ViewVC Help
View File | Revision Log | Show Annotations | View Changeset | Root Listing
root/OpenMD/trunk/src/parallel/ForceMatrixDecomposition.cpp
(Generate patch)

Comparing:
branches/development/src/parallel/ForceDecomposition.cpp (file contents), Revision 1538 by chuckv, Tue Jan 11 18:58:12 2011 UTC vs.
branches/development/src/parallel/ForceMatrixDecomposition.cpp (file contents), Revision 1593 by gezelter, Fri Jul 15 21:35:14 2011 UTC

# Line 1 | Line 1
1 < /**
2 < * @file ForceDecomposition.cpp
3 < * @author Charles Vardeman <cvardema.at.nd.edu>
4 < * @date 08/18/2010
5 < * @time 11:56am
6 < * @version 1.0
1 > /*
2 > * Copyright (c) 2005 The University of Notre Dame. All Rights Reserved.
3   *
8 – * @section LICENSE
9 – * Copyright (c) 2010 The University of Notre Dame. All Rights Reserved.
10 – *
4   * The University of Notre Dame grants you ("Licensee") a
5   * non-exclusive, royalty free, license to use, modify and
6   * redistribute this software in source and binary code form, provided
# Line 45 | Line 38
38   * [3]  Sun, Lin & Gezelter, J. Chem. Phys. 128, 24107 (2008).          
39   * [4]  Vardeman & Gezelter, in progress (2009).                        
40   */
41 + #include "parallel/ForceMatrixDecomposition.hpp"
42 + #include "math/SquareMatrix3.hpp"
43 + #include "nonbonded/NonBondedInteraction.hpp"
44 + #include "brains/SnapshotManager.hpp"
45 + #include "brains/PairList.hpp"
46  
47 + using namespace std;
48 + namespace OpenMD {
49  
50 +  ForceMatrixDecomposition::ForceMatrixDecomposition(SimInfo* info, InteractionManager* iMan) : ForceDecomposition(info, iMan) {
51  
52 < /*  -*- c++ -*-  */
53 < #include "config.h"
54 < #include <stdlib.h>
52 >    // In a parallel computation, row and colum scans must visit all
53 >    // surrounding cells (not just the 14 upper triangular blocks that
54 >    // are used when the processor can see all pairs)
55   #ifdef IS_MPI
56 < #include <mpi.h>
57 < #endif
56 >    cellOffsets_.push_back( Vector3i(-1, 0, 0) );
57 >    cellOffsets_.push_back( Vector3i(-1,-1, 0) );
58 >    cellOffsets_.push_back( Vector3i( 0,-1, 0) );
59 >    cellOffsets_.push_back( Vector3i( 1,-1, 0) );
60 >    cellOffsets_.push_back( Vector3i( 0, 0,-1) );
61 >    cellOffsets_.push_back( Vector3i(-1, 0, 1) );
62 >    cellOffsets_.push_back( Vector3i(-1,-1,-1) );
63 >    cellOffsets_.push_back( Vector3i( 0,-1,-1) );
64 >    cellOffsets_.push_back( Vector3i( 1,-1,-1) );
65 >    cellOffsets_.push_back( Vector3i( 1, 0,-1) );
66 >    cellOffsets_.push_back( Vector3i( 1, 1,-1) );
67 >    cellOffsets_.push_back( Vector3i( 0, 1,-1) );
68 >    cellOffsets_.push_back( Vector3i(-1, 1,-1) );
69 > #endif    
70 >  }
71  
58 – #include <iostream>
59 – #include <vector>
60 – #include <algorithm>
61 – #include <cmath>
62 – #include "parallel/ForceDecomposition.hpp"
72  
73 +  /**
74 +   * distributeInitialData is essentially a copy of the older fortran
75 +   * SimulationSetup
76 +   */
77 +  void ForceMatrixDecomposition::distributeInitialData() {
78 +    snap_ = sman_->getCurrentSnapshot();
79 +    storageLayout_ = sman_->getStorageLayout();
80 +    ff_ = info_->getForceField();
81 +    nLocal_ = snap_->getNumberOfAtoms();
82 +    
83 +    nGroups_ = info_->getNLocalCutoffGroups();
84 +    // gather the information for atomtype IDs (atids):
85 +    idents = info_->getIdentArray();
86 +    AtomLocalToGlobal = info_->getGlobalAtomIndices();
87 +    cgLocalToGlobal = info_->getGlobalGroupIndices();
88 +    vector<int> globalGroupMembership = info_->getGlobalGroupMembership();
89  
90 < using namespace std;
66 < using namespace OpenMD;
90 >    massFactors = info_->getMassFactors();
91  
92 < //__static
92 >    PairList* excludes = info_->getExcludedInteractions();
93 >    PairList* oneTwo = info_->getOneTwoInteractions();
94 >    PairList* oneThree = info_->getOneThreeInteractions();
95 >    PairList* oneFour = info_->getOneFourInteractions();
96 >
97   #ifdef IS_MPI
98 < static vector<MPI:Comm> communictors;
99 < #endif
98 >
99 >    MPI::Intracomm row = rowComm.getComm();
100 >    MPI::Intracomm col = colComm.getComm();
101  
102 < //____ MPITypeTraits
103 < template<typename T>
104 < struct MPITypeTraits;
102 >    AtomPlanIntRow = new Plan<int>(row, nLocal_);
103 >    AtomPlanRealRow = new Plan<RealType>(row, nLocal_);
104 >    AtomPlanVectorRow = new Plan<Vector3d>(row, nLocal_);
105 >    AtomPlanMatrixRow = new Plan<Mat3x3d>(row, nLocal_);
106 >    AtomPlanPotRow = new Plan<potVec>(row, nLocal_);
107  
108 < #ifdef IS_MPI
109 < template<>
110 < struct MPITypeTraits<RealType> {
111 <  static const MPI::Datatype datatype;
112 < };
82 < const MPI_Datatype MPITypeTraits<RealType>::datatype = MY_MPI_REAL;
108 >    AtomPlanIntColumn = new Plan<int>(col, nLocal_);
109 >    AtomPlanRealColumn = new Plan<RealType>(col, nLocal_);
110 >    AtomPlanVectorColumn = new Plan<Vector3d>(col, nLocal_);
111 >    AtomPlanMatrixColumn = new Plan<Mat3x3d>(col, nLocal_);
112 >    AtomPlanPotColumn = new Plan<potVec>(col, nLocal_);
113  
114 < template<>
115 < struct MPITypeTraits<int> {
116 <  static const MPI::Datatype datatype;
117 < };
88 < const MPI::Datatype MPITypeTraits<int>::datatype = MPI_INT;
89 < #endif
114 >    cgPlanIntRow = new Plan<int>(row, nGroups_);
115 >    cgPlanVectorRow = new Plan<Vector3d>(row, nGroups_);
116 >    cgPlanIntColumn = new Plan<int>(col, nGroups_);
117 >    cgPlanVectorColumn = new Plan<Vector3d>(col, nGroups_);
118  
119 < /**
120 < * Constructor for ForceDecomposition Parallel Decomposition Method
121 < * Will try to construct a symmetric grid of processors. Ideally, the
122 < * number of processors will be a square ex: 4, 9, 16, 25.
95 < *
96 < */
119 >    nAtomsInRow_ = AtomPlanIntRow->getSize();
120 >    nAtomsInCol_ = AtomPlanIntColumn->getSize();
121 >    nGroupsInRow_ = cgPlanIntRow->getSize();
122 >    nGroupsInCol_ = cgPlanIntColumn->getSize();
123  
124 < ForceDecomposition::ForceDecomposition() {
124 >    // Modify the data storage objects with the correct layouts and sizes:
125 >    atomRowData.resize(nAtomsInRow_);
126 >    atomRowData.setStorageLayout(storageLayout_);
127 >    atomColData.resize(nAtomsInCol_);
128 >    atomColData.setStorageLayout(storageLayout_);
129 >    cgRowData.resize(nGroupsInRow_);
130 >    cgRowData.setStorageLayout(DataStorage::dslPosition);
131 >    cgColData.resize(nGroupsInCol_);
132 >    cgColData.setStorageLayout(DataStorage::dslPosition);
133 >        
134 >    identsRow.resize(nAtomsInRow_);
135 >    identsCol.resize(nAtomsInCol_);
136 >    
137 >    AtomPlanIntRow->gather(idents, identsRow);
138 >    AtomPlanIntColumn->gather(idents, identsCol);
139 >    
140 >    // allocate memory for the parallel objects
141 >    atypesRow.resize(nAtomsInRow_);
142 >    atypesCol.resize(nAtomsInCol_);
143  
144 < #ifdef IS_MPI
145 <  int nProcs = MPI::COMM_WORLD.Get_size();
146 <  int worldRank = MPI::COMM_WORLD.Get_rank();
144 >    for (int i = 0; i < nAtomsInRow_; i++)
145 >      atypesRow[i] = ff_->getAtomType(identsRow[i]);
146 >    for (int i = 0; i < nAtomsInCol_; i++)
147 >      atypesCol[i] = ff_->getAtomType(identsCol[i]);        
148 >
149 >    pot_row.resize(nAtomsInRow_);
150 >    pot_col.resize(nAtomsInCol_);
151 >
152 >    AtomRowToGlobal.resize(nAtomsInRow_);
153 >    AtomColToGlobal.resize(nAtomsInCol_);
154 >    AtomPlanIntRow->gather(AtomLocalToGlobal, AtomRowToGlobal);
155 >    AtomPlanIntColumn->gather(AtomLocalToGlobal, AtomColToGlobal);
156 >
157 >    cerr << "Atoms in Local:\n";
158 >    for (int i = 0; i < AtomLocalToGlobal.size(); i++) {
159 >      cerr << "i =\t" << i << "\t localAt =\t" << AtomLocalToGlobal[i] << "\n";
160 >    }
161 >    cerr << "Atoms in Row:\n";
162 >    for (int i = 0; i < AtomRowToGlobal.size(); i++) {
163 >      cerr << "i =\t" << i << "\t rowAt =\t" << AtomRowToGlobal[i] << "\n";
164 >    }
165 >    cerr << "Atoms in Col:\n";
166 >    for (int i = 0; i < AtomColToGlobal.size(); i++) {
167 >      cerr << "i =\t" << i << "\t colAt =\t" << AtomColToGlobal[i] << "\n";
168 >    }
169 >
170 >    cgRowToGlobal.resize(nGroupsInRow_);
171 >    cgColToGlobal.resize(nGroupsInCol_);
172 >    cgPlanIntRow->gather(cgLocalToGlobal, cgRowToGlobal);
173 >    cgPlanIntColumn->gather(cgLocalToGlobal, cgColToGlobal);
174 >
175 >    cerr << "Gruops in Local:\n";
176 >    for (int i = 0; i < cgLocalToGlobal.size(); i++) {
177 >      cerr << "i =\t" << i << "\t localCG =\t" << cgLocalToGlobal[i] << "\n";
178 >    }
179 >    cerr << "Groups in Row:\n";
180 >    for (int i = 0; i < cgRowToGlobal.size(); i++) {
181 >      cerr << "i =\t" << i << "\t rowCG =\t" << cgRowToGlobal[i] << "\n";
182 >    }
183 >    cerr << "Groups in Col:\n";
184 >    for (int i = 0; i < cgColToGlobal.size(); i++) {
185 >      cerr << "i =\t" << i << "\t colCG =\t" << cgColToGlobal[i] << "\n";
186 >    }
187 >
188 >
189 >    massFactorsRow.resize(nAtomsInRow_);
190 >    massFactorsCol.resize(nAtomsInCol_);
191 >    AtomPlanRealRow->gather(massFactors, massFactorsRow);
192 >    AtomPlanRealColumn->gather(massFactors, massFactorsCol);
193 >
194 >    groupListRow_.clear();
195 >    groupListRow_.resize(nGroupsInRow_);
196 >    for (int i = 0; i < nGroupsInRow_; i++) {
197 >      int gid = cgRowToGlobal[i];
198 >      for (int j = 0; j < nAtomsInRow_; j++) {
199 >        int aid = AtomRowToGlobal[j];
200 >        if (globalGroupMembership[aid] == gid)
201 >          groupListRow_[i].push_back(j);
202 >      }      
203 >    }
204 >
205 >    groupListCol_.clear();
206 >    groupListCol_.resize(nGroupsInCol_);
207 >    for (int i = 0; i < nGroupsInCol_; i++) {
208 >      int gid = cgColToGlobal[i];
209 >      for (int j = 0; j < nAtomsInCol_; j++) {
210 >        int aid = AtomColToGlobal[j];
211 >        if (globalGroupMembership[aid] == gid)
212 >          groupListCol_[i].push_back(j);
213 >      }      
214 >    }
215 >
216 >    excludesForAtom.clear();
217 >    excludesForAtom.resize(nAtomsInRow_);
218 >    toposForAtom.clear();
219 >    toposForAtom.resize(nAtomsInRow_);
220 >    topoDist.clear();
221 >    topoDist.resize(nAtomsInRow_);
222 >    for (int i = 0; i < nAtomsInRow_; i++) {
223 >      int iglob = AtomRowToGlobal[i];
224 >
225 >      for (int j = 0; j < nAtomsInCol_; j++) {
226 >        int jglob = AtomColToGlobal[j];
227 >
228 >        if (excludes->hasPair(iglob, jglob))
229 >          excludesForAtom[i].push_back(j);      
230 >        
231 >        if (oneTwo->hasPair(iglob, jglob)) {
232 >          toposForAtom[i].push_back(j);
233 >          topoDist[i].push_back(1);
234 >        } else {
235 >          if (oneThree->hasPair(iglob, jglob)) {
236 >            toposForAtom[i].push_back(j);
237 >            topoDist[i].push_back(2);
238 >          } else {
239 >            if (oneFour->hasPair(iglob, jglob)) {
240 >              toposForAtom[i].push_back(j);
241 >              topoDist[i].push_back(3);
242 >            }
243 >          }
244 >        }
245 >      }      
246 >    }
247 >
248   #endif
249  
250 <  // First time through, construct column stride.
251 <  if (communicators.size() == 0)
252 <  {
253 <    int nColumnsMax = (int) round(sqrt((float) nProcs));
254 <    for (int i = 0; i < nProcs; ++i)
255 <    {
256 <      if (nProcs%i==0) nColumns=i;
250 >    // allocate memory for the parallel objects
251 >    atypesLocal.resize(nLocal_);
252 >
253 >    for (int i = 0; i < nLocal_; i++)
254 >      atypesLocal[i] = ff_->getAtomType(idents[i]);
255 >
256 >    groupList_.clear();
257 >    groupList_.resize(nGroups_);
258 >    for (int i = 0; i < nGroups_; i++) {
259 >      int gid = cgLocalToGlobal[i];
260 >      for (int j = 0; j < nLocal_; j++) {
261 >        int aid = AtomLocalToGlobal[j];
262 >        if (globalGroupMembership[aid] == gid) {
263 >          groupList_[i].push_back(j);
264 >        }
265 >      }      
266      }
267  
268 <    int nRows = nProcs/nColumns;    
269 <    myRank_ = (int) worldRank%nColumns;
270 <  }
271 <  else
272 <  {
273 <    myRank_ = myRank/nColumns;
120 <  }
121 <  MPI::Comm newComm = MPI:COMM_WORLD.Split(myRank_,0);
122 <  
123 <  isColumn_ = false;
124 <  
125 < }
268 >    excludesForAtom.clear();
269 >    excludesForAtom.resize(nLocal_);
270 >    toposForAtom.clear();
271 >    toposForAtom.resize(nLocal_);
272 >    topoDist.clear();
273 >    topoDist.resize(nLocal_);
274  
275 < ForceDecomposition::gather(sendbuf, receivebuf){
276 <  communicators(myIndex_).Allgatherv();
129 < }
275 >    for (int i = 0; i < nLocal_; i++) {
276 >      int iglob = AtomLocalToGlobal[i];
277  
278 +      for (int j = 0; j < nLocal_; j++) {
279 +        int jglob = AtomLocalToGlobal[j];
280  
281 +        if (excludes->hasPair(iglob, jglob))
282 +          excludesForAtom[i].push_back(j);              
283 +        
284 +        if (oneTwo->hasPair(iglob, jglob)) {
285 +          toposForAtom[i].push_back(j);
286 +          topoDist[i].push_back(1);
287 +        } else {
288 +          if (oneThree->hasPair(iglob, jglob)) {
289 +            toposForAtom[i].push_back(j);
290 +            topoDist[i].push_back(2);
291 +          } else {
292 +            if (oneFour->hasPair(iglob, jglob)) {
293 +              toposForAtom[i].push_back(j);
294 +              topoDist[i].push_back(3);
295 +            }
296 +          }
297 +        }
298 +      }      
299 +    }
300 +    
301 +    createGtypeCutoffMap();
302  
303 < ForceDecomposition::scatter(sbuffer, rbuffer){
304 <  communicators(myIndex_).Reduce_scatter(sbuffer, recevbuf. recvcounts, MPI::DOUBLE, MPI::SUM);
305 < }
303 >  }
304 >  
305 >  void ForceMatrixDecomposition::createGtypeCutoffMap() {
306 >    
307 >    RealType tol = 1e-6;
308 >    largestRcut_ = 0.0;
309 >    RealType rc;
310 >    int atid;
311 >    set<AtomType*> atypes = info_->getSimulatedAtomTypes();
312 >    
313 >    map<int, RealType> atypeCutoff;
314 >      
315 >    for (set<AtomType*>::iterator at = atypes.begin();
316 >         at != atypes.end(); ++at){
317 >      atid = (*at)->getIdent();
318 >      if (userChoseCutoff_)
319 >        atypeCutoff[atid] = userCutoff_;
320 >      else
321 >        atypeCutoff[atid] = interactionMan_->getSuggestedCutoffRadius(*at);
322 >    }
323 >    
324 >    vector<RealType> gTypeCutoffs;
325 >    // first we do a single loop over the cutoff groups to find the
326 >    // largest cutoff for any atypes present in this group.
327 > #ifdef IS_MPI
328 >    vector<RealType> groupCutoffRow(nGroupsInRow_, 0.0);
329 >    groupRowToGtype.resize(nGroupsInRow_);
330 >    for (int cg1 = 0; cg1 < nGroupsInRow_; cg1++) {
331 >      vector<int> atomListRow = getAtomsInGroupRow(cg1);
332 >      for (vector<int>::iterator ia = atomListRow.begin();
333 >           ia != atomListRow.end(); ++ia) {            
334 >        int atom1 = (*ia);
335 >        atid = identsRow[atom1];
336 >        if (atypeCutoff[atid] > groupCutoffRow[cg1]) {
337 >          groupCutoffRow[cg1] = atypeCutoff[atid];
338 >        }
339 >      }
340 >
341 >      bool gTypeFound = false;
342 >      for (int gt = 0; gt < gTypeCutoffs.size(); gt++) {
343 >        if (abs(groupCutoffRow[cg1] - gTypeCutoffs[gt]) < tol) {
344 >          groupRowToGtype[cg1] = gt;
345 >          gTypeFound = true;
346 >        }
347 >      }
348 >      if (!gTypeFound) {
349 >        gTypeCutoffs.push_back( groupCutoffRow[cg1] );
350 >        groupRowToGtype[cg1] = gTypeCutoffs.size() - 1;
351 >      }
352 >      
353 >    }
354 >    vector<RealType> groupCutoffCol(nGroupsInCol_, 0.0);
355 >    groupColToGtype.resize(nGroupsInCol_);
356 >    for (int cg2 = 0; cg2 < nGroupsInCol_; cg2++) {
357 >      vector<int> atomListCol = getAtomsInGroupColumn(cg2);
358 >      for (vector<int>::iterator jb = atomListCol.begin();
359 >           jb != atomListCol.end(); ++jb) {            
360 >        int atom2 = (*jb);
361 >        atid = identsCol[atom2];
362 >        if (atypeCutoff[atid] > groupCutoffCol[cg2]) {
363 >          groupCutoffCol[cg2] = atypeCutoff[atid];
364 >        }
365 >      }
366 >      bool gTypeFound = false;
367 >      for (int gt = 0; gt < gTypeCutoffs.size(); gt++) {
368 >        if (abs(groupCutoffCol[cg2] - gTypeCutoffs[gt]) < tol) {
369 >          groupColToGtype[cg2] = gt;
370 >          gTypeFound = true;
371 >        }
372 >      }
373 >      if (!gTypeFound) {
374 >        gTypeCutoffs.push_back( groupCutoffCol[cg2] );
375 >        groupColToGtype[cg2] = gTypeCutoffs.size() - 1;
376 >      }
377 >    }
378 > #else
379 >
380 >    vector<RealType> groupCutoff(nGroups_, 0.0);
381 >    groupToGtype.resize(nGroups_);
382 >    for (int cg1 = 0; cg1 < nGroups_; cg1++) {
383 >      groupCutoff[cg1] = 0.0;
384 >      vector<int> atomList = getAtomsInGroupRow(cg1);
385 >      for (vector<int>::iterator ia = atomList.begin();
386 >           ia != atomList.end(); ++ia) {            
387 >        int atom1 = (*ia);
388 >        atid = idents[atom1];
389 >        if (atypeCutoff[atid] > groupCutoff[cg1])
390 >          groupCutoff[cg1] = atypeCutoff[atid];
391 >      }
392 >      
393 >      bool gTypeFound = false;
394 >      for (int gt = 0; gt < gTypeCutoffs.size(); gt++) {
395 >        if (abs(groupCutoff[cg1] - gTypeCutoffs[gt]) < tol) {
396 >          groupToGtype[cg1] = gt;
397 >          gTypeFound = true;
398 >        }
399 >      }
400 >      if (!gTypeFound) {      
401 >        gTypeCutoffs.push_back( groupCutoff[cg1] );
402 >        groupToGtype[cg1] = gTypeCutoffs.size() - 1;
403 >      }      
404 >    }
405 > #endif
406 >
407 >    // Now we find the maximum group cutoff value present in the simulation
408 >
409 >    RealType groupMax = *max_element(gTypeCutoffs.begin(),
410 >                                     gTypeCutoffs.end());
411 >
412 > #ifdef IS_MPI
413 >    MPI::COMM_WORLD.Allreduce(&groupMax, &groupMax, 1, MPI::REALTYPE,
414 >                              MPI::MAX);
415 > #endif
416 >    
417 >    RealType tradRcut = groupMax;
418 >
419 >    for (int i = 0; i < gTypeCutoffs.size();  i++) {
420 >      for (int j = 0; j < gTypeCutoffs.size();  j++) {      
421 >        RealType thisRcut;
422 >        switch(cutoffPolicy_) {
423 >        case TRADITIONAL:
424 >          thisRcut = tradRcut;
425 >          break;
426 >        case MIX:
427 >          thisRcut = 0.5 * (gTypeCutoffs[i] + gTypeCutoffs[j]);
428 >          break;
429 >        case MAX:
430 >          thisRcut = max(gTypeCutoffs[i], gTypeCutoffs[j]);
431 >          break;
432 >        default:
433 >          sprintf(painCave.errMsg,
434 >                  "ForceMatrixDecomposition::createGtypeCutoffMap "
435 >                  "hit an unknown cutoff policy!\n");
436 >          painCave.severity = OPENMD_ERROR;
437 >          painCave.isFatal = 1;
438 >          simError();
439 >          break;
440 >        }
441 >
442 >        pair<int,int> key = make_pair(i,j);
443 >        gTypeCutoffMap[key].first = thisRcut;
444 >        if (thisRcut > largestRcut_) largestRcut_ = thisRcut;
445 >        gTypeCutoffMap[key].second = thisRcut*thisRcut;
446 >        gTypeCutoffMap[key].third = pow(thisRcut + skinThickness_, 2);
447 >        // sanity check
448 >        
449 >        if (userChoseCutoff_) {
450 >          if (abs(gTypeCutoffMap[key].first - userCutoff_) > 0.0001) {
451 >            sprintf(painCave.errMsg,
452 >                    "ForceMatrixDecomposition::createGtypeCutoffMap "
453 >                    "user-specified rCut (%lf) does not match computed group Cutoff\n", userCutoff_);
454 >            painCave.severity = OPENMD_ERROR;
455 >            painCave.isFatal = 1;
456 >            simError();            
457 >          }
458 >        }
459 >      }
460 >    }
461 >  }
462 >
463 >
464 >  groupCutoffs ForceMatrixDecomposition::getGroupCutoffs(int cg1, int cg2) {
465 >    int i, j;  
466 > #ifdef IS_MPI
467 >    i = groupRowToGtype[cg1];
468 >    j = groupColToGtype[cg2];
469 > #else
470 >    i = groupToGtype[cg1];
471 >    j = groupToGtype[cg2];
472 > #endif    
473 >    return gTypeCutoffMap[make_pair(i,j)];
474 >  }
475 >
476 >  int ForceMatrixDecomposition::getTopologicalDistance(int atom1, int atom2) {
477 >    for (int j = 0; j < toposForAtom[atom1].size(); j++) {
478 >      if (toposForAtom[atom1][j] == atom2)
479 >        return topoDist[atom1][j];
480 >    }
481 >    return 0;
482 >  }
483 >
484 >  void ForceMatrixDecomposition::zeroWorkArrays() {
485 >    pairwisePot = 0.0;
486 >    embeddingPot = 0.0;
487 >
488 > #ifdef IS_MPI
489 >    if (storageLayout_ & DataStorage::dslForce) {
490 >      fill(atomRowData.force.begin(), atomRowData.force.end(), V3Zero);
491 >      fill(atomColData.force.begin(), atomColData.force.end(), V3Zero);
492 >    }
493 >
494 >    if (storageLayout_ & DataStorage::dslTorque) {
495 >      fill(atomRowData.torque.begin(), atomRowData.torque.end(), V3Zero);
496 >      fill(atomColData.torque.begin(), atomColData.torque.end(), V3Zero);
497 >    }
498 >    
499 >    fill(pot_row.begin(), pot_row.end(),
500 >         Vector<RealType, N_INTERACTION_FAMILIES> (0.0));
501 >
502 >    fill(pot_col.begin(), pot_col.end(),
503 >         Vector<RealType, N_INTERACTION_FAMILIES> (0.0));  
504 >
505 >    if (storageLayout_ & DataStorage::dslParticlePot) {    
506 >      fill(atomRowData.particlePot.begin(), atomRowData.particlePot.end(),
507 >           0.0);
508 >      fill(atomColData.particlePot.begin(), atomColData.particlePot.end(),
509 >           0.0);
510 >    }
511 >
512 >    if (storageLayout_ & DataStorage::dslDensity) {      
513 >      fill(atomRowData.density.begin(), atomRowData.density.end(), 0.0);
514 >      fill(atomColData.density.begin(), atomColData.density.end(), 0.0);
515 >    }
516 >
517 >    if (storageLayout_ & DataStorage::dslFunctional) {  
518 >      fill(atomRowData.functional.begin(), atomRowData.functional.end(),
519 >           0.0);
520 >      fill(atomColData.functional.begin(), atomColData.functional.end(),
521 >           0.0);
522 >    }
523 >
524 >    if (storageLayout_ & DataStorage::dslFunctionalDerivative) {      
525 >      fill(atomRowData.functionalDerivative.begin(),
526 >           atomRowData.functionalDerivative.end(), 0.0);
527 >      fill(atomColData.functionalDerivative.begin(),
528 >           atomColData.functionalDerivative.end(), 0.0);
529 >    }
530 >
531 >    if (storageLayout_ & DataStorage::dslSkippedCharge) {      
532 >      fill(atomRowData.skippedCharge.begin(),
533 >           atomRowData.skippedCharge.end(), 0.0);
534 >      fill(atomColData.skippedCharge.begin(),
535 >           atomColData.skippedCharge.end(), 0.0);
536 >    }
537 >
538 > #endif
539 >    // even in parallel, we need to zero out the local arrays:
540 >
541 >    if (storageLayout_ & DataStorage::dslParticlePot) {      
542 >      fill(snap_->atomData.particlePot.begin(),
543 >           snap_->atomData.particlePot.end(), 0.0);
544 >    }
545 >    
546 >    if (storageLayout_ & DataStorage::dslDensity) {      
547 >      fill(snap_->atomData.density.begin(),
548 >           snap_->atomData.density.end(), 0.0);
549 >    }
550 >    if (storageLayout_ & DataStorage::dslFunctional) {
551 >      fill(snap_->atomData.functional.begin(),
552 >           snap_->atomData.functional.end(), 0.0);
553 >    }
554 >    if (storageLayout_ & DataStorage::dslFunctionalDerivative) {      
555 >      fill(snap_->atomData.functionalDerivative.begin(),
556 >           snap_->atomData.functionalDerivative.end(), 0.0);
557 >    }
558 >    if (storageLayout_ & DataStorage::dslSkippedCharge) {      
559 >      fill(snap_->atomData.skippedCharge.begin(),
560 >           snap_->atomData.skippedCharge.end(), 0.0);
561 >    }
562 >    
563 >  }
564 >
565 >
566 >  void ForceMatrixDecomposition::distributeData()  {
567 >    snap_ = sman_->getCurrentSnapshot();
568 >    storageLayout_ = sman_->getStorageLayout();
569 > #ifdef IS_MPI
570 >    
571 >    // gather up the atomic positions
572 >    AtomPlanVectorRow->gather(snap_->atomData.position,
573 >                              atomRowData.position);
574 >    AtomPlanVectorColumn->gather(snap_->atomData.position,
575 >                                 atomColData.position);
576 >    
577 >    // gather up the cutoff group positions
578 >
579 >    cerr  << "before gather\n";
580 >    for (int i = 0; i < snap_->cgData.position.size(); i++) {
581 >      cerr << "cgpos = " << snap_->cgData.position[i] << "\n";
582 >    }
583 >
584 >    cgPlanVectorRow->gather(snap_->cgData.position,
585 >                            cgRowData.position);
586 >
587 >    cerr  << "after gather\n";
588 >    for (int i = 0; i < cgRowData.position.size(); i++) {
589 >      cerr << "cgRpos = " << cgRowData.position[i] << "\n";
590 >    }
591 >
592 >    cgPlanVectorColumn->gather(snap_->cgData.position,
593 >                               cgColData.position);
594 >    for (int i = 0; i < cgColData.position.size(); i++) {
595 >      cerr << "cgCpos = " << cgColData.position[i] << "\n";
596 >    }
597 >
598 >    
599 >    // if needed, gather the atomic rotation matrices
600 >    if (storageLayout_ & DataStorage::dslAmat) {
601 >      AtomPlanMatrixRow->gather(snap_->atomData.aMat,
602 >                                atomRowData.aMat);
603 >      AtomPlanMatrixColumn->gather(snap_->atomData.aMat,
604 >                                   atomColData.aMat);
605 >    }
606 >    
607 >    // if needed, gather the atomic eletrostatic frames
608 >    if (storageLayout_ & DataStorage::dslElectroFrame) {
609 >      AtomPlanMatrixRow->gather(snap_->atomData.electroFrame,
610 >                                atomRowData.electroFrame);
611 >      AtomPlanMatrixColumn->gather(snap_->atomData.electroFrame,
612 >                                   atomColData.electroFrame);
613 >    }
614 >
615 > #endif      
616 >  }
617 >  
618 >  /* collects information obtained during the pre-pair loop onto local
619 >   * data structures.
620 >   */
621 >  void ForceMatrixDecomposition::collectIntermediateData() {
622 >    snap_ = sman_->getCurrentSnapshot();
623 >    storageLayout_ = sman_->getStorageLayout();
624 > #ifdef IS_MPI
625 >    
626 >    if (storageLayout_ & DataStorage::dslDensity) {
627 >      
628 >      AtomPlanRealRow->scatter(atomRowData.density,
629 >                               snap_->atomData.density);
630 >      
631 >      int n = snap_->atomData.density.size();
632 >      vector<RealType> rho_tmp(n, 0.0);
633 >      AtomPlanRealColumn->scatter(atomColData.density, rho_tmp);
634 >      for (int i = 0; i < n; i++)
635 >        snap_->atomData.density[i] += rho_tmp[i];
636 >    }
637 > #endif
638 >  }
639  
640 +  /*
641 +   * redistributes information obtained during the pre-pair loop out to
642 +   * row and column-indexed data structures
643 +   */
644 +  void ForceMatrixDecomposition::distributeIntermediateData() {
645 +    snap_ = sman_->getCurrentSnapshot();
646 +    storageLayout_ = sman_->getStorageLayout();
647 + #ifdef IS_MPI
648 +    if (storageLayout_ & DataStorage::dslFunctional) {
649 +      AtomPlanRealRow->gather(snap_->atomData.functional,
650 +                              atomRowData.functional);
651 +      AtomPlanRealColumn->gather(snap_->atomData.functional,
652 +                                 atomColData.functional);
653 +    }
654 +    
655 +    if (storageLayout_ & DataStorage::dslFunctionalDerivative) {
656 +      AtomPlanRealRow->gather(snap_->atomData.functionalDerivative,
657 +                              atomRowData.functionalDerivative);
658 +      AtomPlanRealColumn->gather(snap_->atomData.functionalDerivative,
659 +                                 atomColData.functionalDerivative);
660 +    }
661 + #endif
662 +  }
663 +  
664 +  
665 +  void ForceMatrixDecomposition::collectData() {
666 +    snap_ = sman_->getCurrentSnapshot();
667 +    storageLayout_ = sman_->getStorageLayout();
668 + #ifdef IS_MPI    
669 +    int n = snap_->atomData.force.size();
670 +    vector<Vector3d> frc_tmp(n, V3Zero);
671 +    
672 +    AtomPlanVectorRow->scatter(atomRowData.force, frc_tmp);
673 +    for (int i = 0; i < n; i++) {
674 +      snap_->atomData.force[i] += frc_tmp[i];
675 +      frc_tmp[i] = 0.0;
676 +    }
677 +    
678 +    AtomPlanVectorColumn->scatter(atomColData.force, frc_tmp);
679 +    for (int i = 0; i < n; i++) {
680 +      snap_->atomData.force[i] += frc_tmp[i];
681 +    }
682 +        
683 +    if (storageLayout_ & DataStorage::dslTorque) {
684  
685 +      int nt = snap_->atomData.torque.size();
686 +      vector<Vector3d> trq_tmp(nt, V3Zero);
687 +
688 +      AtomPlanVectorRow->scatter(atomRowData.torque, trq_tmp);
689 +      for (int i = 0; i < nt; i++) {
690 +        snap_->atomData.torque[i] += trq_tmp[i];
691 +        trq_tmp[i] = 0.0;
692 +      }
693 +      
694 +      AtomPlanVectorColumn->scatter(atomColData.torque, trq_tmp);
695 +      for (int i = 0; i < nt; i++)
696 +        snap_->atomData.torque[i] += trq_tmp[i];
697 +    }
698 +
699 +    if (storageLayout_ & DataStorage::dslSkippedCharge) {
700 +
701 +      int ns = snap_->atomData.skippedCharge.size();
702 +      vector<RealType> skch_tmp(ns, 0.0);
703 +
704 +      AtomPlanRealRow->scatter(atomRowData.skippedCharge, skch_tmp);
705 +      for (int i = 0; i < ns; i++) {
706 +        snap_->atomData.skippedCharge[i] += skch_tmp[i];
707 +        skch_tmp[i] = 0.0;
708 +      }
709 +      
710 +      AtomPlanRealColumn->scatter(atomColData.skippedCharge, skch_tmp);
711 +      for (int i = 0; i < ns; i++)
712 +        snap_->atomData.skippedCharge[i] += skch_tmp[i];
713 +    }
714 +    
715 +    nLocal_ = snap_->getNumberOfAtoms();
716 +
717 +    vector<potVec> pot_temp(nLocal_,
718 +                            Vector<RealType, N_INTERACTION_FAMILIES> (0.0));
719 +
720 +    // scatter/gather pot_row into the members of my column
721 +          
722 +    AtomPlanPotRow->scatter(pot_row, pot_temp);
723 +
724 +    for (int ii = 0;  ii < pot_temp.size(); ii++ )
725 +      pairwisePot += pot_temp[ii];
726 +    
727 +    fill(pot_temp.begin(), pot_temp.end(),
728 +         Vector<RealType, N_INTERACTION_FAMILIES> (0.0));
729 +      
730 +    AtomPlanPotColumn->scatter(pot_col, pot_temp);    
731 +    
732 +    for (int ii = 0;  ii < pot_temp.size(); ii++ )
733 +      pairwisePot += pot_temp[ii];    
734 + #endif
735 +
736 +    cerr << "pairwisePot = " <<  pairwisePot << "\n";
737 +  }
738 +
739 +  int ForceMatrixDecomposition::getNAtomsInRow() {  
740 + #ifdef IS_MPI
741 +    return nAtomsInRow_;
742 + #else
743 +    return nLocal_;
744 + #endif
745 +  }
746 +
747 +  /**
748 +   * returns the list of atoms belonging to this group.  
749 +   */
750 +  vector<int> ForceMatrixDecomposition::getAtomsInGroupRow(int cg1){
751 + #ifdef IS_MPI
752 +    return groupListRow_[cg1];
753 + #else
754 +    return groupList_[cg1];
755 + #endif
756 +  }
757 +
758 +  vector<int> ForceMatrixDecomposition::getAtomsInGroupColumn(int cg2){
759 + #ifdef IS_MPI
760 +    return groupListCol_[cg2];
761 + #else
762 +    return groupList_[cg2];
763 + #endif
764 +  }
765 +  
766 +  Vector3d ForceMatrixDecomposition::getIntergroupVector(int cg1, int cg2){
767 +    Vector3d d;
768 +    
769 + #ifdef IS_MPI
770 +    d = cgColData.position[cg2] - cgRowData.position[cg1];
771 +    cerr << "cg1 = " << cg1 << "\tcg1p = " << cgRowData.position[cg1] << "\n";
772 +    cerr << "cg2 = " << cg2 << "\tcg2p = " << cgColData.position[cg2] << "\n";
773 + #else
774 +    d = snap_->cgData.position[cg2] - snap_->cgData.position[cg1];
775 +    cerr << "cg1 = " << cg1 << "\tcg1p = " << snap_->cgData.position[cg1] << "\n";
776 +    cerr << "cg2 = " << cg2 << "\tcg2p = " << snap_->cgData.position[cg2] << "\n";
777 + #endif
778 +    
779 +    snap_->wrapVector(d);
780 +    return d;    
781 +  }
782 +
783 +
784 +  Vector3d ForceMatrixDecomposition::getAtomToGroupVectorRow(int atom1, int cg1){
785 +
786 +    Vector3d d;
787 +    
788 + #ifdef IS_MPI
789 +    d = cgRowData.position[cg1] - atomRowData.position[atom1];
790 + #else
791 +    d = snap_->cgData.position[cg1] - snap_->atomData.position[atom1];
792 + #endif
793 +
794 +    snap_->wrapVector(d);
795 +    return d;    
796 +  }
797 +  
798 +  Vector3d ForceMatrixDecomposition::getAtomToGroupVectorColumn(int atom2, int cg2){
799 +    Vector3d d;
800 +    
801 + #ifdef IS_MPI
802 +    d = cgColData.position[cg2] - atomColData.position[atom2];
803 + #else
804 +    d = snap_->cgData.position[cg2] - snap_->atomData.position[atom2];
805 + #endif
806 +    
807 +    snap_->wrapVector(d);
808 +    return d;    
809 +  }
810 +
811 +  RealType ForceMatrixDecomposition::getMassFactorRow(int atom1) {
812 + #ifdef IS_MPI
813 +    return massFactorsRow[atom1];
814 + #else
815 +    return massFactors[atom1];
816 + #endif
817 +  }
818 +
819 +  RealType ForceMatrixDecomposition::getMassFactorColumn(int atom2) {
820 + #ifdef IS_MPI
821 +    return massFactorsCol[atom2];
822 + #else
823 +    return massFactors[atom2];
824 + #endif
825 +
826 +  }
827 +    
828 +  Vector3d ForceMatrixDecomposition::getInteratomicVector(int atom1, int atom2){
829 +    Vector3d d;
830 +    
831 + #ifdef IS_MPI
832 +    d = atomColData.position[atom2] - atomRowData.position[atom1];
833 + #else
834 +    d = snap_->atomData.position[atom2] - snap_->atomData.position[atom1];
835 + #endif
836 +
837 +    snap_->wrapVector(d);
838 +    return d;    
839 +  }
840 +
841 +  vector<int> ForceMatrixDecomposition::getExcludesForAtom(int atom1) {
842 +    return excludesForAtom[atom1];
843 +  }
844 +
845 +  /**
846 +   * We need to exclude some overcounted interactions that result from
847 +   * the parallel decomposition.
848 +   */
849 +  bool ForceMatrixDecomposition::skipAtomPair(int atom1, int atom2) {
850 +    int unique_id_1, unique_id_2;
851 +    
852 +
853 +    cerr << "sap with atom1, atom2 =\t" << atom1 << "\t" << atom2 << "\n";
854 + #ifdef IS_MPI
855 +    // in MPI, we have to look up the unique IDs for each atom
856 +    unique_id_1 = AtomRowToGlobal[atom1];
857 +    unique_id_2 = AtomColToGlobal[atom2];
858 +
859 +    cerr << "sap with uid1, uid2 =\t" << unique_id_1 << "\t" << unique_id_2 << "\n";
860 +    // this situation should only arise in MPI simulations
861 +    if (unique_id_1 == unique_id_2) return true;
862 +    
863 +    // this prevents us from doing the pair on multiple processors
864 +    if (unique_id_1 < unique_id_2) {
865 +      if ((unique_id_1 + unique_id_2) % 2 == 0) return true;
866 +    } else {
867 +      if ((unique_id_1 + unique_id_2) % 2 == 1) return true;
868 +    }
869 + #endif
870 +    return false;
871 +  }
872 +
873 +  /**
874 +   * We need to handle the interactions for atoms who are involved in
875 +   * the same rigid body as well as some short range interactions
876 +   * (bonds, bends, torsions) differently from other interactions.
877 +   * We'll still visit the pairwise routines, but with a flag that
878 +   * tells those routines to exclude the pair from direct long range
879 +   * interactions.  Some indirect interactions (notably reaction
880 +   * field) must still be handled for these pairs.
881 +   */
882 +  bool ForceMatrixDecomposition::excludeAtomPair(int atom1, int atom2) {
883 +    int unique_id_2;
884 + #ifdef IS_MPI
885 +    // in MPI, we have to look up the unique IDs for the row atom.
886 +    unique_id_2 = AtomColToGlobal[atom2];
887 + #else
888 +    // in the normal loop, the atom numbers are unique
889 +    unique_id_2 = atom2;
890 + #endif
891 +    
892 +    for (vector<int>::iterator i = excludesForAtom[atom1].begin();
893 +         i != excludesForAtom[atom1].end(); ++i) {
894 +      if ( (*i) == unique_id_2 ) return true;
895 +    }
896 +
897 +    return false;
898 +  }
899 +
900 +
901 +  void ForceMatrixDecomposition::addForceToAtomRow(int atom1, Vector3d fg){
902 + #ifdef IS_MPI
903 +    atomRowData.force[atom1] += fg;
904 + #else
905 +    snap_->atomData.force[atom1] += fg;
906 + #endif
907 +  }
908 +
909 +  void ForceMatrixDecomposition::addForceToAtomColumn(int atom2, Vector3d fg){
910 + #ifdef IS_MPI
911 +    atomColData.force[atom2] += fg;
912 + #else
913 +    snap_->atomData.force[atom2] += fg;
914 + #endif
915 +  }
916 +
917 +    // filling interaction blocks with pointers
918 +  void ForceMatrixDecomposition::fillInteractionData(InteractionData &idat,
919 +                                                     int atom1, int atom2) {
920 +
921 +    idat.excluded = excludeAtomPair(atom1, atom2);
922 +  
923 + #ifdef IS_MPI
924 +    idat.atypes = make_pair( atypesRow[atom1], atypesCol[atom2]);
925 +    //idat.atypes = make_pair( ff_->getAtomType(identsRow[atom1]),
926 +    //                         ff_->getAtomType(identsCol[atom2]) );
927 +    
928 +    if (storageLayout_ & DataStorage::dslAmat) {
929 +      idat.A1 = &(atomRowData.aMat[atom1]);
930 +      idat.A2 = &(atomColData.aMat[atom2]);
931 +    }
932 +    
933 +    if (storageLayout_ & DataStorage::dslElectroFrame) {
934 +      idat.eFrame1 = &(atomRowData.electroFrame[atom1]);
935 +      idat.eFrame2 = &(atomColData.electroFrame[atom2]);
936 +    }
937 +
938 +    if (storageLayout_ & DataStorage::dslTorque) {
939 +      idat.t1 = &(atomRowData.torque[atom1]);
940 +      idat.t2 = &(atomColData.torque[atom2]);
941 +    }
942 +
943 +    if (storageLayout_ & DataStorage::dslDensity) {
944 +      idat.rho1 = &(atomRowData.density[atom1]);
945 +      idat.rho2 = &(atomColData.density[atom2]);
946 +    }
947 +
948 +    if (storageLayout_ & DataStorage::dslFunctional) {
949 +      idat.frho1 = &(atomRowData.functional[atom1]);
950 +      idat.frho2 = &(atomColData.functional[atom2]);
951 +    }
952 +
953 +    if (storageLayout_ & DataStorage::dslFunctionalDerivative) {
954 +      idat.dfrho1 = &(atomRowData.functionalDerivative[atom1]);
955 +      idat.dfrho2 = &(atomColData.functionalDerivative[atom2]);
956 +    }
957 +
958 +    if (storageLayout_ & DataStorage::dslParticlePot) {
959 +      idat.particlePot1 = &(atomRowData.particlePot[atom1]);
960 +      idat.particlePot2 = &(atomColData.particlePot[atom2]);
961 +    }
962 +
963 +    if (storageLayout_ & DataStorage::dslSkippedCharge) {              
964 +      idat.skippedCharge1 = &(atomRowData.skippedCharge[atom1]);
965 +      idat.skippedCharge2 = &(atomColData.skippedCharge[atom2]);
966 +    }
967 +
968 + #else
969 +
970 +    idat.atypes = make_pair( atypesLocal[atom1], atypesLocal[atom2]);
971 +    //idat.atypes = make_pair( ff_->getAtomType(idents[atom1]),
972 +    //                         ff_->getAtomType(idents[atom2]) );
973 +
974 +    if (storageLayout_ & DataStorage::dslAmat) {
975 +      idat.A1 = &(snap_->atomData.aMat[atom1]);
976 +      idat.A2 = &(snap_->atomData.aMat[atom2]);
977 +    }
978 +
979 +    if (storageLayout_ & DataStorage::dslElectroFrame) {
980 +      idat.eFrame1 = &(snap_->atomData.electroFrame[atom1]);
981 +      idat.eFrame2 = &(snap_->atomData.electroFrame[atom2]);
982 +    }
983 +
984 +    if (storageLayout_ & DataStorage::dslTorque) {
985 +      idat.t1 = &(snap_->atomData.torque[atom1]);
986 +      idat.t2 = &(snap_->atomData.torque[atom2]);
987 +    }
988 +
989 +    if (storageLayout_ & DataStorage::dslDensity) {    
990 +      idat.rho1 = &(snap_->atomData.density[atom1]);
991 +      idat.rho2 = &(snap_->atomData.density[atom2]);
992 +    }
993 +
994 +    if (storageLayout_ & DataStorage::dslFunctional) {
995 +      idat.frho1 = &(snap_->atomData.functional[atom1]);
996 +      idat.frho2 = &(snap_->atomData.functional[atom2]);
997 +    }
998 +
999 +    if (storageLayout_ & DataStorage::dslFunctionalDerivative) {
1000 +      idat.dfrho1 = &(snap_->atomData.functionalDerivative[atom1]);
1001 +      idat.dfrho2 = &(snap_->atomData.functionalDerivative[atom2]);
1002 +    }
1003 +
1004 +    if (storageLayout_ & DataStorage::dslParticlePot) {
1005 +      idat.particlePot1 = &(snap_->atomData.particlePot[atom1]);
1006 +      idat.particlePot2 = &(snap_->atomData.particlePot[atom2]);
1007 +    }
1008 +
1009 +    if (storageLayout_ & DataStorage::dslSkippedCharge) {
1010 +      idat.skippedCharge1 = &(snap_->atomData.skippedCharge[atom1]);
1011 +      idat.skippedCharge2 = &(snap_->atomData.skippedCharge[atom2]);
1012 +    }
1013 + #endif
1014 +  }
1015 +
1016 +  
1017 +  void ForceMatrixDecomposition::unpackInteractionData(InteractionData &idat, int atom1, int atom2) {    
1018 + #ifdef IS_MPI
1019 +    pot_row[atom1] += 0.5 *  *(idat.pot);
1020 +    pot_col[atom2] += 0.5 *  *(idat.pot);
1021 +
1022 +    atomRowData.force[atom1] += *(idat.f1);
1023 +    atomColData.force[atom2] -= *(idat.f1);
1024 + #else
1025 +    pairwisePot += *(idat.pot);
1026 +
1027 +    snap_->atomData.force[atom1] += *(idat.f1);
1028 +    snap_->atomData.force[atom2] -= *(idat.f1);
1029 + #endif
1030 +    
1031 +  }
1032 +
1033 +  /*
1034 +   * buildNeighborList
1035 +   *
1036 +   * first element of pair is row-indexed CutoffGroup
1037 +   * second element of pair is column-indexed CutoffGroup
1038 +   */
1039 +  vector<pair<int, int> > ForceMatrixDecomposition::buildNeighborList() {
1040 +      
1041 +    vector<pair<int, int> > neighborList;
1042 +    groupCutoffs cuts;
1043 +    bool doAllPairs = false;
1044 +
1045 + #ifdef IS_MPI
1046 +    cellListRow_.clear();
1047 +    cellListCol_.clear();
1048 + #else
1049 +    cellList_.clear();
1050 + #endif
1051 +
1052 +    RealType rList_ = (largestRcut_ + skinThickness_);
1053 +    RealType rl2 = rList_ * rList_;
1054 +    Snapshot* snap_ = sman_->getCurrentSnapshot();
1055 +    Mat3x3d Hmat = snap_->getHmat();
1056 +    Vector3d Hx = Hmat.getColumn(0);
1057 +    Vector3d Hy = Hmat.getColumn(1);
1058 +    Vector3d Hz = Hmat.getColumn(2);
1059 +
1060 +    nCells_.x() = (int) ( Hx.length() )/ rList_;
1061 +    nCells_.y() = (int) ( Hy.length() )/ rList_;
1062 +    nCells_.z() = (int) ( Hz.length() )/ rList_;
1063 +
1064 +    // handle small boxes where the cell offsets can end up repeating cells
1065 +    
1066 +    if (nCells_.x() < 3) doAllPairs = true;
1067 +    if (nCells_.y() < 3) doAllPairs = true;
1068 +    if (nCells_.z() < 3) doAllPairs = true;
1069 +
1070 +    Mat3x3d invHmat = snap_->getInvHmat();
1071 +    Vector3d rs, scaled, dr;
1072 +    Vector3i whichCell;
1073 +    int cellIndex;
1074 +    int nCtot = nCells_.x() * nCells_.y() * nCells_.z();
1075 +
1076 + #ifdef IS_MPI
1077 +    cellListRow_.resize(nCtot);
1078 +    cellListCol_.resize(nCtot);
1079 + #else
1080 +    cellList_.resize(nCtot);
1081 + #endif
1082 +
1083 +    if (!doAllPairs) {
1084 + #ifdef IS_MPI
1085 +
1086 +      for (int i = 0; i < nGroupsInRow_; i++) {
1087 +        rs = cgRowData.position[i];
1088 +        
1089 +        // scaled positions relative to the box vectors
1090 +        scaled = invHmat * rs;
1091 +        
1092 +        // wrap the vector back into the unit box by subtracting integer box
1093 +        // numbers
1094 +        for (int j = 0; j < 3; j++) {
1095 +          scaled[j] -= roundMe(scaled[j]);
1096 +          scaled[j] += 0.5;
1097 +        }
1098 +        
1099 +        // find xyz-indices of cell that cutoffGroup is in.
1100 +        whichCell.x() = nCells_.x() * scaled.x();
1101 +        whichCell.y() = nCells_.y() * scaled.y();
1102 +        whichCell.z() = nCells_.z() * scaled.z();
1103 +        
1104 +        // find single index of this cell:
1105 +        cellIndex = Vlinear(whichCell, nCells_);
1106 +        
1107 +        // add this cutoff group to the list of groups in this cell;
1108 +        cellListRow_[cellIndex].push_back(i);
1109 +      }
1110 +      for (int i = 0; i < nGroupsInCol_; i++) {
1111 +        rs = cgColData.position[i];
1112 +        
1113 +        // scaled positions relative to the box vectors
1114 +        scaled = invHmat * rs;
1115 +        
1116 +        // wrap the vector back into the unit box by subtracting integer box
1117 +        // numbers
1118 +        for (int j = 0; j < 3; j++) {
1119 +          scaled[j] -= roundMe(scaled[j]);
1120 +          scaled[j] += 0.5;
1121 +        }
1122 +        
1123 +        // find xyz-indices of cell that cutoffGroup is in.
1124 +        whichCell.x() = nCells_.x() * scaled.x();
1125 +        whichCell.y() = nCells_.y() * scaled.y();
1126 +        whichCell.z() = nCells_.z() * scaled.z();
1127 +        
1128 +        // find single index of this cell:
1129 +        cellIndex = Vlinear(whichCell, nCells_);
1130 +        
1131 +        // add this cutoff group to the list of groups in this cell;
1132 +        cellListCol_[cellIndex].push_back(i);
1133 +      }
1134 + #else
1135 +      for (int i = 0; i < nGroups_; i++) {
1136 +        rs = snap_->cgData.position[i];
1137 +        
1138 +        // scaled positions relative to the box vectors
1139 +        scaled = invHmat * rs;
1140 +        
1141 +        // wrap the vector back into the unit box by subtracting integer box
1142 +        // numbers
1143 +        for (int j = 0; j < 3; j++) {
1144 +          scaled[j] -= roundMe(scaled[j]);
1145 +          scaled[j] += 0.5;
1146 +        }
1147 +        
1148 +        // find xyz-indices of cell that cutoffGroup is in.
1149 +        whichCell.x() = nCells_.x() * scaled.x();
1150 +        whichCell.y() = nCells_.y() * scaled.y();
1151 +        whichCell.z() = nCells_.z() * scaled.z();
1152 +        
1153 +        // find single index of this cell:
1154 +        cellIndex = Vlinear(whichCell, nCells_);
1155 +        
1156 +        // add this cutoff group to the list of groups in this cell;
1157 +        cellList_[cellIndex].push_back(i);
1158 +      }
1159 + #endif
1160 +
1161 +      for (int m1z = 0; m1z < nCells_.z(); m1z++) {
1162 +        for (int m1y = 0; m1y < nCells_.y(); m1y++) {
1163 +          for (int m1x = 0; m1x < nCells_.x(); m1x++) {
1164 +            Vector3i m1v(m1x, m1y, m1z);
1165 +            int m1 = Vlinear(m1v, nCells_);
1166 +            
1167 +            for (vector<Vector3i>::iterator os = cellOffsets_.begin();
1168 +                 os != cellOffsets_.end(); ++os) {
1169 +              
1170 +              Vector3i m2v = m1v + (*os);
1171 +              
1172 +              if (m2v.x() >= nCells_.x()) {
1173 +                m2v.x() = 0;          
1174 +              } else if (m2v.x() < 0) {
1175 +                m2v.x() = nCells_.x() - 1;
1176 +              }
1177 +              
1178 +              if (m2v.y() >= nCells_.y()) {
1179 +                m2v.y() = 0;          
1180 +              } else if (m2v.y() < 0) {
1181 +                m2v.y() = nCells_.y() - 1;
1182 +              }
1183 +              
1184 +              if (m2v.z() >= nCells_.z()) {
1185 +                m2v.z() = 0;          
1186 +              } else if (m2v.z() < 0) {
1187 +                m2v.z() = nCells_.z() - 1;
1188 +              }
1189 +              
1190 +              int m2 = Vlinear (m2v, nCells_);
1191 +              
1192 + #ifdef IS_MPI
1193 +              for (vector<int>::iterator j1 = cellListRow_[m1].begin();
1194 +                   j1 != cellListRow_[m1].end(); ++j1) {
1195 +                for (vector<int>::iterator j2 = cellListCol_[m2].begin();
1196 +                     j2 != cellListCol_[m2].end(); ++j2) {
1197 +                  
1198 +                  // In parallel, we need to visit *all* pairs of row &
1199 +                  // column indicies and will truncate later on.
1200 +                  dr = cgColData.position[(*j2)] - cgRowData.position[(*j1)];
1201 +                  snap_->wrapVector(dr);
1202 +                  cuts = getGroupCutoffs( (*j1), (*j2) );
1203 +                  if (dr.lengthSquare() < cuts.third) {
1204 +                    neighborList.push_back(make_pair((*j1), (*j2)));
1205 +                  }                  
1206 +                }
1207 +              }
1208 + #else
1209 +              
1210 +              for (vector<int>::iterator j1 = cellList_[m1].begin();
1211 +                   j1 != cellList_[m1].end(); ++j1) {
1212 +                for (vector<int>::iterator j2 = cellList_[m2].begin();
1213 +                     j2 != cellList_[m2].end(); ++j2) {
1214 +                  
1215 +                  // Always do this if we're in different cells or if
1216 +                  // we're in the same cell and the global index of the
1217 +                  // j2 cutoff group is less than the j1 cutoff group
1218 +                  
1219 +                  if (m2 != m1 || (*j2) < (*j1)) {
1220 +                    dr = snap_->cgData.position[(*j2)] - snap_->cgData.position[(*j1)];
1221 +                    snap_->wrapVector(dr);
1222 +                    cuts = getGroupCutoffs( (*j1), (*j2) );
1223 +                    if (dr.lengthSquare() < cuts.third) {
1224 +                      neighborList.push_back(make_pair((*j1), (*j2)));
1225 +                    }
1226 +                  }
1227 +                }
1228 +              }
1229 + #endif
1230 +            }
1231 +          }
1232 +        }
1233 +      }
1234 +    } else {
1235 +      // branch to do all cutoff group pairs
1236 + #ifdef IS_MPI
1237 +      for (int j1 = 0; j1 < nGroupsInRow_; j1++) {
1238 +        for (int j2 = 0; j2 < nGroupsInCol_; j2++) {      
1239 +          dr = cgColData.position[j2] - cgRowData.position[j1];
1240 +          snap_->wrapVector(dr);
1241 +          cuts = getGroupCutoffs( j1, j2 );
1242 +          if (dr.lengthSquare() < cuts.third) {
1243 +            neighborList.push_back(make_pair(j1, j2));
1244 +          }
1245 +        }
1246 +      }
1247 + #else
1248 +      for (int j1 = 0; j1 < nGroups_ - 1; j1++) {
1249 +        for (int j2 = j1 + 1; j2 < nGroups_; j2++) {
1250 +          dr = snap_->cgData.position[j2] - snap_->cgData.position[j1];
1251 +          snap_->wrapVector(dr);
1252 +          cuts = getGroupCutoffs( j1, j2 );
1253 +          if (dr.lengthSquare() < cuts.third) {
1254 +            neighborList.push_back(make_pair(j1, j2));
1255 +          }
1256 +        }
1257 +      }        
1258 + #endif
1259 +    }
1260 +      
1261 +    // save the local cutoff group positions for the check that is
1262 +    // done on each loop:
1263 +    saved_CG_positions_.clear();
1264 +    for (int i = 0; i < nGroups_; i++)
1265 +      saved_CG_positions_.push_back(snap_->cgData.position[i]);
1266 +    
1267 +    return neighborList;
1268 +  }
1269 + } //end namespace OpenMD

Diff Legend

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