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 1612 by gezelter, Fri Aug 12 19:59:56 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_.clear();
57 >    cellOffsets_.push_back( Vector3i(-1,-1,-1) );
58 >    cellOffsets_.push_back( Vector3i( 0,-1,-1) );
59 >    cellOffsets_.push_back( Vector3i( 1,-1,-1) );                          
60 >    cellOffsets_.push_back( Vector3i(-1, 0,-1) );
61 >    cellOffsets_.push_back( Vector3i( 0, 0,-1) );
62 >    cellOffsets_.push_back( Vector3i( 1, 0,-1) );
63 >    cellOffsets_.push_back( Vector3i(-1, 1,-1) );
64 >    cellOffsets_.push_back( Vector3i( 0, 1,-1) );      
65 >    cellOffsets_.push_back( Vector3i( 1, 1,-1) );
66 >    cellOffsets_.push_back( Vector3i(-1,-1, 0) );
67 >    cellOffsets_.push_back( Vector3i( 0,-1, 0) );
68 >    cellOffsets_.push_back( Vector3i( 1,-1, 0) );
69 >    cellOffsets_.push_back( Vector3i(-1, 0, 0) );      
70 >    cellOffsets_.push_back( Vector3i( 0, 0, 0) );
71 >    cellOffsets_.push_back( Vector3i( 1, 0, 0) );
72 >    cellOffsets_.push_back( Vector3i(-1, 1, 0) );
73 >    cellOffsets_.push_back( Vector3i( 0, 1, 0) );
74 >    cellOffsets_.push_back( Vector3i( 1, 1, 0) );
75 >    cellOffsets_.push_back( Vector3i(-1,-1, 1) );
76 >    cellOffsets_.push_back( Vector3i( 0,-1, 1) );
77 >    cellOffsets_.push_back( Vector3i( 1,-1, 1) );
78 >    cellOffsets_.push_back( Vector3i(-1, 0, 1) );
79 >    cellOffsets_.push_back( Vector3i( 0, 0, 1) );
80 >    cellOffsets_.push_back( Vector3i( 1, 0, 1) );
81 >    cellOffsets_.push_back( Vector3i(-1, 1, 1) );
82 >    cellOffsets_.push_back( Vector3i( 0, 1, 1) );
83 >    cellOffsets_.push_back( Vector3i( 1, 1, 1) );
84 > #endif    
85 >  }
86  
58 #include <iostream>
59 #include <vector>
60 #include <algorithm>
61 #include <cmath>
62 #include "parallel/ForceDecomposition.hpp"
87  
88 +  /**
89 +   * distributeInitialData is essentially a copy of the older fortran
90 +   * SimulationSetup
91 +   */
92 +  void ForceMatrixDecomposition::distributeInitialData() {
93 +    snap_ = sman_->getCurrentSnapshot();
94 +    storageLayout_ = sman_->getStorageLayout();
95 +    ff_ = info_->getForceField();
96 +    nLocal_ = snap_->getNumberOfAtoms();
97 +    
98 +    nGroups_ = info_->getNLocalCutoffGroups();
99 +    // gather the information for atomtype IDs (atids):
100 +    idents = info_->getIdentArray();
101 +    AtomLocalToGlobal = info_->getGlobalAtomIndices();
102 +    cgLocalToGlobal = info_->getGlobalGroupIndices();
103 +    vector<int> globalGroupMembership = info_->getGlobalGroupMembership();
104  
105 < using namespace std;
66 < using namespace OpenMD;
105 >    massFactors = info_->getMassFactors();
106  
107 < //__static
107 >    PairList* excludes = info_->getExcludedInteractions();
108 >    PairList* oneTwo = info_->getOneTwoInteractions();
109 >    PairList* oneThree = info_->getOneThreeInteractions();
110 >    PairList* oneFour = info_->getOneFourInteractions();
111 >
112   #ifdef IS_MPI
113 < static vector<MPI:Comm> communictors;
114 < #endif
113 >
114 >    MPI::Intracomm row = rowComm.getComm();
115 >    MPI::Intracomm col = colComm.getComm();
116  
117 < //____ MPITypeTraits
118 < template<typename T>
119 < struct MPITypeTraits;
117 >    AtomPlanIntRow = new Plan<int>(row, nLocal_);
118 >    AtomPlanRealRow = new Plan<RealType>(row, nLocal_);
119 >    AtomPlanVectorRow = new Plan<Vector3d>(row, nLocal_);
120 >    AtomPlanMatrixRow = new Plan<Mat3x3d>(row, nLocal_);
121 >    AtomPlanPotRow = new Plan<potVec>(row, nLocal_);
122  
123 < #ifdef IS_MPI
124 < template<>
125 < struct MPITypeTraits<RealType> {
126 <  static const MPI::Datatype datatype;
127 < };
82 < const MPI_Datatype MPITypeTraits<RealType>::datatype = MY_MPI_REAL;
123 >    AtomPlanIntColumn = new Plan<int>(col, nLocal_);
124 >    AtomPlanRealColumn = new Plan<RealType>(col, nLocal_);
125 >    AtomPlanVectorColumn = new Plan<Vector3d>(col, nLocal_);
126 >    AtomPlanMatrixColumn = new Plan<Mat3x3d>(col, nLocal_);
127 >    AtomPlanPotColumn = new Plan<potVec>(col, nLocal_);
128  
129 < template<>
130 < struct MPITypeTraits<int> {
131 <  static const MPI::Datatype datatype;
132 < };
88 < const MPI::Datatype MPITypeTraits<int>::datatype = MPI_INT;
89 < #endif
129 >    cgPlanIntRow = new Plan<int>(row, nGroups_);
130 >    cgPlanVectorRow = new Plan<Vector3d>(row, nGroups_);
131 >    cgPlanIntColumn = new Plan<int>(col, nGroups_);
132 >    cgPlanVectorColumn = new Plan<Vector3d>(col, nGroups_);
133  
134 < /**
135 < * Constructor for ForceDecomposition Parallel Decomposition Method
136 < * Will try to construct a symmetric grid of processors. Ideally, the
137 < * number of processors will be a square ex: 4, 9, 16, 25.
95 < *
96 < */
134 >    nAtomsInRow_ = AtomPlanIntRow->getSize();
135 >    nAtomsInCol_ = AtomPlanIntColumn->getSize();
136 >    nGroupsInRow_ = cgPlanIntRow->getSize();
137 >    nGroupsInCol_ = cgPlanIntColumn->getSize();
138  
139 < ForceDecomposition::ForceDecomposition() {
139 >    // Modify the data storage objects with the correct layouts and sizes:
140 >    atomRowData.resize(nAtomsInRow_);
141 >    atomRowData.setStorageLayout(storageLayout_);
142 >    atomColData.resize(nAtomsInCol_);
143 >    atomColData.setStorageLayout(storageLayout_);
144 >    cgRowData.resize(nGroupsInRow_);
145 >    cgRowData.setStorageLayout(DataStorage::dslPosition);
146 >    cgColData.resize(nGroupsInCol_);
147 >    cgColData.setStorageLayout(DataStorage::dslPosition);
148 >        
149 >    identsRow.resize(nAtomsInRow_);
150 >    identsCol.resize(nAtomsInCol_);
151 >    
152 >    AtomPlanIntRow->gather(idents, identsRow);
153 >    AtomPlanIntColumn->gather(idents, identsCol);
154 >    
155 >    // allocate memory for the parallel objects
156 >    atypesRow.resize(nAtomsInRow_);
157 >    atypesCol.resize(nAtomsInCol_);
158  
159 < #ifdef IS_MPI
160 <  int nProcs = MPI::COMM_WORLD.Get_size();
161 <  int worldRank = MPI::COMM_WORLD.Get_rank();
159 >    for (int i = 0; i < nAtomsInRow_; i++)
160 >      atypesRow[i] = ff_->getAtomType(identsRow[i]);
161 >    for (int i = 0; i < nAtomsInCol_; i++)
162 >      atypesCol[i] = ff_->getAtomType(identsCol[i]);        
163 >
164 >    pot_row.resize(nAtomsInRow_);
165 >    pot_col.resize(nAtomsInCol_);
166 >
167 >    AtomRowToGlobal.resize(nAtomsInRow_);
168 >    AtomColToGlobal.resize(nAtomsInCol_);
169 >    AtomPlanIntRow->gather(AtomLocalToGlobal, AtomRowToGlobal);
170 >    AtomPlanIntColumn->gather(AtomLocalToGlobal, AtomColToGlobal);
171 >
172 >    cgRowToGlobal.resize(nGroupsInRow_);
173 >    cgColToGlobal.resize(nGroupsInCol_);
174 >    cgPlanIntRow->gather(cgLocalToGlobal, cgRowToGlobal);
175 >    cgPlanIntColumn->gather(cgLocalToGlobal, cgColToGlobal);
176 >
177 >    massFactorsRow.resize(nAtomsInRow_);
178 >    massFactorsCol.resize(nAtomsInCol_);
179 >    AtomPlanRealRow->gather(massFactors, massFactorsRow);
180 >    AtomPlanRealColumn->gather(massFactors, massFactorsCol);
181 >
182 >    groupListRow_.clear();
183 >    groupListRow_.resize(nGroupsInRow_);
184 >    for (int i = 0; i < nGroupsInRow_; i++) {
185 >      int gid = cgRowToGlobal[i];
186 >      for (int j = 0; j < nAtomsInRow_; j++) {
187 >        int aid = AtomRowToGlobal[j];
188 >        if (globalGroupMembership[aid] == gid)
189 >          groupListRow_[i].push_back(j);
190 >      }      
191 >    }
192 >
193 >    groupListCol_.clear();
194 >    groupListCol_.resize(nGroupsInCol_);
195 >    for (int i = 0; i < nGroupsInCol_; i++) {
196 >      int gid = cgColToGlobal[i];
197 >      for (int j = 0; j < nAtomsInCol_; j++) {
198 >        int aid = AtomColToGlobal[j];
199 >        if (globalGroupMembership[aid] == gid)
200 >          groupListCol_[i].push_back(j);
201 >      }      
202 >    }
203 >
204 >    excludesForAtom.clear();
205 >    excludesForAtom.resize(nAtomsInRow_);
206 >    toposForAtom.clear();
207 >    toposForAtom.resize(nAtomsInRow_);
208 >    topoDist.clear();
209 >    topoDist.resize(nAtomsInRow_);
210 >    for (int i = 0; i < nAtomsInRow_; i++) {
211 >      int iglob = AtomRowToGlobal[i];
212 >
213 >      for (int j = 0; j < nAtomsInCol_; j++) {
214 >        int jglob = AtomColToGlobal[j];
215 >
216 >        if (excludes->hasPair(iglob, jglob))
217 >          excludesForAtom[i].push_back(j);      
218 >        
219 >        if (oneTwo->hasPair(iglob, jglob)) {
220 >          toposForAtom[i].push_back(j);
221 >          topoDist[i].push_back(1);
222 >        } else {
223 >          if (oneThree->hasPair(iglob, jglob)) {
224 >            toposForAtom[i].push_back(j);
225 >            topoDist[i].push_back(2);
226 >          } else {
227 >            if (oneFour->hasPair(iglob, jglob)) {
228 >              toposForAtom[i].push_back(j);
229 >              topoDist[i].push_back(3);
230 >            }
231 >          }
232 >        }
233 >      }      
234 >    }
235 >
236   #endif
237  
238 <  // First time through, construct column stride.
239 <  if (communicators.size() == 0)
240 <  {
241 <    int nColumnsMax = (int) round(sqrt((float) nProcs));
242 <    for (int i = 0; i < nProcs; ++i)
243 <    {
244 <      if (nProcs%i==0) nColumns=i;
238 >    // allocate memory for the parallel objects
239 >    atypesLocal.resize(nLocal_);
240 >
241 >    for (int i = 0; i < nLocal_; i++)
242 >      atypesLocal[i] = ff_->getAtomType(idents[i]);
243 >
244 >    groupList_.clear();
245 >    groupList_.resize(nGroups_);
246 >    for (int i = 0; i < nGroups_; i++) {
247 >      int gid = cgLocalToGlobal[i];
248 >      for (int j = 0; j < nLocal_; j++) {
249 >        int aid = AtomLocalToGlobal[j];
250 >        if (globalGroupMembership[aid] == gid) {
251 >          groupList_[i].push_back(j);
252 >        }
253 >      }      
254      }
255  
256 <    int nRows = nProcs/nColumns;    
257 <    myRank_ = (int) worldRank%nColumns;
258 <  }
259 <  else
260 <  {
261 <    myRank_ = myRank/nColumns;
120 <  }
121 <  MPI::Comm newComm = MPI:COMM_WORLD.Split(myRank_,0);
122 <  
123 <  isColumn_ = false;
124 <  
125 < }
256 >    excludesForAtom.clear();
257 >    excludesForAtom.resize(nLocal_);
258 >    toposForAtom.clear();
259 >    toposForAtom.resize(nLocal_);
260 >    topoDist.clear();
261 >    topoDist.resize(nLocal_);
262  
263 < ForceDecomposition::gather(sendbuf, receivebuf){
264 <  communicators(myIndex_).Allgatherv();
129 < }
263 >    for (int i = 0; i < nLocal_; i++) {
264 >      int iglob = AtomLocalToGlobal[i];
265  
266 +      for (int j = 0; j < nLocal_; j++) {
267 +        int jglob = AtomLocalToGlobal[j];
268  
269 +        if (excludes->hasPair(iglob, jglob))
270 +          excludesForAtom[i].push_back(j);              
271 +        
272 +        if (oneTwo->hasPair(iglob, jglob)) {
273 +          toposForAtom[i].push_back(j);
274 +          topoDist[i].push_back(1);
275 +        } else {
276 +          if (oneThree->hasPair(iglob, jglob)) {
277 +            toposForAtom[i].push_back(j);
278 +            topoDist[i].push_back(2);
279 +          } else {
280 +            if (oneFour->hasPair(iglob, jglob)) {
281 +              toposForAtom[i].push_back(j);
282 +              topoDist[i].push_back(3);
283 +            }
284 +          }
285 +        }
286 +      }      
287 +    }
288 +    
289 +    createGtypeCutoffMap();
290  
291 < ForceDecomposition::scatter(sbuffer, rbuffer){
292 <  communicators(myIndex_).Reduce_scatter(sbuffer, recevbuf. recvcounts, MPI::DOUBLE, MPI::SUM);
293 < }
291 >  }
292 >  
293 >  void ForceMatrixDecomposition::createGtypeCutoffMap() {
294 >    
295 >    RealType tol = 1e-6;
296 >    largestRcut_ = 0.0;
297 >    RealType rc;
298 >    int atid;
299 >    set<AtomType*> atypes = info_->getSimulatedAtomTypes();
300 >    
301 >    map<int, RealType> atypeCutoff;
302 >      
303 >    for (set<AtomType*>::iterator at = atypes.begin();
304 >         at != atypes.end(); ++at){
305 >      atid = (*at)->getIdent();
306 >      if (userChoseCutoff_)
307 >        atypeCutoff[atid] = userCutoff_;
308 >      else
309 >        atypeCutoff[atid] = interactionMan_->getSuggestedCutoffRadius(*at);
310 >    }
311 >    
312 >    vector<RealType> gTypeCutoffs;
313 >    // first we do a single loop over the cutoff groups to find the
314 >    // largest cutoff for any atypes present in this group.
315 > #ifdef IS_MPI
316 >    vector<RealType> groupCutoffRow(nGroupsInRow_, 0.0);
317 >    groupRowToGtype.resize(nGroupsInRow_);
318 >    for (int cg1 = 0; cg1 < nGroupsInRow_; cg1++) {
319 >      vector<int> atomListRow = getAtomsInGroupRow(cg1);
320 >      for (vector<int>::iterator ia = atomListRow.begin();
321 >           ia != atomListRow.end(); ++ia) {            
322 >        int atom1 = (*ia);
323 >        atid = identsRow[atom1];
324 >        if (atypeCutoff[atid] > groupCutoffRow[cg1]) {
325 >          groupCutoffRow[cg1] = atypeCutoff[atid];
326 >        }
327 >      }
328  
329 +      bool gTypeFound = false;
330 +      for (int gt = 0; gt < gTypeCutoffs.size(); gt++) {
331 +        if (abs(groupCutoffRow[cg1] - gTypeCutoffs[gt]) < tol) {
332 +          groupRowToGtype[cg1] = gt;
333 +          gTypeFound = true;
334 +        }
335 +      }
336 +      if (!gTypeFound) {
337 +        gTypeCutoffs.push_back( groupCutoffRow[cg1] );
338 +        groupRowToGtype[cg1] = gTypeCutoffs.size() - 1;
339 +      }
340 +      
341 +    }
342 +    vector<RealType> groupCutoffCol(nGroupsInCol_, 0.0);
343 +    groupColToGtype.resize(nGroupsInCol_);
344 +    for (int cg2 = 0; cg2 < nGroupsInCol_; cg2++) {
345 +      vector<int> atomListCol = getAtomsInGroupColumn(cg2);
346 +      for (vector<int>::iterator jb = atomListCol.begin();
347 +           jb != atomListCol.end(); ++jb) {            
348 +        int atom2 = (*jb);
349 +        atid = identsCol[atom2];
350 +        if (atypeCutoff[atid] > groupCutoffCol[cg2]) {
351 +          groupCutoffCol[cg2] = atypeCutoff[atid];
352 +        }
353 +      }
354 +      bool gTypeFound = false;
355 +      for (int gt = 0; gt < gTypeCutoffs.size(); gt++) {
356 +        if (abs(groupCutoffCol[cg2] - gTypeCutoffs[gt]) < tol) {
357 +          groupColToGtype[cg2] = gt;
358 +          gTypeFound = true;
359 +        }
360 +      }
361 +      if (!gTypeFound) {
362 +        gTypeCutoffs.push_back( groupCutoffCol[cg2] );
363 +        groupColToGtype[cg2] = gTypeCutoffs.size() - 1;
364 +      }
365 +    }
366 + #else
367  
368 +    vector<RealType> groupCutoff(nGroups_, 0.0);
369 +    groupToGtype.resize(nGroups_);
370 +    for (int cg1 = 0; cg1 < nGroups_; cg1++) {
371 +      groupCutoff[cg1] = 0.0;
372 +      vector<int> atomList = getAtomsInGroupRow(cg1);
373 +      for (vector<int>::iterator ia = atomList.begin();
374 +           ia != atomList.end(); ++ia) {            
375 +        int atom1 = (*ia);
376 +        atid = idents[atom1];
377 +        if (atypeCutoff[atid] > groupCutoff[cg1])
378 +          groupCutoff[cg1] = atypeCutoff[atid];
379 +      }
380 +      
381 +      bool gTypeFound = false;
382 +      for (int gt = 0; gt < gTypeCutoffs.size(); gt++) {
383 +        if (abs(groupCutoff[cg1] - gTypeCutoffs[gt]) < tol) {
384 +          groupToGtype[cg1] = gt;
385 +          gTypeFound = true;
386 +        }
387 +      }
388 +      if (!gTypeFound) {      
389 +        gTypeCutoffs.push_back( groupCutoff[cg1] );
390 +        groupToGtype[cg1] = gTypeCutoffs.size() - 1;
391 +      }      
392 +    }
393 + #endif
394 +
395 +    // Now we find the maximum group cutoff value present in the simulation
396 +
397 +    RealType groupMax = *max_element(gTypeCutoffs.begin(),
398 +                                     gTypeCutoffs.end());
399 +
400 + #ifdef IS_MPI
401 +    MPI::COMM_WORLD.Allreduce(&groupMax, &groupMax, 1, MPI::REALTYPE,
402 +                              MPI::MAX);
403 + #endif
404 +    
405 +    RealType tradRcut = groupMax;
406 +
407 +    for (int i = 0; i < gTypeCutoffs.size();  i++) {
408 +      for (int j = 0; j < gTypeCutoffs.size();  j++) {      
409 +        RealType thisRcut;
410 +        switch(cutoffPolicy_) {
411 +        case TRADITIONAL:
412 +          thisRcut = tradRcut;
413 +          break;
414 +        case MIX:
415 +          thisRcut = 0.5 * (gTypeCutoffs[i] + gTypeCutoffs[j]);
416 +          break;
417 +        case MAX:
418 +          thisRcut = max(gTypeCutoffs[i], gTypeCutoffs[j]);
419 +          break;
420 +        default:
421 +          sprintf(painCave.errMsg,
422 +                  "ForceMatrixDecomposition::createGtypeCutoffMap "
423 +                  "hit an unknown cutoff policy!\n");
424 +          painCave.severity = OPENMD_ERROR;
425 +          painCave.isFatal = 1;
426 +          simError();
427 +          break;
428 +        }
429 +
430 +        pair<int,int> key = make_pair(i,j);
431 +        gTypeCutoffMap[key].first = thisRcut;
432 +        if (thisRcut > largestRcut_) largestRcut_ = thisRcut;
433 +        gTypeCutoffMap[key].second = thisRcut*thisRcut;
434 +        gTypeCutoffMap[key].third = pow(thisRcut + skinThickness_, 2);
435 +        // sanity check
436 +        
437 +        if (userChoseCutoff_) {
438 +          if (abs(gTypeCutoffMap[key].first - userCutoff_) > 0.0001) {
439 +            sprintf(painCave.errMsg,
440 +                    "ForceMatrixDecomposition::createGtypeCutoffMap "
441 +                    "user-specified rCut (%lf) does not match computed group Cutoff\n", userCutoff_);
442 +            painCave.severity = OPENMD_ERROR;
443 +            painCave.isFatal = 1;
444 +            simError();            
445 +          }
446 +        }
447 +      }
448 +    }
449 +  }
450 +
451 +
452 +  groupCutoffs ForceMatrixDecomposition::getGroupCutoffs(int cg1, int cg2) {
453 +    int i, j;  
454 + #ifdef IS_MPI
455 +    i = groupRowToGtype[cg1];
456 +    j = groupColToGtype[cg2];
457 + #else
458 +    i = groupToGtype[cg1];
459 +    j = groupToGtype[cg2];
460 + #endif    
461 +    return gTypeCutoffMap[make_pair(i,j)];
462 +  }
463 +
464 +  int ForceMatrixDecomposition::getTopologicalDistance(int atom1, int atom2) {
465 +    for (int j = 0; j < toposForAtom[atom1].size(); j++) {
466 +      if (toposForAtom[atom1][j] == atom2)
467 +        return topoDist[atom1][j];
468 +    }
469 +    return 0;
470 +  }
471 +
472 +  void ForceMatrixDecomposition::zeroWorkArrays() {
473 +    pairwisePot = 0.0;
474 +    embeddingPot = 0.0;
475 +
476 + #ifdef IS_MPI
477 +    if (storageLayout_ & DataStorage::dslForce) {
478 +      fill(atomRowData.force.begin(), atomRowData.force.end(), V3Zero);
479 +      fill(atomColData.force.begin(), atomColData.force.end(), V3Zero);
480 +    }
481 +
482 +    if (storageLayout_ & DataStorage::dslTorque) {
483 +      fill(atomRowData.torque.begin(), atomRowData.torque.end(), V3Zero);
484 +      fill(atomColData.torque.begin(), atomColData.torque.end(), V3Zero);
485 +    }
486 +    
487 +    fill(pot_row.begin(), pot_row.end(),
488 +         Vector<RealType, N_INTERACTION_FAMILIES> (0.0));
489 +
490 +    fill(pot_col.begin(), pot_col.end(),
491 +         Vector<RealType, N_INTERACTION_FAMILIES> (0.0));  
492 +
493 +    if (storageLayout_ & DataStorage::dslParticlePot) {    
494 +      fill(atomRowData.particlePot.begin(), atomRowData.particlePot.end(),
495 +           0.0);
496 +      fill(atomColData.particlePot.begin(), atomColData.particlePot.end(),
497 +           0.0);
498 +    }
499 +
500 +    if (storageLayout_ & DataStorage::dslDensity) {      
501 +      fill(atomRowData.density.begin(), atomRowData.density.end(), 0.0);
502 +      fill(atomColData.density.begin(), atomColData.density.end(), 0.0);
503 +    }
504 +
505 +    if (storageLayout_ & DataStorage::dslFunctional) {  
506 +      fill(atomRowData.functional.begin(), atomRowData.functional.end(),
507 +           0.0);
508 +      fill(atomColData.functional.begin(), atomColData.functional.end(),
509 +           0.0);
510 +    }
511 +
512 +    if (storageLayout_ & DataStorage::dslFunctionalDerivative) {      
513 +      fill(atomRowData.functionalDerivative.begin(),
514 +           atomRowData.functionalDerivative.end(), 0.0);
515 +      fill(atomColData.functionalDerivative.begin(),
516 +           atomColData.functionalDerivative.end(), 0.0);
517 +    }
518 +
519 +    if (storageLayout_ & DataStorage::dslSkippedCharge) {      
520 +      fill(atomRowData.skippedCharge.begin(),
521 +           atomRowData.skippedCharge.end(), 0.0);
522 +      fill(atomColData.skippedCharge.begin(),
523 +           atomColData.skippedCharge.end(), 0.0);
524 +    }
525 +
526 + #endif
527 +    // even in parallel, we need to zero out the local arrays:
528 +
529 +    if (storageLayout_ & DataStorage::dslParticlePot) {      
530 +      fill(snap_->atomData.particlePot.begin(),
531 +           snap_->atomData.particlePot.end(), 0.0);
532 +    }
533 +    
534 +    if (storageLayout_ & DataStorage::dslDensity) {      
535 +      fill(snap_->atomData.density.begin(),
536 +           snap_->atomData.density.end(), 0.0);
537 +    }
538 +    if (storageLayout_ & DataStorage::dslFunctional) {
539 +      fill(snap_->atomData.functional.begin(),
540 +           snap_->atomData.functional.end(), 0.0);
541 +    }
542 +    if (storageLayout_ & DataStorage::dslFunctionalDerivative) {      
543 +      fill(snap_->atomData.functionalDerivative.begin(),
544 +           snap_->atomData.functionalDerivative.end(), 0.0);
545 +    }
546 +    if (storageLayout_ & DataStorage::dslSkippedCharge) {      
547 +      fill(snap_->atomData.skippedCharge.begin(),
548 +           snap_->atomData.skippedCharge.end(), 0.0);
549 +    }
550 +    
551 +  }
552 +
553 +
554 +  void ForceMatrixDecomposition::distributeData()  {
555 +    snap_ = sman_->getCurrentSnapshot();
556 +    storageLayout_ = sman_->getStorageLayout();
557 + #ifdef IS_MPI
558 +    
559 +    // gather up the atomic positions
560 +    AtomPlanVectorRow->gather(snap_->atomData.position,
561 +                              atomRowData.position);
562 +    AtomPlanVectorColumn->gather(snap_->atomData.position,
563 +                                 atomColData.position);
564 +    
565 +    // gather up the cutoff group positions
566 +
567 +    cgPlanVectorRow->gather(snap_->cgData.position,
568 +                            cgRowData.position);
569 +
570 +    cgPlanVectorColumn->gather(snap_->cgData.position,
571 +                               cgColData.position);
572 +
573 +    
574 +    // if needed, gather the atomic rotation matrices
575 +    if (storageLayout_ & DataStorage::dslAmat) {
576 +      AtomPlanMatrixRow->gather(snap_->atomData.aMat,
577 +                                atomRowData.aMat);
578 +      AtomPlanMatrixColumn->gather(snap_->atomData.aMat,
579 +                                   atomColData.aMat);
580 +    }
581 +    
582 +    // if needed, gather the atomic eletrostatic frames
583 +    if (storageLayout_ & DataStorage::dslElectroFrame) {
584 +      AtomPlanMatrixRow->gather(snap_->atomData.electroFrame,
585 +                                atomRowData.electroFrame);
586 +      AtomPlanMatrixColumn->gather(snap_->atomData.electroFrame,
587 +                                   atomColData.electroFrame);
588 +    }
589 +
590 + #endif      
591 +  }
592 +  
593 +  /* collects information obtained during the pre-pair loop onto local
594 +   * data structures.
595 +   */
596 +  void ForceMatrixDecomposition::collectIntermediateData() {
597 +    snap_ = sman_->getCurrentSnapshot();
598 +    storageLayout_ = sman_->getStorageLayout();
599 + #ifdef IS_MPI
600 +    
601 +    if (storageLayout_ & DataStorage::dslDensity) {
602 +      
603 +      AtomPlanRealRow->scatter(atomRowData.density,
604 +                               snap_->atomData.density);
605 +      
606 +      int n = snap_->atomData.density.size();
607 +      vector<RealType> rho_tmp(n, 0.0);
608 +      AtomPlanRealColumn->scatter(atomColData.density, rho_tmp);
609 +      for (int i = 0; i < n; i++)
610 +        snap_->atomData.density[i] += rho_tmp[i];
611 +    }
612 + #endif
613 +  }
614 +
615 +  /*
616 +   * redistributes information obtained during the pre-pair loop out to
617 +   * row and column-indexed data structures
618 +   */
619 +  void ForceMatrixDecomposition::distributeIntermediateData() {
620 +    snap_ = sman_->getCurrentSnapshot();
621 +    storageLayout_ = sman_->getStorageLayout();
622 + #ifdef IS_MPI
623 +    if (storageLayout_ & DataStorage::dslFunctional) {
624 +      AtomPlanRealRow->gather(snap_->atomData.functional,
625 +                              atomRowData.functional);
626 +      AtomPlanRealColumn->gather(snap_->atomData.functional,
627 +                                 atomColData.functional);
628 +    }
629 +    
630 +    if (storageLayout_ & DataStorage::dslFunctionalDerivative) {
631 +      AtomPlanRealRow->gather(snap_->atomData.functionalDerivative,
632 +                              atomRowData.functionalDerivative);
633 +      AtomPlanRealColumn->gather(snap_->atomData.functionalDerivative,
634 +                                 atomColData.functionalDerivative);
635 +    }
636 + #endif
637 +  }
638 +  
639 +  
640 +  void ForceMatrixDecomposition::collectData() {
641 +    snap_ = sman_->getCurrentSnapshot();
642 +    storageLayout_ = sman_->getStorageLayout();
643 + #ifdef IS_MPI    
644 +    int n = snap_->atomData.force.size();
645 +    vector<Vector3d> frc_tmp(n, V3Zero);
646 +    
647 +    AtomPlanVectorRow->scatter(atomRowData.force, frc_tmp);
648 +    for (int i = 0; i < n; i++) {
649 +      snap_->atomData.force[i] += frc_tmp[i];
650 +      frc_tmp[i] = 0.0;
651 +    }
652 +    
653 +    AtomPlanVectorColumn->scatter(atomColData.force, frc_tmp);
654 +    for (int i = 0; i < n; i++) {
655 +      snap_->atomData.force[i] += frc_tmp[i];
656 +    }
657 +        
658 +    if (storageLayout_ & DataStorage::dslTorque) {
659 +
660 +      int nt = snap_->atomData.torque.size();
661 +      vector<Vector3d> trq_tmp(nt, V3Zero);
662 +
663 +      AtomPlanVectorRow->scatter(atomRowData.torque, trq_tmp);
664 +      for (int i = 0; i < nt; i++) {
665 +        snap_->atomData.torque[i] += trq_tmp[i];
666 +        trq_tmp[i] = 0.0;
667 +      }
668 +      
669 +      AtomPlanVectorColumn->scatter(atomColData.torque, trq_tmp);
670 +      for (int i = 0; i < nt; i++)
671 +        snap_->atomData.torque[i] += trq_tmp[i];
672 +    }
673 +
674 +    if (storageLayout_ & DataStorage::dslSkippedCharge) {
675 +
676 +      int ns = snap_->atomData.skippedCharge.size();
677 +      vector<RealType> skch_tmp(ns, 0.0);
678 +
679 +      AtomPlanRealRow->scatter(atomRowData.skippedCharge, skch_tmp);
680 +      for (int i = 0; i < ns; i++) {
681 +        snap_->atomData.skippedCharge[i] += skch_tmp[i];
682 +        skch_tmp[i] = 0.0;
683 +      }
684 +      
685 +      AtomPlanRealColumn->scatter(atomColData.skippedCharge, skch_tmp);
686 +      for (int i = 0; i < ns; i++)
687 +        snap_->atomData.skippedCharge[i] += skch_tmp[i];
688 +    }
689 +    
690 +    nLocal_ = snap_->getNumberOfAtoms();
691 +
692 +    vector<potVec> pot_temp(nLocal_,
693 +                            Vector<RealType, N_INTERACTION_FAMILIES> (0.0));
694 +
695 +    // scatter/gather pot_row into the members of my column
696 +          
697 +    AtomPlanPotRow->scatter(pot_row, pot_temp);
698 +
699 +    for (int ii = 0;  ii < pot_temp.size(); ii++ )
700 +      pairwisePot += pot_temp[ii];
701 +    
702 +    fill(pot_temp.begin(), pot_temp.end(),
703 +         Vector<RealType, N_INTERACTION_FAMILIES> (0.0));
704 +      
705 +    AtomPlanPotColumn->scatter(pot_col, pot_temp);    
706 +    
707 +    for (int ii = 0;  ii < pot_temp.size(); ii++ )
708 +      pairwisePot += pot_temp[ii];    
709 +    
710 +    for (int ii = 0; ii < N_INTERACTION_FAMILIES; ii++) {
711 +      RealType ploc1 = pairwisePot[ii];
712 +      RealType ploc2 = 0.0;
713 +      MPI::COMM_WORLD.Allreduce(&ploc1, &ploc2, 1, MPI::REALTYPE, MPI::SUM);
714 +      pairwisePot[ii] = ploc2;
715 +    }
716 +
717 + #endif
718 +
719 +  }
720 +
721 +  int ForceMatrixDecomposition::getNAtomsInRow() {  
722 + #ifdef IS_MPI
723 +    return nAtomsInRow_;
724 + #else
725 +    return nLocal_;
726 + #endif
727 +  }
728 +
729 +  /**
730 +   * returns the list of atoms belonging to this group.  
731 +   */
732 +  vector<int> ForceMatrixDecomposition::getAtomsInGroupRow(int cg1){
733 + #ifdef IS_MPI
734 +    return groupListRow_[cg1];
735 + #else
736 +    return groupList_[cg1];
737 + #endif
738 +  }
739 +
740 +  vector<int> ForceMatrixDecomposition::getAtomsInGroupColumn(int cg2){
741 + #ifdef IS_MPI
742 +    return groupListCol_[cg2];
743 + #else
744 +    return groupList_[cg2];
745 + #endif
746 +  }
747 +  
748 +  Vector3d ForceMatrixDecomposition::getIntergroupVector(int cg1, int cg2){
749 +    Vector3d d;
750 +    
751 + #ifdef IS_MPI
752 +    d = cgColData.position[cg2] - cgRowData.position[cg1];
753 + #else
754 +    d = snap_->cgData.position[cg2] - snap_->cgData.position[cg1];
755 + #endif
756 +    
757 +    snap_->wrapVector(d);
758 +    return d;    
759 +  }
760 +
761 +
762 +  Vector3d ForceMatrixDecomposition::getAtomToGroupVectorRow(int atom1, int cg1){
763 +
764 +    Vector3d d;
765 +    
766 + #ifdef IS_MPI
767 +    d = cgRowData.position[cg1] - atomRowData.position[atom1];
768 + #else
769 +    d = snap_->cgData.position[cg1] - snap_->atomData.position[atom1];
770 + #endif
771 +
772 +    snap_->wrapVector(d);
773 +    return d;    
774 +  }
775 +  
776 +  Vector3d ForceMatrixDecomposition::getAtomToGroupVectorColumn(int atom2, int cg2){
777 +    Vector3d d;
778 +    
779 + #ifdef IS_MPI
780 +    d = cgColData.position[cg2] - atomColData.position[atom2];
781 + #else
782 +    d = snap_->cgData.position[cg2] - snap_->atomData.position[atom2];
783 + #endif
784 +    
785 +    snap_->wrapVector(d);
786 +    return d;    
787 +  }
788 +
789 +  RealType ForceMatrixDecomposition::getMassFactorRow(int atom1) {
790 + #ifdef IS_MPI
791 +    return massFactorsRow[atom1];
792 + #else
793 +    return massFactors[atom1];
794 + #endif
795 +  }
796 +
797 +  RealType ForceMatrixDecomposition::getMassFactorColumn(int atom2) {
798 + #ifdef IS_MPI
799 +    return massFactorsCol[atom2];
800 + #else
801 +    return massFactors[atom2];
802 + #endif
803 +
804 +  }
805 +    
806 +  Vector3d ForceMatrixDecomposition::getInteratomicVector(int atom1, int atom2){
807 +    Vector3d d;
808 +    
809 + #ifdef IS_MPI
810 +    d = atomColData.position[atom2] - atomRowData.position[atom1];
811 + #else
812 +    d = snap_->atomData.position[atom2] - snap_->atomData.position[atom1];
813 + #endif
814 +
815 +    snap_->wrapVector(d);
816 +    return d;    
817 +  }
818 +
819 +  vector<int> ForceMatrixDecomposition::getExcludesForAtom(int atom1) {
820 +    return excludesForAtom[atom1];
821 +  }
822 +
823 +  /**
824 +   * We need to exclude some overcounted interactions that result from
825 +   * the parallel decomposition.
826 +   */
827 +  bool ForceMatrixDecomposition::skipAtomPair(int atom1, int atom2) {
828 +    int unique_id_1, unique_id_2;
829 +    
830 + #ifdef IS_MPI
831 +    // in MPI, we have to look up the unique IDs for each atom
832 +    unique_id_1 = AtomRowToGlobal[atom1];
833 +    unique_id_2 = AtomColToGlobal[atom2];
834 +
835 +    // this situation should only arise in MPI simulations
836 +    if (unique_id_1 == unique_id_2) return true;
837 +    
838 +    // this prevents us from doing the pair on multiple processors
839 +    if (unique_id_1 < unique_id_2) {
840 +      if ((unique_id_1 + unique_id_2) % 2 == 0) return true;
841 +    } else {
842 +      if ((unique_id_1 + unique_id_2) % 2 == 1) return true;
843 +    }
844 + #endif
845 +    return false;
846 +  }
847 +
848 +  /**
849 +   * We need to handle the interactions for atoms who are involved in
850 +   * the same rigid body as well as some short range interactions
851 +   * (bonds, bends, torsions) differently from other interactions.
852 +   * We'll still visit the pairwise routines, but with a flag that
853 +   * tells those routines to exclude the pair from direct long range
854 +   * interactions.  Some indirect interactions (notably reaction
855 +   * field) must still be handled for these pairs.
856 +   */
857 +  bool ForceMatrixDecomposition::excludeAtomPair(int atom1, int atom2) {
858 +    int unique_id_2;
859 + #ifdef IS_MPI
860 +    // in MPI, we have to look up the unique IDs for the row atom.
861 +    unique_id_2 = AtomColToGlobal[atom2];
862 + #else
863 +    // in the normal loop, the atom numbers are unique
864 +    unique_id_2 = atom2;
865 + #endif
866 +    
867 +    for (vector<int>::iterator i = excludesForAtom[atom1].begin();
868 +         i != excludesForAtom[atom1].end(); ++i) {
869 +      if ( (*i) == unique_id_2 ) return true;
870 +    }
871 +
872 +    return false;
873 +  }
874 +
875 +
876 +  void ForceMatrixDecomposition::addForceToAtomRow(int atom1, Vector3d fg){
877 + #ifdef IS_MPI
878 +    atomRowData.force[atom1] += fg;
879 + #else
880 +    snap_->atomData.force[atom1] += fg;
881 + #endif
882 +  }
883 +
884 +  void ForceMatrixDecomposition::addForceToAtomColumn(int atom2, Vector3d fg){
885 + #ifdef IS_MPI
886 +    atomColData.force[atom2] += fg;
887 + #else
888 +    snap_->atomData.force[atom2] += fg;
889 + #endif
890 +  }
891 +
892 +    // filling interaction blocks with pointers
893 +  void ForceMatrixDecomposition::fillInteractionData(InteractionData &idat,
894 +                                                     int atom1, int atom2) {
895 +
896 +    idat.excluded = excludeAtomPair(atom1, atom2);
897 +  
898 + #ifdef IS_MPI
899 +    idat.atypes = make_pair( atypesRow[atom1], atypesCol[atom2]);
900 +    //idat.atypes = make_pair( ff_->getAtomType(identsRow[atom1]),
901 +    //                         ff_->getAtomType(identsCol[atom2]) );
902 +    
903 +    if (storageLayout_ & DataStorage::dslAmat) {
904 +      idat.A1 = &(atomRowData.aMat[atom1]);
905 +      idat.A2 = &(atomColData.aMat[atom2]);
906 +    }
907 +    
908 +    if (storageLayout_ & DataStorage::dslElectroFrame) {
909 +      idat.eFrame1 = &(atomRowData.electroFrame[atom1]);
910 +      idat.eFrame2 = &(atomColData.electroFrame[atom2]);
911 +    }
912 +
913 +    if (storageLayout_ & DataStorage::dslTorque) {
914 +      idat.t1 = &(atomRowData.torque[atom1]);
915 +      idat.t2 = &(atomColData.torque[atom2]);
916 +    }
917 +
918 +    if (storageLayout_ & DataStorage::dslDensity) {
919 +      idat.rho1 = &(atomRowData.density[atom1]);
920 +      idat.rho2 = &(atomColData.density[atom2]);
921 +    }
922 +
923 +    if (storageLayout_ & DataStorage::dslFunctional) {
924 +      idat.frho1 = &(atomRowData.functional[atom1]);
925 +      idat.frho2 = &(atomColData.functional[atom2]);
926 +    }
927 +
928 +    if (storageLayout_ & DataStorage::dslFunctionalDerivative) {
929 +      idat.dfrho1 = &(atomRowData.functionalDerivative[atom1]);
930 +      idat.dfrho2 = &(atomColData.functionalDerivative[atom2]);
931 +    }
932 +
933 +    if (storageLayout_ & DataStorage::dslParticlePot) {
934 +      idat.particlePot1 = &(atomRowData.particlePot[atom1]);
935 +      idat.particlePot2 = &(atomColData.particlePot[atom2]);
936 +    }
937 +
938 +    if (storageLayout_ & DataStorage::dslSkippedCharge) {              
939 +      idat.skippedCharge1 = &(atomRowData.skippedCharge[atom1]);
940 +      idat.skippedCharge2 = &(atomColData.skippedCharge[atom2]);
941 +    }
942 +
943 + #else
944 +
945 +    idat.atypes = make_pair( atypesLocal[atom1], atypesLocal[atom2]);
946 +    //idat.atypes = make_pair( ff_->getAtomType(idents[atom1]),
947 +    //                         ff_->getAtomType(idents[atom2]) );
948 +
949 +    if (storageLayout_ & DataStorage::dslAmat) {
950 +      idat.A1 = &(snap_->atomData.aMat[atom1]);
951 +      idat.A2 = &(snap_->atomData.aMat[atom2]);
952 +    }
953 +
954 +    if (storageLayout_ & DataStorage::dslElectroFrame) {
955 +      idat.eFrame1 = &(snap_->atomData.electroFrame[atom1]);
956 +      idat.eFrame2 = &(snap_->atomData.electroFrame[atom2]);
957 +    }
958 +
959 +    if (storageLayout_ & DataStorage::dslTorque) {
960 +      idat.t1 = &(snap_->atomData.torque[atom1]);
961 +      idat.t2 = &(snap_->atomData.torque[atom2]);
962 +    }
963 +
964 +    if (storageLayout_ & DataStorage::dslDensity) {    
965 +      idat.rho1 = &(snap_->atomData.density[atom1]);
966 +      idat.rho2 = &(snap_->atomData.density[atom2]);
967 +    }
968 +
969 +    if (storageLayout_ & DataStorage::dslFunctional) {
970 +      idat.frho1 = &(snap_->atomData.functional[atom1]);
971 +      idat.frho2 = &(snap_->atomData.functional[atom2]);
972 +    }
973 +
974 +    if (storageLayout_ & DataStorage::dslFunctionalDerivative) {
975 +      idat.dfrho1 = &(snap_->atomData.functionalDerivative[atom1]);
976 +      idat.dfrho2 = &(snap_->atomData.functionalDerivative[atom2]);
977 +    }
978 +
979 +    if (storageLayout_ & DataStorage::dslParticlePot) {
980 +      idat.particlePot1 = &(snap_->atomData.particlePot[atom1]);
981 +      idat.particlePot2 = &(snap_->atomData.particlePot[atom2]);
982 +    }
983 +
984 +    if (storageLayout_ & DataStorage::dslSkippedCharge) {
985 +      idat.skippedCharge1 = &(snap_->atomData.skippedCharge[atom1]);
986 +      idat.skippedCharge2 = &(snap_->atomData.skippedCharge[atom2]);
987 +    }
988 + #endif
989 +  }
990 +
991 +  
992 +  void ForceMatrixDecomposition::unpackInteractionData(InteractionData &idat, int atom1, int atom2) {    
993 + #ifdef IS_MPI
994 +    pot_row[atom1] += 0.5 *  *(idat.pot);
995 +    pot_col[atom2] += 0.5 *  *(idat.pot);
996 +
997 +    atomRowData.force[atom1] += *(idat.f1);
998 +    atomColData.force[atom2] -= *(idat.f1);
999 + #else
1000 +    pairwisePot += *(idat.pot);
1001 +
1002 +    snap_->atomData.force[atom1] += *(idat.f1);
1003 +    snap_->atomData.force[atom2] -= *(idat.f1);
1004 + #endif
1005 +    
1006 +  }
1007 +
1008 +  /*
1009 +   * buildNeighborList
1010 +   *
1011 +   * first element of pair is row-indexed CutoffGroup
1012 +   * second element of pair is column-indexed CutoffGroup
1013 +   */
1014 +  vector<pair<int, int> > ForceMatrixDecomposition::buildNeighborList() {
1015 +      
1016 +    vector<pair<int, int> > neighborList;
1017 +    groupCutoffs cuts;
1018 +    bool doAllPairs = false;
1019 +
1020 + #ifdef IS_MPI
1021 +    cellListRow_.clear();
1022 +    cellListCol_.clear();
1023 + #else
1024 +    cellList_.clear();
1025 + #endif
1026 +
1027 +    RealType rList_ = (largestRcut_ + skinThickness_);
1028 +    RealType rl2 = rList_ * rList_;
1029 +    Snapshot* snap_ = sman_->getCurrentSnapshot();
1030 +    Mat3x3d Hmat = snap_->getHmat();
1031 +    Vector3d Hx = Hmat.getColumn(0);
1032 +    Vector3d Hy = Hmat.getColumn(1);
1033 +    Vector3d Hz = Hmat.getColumn(2);
1034 +
1035 +    nCells_.x() = (int) ( Hx.length() )/ rList_;
1036 +    nCells_.y() = (int) ( Hy.length() )/ rList_;
1037 +    nCells_.z() = (int) ( Hz.length() )/ rList_;
1038 +
1039 +    // handle small boxes where the cell offsets can end up repeating cells
1040 +    
1041 +    if (nCells_.x() < 3) doAllPairs = true;
1042 +    if (nCells_.y() < 3) doAllPairs = true;
1043 +    if (nCells_.z() < 3) doAllPairs = true;
1044 +
1045 +    Mat3x3d invHmat = snap_->getInvHmat();
1046 +    Vector3d rs, scaled, dr;
1047 +    Vector3i whichCell;
1048 +    int cellIndex;
1049 +    int nCtot = nCells_.x() * nCells_.y() * nCells_.z();
1050 +
1051 + #ifdef IS_MPI
1052 +    cellListRow_.resize(nCtot);
1053 +    cellListCol_.resize(nCtot);
1054 + #else
1055 +    cellList_.resize(nCtot);
1056 + #endif
1057 +
1058 +    if (!doAllPairs) {
1059 + #ifdef IS_MPI
1060 +
1061 +      for (int i = 0; i < nGroupsInRow_; i++) {
1062 +        rs = cgRowData.position[i];
1063 +        
1064 +        // scaled positions relative to the box vectors
1065 +        scaled = invHmat * rs;
1066 +        
1067 +        // wrap the vector back into the unit box by subtracting integer box
1068 +        // numbers
1069 +        for (int j = 0; j < 3; j++) {
1070 +          scaled[j] -= roundMe(scaled[j]);
1071 +          scaled[j] += 0.5;
1072 +        }
1073 +        
1074 +        // find xyz-indices of cell that cutoffGroup is in.
1075 +        whichCell.x() = nCells_.x() * scaled.x();
1076 +        whichCell.y() = nCells_.y() * scaled.y();
1077 +        whichCell.z() = nCells_.z() * scaled.z();
1078 +        
1079 +        // find single index of this cell:
1080 +        cellIndex = Vlinear(whichCell, nCells_);
1081 +        
1082 +        // add this cutoff group to the list of groups in this cell;
1083 +        cellListRow_[cellIndex].push_back(i);
1084 +      }
1085 +      for (int i = 0; i < nGroupsInCol_; i++) {
1086 +        rs = cgColData.position[i];
1087 +        
1088 +        // scaled positions relative to the box vectors
1089 +        scaled = invHmat * rs;
1090 +        
1091 +        // wrap the vector back into the unit box by subtracting integer box
1092 +        // numbers
1093 +        for (int j = 0; j < 3; j++) {
1094 +          scaled[j] -= roundMe(scaled[j]);
1095 +          scaled[j] += 0.5;
1096 +        }
1097 +        
1098 +        // find xyz-indices of cell that cutoffGroup is in.
1099 +        whichCell.x() = nCells_.x() * scaled.x();
1100 +        whichCell.y() = nCells_.y() * scaled.y();
1101 +        whichCell.z() = nCells_.z() * scaled.z();
1102 +        
1103 +        // find single index of this cell:
1104 +        cellIndex = Vlinear(whichCell, nCells_);
1105 +        
1106 +        // add this cutoff group to the list of groups in this cell;
1107 +        cellListCol_[cellIndex].push_back(i);
1108 +      }
1109 +    
1110 + #else
1111 +      for (int i = 0; i < nGroups_; i++) {
1112 +        rs = snap_->cgData.position[i];
1113 +        
1114 +        // scaled positions relative to the box vectors
1115 +        scaled = invHmat * rs;
1116 +        
1117 +        // wrap the vector back into the unit box by subtracting integer box
1118 +        // numbers
1119 +        for (int j = 0; j < 3; j++) {
1120 +          scaled[j] -= roundMe(scaled[j]);
1121 +          scaled[j] += 0.5;
1122 +        }
1123 +        
1124 +        // find xyz-indices of cell that cutoffGroup is in.
1125 +        whichCell.x() = nCells_.x() * scaled.x();
1126 +        whichCell.y() = nCells_.y() * scaled.y();
1127 +        whichCell.z() = nCells_.z() * scaled.z();
1128 +        
1129 +        // find single index of this cell:
1130 +        cellIndex = Vlinear(whichCell, nCells_);
1131 +        
1132 +        // add this cutoff group to the list of groups in this cell;
1133 +        cellList_[cellIndex].push_back(i);
1134 +      }
1135 +
1136 + #endif
1137 +
1138 +      for (int m1z = 0; m1z < nCells_.z(); m1z++) {
1139 +        for (int m1y = 0; m1y < nCells_.y(); m1y++) {
1140 +          for (int m1x = 0; m1x < nCells_.x(); m1x++) {
1141 +            Vector3i m1v(m1x, m1y, m1z);
1142 +            int m1 = Vlinear(m1v, nCells_);
1143 +            
1144 +            for (vector<Vector3i>::iterator os = cellOffsets_.begin();
1145 +                 os != cellOffsets_.end(); ++os) {
1146 +              
1147 +              Vector3i m2v = m1v + (*os);
1148 +            
1149 +
1150 +              if (m2v.x() >= nCells_.x()) {
1151 +                m2v.x() = 0;          
1152 +              } else if (m2v.x() < 0) {
1153 +                m2v.x() = nCells_.x() - 1;
1154 +              }
1155 +              
1156 +              if (m2v.y() >= nCells_.y()) {
1157 +                m2v.y() = 0;          
1158 +              } else if (m2v.y() < 0) {
1159 +                m2v.y() = nCells_.y() - 1;
1160 +              }
1161 +              
1162 +              if (m2v.z() >= nCells_.z()) {
1163 +                m2v.z() = 0;          
1164 +              } else if (m2v.z() < 0) {
1165 +                m2v.z() = nCells_.z() - 1;
1166 +              }
1167 +
1168 +              int m2 = Vlinear (m2v, nCells_);
1169 +              
1170 + #ifdef IS_MPI
1171 +              for (vector<int>::iterator j1 = cellListRow_[m1].begin();
1172 +                   j1 != cellListRow_[m1].end(); ++j1) {
1173 +                for (vector<int>::iterator j2 = cellListCol_[m2].begin();
1174 +                     j2 != cellListCol_[m2].end(); ++j2) {
1175 +                  
1176 +                  // In parallel, we need to visit *all* pairs of row
1177 +                  // & column indicies and will divide labor in the
1178 +                  // force evaluation later.
1179 +                  dr = cgColData.position[(*j2)] - cgRowData.position[(*j1)];
1180 +                  snap_->wrapVector(dr);
1181 +                  cuts = getGroupCutoffs( (*j1), (*j2) );
1182 +                  if (dr.lengthSquare() < cuts.third) {
1183 +                    neighborList.push_back(make_pair((*j1), (*j2)));
1184 +                  }                  
1185 +                }
1186 +              }
1187 + #else
1188 +              
1189 +              for (vector<int>::iterator j1 = cellList_[m1].begin();
1190 +                   j1 != cellList_[m1].end(); ++j1) {
1191 +                for (vector<int>::iterator j2 = cellList_[m2].begin();
1192 +                     j2 != cellList_[m2].end(); ++j2) {
1193 +                  
1194 +                  // Always do this if we're in different cells or if
1195 +                  // we're in the same cell and the global index of the
1196 +                  // j2 cutoff group is less than the j1 cutoff group
1197 +                  
1198 +                  if (m2 != m1 || (*j2) < (*j1)) {
1199 +                    dr = snap_->cgData.position[(*j2)] - snap_->cgData.position[(*j1)];
1200 +                    snap_->wrapVector(dr);
1201 +                    cuts = getGroupCutoffs( (*j1), (*j2) );
1202 +                    if (dr.lengthSquare() < cuts.third) {
1203 +                      neighborList.push_back(make_pair((*j1), (*j2)));
1204 +                    }
1205 +                  }
1206 +                }
1207 +              }
1208 + #endif
1209 +            }
1210 +          }
1211 +        }
1212 +      }
1213 +    } else {
1214 +      // branch to do all cutoff group pairs
1215 + #ifdef IS_MPI
1216 +      for (int j1 = 0; j1 < nGroupsInRow_; j1++) {
1217 +        for (int j2 = 0; j2 < nGroupsInCol_; j2++) {      
1218 +          dr = cgColData.position[j2] - cgRowData.position[j1];
1219 +          snap_->wrapVector(dr);
1220 +          cuts = getGroupCutoffs( j1, j2 );
1221 +          if (dr.lengthSquare() < cuts.third) {
1222 +            neighborList.push_back(make_pair(j1, j2));
1223 +          }
1224 +        }
1225 +      }
1226 + #else
1227 +      for (int j1 = 0; j1 < nGroups_ - 1; j1++) {
1228 +        for (int j2 = j1 + 1; j2 < nGroups_; j2++) {
1229 +          dr = snap_->cgData.position[j2] - snap_->cgData.position[j1];
1230 +          snap_->wrapVector(dr);
1231 +          cuts = getGroupCutoffs( j1, j2 );
1232 +          if (dr.lengthSquare() < cuts.third) {
1233 +            neighborList.push_back(make_pair(j1, j2));
1234 +          }
1235 +        }
1236 +      }        
1237 + #endif
1238 +    }
1239 +      
1240 +    // save the local cutoff group positions for the check that is
1241 +    // done on each loop:
1242 +    saved_CG_positions_.clear();
1243 +    for (int i = 0; i < nGroups_; i++)
1244 +      saved_CG_positions_.push_back(snap_->cgData.position[i]);
1245 +    
1246 +    return neighborList;
1247 +  }
1248 + } //end namespace OpenMD

Diff Legend

Removed lines
+ Added lines
< Changed lines
> Changed lines