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 1330 by skuang, Thu Mar 19 21:03:36 2009 UTC vs.
branches/development/src/integrators/RNEMD.cpp (file contents), Revision 1728 by jmarr, Wed May 30 16:07:03 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 <cmath>
43		#include "integrators/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/PhysicalConstants.hpp"
51	+	#include "utils/Tuple.hpp"
52
53		#ifndef IS_MPI
54		#include "math/SeqRandNumGen.hpp"
55		#else
56		#include "math/ParallelRandNumGen.hpp"
57	+	#include <mpi.h>
58		#endif
59
60	<	/* Remove me after testing*/
54	<	/*
55	<	#include <cstdio>
56	<	#include <iostream>
57	<	*/
58	<	/*End remove me*/
60	>	#define HONKING_LARGE_VALUE 1.0e10
61
62	<	namespace oopse {
62	>	using namespace std;
63	>	namespace OpenMD {
64
65	<	RNEMD::RNEMD(SimInfo* info) : info_(info) {
66	<
65	>	RNEMD::RNEMD(SimInfo* info) : info_(info), evaluator_(info), seleMan_(info),
66	>	usePeriodicBoundaryConditions_(info->getSimParams()->getUsePeriodicBoundaryConditions()) {
67	>
68	>	failTrialCount_ = 0;
69	>	failRootCount_ = 0;
70	>
71		int seedValue;
72		Globals * simParams = info->getSimParams();
73
74	<	stringToEnumMap_["Kinetic"] = rnemdKinetic;
74	>	stringToEnumMap_["KineticSwap"] = rnemdKineticSwap;
75	>	stringToEnumMap_["KineticScale"] = rnemdKineticScale;
76	>	stringToEnumMap_["KineticScaleVAM"] = rnemdKineticScaleVAM;
77	>	stringToEnumMap_["KineticScaleAM"] = rnemdKineticScaleAM;
78	>	stringToEnumMap_["PxScale"] = rnemdPxScale;
79	>	stringToEnumMap_["PyScale"] = rnemdPyScale;
80	>	stringToEnumMap_["PzScale"] = rnemdPzScale;
81		stringToEnumMap_["Px"] = rnemdPx;
82		stringToEnumMap_["Py"] = rnemdPy;
83		stringToEnumMap_["Pz"] = rnemdPz;
84	+	stringToEnumMap_["ShiftScaleV"] = rnemdShiftScaleV;
85	+	stringToEnumMap_["ShiftScaleVAM"] = rnemdShiftScaleVAM;
86		stringToEnumMap_["Unknown"] = rnemdUnknown;
87
88	<	const std::string st = simParams->getRNEMD_swapType();
88	>	runTime_ = simParams->getRunTime();
89	>	statusTime_ = simParams->getStatusTime();
90
91	<	std::map<std::string, RNEMDTypeEnum>::iterator i;
91	>	rnemdObjectSelection_ = simParams->getRNEMD_objectSelection();
92	>	evaluator_.loadScriptString(rnemdObjectSelection_);
93	>	seleMan_.setSelectionSet(evaluator_.evaluate());
94	>
95	>	// do some sanity checking
96	>
97	>	int selectionCount = seleMan_.getSelectionCount();
98	>	int nIntegrable = info->getNGlobalIntegrableObjects();
99	>
100	>	if (selectionCount > nIntegrable) {
101	>	sprintf(painCave.errMsg,
102	>	"RNEMD: The current RNEMD_objectSelection,\n"
103	>	"\t\t%s\n"
104	>	"\thas resulted in %d selected objects.  However,\n"
105	>	"\tthe total number of integrable objects in the system\n"
106	>	"\tis only %d.  This is almost certainly not what you want\n"
107	>	"\tto do.  A likely cause of this is forgetting the _RB_0\n"
108	>	"\tselector in the selection script!\n",
109	>	rnemdObjectSelection_.c_str(),
110	>	selectionCount, nIntegrable);
111	>	painCave.isFatal = 0;
112	>	painCave.severity = OPENMD_WARNING;
113	>	simError();
114	>	}
115	>
116	>	const string st = simParams->getRNEMD_exchangeType();
117	>
118	>	map<string, RNEMDTypeEnum>::iterator i;
119		i = stringToEnumMap_.find(st);
120	<	rnemdType_  = (i == stringToEnumMap_.end()) ? RNEMD::rnemdUnknown : i->second;
120	>	rnemdType_ = (i == stringToEnumMap_.end()) ? RNEMD::rnemdUnknown : i->second;
121	>	if (rnemdType_ == rnemdUnknown) {
122	>	sprintf(painCave.errMsg,
123	>	"RNEMD: The current RNEMD_exchangeType,\n"
124	>	"\t\t%s\n"
125	>	"\tis not one of the recognized exchange types.\n",
126	>	st.c_str());
127	>	painCave.isFatal = 1;
128	>	painCave.severity = OPENMD_ERROR;
129	>	simError();
130	>	}
131	>
132	>	outputTemp_ = false;
133	>	if (simParams->haveRNEMD_outputTemperature()) {
134	>	outputTemp_ = simParams->getRNEMD_outputTemperature();
135	>	} else if ((rnemdType_ == rnemdKineticSwap) ||
136	>	(rnemdType_ == rnemdKineticScale) ||
137	>	(rnemdType_ == rnemdKineticScaleVAM) ||
138	>	(rnemdType_ == rnemdKineticScaleAM)) {
139	>	outputTemp_ = true;
140	>	}
141	>	outputVx_ = false;
142	>	if (simParams->haveRNEMD_outputVx()) {
143	>	outputVx_ = simParams->getRNEMD_outputVx();
144	>	} else if ((rnemdType_ == rnemdPx) || (rnemdType_ == rnemdPxScale)) {
145	>	outputVx_ = true;
146	>	}
147	>	outputVy_ = false;
148	>	if (simParams->haveRNEMD_outputVy()) {
149	>	outputVy_ = simParams->getRNEMD_outputVy();
150	>	} else if ((rnemdType_ == rnemdPy) || (rnemdType_ == rnemdPyScale)) {
151	>	outputVy_ = true;
152	>	}
153	>	output3DTemp_ = false;
154	>	if (simParams->haveRNEMD_outputXyzTemperature()) {
155	>	output3DTemp_ = simParams->getRNEMD_outputXyzTemperature();
156	>	}
157	>	outputRotTemp_ = false;
158	>	if (simParams->haveRNEMD_outputRotTemperature()) {
159	>	outputRotTemp_ = simParams->getRNEMD_outputRotTemperature();
160	>	}
161	>	// James put this in.
162	>	outputDen_ = false;
163	>	if (simParams->haveRNEMD_outputDen()) {
164	>	outputDen_ = simParams->getRNEMD_outputDen();
165	>	}
166	>	outputAh_ = false;
167	>	if (simParams->haveRNEMD_outputAh()) {
168	>	outputAh_ = simParams->getRNEMD_outputAh();
169	>	}
170	>	outputVz_ = false;
171	>	if (simParams->haveRNEMD_outputVz()) {
172	>	outputVz_ = simParams->getRNEMD_outputVz();
173	>	} else if ((rnemdType_ == rnemdPz) || (rnemdType_ == rnemdPzScale)) {
174	>	outputVz_ = true;
175	>	}
176	>
177
178	+	#ifdef IS_MPI
179	+	if (worldRank == 0) {
180	+	#endif
181
182	<	set_RNEMD_swapTime(simParams->getRNEMD_swapTime());
182	>	//may have rnemdWriter separately
183	>	string rnemdFileName;
184	>
185	>	if (outputTemp_) {
186	>	rnemdFileName = "temperature.log";
187	>	tempLog_.open(rnemdFileName.c_str());
188	>	}
189	>	if (outputVx_) {
190	>	rnemdFileName = "velocityX.log";
191	>	vxzLog_.open(rnemdFileName.c_str());
192	>	}
193	>	if (outputVy_) {
194	>	rnemdFileName = "velocityY.log";
195	>	vyzLog_.open(rnemdFileName.c_str());
196	>	}
197	>
198	>	if (output3DTemp_) {
199	>	rnemdFileName = "temperatureX.log";
200	>	xTempLog_.open(rnemdFileName.c_str());
201	>	rnemdFileName = "temperatureY.log";
202	>	yTempLog_.open(rnemdFileName.c_str());
203	>	rnemdFileName = "temperatureZ.log";
204	>	zTempLog_.open(rnemdFileName.c_str());
205	>	}
206	>	if (outputRotTemp_) {
207	>	rnemdFileName = "temperatureR.log";
208	>	rotTempLog_.open(rnemdFileName.c_str());
209	>	}
210	>
211	>	//James put this in
212	>	if (outputDen_) {
213	>	rnemdFileName = "Density.log";
214	>	denLog_.open(rnemdFileName.c_str());
215	>	}
216	>	if (outputAh_) {
217	>	rnemdFileName = "Ah.log";
218	>	AhLog_.open(rnemdFileName.c_str());
219	>	}
220	>	if (outputVz_) {
221	>	rnemdFileName = "velocityZ.log";
222	>	vzzLog_.open(rnemdFileName.c_str());
223	>	}
224	>	logFrameCount_ = 0;
225	>	#ifdef IS_MPI
226	>	}
227	>	#endif
228	>
229	>	set_RNEMD_exchange_time(simParams->getRNEMD_exchangeTime());
230		set_RNEMD_nBins(simParams->getRNEMD_nBins());
231	<	exchangeSum_ = 0.0;
232	<
231	>	midBin_ = nBins_ / 2;
232	>	if (simParams->haveRNEMD_binShift()) {
233	>	if (simParams->getRNEMD_binShift()) {
234	>	zShift_ = 0.5 / (RealType)(nBins_);
235	>	} else {
236	>	zShift_ = 0.0;
237	>	}
238	>	} else {
239	>	zShift_ = 0.0;
240	>	}
241	>	//cerr << "I shift slabs by " << zShift_ << " Lz\n";
242	>	//shift slabs by half slab width, maybe useful in heterogeneous systems
243	>	//set to 0.0 if not using it; N/A in status output yet
244	>	if (simParams->haveRNEMD_logWidth()) {
245	>	set_RNEMD_logWidth(simParams->getRNEMD_logWidth());
246	>	/*arbitary rnemdLogWidth_, no checking;
247	>	if (rnemdLogWidth_ != nBins_ && rnemdLogWidth_ != midBin_ + 1) {
248	>	cerr << "WARNING! RNEMD_logWidth has abnormal value!\n";
249	>	cerr << "Automaically set back to default.\n";
250	>	rnemdLogWidth_ = nBins_;
251	>	}*/
252	>	} else {
253	>	set_RNEMD_logWidth(nBins_);
254	>	}
255	>	tempHist_.resize(rnemdLogWidth_, 0.0);
256	>	tempCount_.resize(rnemdLogWidth_, 0);
257	>	pxzHist_.resize(rnemdLogWidth_, 0.0);
258	>	//vxzCount_.resize(rnemdLogWidth_, 0);
259	>	pyzHist_.resize(rnemdLogWidth_, 0.0);
260	>	//vyzCount_.resize(rnemdLogWidth_, 0);
261	>
262	>	mHist_.resize(rnemdLogWidth_, 0.0);
263	>	xTempHist_.resize(rnemdLogWidth_, 0.0);
264	>	yTempHist_.resize(rnemdLogWidth_, 0.0);
265	>	zTempHist_.resize(rnemdLogWidth_, 0.0);
266	>	xyzTempCount_.resize(rnemdLogWidth_, 0);
267	>	rotTempHist_.resize(rnemdLogWidth_, 0.0);
268	>	rotTempCount_.resize(rnemdLogWidth_, 0);
269	>	// James put this in
270	>	DenHist_.resize(rnemdLogWidth_, 0.0);
271	>	pzzHist_.resize(rnemdLogWidth_, 0.0);
272	>
273	>	set_RNEMD_exchange_total(0.0);
274	>	if (simParams->haveRNEMD_targetFlux()) {
275	>	set_RNEMD_target_flux(simParams->getRNEMD_targetFlux());
276	>	} else {
277	>	set_RNEMD_target_flux(0.0);
278	>	}
279	>	if (simParams->haveRNEMD_targetJzKE()) {
280	>	set_RNEMD_target_JzKE(simParams->getRNEMD_targetJzKE());
281	>	} else {
282	>	set_RNEMD_target_JzKE(0.0);
283	>	}
284	>	if (simParams->haveRNEMD_targetJzpx()) {
285	>	set_RNEMD_target_jzpx(simParams->getRNEMD_targetJzpx());
286	>	} else {
287	>	set_RNEMD_target_jzpx(0.0);
288	>	}
289	>	jzp_.x() = targetJzpx_;
290	>	njzp_.x() = -targetJzpx_;
291	>	if (simParams->haveRNEMD_targetJzpy()) {
292	>	set_RNEMD_target_jzpy(simParams->getRNEMD_targetJzpy());
293	>	} else {
294	>	set_RNEMD_target_jzpy(0.0);
295	>	}
296	>	jzp_.y() = targetJzpy_;
297	>	njzp_.y() = -targetJzpy_;
298	>	if (simParams->haveRNEMD_targetJzpz()) {
299	>	set_RNEMD_target_jzpz(simParams->getRNEMD_targetJzpz());
300	>	} else {
301	>	set_RNEMD_target_jzpz(0.0);
302	>	}
303	>	jzp_.z() = targetJzpz_;
304	>	njzp_.z() = -targetJzpz_;
305	>
306		#ifndef IS_MPI
307		if (simParams->haveSeed()) {
308		seedValue = simParams->getSeed();
#	Line 100 | Line 322 | namespace oopse {
322
323		RNEMD::~RNEMD() {
324		delete randNumGen_;
325	+
326	+	#ifdef IS_MPI
327	+	if (worldRank == 0) {
328	+	#endif
329	+
330	+	sprintf(painCave.errMsg,
331	+	"RNEMD: total failed trials: %d\n",
332	+	failTrialCount_);
333	+	painCave.isFatal = 0;
334	+	painCave.severity = OPENMD_INFO;
335	+	simError();
336	+
337	+	if (outputTemp_) tempLog_.close();
338	+	if (outputVx_)   vxzLog_.close();
339	+	if (outputVy_)   vyzLog_.close();
340	+
341	+	if (rnemdType_ == rnemdKineticScale || rnemdType_ == rnemdPxScale ||
342	+	rnemdType_ == rnemdPyScale) {
343	+	sprintf(painCave.errMsg,
344	+	"RNEMD: total root-checking warnings: %d\n",
345	+	failRootCount_);
346	+	painCave.isFatal = 0;
347	+	painCave.severity = OPENMD_INFO;
348	+	simError();
349	+	}
350	+	if (output3DTemp_) {
351	+	xTempLog_.close();
352	+	yTempLog_.close();
353	+	zTempLog_.close();
354	+	}
355	+	if (outputRotTemp_) rotTempLog_.close();
356	+	// James put this in
357	+	if (outputDen_) denLog_.close();
358	+	if (outputAh_)  AhLog_.close();
359	+	if (outputVz_)  vzzLog_.close();
360	+
361	+	#ifdef IS_MPI
362	+	}
363	+	#endif
364		}
365
366		void RNEMD::doSwap() {
367	<	std::cerr << "in RNEMD!\n";
368	<	std::cerr << "nBins = " << nBins_ << "\n";
369	<	std::cerr << "swapTime = " << swapTime_ << "\n";
370	<	std::cerr << "exchangeSum = " << exchangeSum_ << "\n";
371	<	std::cerr << "swapType = " << rnemdType_ << "\n";
372	<	}
367	>
368	>	Snapshot* currentSnap_ = info_->getSnapshotManager()->getCurrentSnapshot();
369	>	Mat3x3d hmat = currentSnap_->getHmat();
370	>
371	>	seleMan_.setSelectionSet(evaluator_.evaluate());
372	>
373	>	int selei;
374	>	StuntDouble* sd;
375	>	int idx;
376	>
377	>	RealType min_val;
378	>	bool min_found = false;
379	>	StuntDouble* min_sd;
380	>
381	>	RealType max_val;
382	>	bool max_found = false;
383	>	StuntDouble* max_sd;
384	>
385	>	for (sd = seleMan_.beginSelected(selei); sd != NULL;
386	>	sd = seleMan_.nextSelected(selei)) {
387	>
388	>	idx = sd->getLocalIndex();
389	>
390	>	Vector3d pos = sd->getPos();
391	>
392	>	// wrap the stuntdouble's position back into the box:
393	>
394	>	if (usePeriodicBoundaryConditions_)
395	>	currentSnap_->wrapVector(pos);
396	>
397	>	// which bin is this stuntdouble in?
398	>	// wrapped positions are in the range [-0.5*hmat(2,2), +0.5*hmat(2,2)]
399	>
400	>	int binNo = int(nBins_ * (pos.z() / hmat(2,2) + zShift_ + 0.5)) % nBins_;
401	>
402	>
403	>	// if we're in bin 0 or the middleBin
404	>	if (binNo == 0 || binNo == midBin_) {
405	>
406	>	RealType mass = sd->getMass();
407	>	Vector3d vel = sd->getVel();
408	>	RealType value;
409	>
410	>	switch(rnemdType_) {
411	>	case rnemdKineticSwap :
412	>
413	>	value = mass * vel.lengthSquare();
414	>
415	>	if (sd->isDirectional()) {
416	>	Vector3d angMom = sd->getJ();
417	>	Mat3x3d I = sd->getI();
418	>
419	>	if (sd->isLinear()) {
420	>	int i = sd->linearAxis();
421	>	int j = (i + 1) % 3;
422	>	int k = (i + 2) % 3;
423	>	value += angMom[j] * angMom[j] / I(j, j) +
424	>	angMom[k] * angMom[k] / I(k, k);
425	>	} else {
426	>	value += angMom[0]*angMom[0]/I(0, 0)
427	>	+ angMom[1]*angMom[1]/I(1, 1)
428	>	+ angMom[2]*angMom[2]/I(2, 2);
429	>	}
430	>	} //angular momenta exchange enabled
431	>	//energyConvert temporarily disabled
432	>	//make exchangeSum_ comparable between swap & scale
433	>	//value = value * 0.5 / PhysicalConstants::energyConvert;
434	>	value *= 0.5;
435	>	break;
436	>	case rnemdPx :
437	>	value = mass * vel[0];
438	>	break;
439	>	case rnemdPy :
440	>	value = mass * vel[1];
441	>	break;
442	>	case rnemdPz :
443	>	value = mass * vel[2];
444	>	break;
445	>	default :
446	>	break;
447	>	}
448	>
449	>	if (binNo == 0) {
450	>	if (!min_found) {
451	>	min_val = value;
452	>	min_sd = sd;
453	>	min_found = true;
454	>	} else {
455	>	if (min_val > value) {
456	>	min_val = value;
457	>	min_sd = sd;
458	>	}
459	>	}
460	>	} else { //midBin_
461	>	if (!max_found) {
462	>	max_val = value;
463	>	max_sd = sd;
464	>	max_found = true;
465	>	} else {
466	>	if (max_val < value) {
467	>	max_val = value;
468	>	max_sd = sd;
469	>	}
470	>	}
471	>	}
472	>	}
473	>	}
474	>
475	>	#ifdef IS_MPI
476	>	int nProc, worldRank;
477	>
478	>	nProc = MPI::COMM_WORLD.Get_size();
479	>	worldRank = MPI::COMM_WORLD.Get_rank();
480	>
481	>	bool my_min_found = min_found;
482	>	bool my_max_found = max_found;
483	>
484	>	// Even if we didn't find a minimum, did someone else?
485	>	MPI::COMM_WORLD.Allreduce(&my_min_found, &min_found, 1, MPI::BOOL, MPI::LOR);
486	>	// Even if we didn't find a maximum, did someone else?
487	>	MPI::COMM_WORLD.Allreduce(&my_max_found, &max_found, 1, MPI::BOOL, MPI::LOR);
488	>	#endif
489	>
490	>	if (max_found && min_found) {
491	>
492	>	#ifdef IS_MPI
493	>	struct {
494	>	RealType val;
495	>	int rank;
496	>	} max_vals, min_vals;
497	>
498	>	if (my_min_found) {
499	>	min_vals.val = min_val;
500	>	} else {
501	>	min_vals.val = HONKING_LARGE_VALUE;
502	>	}
503	>	min_vals.rank = worldRank;
504	>
505	>	// Who had the minimum?
506	>	MPI::COMM_WORLD.Allreduce(&min_vals, &min_vals,
507	>	1, MPI::REALTYPE_INT, MPI::MINLOC);
508	>	min_val = min_vals.val;
509	>
510	>	if (my_max_found) {
511	>	max_vals.val = max_val;
512	>	} else {
513	>	max_vals.val = -HONKING_LARGE_VALUE;
514	>	}
515	>	max_vals.rank = worldRank;
516	>
517	>	// Who had the maximum?
518	>	MPI::COMM_WORLD.Allreduce(&max_vals, &max_vals,
519	>	1, MPI::REALTYPE_INT, MPI::MAXLOC);
520	>	max_val = max_vals.val;
521	>	#endif
522	>
523	>	if (min_val < max_val) {
524	>
525	>	#ifdef IS_MPI
526	>	if (max_vals.rank == worldRank && min_vals.rank == worldRank) {
527	>	// I have both maximum and minimum, so proceed like a single
528	>	// processor version:
529	>	#endif
530	>
531	>	Vector3d min_vel = min_sd->getVel();
532	>	Vector3d max_vel = max_sd->getVel();
533	>	RealType temp_vel;
534	>
535	>	switch(rnemdType_) {
536	>	case rnemdKineticSwap :
537	>	min_sd->setVel(max_vel);
538	>	max_sd->setVel(min_vel);
539	>	if (min_sd->isDirectional() && max_sd->isDirectional()) {
540	>	Vector3d min_angMom = min_sd->getJ();
541	>	Vector3d max_angMom = max_sd->getJ();
542	>	min_sd->setJ(max_angMom);
543	>	max_sd->setJ(min_angMom);
544	>	}//angular momenta exchange enabled
545	>	//assumes same rigid body identity
546	>	break;
547	>	case rnemdPx :
548	>	temp_vel = min_vel.x();
549	>	min_vel.x() = max_vel.x();
550	>	max_vel.x() = temp_vel;
551	>	min_sd->setVel(min_vel);
552	>	max_sd->setVel(max_vel);
553	>	break;
554	>	case rnemdPy :
555	>	temp_vel = min_vel.y();
556	>	min_vel.y() = max_vel.y();
557	>	max_vel.y() = temp_vel;
558	>	min_sd->setVel(min_vel);
559	>	max_sd->setVel(max_vel);
560	>	break;
561	>	case rnemdPz :
562	>	temp_vel = min_vel.z();
563	>	min_vel.z() = max_vel.z();
564	>	max_vel.z() = temp_vel;
565	>	min_sd->setVel(min_vel);
566	>	max_sd->setVel(max_vel);
567	>	break;
568	>	default :
569	>	break;
570	>	}
571	>
572	>	#ifdef IS_MPI
573	>	// the rest of the cases only apply in parallel simulations:
574	>	} else if (max_vals.rank == worldRank) {
575	>	// I had the max, but not the minimum
576	>
577	>	Vector3d min_vel;
578	>	Vector3d max_vel = max_sd->getVel();
579	>	MPI::Status status;
580	>
581	>	// point-to-point swap of the velocity vector
582	>	MPI::COMM_WORLD.Sendrecv(max_vel.getArrayPointer(), 3, MPI::REALTYPE,
583	>	min_vals.rank, 0,
584	>	min_vel.getArrayPointer(), 3, MPI::REALTYPE,
585	>	min_vals.rank, 0, status);
586	>
587	>	switch(rnemdType_) {
588	>	case rnemdKineticSwap :
589	>	max_sd->setVel(min_vel);
590	>	//angular momenta exchange enabled
591	>	if (max_sd->isDirectional()) {
592	>	Vector3d min_angMom;
593	>	Vector3d max_angMom = max_sd->getJ();
594	>
595	>	// point-to-point swap of the angular momentum vector
596	>	MPI::COMM_WORLD.Sendrecv(max_angMom.getArrayPointer(), 3,
597	>	MPI::REALTYPE, min_vals.rank, 1,
598	>	min_angMom.getArrayPointer(), 3,
599	>	MPI::REALTYPE, min_vals.rank, 1,
600	>	status);
601	>
602	>	max_sd->setJ(min_angMom);
603	>	}
604	>	break;
605	>	case rnemdPx :
606	>	max_vel.x() = min_vel.x();
607	>	max_sd->setVel(max_vel);
608	>	break;
609	>	case rnemdPy :
610	>	max_vel.y() = min_vel.y();
611	>	max_sd->setVel(max_vel);
612	>	break;
613	>	case rnemdPz :
614	>	max_vel.z() = min_vel.z();
615	>	max_sd->setVel(max_vel);
616	>	break;
617	>	default :
618	>	break;
619	>	}
620	>	} else if (min_vals.rank == worldRank) {
621	>	// I had the minimum but not the maximum:
622	>
623	>	Vector3d max_vel;
624	>	Vector3d min_vel = min_sd->getVel();
625	>	MPI::Status status;
626	>
627	>	// point-to-point swap of the velocity vector
628	>	MPI::COMM_WORLD.Sendrecv(min_vel.getArrayPointer(), 3, MPI::REALTYPE,
629	>	max_vals.rank, 0,
630	>	max_vel.getArrayPointer(), 3, MPI::REALTYPE,
631	>	max_vals.rank, 0, status);
632	>
633	>	switch(rnemdType_) {
634	>	case rnemdKineticSwap :
635	>	min_sd->setVel(max_vel);
636	>	//angular momenta exchange enabled
637	>	if (min_sd->isDirectional()) {
638	>	Vector3d min_angMom = min_sd->getJ();
639	>	Vector3d max_angMom;
640	>
641	>	// point-to-point swap of the angular momentum vector
642	>	MPI::COMM_WORLD.Sendrecv(min_angMom.getArrayPointer(), 3,
643	>	MPI::REALTYPE, max_vals.rank, 1,
644	>	max_angMom.getArrayPointer(), 3,
645	>	MPI::REALTYPE, max_vals.rank, 1,
646	>	status);
647	>
648	>	min_sd->setJ(max_angMom);
649	>	}
650	>	break;
651	>	case rnemdPx :
652	>	min_vel.x() = max_vel.x();
653	>	min_sd->setVel(min_vel);
654	>	break;
655	>	case rnemdPy :
656	>	min_vel.y() = max_vel.y();
657	>	min_sd->setVel(min_vel);
658	>	break;
659	>	case rnemdPz :
660	>	min_vel.z() = max_vel.z();
661	>	min_sd->setVel(min_vel);
662	>	break;
663	>	default :
664	>	break;
665	>	}
666	>	}
667	>	#endif
668	>	exchangeSum_ += max_val - min_val;
669	>	} else {
670	>	sprintf(painCave.errMsg,
671	>	"RNEMD: exchange NOT performed because min_val > max_val\n");
672	>	painCave.isFatal = 0;
673	>	painCave.severity = OPENMD_INFO;
674	>	simError();
675	>	failTrialCount_++;
676	>	}
677	>	} else {
678	>	sprintf(painCave.errMsg,
679	>	"RNEMD: exchange NOT performed because selected object\n"
680	>	"\tnot present in at least one of the two slabs.\n");
681	>	painCave.isFatal = 0;
682	>	painCave.severity = OPENMD_INFO;
683	>	simError();
684	>	failTrialCount_++;
685	>	}
686	>
687	>	}
688	>
689	>	void RNEMD::doScale() {
690	>
691	>	Snapshot* currentSnap_ = info_->getSnapshotManager()->getCurrentSnapshot();
692	>	Mat3x3d hmat = currentSnap_->getHmat();
693	>
694	>	seleMan_.setSelectionSet(evaluator_.evaluate());
695	>
696	>	int selei;
697	>	StuntDouble* sd;
698	>	int idx;
699	>
700	>	vector<StuntDouble*> hotBin, coldBin;
701	>
702	>	RealType Phx = 0.0;
703	>	RealType Phy = 0.0;
704	>	RealType Phz = 0.0;
705	>	RealType Khx = 0.0;
706	>	RealType Khy = 0.0;
707	>	RealType Khz = 0.0;
708	>	RealType Khw = 0.0;
709	>	RealType Pcx = 0.0;
710	>	RealType Pcy = 0.0;
711	>	RealType Pcz = 0.0;
712	>	RealType Kcx = 0.0;
713	>	RealType Kcy = 0.0;
714	>	RealType Kcz = 0.0;
715	>	RealType Kcw = 0.0;
716	>
717	>	for (sd = seleMan_.beginSelected(selei); sd != NULL;
718	>	sd = seleMan_.nextSelected(selei)) {
719	>
720	>	idx = sd->getLocalIndex();
721	>
722	>	Vector3d pos = sd->getPos();
723	>
724	>	// wrap the stuntdouble's position back into the box:
725	>
726	>	if (usePeriodicBoundaryConditions_)
727	>	currentSnap_->wrapVector(pos);
728	>
729	>	// which bin is this stuntdouble in?
730	>	// wrapped positions are in the range [-0.5*hmat(2,2), +0.5*hmat(2,2)]
731	>
732	>	int binNo = int(nBins_ * (pos.z() / hmat(2,2) + zShift_ + 0.5)) % nBins_;
733	>
734	>	// if we're in bin 0 or the middleBin
735	>	if (binNo == 0 || binNo == midBin_) {
736	>
737	>	RealType mass = sd->getMass();
738	>	Vector3d vel = sd->getVel();
739	>
740	>	if (binNo == 0) {
741	>	hotBin.push_back(sd);
742	>	Phx += mass * vel.x();
743	>	Phy += mass * vel.y();
744	>	Phz += mass * vel.z();
745	>	Khx += mass * vel.x() * vel.x();
746	>	Khy += mass * vel.y() * vel.y();
747	>	Khz += mass * vel.z() * vel.z();
748	>	//if (rnemdType_ == rnemdKineticScaleVAM) {
749	>	if (sd->isDirectional()) {
750	>	Vector3d angMom = sd->getJ();
751	>	Mat3x3d I = sd->getI();
752	>	if (sd->isLinear()) {
753	>	int i = sd->linearAxis();
754	>	int j = (i + 1) % 3;
755	>	int k = (i + 2) % 3;
756	>	Khw += angMom[j] * angMom[j] / I(j, j) +
757	>	angMom[k] * angMom[k] / I(k, k);
758	>	} else {
759	>	Khw += angMom[0]*angMom[0]/I(0, 0)
760	>	+ angMom[1]*angMom[1]/I(1, 1)
761	>	+ angMom[2]*angMom[2]/I(2, 2);
762	>	}
763	>	}
764	>	//}
765	>	} else { //midBin_
766	>	coldBin.push_back(sd);
767	>	Pcx += mass * vel.x();
768	>	Pcy += mass * vel.y();
769	>	Pcz += mass * vel.z();
770	>	Kcx += mass * vel.x() * vel.x();
771	>	Kcy += mass * vel.y() * vel.y();
772	>	Kcz += mass * vel.z() * vel.z();
773	>	//if (rnemdType_ == rnemdKineticScaleVAM) {
774	>	if (sd->isDirectional()) {
775	>	Vector3d angMom = sd->getJ();
776	>	Mat3x3d I = sd->getI();
777	>	if (sd->isLinear()) {
778	>	int i = sd->linearAxis();
779	>	int j = (i + 1) % 3;
780	>	int k = (i + 2) % 3;
781	>	Kcw += angMom[j] * angMom[j] / I(j, j) +
782	>	angMom[k] * angMom[k] / I(k, k);
783	>	} else {
784	>	Kcw += angMom[0]*angMom[0]/I(0, 0)
785	>	+ angMom[1]*angMom[1]/I(1, 1)
786	>	+ angMom[2]*angMom[2]/I(2, 2);
787	>	}
788	>	}
789	>	//}
790	>	}
791	>	}
792	>	}
793	>
794	>	Khx *= 0.5;
795	>	Khy *= 0.5;
796	>	Khz *= 0.5;
797	>	Khw *= 0.5;
798	>	Kcx *= 0.5;
799	>	Kcy *= 0.5;
800	>	Kcz *= 0.5;
801	>	Kcw *= 0.5;
802	>
803	>	// std::cerr << "Khx= " << Khx << "\tKhy= " << Khy << "\tKhz= " << Khz
804	>	//        << "\tKhw= " << Khw << "\tKcx= " << Kcx << "\tKcy= " << Kcy
805	>	//        << "\tKcz= " << Kcz << "\tKcw= " << Kcw << "\n";
806	>	// std::cerr << "Phx= " << Phx << "\tPhy= " << Phy << "\tPhz= " << Phz
807	>	//        << "\tPcx= " << Pcx << "\tPcy= " << Pcy << "\tPcz= " <<Pcz<<"\n";
808	>
809	>	#ifdef IS_MPI
810	>	MPI::COMM_WORLD.Allreduce(MPI::IN_PLACE, &Phx, 1, MPI::REALTYPE, MPI::SUM);
811	>	MPI::COMM_WORLD.Allreduce(MPI::IN_PLACE, &Phy, 1, MPI::REALTYPE, MPI::SUM);
812	>	MPI::COMM_WORLD.Allreduce(MPI::IN_PLACE, &Phz, 1, MPI::REALTYPE, MPI::SUM);
813	>	MPI::COMM_WORLD.Allreduce(MPI::IN_PLACE, &Pcx, 1, MPI::REALTYPE, MPI::SUM);
814	>	MPI::COMM_WORLD.Allreduce(MPI::IN_PLACE, &Pcy, 1, MPI::REALTYPE, MPI::SUM);
815	>	MPI::COMM_WORLD.Allreduce(MPI::IN_PLACE, &Pcz, 1, MPI::REALTYPE, MPI::SUM);
816	>
817	>	MPI::COMM_WORLD.Allreduce(MPI::IN_PLACE, &Khx, 1, MPI::REALTYPE, MPI::SUM);
818	>	MPI::COMM_WORLD.Allreduce(MPI::IN_PLACE, &Khy, 1, MPI::REALTYPE, MPI::SUM);
819	>	MPI::COMM_WORLD.Allreduce(MPI::IN_PLACE, &Khz, 1, MPI::REALTYPE, MPI::SUM);
820	>	MPI::COMM_WORLD.Allreduce(MPI::IN_PLACE, &Khw, 1, MPI::REALTYPE, MPI::SUM);
821	>
822	>	MPI::COMM_WORLD.Allreduce(MPI::IN_PLACE, &Kcx, 1, MPI::REALTYPE, MPI::SUM);
823	>	MPI::COMM_WORLD.Allreduce(MPI::IN_PLACE, &Kcy, 1, MPI::REALTYPE, MPI::SUM);
824	>	MPI::COMM_WORLD.Allreduce(MPI::IN_PLACE, &Kcz, 1, MPI::REALTYPE, MPI::SUM);
825	>	MPI::COMM_WORLD.Allreduce(MPI::IN_PLACE, &Kcw, 1, MPI::REALTYPE, MPI::SUM);
826	>	#endif
827	>
828	>	//solve coldBin coeff's first
829	>	RealType px = Pcx / Phx;
830	>	RealType py = Pcy / Phy;
831	>	RealType pz = Pcz / Phz;
832	>	RealType c, x, y, z;
833	>	bool successfulScale = false;
834	>	if ((rnemdType_ == rnemdKineticScaleVAM) ||
835	>	(rnemdType_ == rnemdKineticScaleAM)) {
836	>	//may need sanity check Khw & Kcw > 0
837	>
838	>	if (rnemdType_ == rnemdKineticScaleVAM) {
839	>	c = 1.0 - targetFlux_ / (Kcx + Kcy + Kcz + Kcw);
840	>	} else {
841	>	c = 1.0 - targetFlux_ / Kcw;
842	>	}
843	>
844	>	if ((c > 0.81) && (c < 1.21)) {//restrict scaling coefficients
845	>	c = sqrt(c);
846	>	std::cerr << "cold slab scaling coefficient: " << c << endl;
847	>	//now convert to hotBin coefficient
848	>	RealType w = 0.0;
849	>	if (rnemdType_ ==  rnemdKineticScaleVAM) {
850	>	x = 1.0 + px * (1.0 - c);
851	>	y = 1.0 + py * (1.0 - c);
852	>	z = 1.0 + pz * (1.0 - c);
853	>	/* more complicated way
854	>	w = 1.0 + (Kcw - Kcw * c * c - (c * c * (Kcx + Kcy + Kcz
855	>	+ Khx * px * px + Khy * py * py + Khz * pz * pz)
856	>	- 2.0 * c * (Khx * px * (1.0 + px) + Khy * py * (1.0 + py)
857	>	+ Khz * pz * (1.0 + pz)) + Khx * px * (2.0 + px)
858	>	+ Khy * py * (2.0 + py) + Khz * pz * (2.0 + pz)
859	>	- Kcx - Kcy - Kcz)) / Khw; the following is simpler
860	>	*/
861	>	if ((fabs(x - 1.0) < 0.1) && (fabs(y - 1.0) < 0.1) &&
862	>	(fabs(z - 1.0) < 0.1)) {
863	>	w = 1.0 + (targetFlux_ + Khx * (1.0 - x * x) + Khy * (1.0 - y * y)
864	>	+ Khz * (1.0 - z * z)) / Khw;
865	>	}//no need to calculate w if x, y or z is out of range
866	>	} else {
867	>	w = 1.0 + targetFlux_ / Khw;
868	>	}
869	>	if ((w > 0.81) && (w < 1.21)) {//restrict scaling coefficients
870	>	//if w is in the right range, so should be x, y, z.
871	>	vector<StuntDouble*>::iterator sdi;
872	>	Vector3d vel;
873	>	for (sdi = coldBin.begin(); sdi != coldBin.end(); sdi++) {
874	>	if (rnemdType_ == rnemdKineticScaleVAM) {
875	>	vel = (*sdi)->getVel() * c;
876	>	//vel.x() *= c;
877	>	//vel.y() *= c;
878	>	//vel.z() *= c;
879	>	(*sdi)->setVel(vel);
880	>	}
881	>	if ((*sdi)->isDirectional()) {
882	>	Vector3d angMom = (*sdi)->getJ() * c;
883	>	//angMom[0] *= c;
884	>	//angMom[1] *= c;
885	>	//angMom[2] *= c;
886	>	(*sdi)->setJ(angMom);
887	>	}
888	>	}
889	>	w = sqrt(w);
890	>	std::cerr << "xh= " << x << "\tyh= " << y << "\tzh= " << z
891	>	<< "\twh= " << w << endl;
892	>	for (sdi = hotBin.begin(); sdi != hotBin.end(); sdi++) {
893	>	if (rnemdType_ == rnemdKineticScaleVAM) {
894	>	vel = (*sdi)->getVel();
895	>	vel.x() *= x;
896	>	vel.y() *= y;
897	>	vel.z() *= z;
898	>	(*sdi)->setVel(vel);
899	>	}
900	>	if ((*sdi)->isDirectional()) {
901	>	Vector3d angMom = (*sdi)->getJ() * w;
902	>	//angMom[0] *= w;
903	>	//angMom[1] *= w;
904	>	//angMom[2] *= w;
905	>	(*sdi)->setJ(angMom);
906	>	}
907	>	}
908	>	successfulScale = true;
909	>	exchangeSum_ += targetFlux_;
910	>	}
911	>	}
912	>	} else {
913	>	RealType a000, a110, c0, a001, a111, b01, b11, c1;
914	>	switch(rnemdType_) {
915	>	case rnemdKineticScale :
916	>	/* used hotBin coeff's & only scale x & y dimensions
917	>	RealType px = Phx / Pcx;
918	>	RealType py = Phy / Pcy;
919	>	a110 = Khy;
920	>	c0 = - Khx - Khy - targetFlux_;
921	>	a000 = Khx;
922	>	a111 = Kcy * py * py;
923	>	b11 = -2.0 * Kcy * py * (1.0 + py);
924	>	c1 = Kcy * py * (2.0 + py) + Kcx * px * ( 2.0 + px) + targetFlux_;
925	>	b01 = -2.0 * Kcx * px * (1.0 + px);
926	>	a001 = Kcx * px * px;
927	>	*/
928	>	//scale all three dimensions, let c_x = c_y
929	>	a000 = Kcx + Kcy;
930	>	a110 = Kcz;
931	>	c0 = targetFlux_ - Kcx - Kcy - Kcz;
932	>	a001 = Khx * px * px + Khy * py * py;
933	>	a111 = Khz * pz * pz;
934	>	b01 = -2.0 * (Khx * px * (1.0 + px) + Khy * py * (1.0 + py));
935	>	b11 = -2.0 * Khz * pz * (1.0 + pz);
936	>	c1 = Khx * px * (2.0 + px) + Khy * py * (2.0 + py)
937	>	+ Khz * pz * (2.0 + pz) - targetFlux_;
938	>	break;
939	>	case rnemdPxScale :
940	>	c = 1 - targetFlux_ / Pcx;
941	>	a000 = Kcy;
942	>	a110 = Kcz;
943	>	c0 = Kcx * c * c - Kcx - Kcy - Kcz;
944	>	a001 = py * py * Khy;
945	>	a111 = pz * pz * Khz;
946	>	b01 = -2.0 * Khy * py * (1.0 + py);
947	>	b11 = -2.0 * Khz * pz * (1.0 + pz);
948	>	c1 = Khy * py * (2.0 + py) + Khz * pz * (2.0 + pz)
949	>	+ Khx * (fastpow(c * px - px - 1.0, 2) - 1.0);
950	>	break;
951	>	case rnemdPyScale :
952	>	c = 1 - targetFlux_ / Pcy;
953	>	a000 = Kcx;
954	>	a110 = Kcz;
955	>	c0 = Kcy * c * c - Kcx - Kcy - Kcz;
956	>	a001 = px * px * Khx;
957	>	a111 = pz * pz * Khz;
958	>	b01 = -2.0 * Khx * px * (1.0 + px);
959	>	b11 = -2.0 * Khz * pz * (1.0 + pz);
960	>	c1 = Khx * px * (2.0 + px) + Khz * pz * (2.0 + pz)
961	>	+ Khy * (fastpow(c * py - py - 1.0, 2) - 1.0);
962	>	break;
963	>	case rnemdPzScale ://we don't really do this, do we?
964	>	c = 1 - targetFlux_ / Pcz;
965	>	a000 = Kcx;
966	>	a110 = Kcy;
967	>	c0 = Kcz * c * c - Kcx - Kcy - Kcz;
968	>	a001 = px * px * Khx;
969	>	a111 = py * py * Khy;
970	>	b01 = -2.0 * Khx * px * (1.0 + px);
971	>	b11 = -2.0 * Khy * py * (1.0 + py);
972	>	c1 = Khx * px * (2.0 + px) + Khy * py * (2.0 + py)
973	>	+ Khz * (fastpow(c * pz - pz - 1.0, 2) - 1.0);
974	>	break;
975	>	default :
976	>	break;
977	>	}
978	>
979	>	RealType v1 = a000 * a111 - a001 * a110;
980	>	RealType v2 = a000 * b01;
981	>	RealType v3 = a000 * b11;
982	>	RealType v4 = a000 * c1 - a001 * c0;
983	>	RealType v8 = a110 * b01;
984	>	RealType v10 = - b01 * c0;
985	>
986	>	RealType u0 = v2 * v10 - v4 * v4;
987	>	RealType u1 = -2.0 * v3 * v4;
988	>	RealType u2 = -v2 * v8 - v3 * v3 - 2.0 * v1 * v4;
989	>	RealType u3 = -2.0 * v1 * v3;
990	>	RealType u4 = - v1 * v1;
991	>	//rescale coefficients
992	>	RealType maxAbs = fabs(u0);
993	>	if (maxAbs < fabs(u1)) maxAbs = fabs(u1);
994	>	if (maxAbs < fabs(u2)) maxAbs = fabs(u2);
995	>	if (maxAbs < fabs(u3)) maxAbs = fabs(u3);
996	>	if (maxAbs < fabs(u4)) maxAbs = fabs(u4);
997	>	u0 /= maxAbs;
998	>	u1 /= maxAbs;
999	>	u2 /= maxAbs;
1000	>	u3 /= maxAbs;
1001	>	u4 /= maxAbs;
1002	>	//max_element(start, end) is also available.
1003	>	Polynomial<RealType> poly; //same as DoublePolynomial poly;
1004	>	poly.setCoefficient(4, u4);
1005	>	poly.setCoefficient(3, u3);
1006	>	poly.setCoefficient(2, u2);
1007	>	poly.setCoefficient(1, u1);
1008	>	poly.setCoefficient(0, u0);
1009	>	vector<RealType> realRoots = poly.FindRealRoots();
1010	>
1011	>	vector<RealType>::iterator ri;
1012	>	RealType r1, r2, alpha0;
1013	>	vector<pair<RealType,RealType> > rps;
1014	>	for (ri = realRoots.begin(); ri !=realRoots.end(); ri++) {
1015	>	r2 = *ri;
1016	>	//check if FindRealRoots() give the right answer
1017	>	if ( fabs(u0 + r2 * (u1 + r2 * (u2 + r2 * (u3 + r2 * u4)))) > 1e-6 ) {
1018	>	sprintf(painCave.errMsg,
1019	>	"RNEMD Warning: polynomial solve seems to have an error!");
1020	>	painCave.isFatal = 0;
1021	>	simError();
1022	>	failRootCount_++;
1023	>	}
1024	>	//might not be useful w/o rescaling coefficients
1025	>	alpha0 = -c0 - a110 * r2 * r2;
1026	>	if (alpha0 >= 0.0) {
1027	>	r1 = sqrt(alpha0 / a000);
1028	>	if (fabs(c1 + r1 * (b01 + r1 * a001) + r2 * (b11 + r2 * a111))
1029	>	< 1e-6)
1030	>	{ rps.push_back(make_pair(r1, r2)); }
1031	>	if (r1 > 1e-6) { //r1 non-negative
1032	>	r1 = -r1;
1033	>	if (fabs(c1 + r1 * (b01 + r1 * a001) + r2 * (b11 + r2 * a111))
1034	>	< 1e-6)
1035	>	{ rps.push_back(make_pair(r1, r2)); }
1036	>	}
1037	>	}
1038	>	}
1039	>	// Consider combining together the solving pair part w/ the searching
1040	>	// best solution part so that we don't need the pairs vector
1041	>	if (!rps.empty()) {
1042	>	RealType smallestDiff = HONKING_LARGE_VALUE;
1043	>	RealType diff;
1044	>	pair<RealType,RealType> bestPair = make_pair(1.0, 1.0);
1045	>	vector<pair<RealType,RealType> >::iterator rpi;
1046	>	for (rpi = rps.begin(); rpi != rps.end(); rpi++) {
1047	>	r1 = (*rpi).first;
1048	>	r2 = (*rpi).second;
1049	>	switch(rnemdType_) {
1050	>	case rnemdKineticScale :
1051	>	diff = fastpow(1.0 - r1, 2) + fastpow(1.0 - r2, 2)
1052	>	+ fastpow(r1 * r1 / r2 / r2 - Kcz/Kcx, 2)
1053	>	+ fastpow(r1 * r1 / r2 / r2 - Kcz/Kcy, 2);
1054	>	break;
1055	>	case rnemdPxScale :
1056	>	diff = fastpow(1.0 - r1, 2) + fastpow(1.0 - r2, 2)
1057	>	+ fastpow(r1 * r1 / r2 / r2 - Kcz/Kcy, 2);
1058	>	break;
1059	>	case rnemdPyScale :
1060	>	diff = fastpow(1.0 - r1, 2) + fastpow(1.0 - r2, 2)
1061	>	+ fastpow(r1 * r1 / r2 / r2 - Kcz/Kcx, 2);
1062	>	break;
1063	>	case rnemdPzScale :
1064	>	diff = fastpow(1.0 - r1, 2) + fastpow(1.0 - r2, 2)
1065	>	+ fastpow(r1 * r1 / r2 / r2 - Kcy/Kcx, 2);
1066	>	default :
1067	>	break;
1068	>	}
1069	>	if (diff < smallestDiff) {
1070	>	smallestDiff = diff;
1071	>	bestPair = *rpi;
1072	>	}
1073	>	}
1074	>	#ifdef IS_MPI
1075	>	if (worldRank == 0) {
1076	>	#endif
1077	>	sprintf(painCave.errMsg,
1078	>	"RNEMD: roots r1= %lf\tr2 = %lf\n",
1079	>	bestPair.first, bestPair.second);
1080	>	painCave.isFatal = 0;
1081	>	painCave.severity = OPENMD_INFO;
1082	>	simError();
1083	>	#ifdef IS_MPI
1084	>	}
1085	>	#endif
1086	>
1087	>	switch(rnemdType_) {
1088	>	case rnemdKineticScale :
1089	>	x = bestPair.first;
1090	>	y = bestPair.first;
1091	>	z = bestPair.second;
1092	>	break;
1093	>	case rnemdPxScale :
1094	>	x = c;
1095	>	y = bestPair.first;
1096	>	z = bestPair.second;
1097	>	break;
1098	>	case rnemdPyScale :
1099	>	x = bestPair.first;
1100	>	y = c;
1101	>	z = bestPair.second;
1102	>	break;
1103	>	case rnemdPzScale :
1104	>	x = bestPair.first;
1105	>	y = bestPair.second;
1106	>	z = c;
1107	>	break;
1108	>	default :
1109	>	break;
1110	>	}
1111	>	vector<StuntDouble*>::iterator sdi;
1112	>	Vector3d vel;
1113	>	for (sdi = coldBin.begin(); sdi != coldBin.end(); sdi++) {
1114	>	vel = (*sdi)->getVel();
1115	>	vel.x() *= x;
1116	>	vel.y() *= y;
1117	>	vel.z() *= z;
1118	>	(*sdi)->setVel(vel);
1119	>	}
1120	>	//convert to hotBin coefficient
1121	>	x = 1.0 + px * (1.0 - x);
1122	>	y = 1.0 + py * (1.0 - y);
1123	>	z = 1.0 + pz * (1.0 - z);
1124	>	for (sdi = hotBin.begin(); sdi != hotBin.end(); sdi++) {
1125	>	vel = (*sdi)->getVel();
1126	>	vel.x() *= x;
1127	>	vel.y() *= y;
1128	>	vel.z() *= z;
1129	>	(*sdi)->setVel(vel);
1130	>	}
1131	>	successfulScale = true;
1132	>	exchangeSum_ += targetFlux_;
1133	>	}
1134	>	}
1135	>	if (successfulScale != true) {
1136	>	sprintf(painCave.errMsg,
1137	>	"RNEMD: exchange NOT performed!\n");
1138	>	painCave.isFatal = 0;
1139	>	painCave.severity = OPENMD_INFO;
1140	>	simError();
1141	>	failTrialCount_++;
1142	>	}
1143	>	}
1144	>
1145	>	void RNEMD::doShiftScale() {
1146	>
1147	>	Snapshot* currentSnap_ = info_->getSnapshotManager()->getCurrentSnapshot();
1148	>	RealType time = currentSnap_->getTime();
1149	>	Mat3x3d hmat = currentSnap_->getHmat();
1150	>
1151	>	seleMan_.setSelectionSet(evaluator_.evaluate());
1152	>
1153	>	int selei;
1154	>	StuntDouble* sd;
1155	>	int idx;
1156	>
1157	>	vector<StuntDouble*> hotBin, coldBin;
1158	>
1159	>	Vector3d Ph(V3Zero);
1160	>	RealType Mh = 0.0;
1161	>	RealType Kh = 0.0;
1162	>	Vector3d Pc(V3Zero);
1163	>	RealType Mc = 0.0;
1164	>	RealType Kc = 0.0;
1165	>
1166	>
1167	>	for (sd = seleMan_.beginSelected(selei); sd != NULL;
1168	>	sd = seleMan_.nextSelected(selei)) {
1169	>
1170	>	idx = sd->getLocalIndex();
1171	>
1172	>	Vector3d pos = sd->getPos();
1173	>
1174	>	// wrap the stuntdouble's position back into the box:
1175	>
1176	>	if (usePeriodicBoundaryConditions_)
1177	>	currentSnap_->wrapVector(pos);
1178	>
1179	>	// which bin is this stuntdouble in?
1180	>	// wrapped positions are in the range [-0.5*hmat(2,2), +0.5*hmat(2,2)]
1181	>
1182	>	int binNo = int(nBins_ * (pos.z() / hmat(2,2) + zShift_ + 0.5)) % nBins_;
1183	>
1184	>	// if we're in bin 0 or the middleBin
1185	>	if (binNo == 0 || binNo == midBin_) {
1186	>
1187	>	RealType mass = sd->getMass();
1188	>	Vector3d vel = sd->getVel();
1189	>
1190	>	if (binNo == 0) {
1191	>	hotBin.push_back(sd);
1192	>	//std::cerr << "before, velocity = " << vel << endl;
1193	>	Ph += mass * vel;
1194	>	//std::cerr << "after, velocity = " << vel << endl;
1195	>	Mh += mass;
1196	>	Kh += mass * vel.lengthSquare();
1197	>	if (rnemdType_ == rnemdShiftScaleVAM) {
1198	>	if (sd->isDirectional()) {
1199	>	Vector3d angMom = sd->getJ();
1200	>	Mat3x3d I = sd->getI();
1201	>	if (sd->isLinear()) {
1202	>	int i = sd->linearAxis();
1203	>	int j = (i + 1) % 3;
1204	>	int k = (i + 2) % 3;
1205	>	Kh += angMom[j] * angMom[j] / I(j, j) +
1206	>	angMom[k] * angMom[k] / I(k, k);
1207	>	} else {
1208	>	Kh += angMom[0] * angMom[0] / I(0, 0) +
1209	>	angMom[1] * angMom[1] / I(1, 1) +
1210	>	angMom[2] * angMom[2] / I(2, 2);
1211	>	}
1212	>	}
1213	>	}
1214	>	} else { //midBin_
1215	>	coldBin.push_back(sd);
1216	>	Pc += mass * vel;
1217	>	Mc += mass;
1218	>	Kc += mass * vel.lengthSquare();
1219	>	if (rnemdType_ == rnemdShiftScaleVAM) {
1220	>	if (sd->isDirectional()) {
1221	>	Vector3d angMom = sd->getJ();
1222	>	Mat3x3d I = sd->getI();
1223	>	if (sd->isLinear()) {
1224	>	int i = sd->linearAxis();
1225	>	int j = (i + 1) % 3;
1226	>	int k = (i + 2) % 3;
1227	>	Kc += angMom[j] * angMom[j] / I(j, j) +
1228	>	angMom[k] * angMom[k] / I(k, k);
1229	>	} else {
1230	>	Kc += angMom[0] * angMom[0] / I(0, 0) +
1231	>	angMom[1] * angMom[1] / I(1, 1) +
1232	>	angMom[2] * angMom[2] / I(2, 2);
1233	>	}
1234	>	}
1235	>	}
1236	>	}
1237	>	}
1238	>	}
1239	>
1240	>	Kh *= 0.5;
1241	>	Kc *= 0.5;
1242	>
1243	>	// std::cerr << "Mh= " << Mh << "\tKh= " << Kh << "\tMc= " << Mc
1244	>	//        << "\tKc= " << Kc << endl;
1245	>	// std::cerr << "Ph= " << Ph << "\tPc= " << Pc << endl;
1246	>
1247	>	#ifdef IS_MPI
1248	>	MPI::COMM_WORLD.Allreduce(MPI::IN_PLACE, &Ph[0], 3, MPI::REALTYPE, MPI::SUM);
1249	>	MPI::COMM_WORLD.Allreduce(MPI::IN_PLACE, &Pc[0], 3, MPI::REALTYPE, MPI::SUM);
1250	>	MPI::COMM_WORLD.Allreduce(MPI::IN_PLACE, &Mh, 1, MPI::REALTYPE, MPI::SUM);
1251	>	MPI::COMM_WORLD.Allreduce(MPI::IN_PLACE, &Kh, 1, MPI::REALTYPE, MPI::SUM);
1252	>	MPI::COMM_WORLD.Allreduce(MPI::IN_PLACE, &Mc, 1, MPI::REALTYPE, MPI::SUM);
1253	>	MPI::COMM_WORLD.Allreduce(MPI::IN_PLACE, &Kc, 1, MPI::REALTYPE, MPI::SUM);
1254	>	#endif
1255	>
1256	>	bool successfulExchange = false;
1257	>	if ((Mh > 0.0) && (Mc > 0.0)) {//both slabs are not empty
1258	>	Vector3d vc = Pc / Mc;
1259	>	Vector3d ac = njzp_ / Mc + vc;
1260	>	Vector3d acrec = njzp_ / Mc;
1261	>	RealType cNumerator = Kc - targetJzKE_ - 0.5 * Mc * ac.lengthSquare();
1262	>	if (cNumerator > 0.0) {
1263	>	RealType cDenominator = Kc - 0.5 * Mc * vc.lengthSquare();
1264	>	if (cDenominator > 0.0) {
1265	>	RealType c = sqrt(cNumerator / cDenominator);
1266	>	if ((c > 0.9) && (c < 1.1)) {//restrict scaling coefficients
1267	>	Vector3d vh = Ph / Mh;
1268	>	Vector3d ah = jzp_ / Mh + vh;
1269	>	Vector3d ahrec = jzp_ / Mh;
1270	>	RealType hNumerator = Kh + targetJzKE_
1271	>	- 0.5 * Mh * ah.lengthSquare();
1272	>	if (hNumerator > 0.0) {
1273	>	RealType hDenominator = Kh - 0.5 * Mh * vh.lengthSquare();
1274	>	if (hDenominator > 0.0) {
1275	>	RealType h = sqrt(hNumerator / hDenominator);
1276	>	if ((h > 0.9) && (h < 1.1)) {
1277	>	// std::cerr << "cold slab scaling coefficient: " << c << "\n";
1278	>	// std::cerr << "hot slab scaling coefficient: " << h <<  "\n";
1279	>	vector<StuntDouble*>::iterator sdi;
1280	>	Vector3d vel;
1281	>	for (sdi = coldBin.begin(); sdi != coldBin.end(); sdi++) {
1282	>	//vel = (*sdi)->getVel();
1283	>	vel = ((*sdi)->getVel() - vc) * c + ac;
1284	>	(*sdi)->setVel(vel);
1285	>	if (rnemdType_ == rnemdShiftScaleVAM) {
1286	>	if ((*sdi)->isDirectional()) {
1287	>	Vector3d angMom = (*sdi)->getJ() * c;
1288	>	(*sdi)->setJ(angMom);
1289	>	}
1290	>	}
1291	>	}
1292	>	for (sdi = hotBin.begin(); sdi != hotBin.end(); sdi++) {
1293	>	//vel = (*sdi)->getVel();
1294	>	vel = ((*sdi)->getVel() - vh) * h + ah;
1295	>	(*sdi)->setVel(vel);
1296	>	if (rnemdType_ == rnemdShiftScaleVAM) {
1297	>	if ((*sdi)->isDirectional()) {
1298	>	Vector3d angMom = (*sdi)->getJ() * h;
1299	>	(*sdi)->setJ(angMom);
1300	>	}
1301	>	}
1302	>	}
1303	>	successfulExchange = true;
1304	>	exchangeSum_ += targetFlux_;
1305	>	// this is a redundant variable for doShiftScale() so that
1306	>	// RNEMD can output one exchange quantity needed in a job.
1307	>	// need a better way to do this.
1308	>	//cerr << "acx =" << ac.x() << "ahx =" << ah.x() << '\n';
1309	>	//cerr << "acy =" << ac.y() << "ahy =" << ah.y() << '\n';
1310	>	//cerr << "acz =" << ac.z() << "ahz =" << ah.z() << '\n';
1311	>	Asum_ += (ahrec.z() - acrec.z());
1312	>	Jsum_ += (jzp_.z()*((1/Mh)+(1/Mc)));
1313	>	AhCount_ = ahrec.z();
1314	>	if (outputAh_) {
1315	>	AhLog_ << time << "   ";
1316	>	AhLog_ << AhCount_;
1317	>	AhLog_ << endl;
1318	>	}
1319	>	}
1320	>	}
1321	>	}
1322	>	}
1323	>	}
1324	>	}
1325	>	}
1326	>	if (successfulExchange != true) {
1327	>	//   sprintf(painCave.errMsg,
1328	>	//              "RNEMD: exchange NOT performed!\n");
1329	>	//   painCave.isFatal = 0;
1330	>	//   painCave.severity = OPENMD_INFO;
1331	>	//   simError();
1332	>	failTrialCount_++;
1333	>	}
1334	>	}
1335	>
1336	>	void RNEMD::doRNEMD() {
1337	>
1338	>	switch(rnemdType_) {
1339	>	case rnemdKineticScale :
1340	>	case rnemdKineticScaleVAM :
1341	>	case rnemdKineticScaleAM :
1342	>	case rnemdPxScale :
1343	>	case rnemdPyScale :
1344	>	case rnemdPzScale :
1345	>	doScale();
1346	>	break;
1347	>	case rnemdKineticSwap :
1348	>	case rnemdPx :
1349	>	case rnemdPy :
1350	>	case rnemdPz :
1351	>	doSwap();
1352	>	break;
1353	>	case rnemdShiftScaleV :
1354	>	case rnemdShiftScaleVAM :
1355	>	doShiftScale();
1356	>	break;
1357	>	case rnemdUnknown :
1358	>	default :
1359	>	break;
1360	>	}
1361	>	}
1362	>
1363	>	void RNEMD::collectData() {
1364	>
1365	>	Snapshot* currentSnap_ = info_->getSnapshotManager()->getCurrentSnapshot();
1366	>	Mat3x3d hmat = currentSnap_->getHmat();
1367	>
1368	>	seleMan_.setSelectionSet(evaluator_.evaluate());
1369	>
1370	>	int selei;
1371	>	StuntDouble* sd;
1372	>	int idx;
1373	>
1374	>	logFrameCount_++;
1375	>
1376	>	// alternative approach, track all molecules instead of only those
1377	>	// selected for scaling/swapping:
1378	>	/*
1379	>	SimInfo::MoleculeIterator miter;
1380	>	vector<StuntDouble*>::iterator iiter;
1381	>	Molecule* mol;
1382	>	StuntDouble* integrableObject;
1383	>	for (mol = info_->beginMolecule(miter); mol != NULL;
1384	>	mol = info_->nextMolecule(miter))
1385	>	integrableObject is essentially sd
1386	>	for (integrableObject = mol->beginIntegrableObject(iiter);
1387	>	integrableObject != NULL;
1388	>	integrableObject = mol->nextIntegrableObject(iiter))
1389	>	*/
1390	>	for (sd = seleMan_.beginSelected(selei); sd != NULL;
1391	>	sd = seleMan_.nextSelected(selei)) {
1392	>
1393	>	idx = sd->getLocalIndex();
1394	>
1395	>	Vector3d pos = sd->getPos();
1396	>
1397	>	// wrap the stuntdouble's position back into the box:
1398	>
1399	>	if (usePeriodicBoundaryConditions_)
1400	>	currentSnap_->wrapVector(pos);
1401	>
1402	>	// which bin is this stuntdouble in?
1403	>	// wrapped positions are in the range [-0.5*hmat(2,2), +0.5*hmat(2,2)]
1404	>
1405	>	int binNo = int(rnemdLogWidth_ * (pos.z() / hmat(2,2) + 0.5)) %
1406	>	rnemdLogWidth_;
1407	>	// no symmetrization allowed due to arbitary rnemdLogWidth_
1408	>	/*
1409	>	if (rnemdLogWidth_ == midBin_ + 1)
1410	>	if (binNo > midBin_)
1411	>	binNo = nBins_ - binNo;
1412	>	*/
1413	>	RealType mass = sd->getMass();
1414	>	mHist_[binNo] += mass;
1415	>	Vector3d vel = sd->getVel();
1416	>	RealType value;
1417	>	//RealType xVal, yVal, zVal;
1418	>
1419	>	if (outputTemp_) {
1420	>	value = mass * vel.lengthSquare();
1421	>	tempCount_[binNo] += 3;
1422	>	if (sd->isDirectional()) {
1423	>	Vector3d angMom = sd->getJ();
1424	>	Mat3x3d I = sd->getI();
1425	>	if (sd->isLinear()) {
1426	>	int i = sd->linearAxis();
1427	>	int j = (i + 1) % 3;
1428	>	int k = (i + 2) % 3;
1429	>	value += angMom[j] * angMom[j] / I(j, j) +
1430	>	angMom[k] * angMom[k] / I(k, k);
1431	>	tempCount_[binNo] +=2;
1432	>	} else {
1433	>	value += angMom[0] * angMom[0] / I(0, 0) +
1434	>	angMom[1]*angMom[1]/I(1, 1) +
1435	>	angMom[2]*angMom[2]/I(2, 2);
1436	>	tempCount_[binNo] +=3;
1437	>	}
1438	>	}
1439	>	value = value / PhysicalConstants::energyConvert
1440	>	/ PhysicalConstants::kb;//may move to getStatus()
1441	>	tempHist_[binNo] += value;
1442	>	}
1443	>	if (outputVx_) {
1444	>	value = mass * vel[0];
1445	>	//vxzCount_[binNo]++;
1446	>	pxzHist_[binNo] += value;
1447	>	}
1448	>	if (outputVy_) {
1449	>	value = mass * vel[1];
1450	>	//vyzCount_[binNo]++;
1451	>	pyzHist_[binNo] += value;
1452	>	}
1453	>
1454	>	if (output3DTemp_) {
1455	>	value = mass * vel.x() * vel.x();
1456	>	xTempHist_[binNo] += value;
1457	>	value = mass * vel.y() * vel.y() / PhysicalConstants::energyConvert
1458	>	/ PhysicalConstants::kb;
1459	>	yTempHist_[binNo] += value;
1460	>	value = mass * vel.z() * vel.z() / PhysicalConstants::energyConvert
1461	>	/ PhysicalConstants::kb;
1462	>	zTempHist_[binNo] += value;
1463	>	xyzTempCount_[binNo]++;
1464	>	}
1465	>	if (outputRotTemp_) {
1466	>	if (sd->isDirectional()) {
1467	>	Vector3d angMom = sd->getJ();
1468	>	Mat3x3d I = sd->getI();
1469	>	if (sd->isLinear()) {
1470	>	int i = sd->linearAxis();
1471	>	int j = (i + 1) % 3;
1472	>	int k = (i + 2) % 3;
1473	>	value = angMom[j] * angMom[j] / I(j, j) +
1474	>	angMom[k] * angMom[k] / I(k, k);
1475	>	rotTempCount_[binNo] +=2;
1476	>	} else {
1477	>	value = angMom[0] * angMom[0] / I(0, 0) +
1478	>	angMom[1] * angMom[1] / I(1, 1) +
1479	>	angMom[2] * angMom[2] / I(2, 2);
1480	>	rotTempCount_[binNo] +=3;
1481	>	}
1482	>	}
1483	>	value = value / PhysicalConstants::energyConvert
1484	>	/ PhysicalConstants::kb;//may move to getStatus()
1485	>	rotTempHist_[binNo] += value;
1486	>	}
1487	>	// James put this in.
1488	>	if (outputDen_) {
1489	>	//value = 1.0;
1490	>	DenHist_[binNo] += 1;
1491	>	}
1492	>	if (outputVz_) {
1493	>	value = mass * vel[2];
1494	>	//vyzCount_[binNo]++;
1495	>	pzzHist_[binNo] += value;
1496	>	}
1497	>	}
1498	>	}
1499	>
1500	>	void RNEMD::getStarted() {
1501	>	collectData();
1502	>	/*now can output profile in step 0, but might not be useful;
1503	>	Snapshot* currentSnap_ = info_->getSnapshotManager()->getCurrentSnapshot();
1504	>	Stats& stat = currentSnap_->statData;
1505	>	stat[Stats::RNEMD_EXCHANGE_TOTAL] = exchangeSum_;
1506	>	*/
1507	>	//may output a header for the log file here
1508	>	getStatus();
1509	>	}
1510	>
1511	>	void RNEMD::getStatus() {
1512	>
1513	>	Snapshot* currentSnap_ = info_->getSnapshotManager()->getCurrentSnapshot();
1514	>	Stats& stat = currentSnap_->statData;
1515	>	RealType time = currentSnap_->getTime();
1516	>
1517	>	stat[Stats::RNEMD_EXCHANGE_TOTAL] = exchangeSum_;
1518	>	//or to be more meaningful, define another item as exchangeSum_ / time
1519	>	int j;
1520	>
1521	>	#ifdef IS_MPI
1522	>
1523	>	// all processors have the same number of bins, and STL vectors pack their
1524	>	// arrays, so in theory, this should be safe:
1525	>
1526	>	MPI::COMM_WORLD.Allreduce(MPI::IN_PLACE, &mHist_[0],
1527	>	rnemdLogWidth_, MPI::REALTYPE, MPI::SUM);
1528	>	if (outputTemp_) {
1529	>	MPI::COMM_WORLD.Allreduce(MPI::IN_PLACE, &tempHist_[0],
1530	>	rnemdLogWidth_, MPI::REALTYPE, MPI::SUM);
1531	>	MPI::COMM_WORLD.Allreduce(MPI::IN_PLACE, &tempCount_[0],
1532	>	rnemdLogWidth_, MPI::INT, MPI::SUM);
1533	>	}
1534	>	if (outputVx_) {
1535	>	MPI::COMM_WORLD.Allreduce(MPI::IN_PLACE, &pxzHist_[0],
1536	>	rnemdLogWidth_, MPI::REALTYPE, MPI::SUM);
1537	>	//MPI::COMM_WORLD.Allreduce(MPI::IN_PLACE, &vxzCount_[0],
1538	>	//                        rnemdLogWidth_, MPI::INT, MPI::SUM);
1539	>	}
1540	>	if (outputVy_) {
1541	>	MPI::COMM_WORLD.Allreduce(MPI::IN_PLACE, &pyzHist_[0],
1542	>	rnemdLogWidth_, MPI::REALTYPE, MPI::SUM);
1543	>	//MPI::COMM_WORLD.Allreduce(MPI::IN_PLACE, &vyzCount_[0],
1544	>	//                        rnemdLogWidth_, MPI::INT, MPI::SUM);
1545	>	}
1546	>	if (output3DTemp_) {
1547	>	MPI::COMM_WORLD.Allreduce(MPI::IN_PLACE, &xTempHist_[0],
1548	>	rnemdLogWidth_, MPI::REALTYPE, MPI::SUM);
1549	>	MPI::COMM_WORLD.Allreduce(MPI::IN_PLACE, &yTempHist_[0],
1550	>	rnemdLogWidth_, MPI::REALTYPE, MPI::SUM);
1551	>	MPI::COMM_WORLD.Allreduce(MPI::IN_PLACE, &zTempHist_[0],
1552	>	rnemdLogWidth_, MPI::REALTYPE, MPI::SUM);
1553	>	MPI::COMM_WORLD.Allreduce(MPI::IN_PLACE, &xyzTempCount_[0],
1554	>	rnemdLogWidth_, MPI::INT, MPI::SUM);
1555	>	}
1556	>	if (outputRotTemp_) {
1557	>	MPI::COMM_WORLD.Allreduce(MPI::IN_PLACE, &rotTempHist_[0],
1558	>	rnemdLogWidth_, MPI::REALTYPE, MPI::SUM);
1559	>	MPI::COMM_WORLD.Allreduce(MPI::IN_PLACE, &rotTempCount_[0],
1560	>	rnemdLogWidth_, MPI::INT, MPI::SUM);
1561	>	}
1562	>	// James put this in
1563	>	if (outputDen_) {
1564	>	MPI::COMM_WORLD.Allreduce(MPI::IN_PLACE, &DenHist_[0],
1565	>	rnemdLogWidth_, MPI::REALTYPE, MPI::SUM);
1566	>	}
1567	>	if (outputAh_) {
1568	>	MPI::COMM_WORLD.Allreduce(MPI::IN_PLACE, &AhCount_,
1569	>	1, MPI::REALTYPE, MPI::SUM);
1570	>	}
1571	>	if (outputVz_) {
1572	>	MPI::COMM_WORLD.Allreduce(MPI::IN_PLACE, &pzzHist_[0],
1573	>	rnemdLogWidth_, MPI::REALTYPE, MPI::SUM);
1574	>	}
1575	>
1576	>	// If we're the root node, should we print out the results
1577	>	int worldRank = MPI::COMM_WORLD.Get_rank();
1578	>	if (worldRank == 0) {
1579	>	#endif
1580	>
1581	>	if (outputTemp_) {
1582	>	tempLog_ << time;
1583	>	for (j = 0; j < rnemdLogWidth_; j++) {
1584	>	tempLog_ << "\t" << tempHist_[j] / (RealType)tempCount_[j];
1585	>	}
1586	>	tempLog_ << endl;
1587	>	}
1588	>	if (outputVx_) {
1589	>	vxzLog_ << time;
1590	>	for (j = 0; j < rnemdLogWidth_; j++) {
1591	>	vxzLog_ << "\t" << pxzHist_[j] / mHist_[j];
1592	>	}
1593	>	vxzLog_ << endl;
1594	>	}
1595	>	if (outputVy_) {
1596	>	vyzLog_ << time;
1597	>	for (j = 0; j < rnemdLogWidth_; j++) {
1598	>	vyzLog_ << "\t" << pyzHist_[j] / mHist_[j];
1599	>	}
1600	>	vyzLog_ << endl;
1601	>	}
1602	>
1603	>	if (output3DTemp_) {
1604	>	RealType temp;
1605	>	xTempLog_ << time;
1606	>	for (j = 0; j < rnemdLogWidth_; j++) {
1607	>	if (outputVx_)
1608	>	xTempHist_[j] -= pxzHist_[j] * pxzHist_[j] / mHist_[j];
1609	>	temp = xTempHist_[j] / (RealType)xyzTempCount_[j]
1610	>	/ PhysicalConstants::energyConvert / PhysicalConstants::kb;
1611	>	xTempLog_ << "\t" << temp;
1612	>	}
1613	>	xTempLog_ << endl;
1614	>	yTempLog_ << time;
1615	>	for (j = 0; j < rnemdLogWidth_; j++) {
1616	>	yTempLog_ << "\t" << yTempHist_[j] / (RealType)xyzTempCount_[j];
1617	>	}
1618	>	yTempLog_ << endl;
1619	>	zTempLog_ << time;
1620	>	for (j = 0; j < rnemdLogWidth_; j++) {
1621	>	zTempLog_ << "\t" << zTempHist_[j] / (RealType)xyzTempCount_[j];
1622	>	}
1623	>	zTempLog_ << endl;
1624	>	}
1625	>	if (outputRotTemp_) {
1626	>	rotTempLog_ << time;
1627	>	for (j = 0; j < rnemdLogWidth_; j++) {
1628	>	rotTempLog_ << "\t" << rotTempHist_[j] / (RealType)rotTempCount_[j];
1629	>	}
1630	>	rotTempLog_ << endl;
1631	>	}
1632	>	// James put this in.
1633	>	Mat3x3d hmat = currentSnap_->getHmat();
1634	>	if (outputDen_) {
1635	>	denLog_ << time;
1636	>	for (j = 0; j < rnemdLogWidth_; j++) {
1637	>
1638	>	RealType binVol = hmat(0,0) * hmat(1,1) * (hmat(2,2) / float(nBins_));
1639	>	denLog_ << "\t" << DenHist_[j] / (float(logFrameCount_) * binVol);
1640	>	}
1641	>	denLog_ << endl;
1642	>	}
1643	>	if (outputVz_) {
1644	>	vzzLog_ << time;
1645	>	for (j = 0; j < rnemdLogWidth_; j++) {
1646	>	vzzLog_ << "\t" << pzzHist_[j] / mHist_[j];
1647	>	}
1648	>	vzzLog_ << endl;
1649	>	}
1650	>	#ifdef IS_MPI
1651	>	}
1652	>	#endif
1653	>
1654	>	for (j = 0; j < rnemdLogWidth_; j++) {
1655	>	mHist_[j] = 0.0;
1656	>	}
1657	>	if (outputTemp_)
1658	>	for (j = 0; j < rnemdLogWidth_; j++) {
1659	>	tempCount_[j] = 0;
1660	>	tempHist_[j] = 0.0;
1661	>	}
1662	>	if (outputVx_)
1663	>	for (j = 0; j < rnemdLogWidth_; j++) {
1664	>	//pxzCount_[j] = 0;
1665	>	pxzHist_[j] = 0.0;
1666	>	}
1667	>	if (outputVy_)
1668	>	for (j = 0; j < rnemdLogWidth_; j++) {
1669	>	//pyzCount_[j] = 0;
1670	>	pyzHist_[j] = 0.0;
1671	>	}
1672	>
1673	>	if (output3DTemp_)
1674	>	for (j = 0; j < rnemdLogWidth_; j++) {
1675	>	xTempHist_[j] = 0.0;
1676	>	yTempHist_[j] = 0.0;
1677	>	zTempHist_[j] = 0.0;
1678	>	xyzTempCount_[j] = 0;
1679	>	}
1680	>	if (outputRotTemp_)
1681	>	for (j = 0; j < rnemdLogWidth_; j++) {
1682	>	rotTempCount_[j] = 0;
1683	>	rotTempHist_[j] = 0.0;
1684	>	}
1685	>	// James put this in
1686	>	if (outputDen_)
1687	>	for (j = 0; j < rnemdLogWidth_; j++) {
1688	>	//pyzCount_[j] = 0;
1689	>	DenHist_[j] = 0.0;
1690	>	}
1691	>	if (outputVz_)
1692	>	for (j = 0; j < rnemdLogWidth_; j++) {
1693	>	//pyzCount_[j] = 0;
1694	>	pzzHist_[j] = 0.0;
1695	>	}
1696	>	// reset the counter
1697	>
1698	>	Numcount_++;
1699	>	if (Numcount_ > int(runTime_/statusTime_))
1700	>	cerr << "time =" << time << "  Asum =" << Asum_ << '\n';
1701	>	if (Numcount_ > int(runTime_/statusTime_))
1702	>	cerr << "time =" << time << "  Jsum =" << Jsum_ << '\n';
1703	>
1704	>	logFrameCount_ = 0;
1705	>	}
1706		}
1707	+

Comparing:
trunk/src/integrators/RNEMD.cpp (property svn:keywords), Revision 1330 by skuang, Thu Mar 19 21:03:36 2009 UTC vs.
branches/development/src/integrators/RNEMD.cpp (property svn:keywords), Revision 1728 by jmarr, Wed May 30 16:07:03 2012 UTC

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

Diff Legend

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