdFile_ << "#         kinetic = " << kineticFlux_ << "\n";
1629	+	rnemdFile_ << "#         momentum = " << momentumFluxVector_ << "\n";
1630	+	rnemdFile_ << "#     target one-time exchanges:\n";
1631	+	rnemdFile_ << "#         kinetic = " << kineticTarget_ << "\n";
1632	+	rnemdFile_ << "#         momentum = " << momentumTarget_ << "\n";
1633	+	rnemdFile_ << "#     actual exchange totals:\n";
1634	+	rnemdFile_ << "#         kinetic = " << kineticExchange_ << "\n";
1635	+	rnemdFile_ << "#         momentum = " << momentumExchange_  << "\n";
1636	+	rnemdFile_ << "#     actual flux:\n";
1637	+	rnemdFile_ << "#         kinetic = " << Jz << "\n";
1638	+	rnemdFile_ << "#         momentum = " << JzP  << "\n";
1639	+	rnemdFile_ << "#     exchange statistics:\n";
1640	+	rnemdFile_ << "#         attempted = " << trialCount_ << "\n";
1641	+	rnemdFile_ << "#         failed = " << failTrialCount_ << "\n";
1642	+	if (rnemdMethod_ == rnemdNIVS) {
1643	+	rnemdFile_ << "#         NIVS root-check errors = "
1644	+	<< failRootCount_ << "\n";
1645	+	}
1646	+	rnemdFile_ << "#######################################################\n";
1647	+
1648	+
1649	+
1650	+	//write title
1651	+	rnemdFile_ << "#";
1652	+	for (unsigned int i = 0; i < outputMask_.size(); ++i) {
1653	+	if (outputMask_[i]) {
1654	+	rnemdFile_ << "\t" << data_[i].title <<
1655	+	"(" << data_[i].units << ")";
1656	+	}
1657	+	}
1658	+	rnemdFile_ << std::endl;
1659	+
1660	+	rnemdFile_.precision(8);
1661	+
1662	+	for (unsigned int j = 0; j < nBins_; j++) {
1663	+
1664	+	for (unsigned int i = 0; i < outputMask_.size(); ++i) {
1665	+	if (outputMask_[i]) {
1666	+	if (data_[i].dataType == "RealType")
1667	+	writeReal(i,j);
1668	+	else if (data_[i].dataType == "Vector3d")
1669	+	writeVector(i,j);
1670	+	else {
1671	+	sprintf( painCave.errMsg,
1672	+	"RNEMD found an unknown data type for: %s ",
1673	+	data_[i].title.c_str());
1674	+	painCave.isFatal = 1;
1675	+	simError();
1676	+	}
1677	+	}
1678	+	}
1679	+	rnemdFile_ << std::endl;
1680	+
1681	+	}
1682	+
1683	+	rnemdFile_ << "#######################################################\n";
1684	+	rnemdFile_ << "# Standard Deviations in those quantities follow:\n";
1685	+	rnemdFile_ << "#######################################################\n";
1686	+
1687	+
1688	+	for (unsigned int j = 0; j < nBins_; j++) {
1689	+	rnemdFile_ << "#";
1690	+	for (unsigned int i = 0; i < outputMask_.size(); ++i) {
1691	+	if (outputMask_[i]) {
1692	+	if (data_[i].dataType == "RealType")
1693	+	writeRealStdDev(i,j);
1694	+	else if (data_[i].dataType == "Vector3d")
1695	+	writeVectorStdDev(i,j);
1696	+	else {
1697	+	sprintf( painCave.errMsg,
1698	+	"RNEMD found an unknown data type for: %s ",
1699	+	data_[i].title.c_str());
1700	+	painCave.isFatal = 1;
1701	+	simError();
1702	+	}
1703	+	}
1704	+	}
1705	+	rnemdFile_ << std::endl;
1706	+
1707	+	}
1708	+
1709	+	rnemdFile_.flush();
1710	+	rnemdFile_.close();
1711	+
1712		#ifdef IS_MPI
1713		}
1714		#endif
1715	+
1716		}
1717	+
1718	+	void RNEMD::writeReal(int index, unsigned int bin) {
1719	+	assert(index >=0 && index < ENDINDEX);
1720	+	assert(bin < nBins_);
1721	+	RealType s;
1722	+
1723	+	data_[index].accumulator[bin]->getAverage(s);
1724	+
1725	+	if (! isinf(s) && ! isnan(s)) {
1726	+	rnemdFile_ << "\t" << s;
1727	+	} else{
1728	+	sprintf( painCave.errMsg,
1729	+	"RNEMD detected a numerical error writing: %s for bin %d",
1730	+	data_[index].title.c_str(), bin);
1731	+	painCave.isFatal = 1;
1732	+	simError();
1733	+	}
1734	+	}
1735	+
1736	+	void RNEMD::writeVector(int index, unsigned int bin) {
1737	+	assert(index >=0 && index < ENDINDEX);
1738	+	assert(bin < nBins_);
1739	+	Vector3d s;
1740	+	dynamic_cast<VectorAccumulator*>(data_[index].accumulator[bin])->getAverage(s);
1741	+	if (isinf(s[0]) || isnan(s[0]) ||
1742	+	isinf(s[1]) || isnan(s[1]) ||
1743	+	isinf(s[2]) || isnan(s[2]) ) {
1744	+	sprintf( painCave.errMsg,
1745	+	"RNEMD detected a numerical error writing: %s for bin %d",
1746	+	data_[index].title.c_str(), bin);
1747	+	painCave.isFatal = 1;
1748	+	simError();
1749	+	} else {
1750	+	rnemdFile_ << "\t" << s[0] << "\t" << s[1] << "\t" << s[2];
1751	+	}
1752	+	}
1753	+
1754	+	void RNEMD::writeRealStdDev(int index, unsigned int bin) {
1755	+	assert(index >=0 && index < ENDINDEX);
1756	+	assert(bin < nBins_);
1757	+	RealType s;
1758	+
1759	+	data_[index].accumulator[bin]->getStdDev(s);
1760	+
1761	+	if (! isinf(s) && ! isnan(s)) {
1762	+	rnemdFile_ << "\t" << s;
1763	+	} else{
1764	+	sprintf( painCave.errMsg,
1765	+	"RNEMD detected a numerical error writing: %s std. dev. for bin %d",
1766	+	data_[index].title.c_str(), bin);
1767	+	painCave.isFatal = 1;
1768	+	simError();
1769	+	}
1770	+	}
1771	+
1772	+	void RNEMD::writeVectorStdDev(int index, unsigned int bin) {
1773	+	assert(index >=0 && index < ENDINDEX);
1774	+	assert(bin < nBins_);
1775	+	Vector3d s;
1776	+	dynamic_cast<VectorAccumulator*>(data_[index].accumulator[bin])->getStdDev(s);
1777	+	if (isinf(s[0]) || isnan(s[0]) ||
1778	+	isinf(s[1]) || isnan(s[1]) ||
1779	+	isinf(s[2]) || isnan(s[2]) ) {
1780	+	sprintf( painCave.errMsg,
1781	+	"RNEMD detected a numerical error writing: %s std. dev. for bin %d",
1782	+	data_[index].title.c_str(), bin);
1783	+	painCave.isFatal = 1;
1784	+	simError();
1785	+	} else {
1786	+	rnemdFile_ << "\t" << s[0] << "\t" << s[1] << "\t" << s[2];
1787	+	}
1788	+	}
1789		}
1790	+

Comparing:
trunk/src/integrators/RNEMD.cpp (property svn:keywords), Revision 1350 by gezelter, Thu May 21 18:56:45 2009 UTC vs.
branches/development/src/rnemd/RNEMD.cpp (property svn:keywords), Revision 1775 by gezelter, Wed Aug 8 18:45:52 2012 UTC

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

Diff Legend

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