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 1547 by gezelter, Mon Apr 11 18:44:16 2011 UTC vs.
branches/development/src/parallel/ForceMatrixDecomposition.cpp (file contents), Revision 1601 by gezelter, Thu Aug 4 20:04:35 2011 UTC

# Line 38 | Line 38
38   * [3]  Sun, Lin & Gezelter, J. Chem. Phys. 128, 24107 (2008).          
39   * [4]  Vardeman & Gezelter, in progress (2009).                        
40   */
41 < #include "parallel/ForceDecomposition.hpp"
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 +    // 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 +    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 +
72 +
73    /**
74     * distributeInitialData is essentially a copy of the older fortran
75     * SimulationSetup
76     */
77 <  
78 <  void ForceDecomposition::distributeInitialData() {
79 < #ifdef IS_MPI    
80 <    Snapshot* snap = sman_->getCurrentSnapshot();
81 <    int nLocal = snap->getNumberOfAtoms();
82 <    int nGroups = snap->getNumberOfCutoffGroups();
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 <    AtomCommIntI = new Communicator<Row,int>(nLocal);
61 <    AtomCommRealI = new Communicator<Row,RealType>(nLocal);
62 <    AtomCommVectorI = new Communicator<Row,Vector3d>(nLocal);
63 <    AtomCommMatrixI = new Communicator<Row,Mat3x3d>(nLocal);
90 >    massFactors = info_->getMassFactors();
91  
92 <    AtomCommIntJ = new Communicator<Column,int>(nLocal);
93 <    AtomCommRealJ = new Communicator<Column,RealType>(nLocal);
94 <    AtomCommVectorJ = new Communicator<Column,Vector3d>(nLocal);
95 <    AtomCommMatrixJ = new Communicator<Column,Mat3x3d>(nLocal);
92 >    PairList* excludes = info_->getExcludedInteractions();
93 >    PairList* oneTwo = info_->getOneTwoInteractions();
94 >    PairList* oneThree = info_->getOneThreeInteractions();
95 >    PairList* oneFour = info_->getOneFourInteractions();
96  
97 <    cgCommIntI = new Communicator<Row,int>(nGroups);
98 <    cgCommVectorI = new Communicator<Row,Vector3d>(nGroups);
99 <    cgCommIntJ = new Communicator<Column,int>(nGroups);
100 <    cgCommVectorJ = new Communicator<Column,Vector3d>(nGroups);
97 > #ifdef IS_MPI
98 >
99 >    MPI::Intracomm row = rowComm.getComm();
100 >    MPI::Intracomm col = colComm.getComm();
101  
102 <    int nAtomsInRow = AtomCommIntI->getSize();
103 <    int nAtomsInCol = AtomCommIntJ->getSize();
104 <    int nGroupsInRow = cgCommIntI->getSize();
105 <    int nGroupsInCol = cgCommIntJ->getSize();
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 <    vector<vector<RealType> > pot_row(N_INTERACTION_FAMILIES,
109 <                                      vector<RealType> (nAtomsInRow, 0.0));
110 <    vector<vector<RealType> > pot_col(N_INTERACTION_FAMILIES,
111 <                                      vector<RealType> (nAtomsInCol, 0.0));
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 >    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 >    nAtomsInRow_ = AtomPlanIntRow->getSize();
120 >    nAtomsInCol_ = AtomPlanIntColumn->getSize();
121 >    nGroupsInRow_ = cgPlanIntRow->getSize();
122 >    nGroupsInCol_ = cgPlanIntColumn->getSize();
123 >
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 <    vector<RealType> pot_local(N_INTERACTION_FAMILIES, 0.0);
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 <    // gather the information for atomtype IDs (atids):
145 <    vector<int> identsLocal = info_->getIdentArray();
146 <    identsRow.reserve(nAtomsInRow);
147 <    identsCol.reserve(nAtomsInCol);
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 <    AtomCommIntI->gather(identsLocal, identsRow);
150 <    AtomCommIntJ->gather(identsLocal, identsCol);
149 >    pot_row.resize(nAtomsInRow_);
150 >    pot_col.resize(nAtomsInCol_);
151  
152 <    AtomLocalToGlobal = info_->getLocalToGlobalAtomIndex();
153 <    AtomCommIntI->gather(AtomLocalToGlobal, AtomRowToGlobal);
154 <    AtomCommIntJ->gather(AtomLocalToGlobal, AtomColToGlobal);
152 >    AtomRowToGlobal.resize(nAtomsInRow_);
153 >    AtomColToGlobal.resize(nAtomsInCol_);
154 >    AtomPlanIntRow->gather(AtomLocalToGlobal, AtomRowToGlobal);
155 >    AtomPlanIntColumn->gather(AtomLocalToGlobal, AtomColToGlobal);
156  
157 <    cgLocalToGlobal = info_->getLocalToGlobalCutoffGroupIndex();
158 <    cgCommIntI->gather(cgLocalToGlobal, cgRowToGlobal);
159 <    cgCommIntJ->gather(cgLocalToGlobal, cgColToGlobal);
157 >    cgRowToGlobal.resize(nGroupsInRow_);
158 >    cgColToGlobal.resize(nGroupsInCol_);
159 >    cgPlanIntRow->gather(cgLocalToGlobal, cgRowToGlobal);
160 >    cgPlanIntColumn->gather(cgLocalToGlobal, cgColToGlobal);
161  
162 <    
162 >    massFactorsRow.resize(nAtomsInRow_);
163 >    massFactorsCol.resize(nAtomsInCol_);
164 >    AtomPlanRealRow->gather(massFactors, massFactorsRow);
165 >    AtomPlanRealColumn->gather(massFactors, massFactorsCol);
166  
167 <    // still need:
168 <    // topoDist
169 <    // exclude
167 >    groupListRow_.clear();
168 >    groupListRow_.resize(nGroupsInRow_);
169 >    for (int i = 0; i < nGroupsInRow_; i++) {
170 >      int gid = cgRowToGlobal[i];
171 >      for (int j = 0; j < nAtomsInRow_; j++) {
172 >        int aid = AtomRowToGlobal[j];
173 >        if (globalGroupMembership[aid] == gid)
174 >          groupListRow_[i].push_back(j);
175 >      }      
176 >    }
177 >
178 >    groupListCol_.clear();
179 >    groupListCol_.resize(nGroupsInCol_);
180 >    for (int i = 0; i < nGroupsInCol_; i++) {
181 >      int gid = cgColToGlobal[i];
182 >      for (int j = 0; j < nAtomsInCol_; j++) {
183 >        int aid = AtomColToGlobal[j];
184 >        if (globalGroupMembership[aid] == gid)
185 >          groupListCol_[i].push_back(j);
186 >      }      
187 >    }
188 >
189 >    excludesForAtom.clear();
190 >    excludesForAtom.resize(nAtomsInRow_);
191 >    toposForAtom.clear();
192 >    toposForAtom.resize(nAtomsInRow_);
193 >    topoDist.clear();
194 >    topoDist.resize(nAtomsInRow_);
195 >    for (int i = 0; i < nAtomsInRow_; i++) {
196 >      int iglob = AtomRowToGlobal[i];
197 >
198 >      for (int j = 0; j < nAtomsInCol_; j++) {
199 >        int jglob = AtomColToGlobal[j];
200 >
201 >        if (excludes->hasPair(iglob, jglob))
202 >          excludesForAtom[i].push_back(j);      
203 >        
204 >        if (oneTwo->hasPair(iglob, jglob)) {
205 >          toposForAtom[i].push_back(j);
206 >          topoDist[i].push_back(1);
207 >        } else {
208 >          if (oneThree->hasPair(iglob, jglob)) {
209 >            toposForAtom[i].push_back(j);
210 >            topoDist[i].push_back(2);
211 >          } else {
212 >            if (oneFour->hasPair(iglob, jglob)) {
213 >              toposForAtom[i].push_back(j);
214 >              topoDist[i].push_back(3);
215 >            }
216 >          }
217 >        }
218 >      }      
219 >    }
220 >
221   #endif
222 +
223 +    // allocate memory for the parallel objects
224 +    atypesLocal.resize(nLocal_);
225 +
226 +    for (int i = 0; i < nLocal_; i++)
227 +      atypesLocal[i] = ff_->getAtomType(idents[i]);
228 +
229 +    groupList_.clear();
230 +    groupList_.resize(nGroups_);
231 +    for (int i = 0; i < nGroups_; i++) {
232 +      int gid = cgLocalToGlobal[i];
233 +      for (int j = 0; j < nLocal_; j++) {
234 +        int aid = AtomLocalToGlobal[j];
235 +        if (globalGroupMembership[aid] == gid) {
236 +          groupList_[i].push_back(j);
237 +        }
238 +      }      
239 +    }
240 +
241 +    excludesForAtom.clear();
242 +    excludesForAtom.resize(nLocal_);
243 +    toposForAtom.clear();
244 +    toposForAtom.resize(nLocal_);
245 +    topoDist.clear();
246 +    topoDist.resize(nLocal_);
247 +
248 +    for (int i = 0; i < nLocal_; i++) {
249 +      int iglob = AtomLocalToGlobal[i];
250 +
251 +      for (int j = 0; j < nLocal_; j++) {
252 +        int jglob = AtomLocalToGlobal[j];
253 +
254 +        if (excludes->hasPair(iglob, jglob))
255 +          excludesForAtom[i].push_back(j);              
256 +        
257 +        if (oneTwo->hasPair(iglob, jglob)) {
258 +          toposForAtom[i].push_back(j);
259 +          topoDist[i].push_back(1);
260 +        } else {
261 +          if (oneThree->hasPair(iglob, jglob)) {
262 +            toposForAtom[i].push_back(j);
263 +            topoDist[i].push_back(2);
264 +          } else {
265 +            if (oneFour->hasPair(iglob, jglob)) {
266 +              toposForAtom[i].push_back(j);
267 +              topoDist[i].push_back(3);
268 +            }
269 +          }
270 +        }
271 +      }      
272 +    }
273 +    
274 +    createGtypeCutoffMap();
275 +
276    }
277 +  
278 +  void ForceMatrixDecomposition::createGtypeCutoffMap() {
279      
280 +    RealType tol = 1e-6;
281 +    largestRcut_ = 0.0;
282 +    RealType rc;
283 +    int atid;
284 +    set<AtomType*> atypes = info_->getSimulatedAtomTypes();
285 +    
286 +    map<int, RealType> atypeCutoff;
287 +      
288 +    for (set<AtomType*>::iterator at = atypes.begin();
289 +         at != atypes.end(); ++at){
290 +      atid = (*at)->getIdent();
291 +      if (userChoseCutoff_)
292 +        atypeCutoff[atid] = userCutoff_;
293 +      else
294 +        atypeCutoff[atid] = interactionMan_->getSuggestedCutoffRadius(*at);
295 +    }
296 +    
297 +    vector<RealType> gTypeCutoffs;
298 +    // first we do a single loop over the cutoff groups to find the
299 +    // largest cutoff for any atypes present in this group.
300 + #ifdef IS_MPI
301 +    vector<RealType> groupCutoffRow(nGroupsInRow_, 0.0);
302 +    groupRowToGtype.resize(nGroupsInRow_);
303 +    for (int cg1 = 0; cg1 < nGroupsInRow_; cg1++) {
304 +      vector<int> atomListRow = getAtomsInGroupRow(cg1);
305 +      for (vector<int>::iterator ia = atomListRow.begin();
306 +           ia != atomListRow.end(); ++ia) {            
307 +        int atom1 = (*ia);
308 +        atid = identsRow[atom1];
309 +        if (atypeCutoff[atid] > groupCutoffRow[cg1]) {
310 +          groupCutoffRow[cg1] = atypeCutoff[atid];
311 +        }
312 +      }
313  
314 +      bool gTypeFound = false;
315 +      for (int gt = 0; gt < gTypeCutoffs.size(); gt++) {
316 +        if (abs(groupCutoffRow[cg1] - gTypeCutoffs[gt]) < tol) {
317 +          groupRowToGtype[cg1] = gt;
318 +          gTypeFound = true;
319 +        }
320 +      }
321 +      if (!gTypeFound) {
322 +        gTypeCutoffs.push_back( groupCutoffRow[cg1] );
323 +        groupRowToGtype[cg1] = gTypeCutoffs.size() - 1;
324 +      }
325 +      
326 +    }
327 +    vector<RealType> groupCutoffCol(nGroupsInCol_, 0.0);
328 +    groupColToGtype.resize(nGroupsInCol_);
329 +    for (int cg2 = 0; cg2 < nGroupsInCol_; cg2++) {
330 +      vector<int> atomListCol = getAtomsInGroupColumn(cg2);
331 +      for (vector<int>::iterator jb = atomListCol.begin();
332 +           jb != atomListCol.end(); ++jb) {            
333 +        int atom2 = (*jb);
334 +        atid = identsCol[atom2];
335 +        if (atypeCutoff[atid] > groupCutoffCol[cg2]) {
336 +          groupCutoffCol[cg2] = atypeCutoff[atid];
337 +        }
338 +      }
339 +      bool gTypeFound = false;
340 +      for (int gt = 0; gt < gTypeCutoffs.size(); gt++) {
341 +        if (abs(groupCutoffCol[cg2] - gTypeCutoffs[gt]) < tol) {
342 +          groupColToGtype[cg2] = gt;
343 +          gTypeFound = true;
344 +        }
345 +      }
346 +      if (!gTypeFound) {
347 +        gTypeCutoffs.push_back( groupCutoffCol[cg2] );
348 +        groupColToGtype[cg2] = gTypeCutoffs.size() - 1;
349 +      }
350 +    }
351 + #else
352  
353 <  void ForceDecomposition::distributeData()  {
353 >    vector<RealType> groupCutoff(nGroups_, 0.0);
354 >    groupToGtype.resize(nGroups_);
355 >    for (int cg1 = 0; cg1 < nGroups_; cg1++) {
356 >      groupCutoff[cg1] = 0.0;
357 >      vector<int> atomList = getAtomsInGroupRow(cg1);
358 >      for (vector<int>::iterator ia = atomList.begin();
359 >           ia != atomList.end(); ++ia) {            
360 >        int atom1 = (*ia);
361 >        atid = idents[atom1];
362 >        if (atypeCutoff[atid] > groupCutoff[cg1])
363 >          groupCutoff[cg1] = atypeCutoff[atid];
364 >      }
365 >      
366 >      bool gTypeFound = false;
367 >      for (int gt = 0; gt < gTypeCutoffs.size(); gt++) {
368 >        if (abs(groupCutoff[cg1] - gTypeCutoffs[gt]) < tol) {
369 >          groupToGtype[cg1] = gt;
370 >          gTypeFound = true;
371 >        }
372 >      }
373 >      if (!gTypeFound) {      
374 >        gTypeCutoffs.push_back( groupCutoff[cg1] );
375 >        groupToGtype[cg1] = gTypeCutoffs.size() - 1;
376 >      }      
377 >    }
378 > #endif
379 >
380 >    // Now we find the maximum group cutoff value present in the simulation
381 >
382 >    RealType groupMax = *max_element(gTypeCutoffs.begin(),
383 >                                     gTypeCutoffs.end());
384 >
385   #ifdef IS_MPI
386 <    Snapshot* snap = sman_->getCurrentSnapshot();
386 >    MPI::COMM_WORLD.Allreduce(&groupMax, &groupMax, 1, MPI::REALTYPE,
387 >                              MPI::MAX);
388 > #endif
389      
390 +    RealType tradRcut = groupMax;
391 +
392 +    for (int i = 0; i < gTypeCutoffs.size();  i++) {
393 +      for (int j = 0; j < gTypeCutoffs.size();  j++) {      
394 +        RealType thisRcut;
395 +        switch(cutoffPolicy_) {
396 +        case TRADITIONAL:
397 +          thisRcut = tradRcut;
398 +          break;
399 +        case MIX:
400 +          thisRcut = 0.5 * (gTypeCutoffs[i] + gTypeCutoffs[j]);
401 +          break;
402 +        case MAX:
403 +          thisRcut = max(gTypeCutoffs[i], gTypeCutoffs[j]);
404 +          break;
405 +        default:
406 +          sprintf(painCave.errMsg,
407 +                  "ForceMatrixDecomposition::createGtypeCutoffMap "
408 +                  "hit an unknown cutoff policy!\n");
409 +          painCave.severity = OPENMD_ERROR;
410 +          painCave.isFatal = 1;
411 +          simError();
412 +          break;
413 +        }
414 +
415 +        pair<int,int> key = make_pair(i,j);
416 +        gTypeCutoffMap[key].first = thisRcut;
417 +        if (thisRcut > largestRcut_) largestRcut_ = thisRcut;
418 +        gTypeCutoffMap[key].second = thisRcut*thisRcut;
419 +        gTypeCutoffMap[key].third = pow(thisRcut + skinThickness_, 2);
420 +        // sanity check
421 +        
422 +        if (userChoseCutoff_) {
423 +          if (abs(gTypeCutoffMap[key].first - userCutoff_) > 0.0001) {
424 +            sprintf(painCave.errMsg,
425 +                    "ForceMatrixDecomposition::createGtypeCutoffMap "
426 +                    "user-specified rCut (%lf) does not match computed group Cutoff\n", userCutoff_);
427 +            painCave.severity = OPENMD_ERROR;
428 +            painCave.isFatal = 1;
429 +            simError();            
430 +          }
431 +        }
432 +      }
433 +    }
434 +  }
435 +
436 +
437 +  groupCutoffs ForceMatrixDecomposition::getGroupCutoffs(int cg1, int cg2) {
438 +    int i, j;  
439 + #ifdef IS_MPI
440 +    i = groupRowToGtype[cg1];
441 +    j = groupColToGtype[cg2];
442 + #else
443 +    i = groupToGtype[cg1];
444 +    j = groupToGtype[cg2];
445 + #endif    
446 +    return gTypeCutoffMap[make_pair(i,j)];
447 +  }
448 +
449 +  int ForceMatrixDecomposition::getTopologicalDistance(int atom1, int atom2) {
450 +    for (int j = 0; j < toposForAtom[atom1].size(); j++) {
451 +      if (toposForAtom[atom1][j] == atom2)
452 +        return topoDist[atom1][j];
453 +    }
454 +    return 0;
455 +  }
456 +
457 +  void ForceMatrixDecomposition::zeroWorkArrays() {
458 +    pairwisePot = 0.0;
459 +    embeddingPot = 0.0;
460 +
461 + #ifdef IS_MPI
462 +    if (storageLayout_ & DataStorage::dslForce) {
463 +      fill(atomRowData.force.begin(), atomRowData.force.end(), V3Zero);
464 +      fill(atomColData.force.begin(), atomColData.force.end(), V3Zero);
465 +    }
466 +
467 +    if (storageLayout_ & DataStorage::dslTorque) {
468 +      fill(atomRowData.torque.begin(), atomRowData.torque.end(), V3Zero);
469 +      fill(atomColData.torque.begin(), atomColData.torque.end(), V3Zero);
470 +    }
471 +    
472 +    fill(pot_row.begin(), pot_row.end(),
473 +         Vector<RealType, N_INTERACTION_FAMILIES> (0.0));
474 +
475 +    fill(pot_col.begin(), pot_col.end(),
476 +         Vector<RealType, N_INTERACTION_FAMILIES> (0.0));  
477 +
478 +    if (storageLayout_ & DataStorage::dslParticlePot) {    
479 +      fill(atomRowData.particlePot.begin(), atomRowData.particlePot.end(),
480 +           0.0);
481 +      fill(atomColData.particlePot.begin(), atomColData.particlePot.end(),
482 +           0.0);
483 +    }
484 +
485 +    if (storageLayout_ & DataStorage::dslDensity) {      
486 +      fill(atomRowData.density.begin(), atomRowData.density.end(), 0.0);
487 +      fill(atomColData.density.begin(), atomColData.density.end(), 0.0);
488 +    }
489 +
490 +    if (storageLayout_ & DataStorage::dslFunctional) {  
491 +      fill(atomRowData.functional.begin(), atomRowData.functional.end(),
492 +           0.0);
493 +      fill(atomColData.functional.begin(), atomColData.functional.end(),
494 +           0.0);
495 +    }
496 +
497 +    if (storageLayout_ & DataStorage::dslFunctionalDerivative) {      
498 +      fill(atomRowData.functionalDerivative.begin(),
499 +           atomRowData.functionalDerivative.end(), 0.0);
500 +      fill(atomColData.functionalDerivative.begin(),
501 +           atomColData.functionalDerivative.end(), 0.0);
502 +    }
503 +
504 +    if (storageLayout_ & DataStorage::dslSkippedCharge) {      
505 +      fill(atomRowData.skippedCharge.begin(),
506 +           atomRowData.skippedCharge.end(), 0.0);
507 +      fill(atomColData.skippedCharge.begin(),
508 +           atomColData.skippedCharge.end(), 0.0);
509 +    }
510 +
511 + #endif
512 +    // even in parallel, we need to zero out the local arrays:
513 +
514 +    if (storageLayout_ & DataStorage::dslParticlePot) {      
515 +      fill(snap_->atomData.particlePot.begin(),
516 +           snap_->atomData.particlePot.end(), 0.0);
517 +    }
518 +    
519 +    if (storageLayout_ & DataStorage::dslDensity) {      
520 +      fill(snap_->atomData.density.begin(),
521 +           snap_->atomData.density.end(), 0.0);
522 +    }
523 +    if (storageLayout_ & DataStorage::dslFunctional) {
524 +      fill(snap_->atomData.functional.begin(),
525 +           snap_->atomData.functional.end(), 0.0);
526 +    }
527 +    if (storageLayout_ & DataStorage::dslFunctionalDerivative) {      
528 +      fill(snap_->atomData.functionalDerivative.begin(),
529 +           snap_->atomData.functionalDerivative.end(), 0.0);
530 +    }
531 +    if (storageLayout_ & DataStorage::dslSkippedCharge) {      
532 +      fill(snap_->atomData.skippedCharge.begin(),
533 +           snap_->atomData.skippedCharge.end(), 0.0);
534 +    }
535 +    
536 +  }
537 +
538 +
539 +  void ForceMatrixDecomposition::distributeData()  {
540 +    snap_ = sman_->getCurrentSnapshot();
541 +    storageLayout_ = sman_->getStorageLayout();
542 + #ifdef IS_MPI
543 +    
544      // gather up the atomic positions
545 <    AtomCommVectorI->gather(snap->atomData.position,
546 <                            snap->atomIData.position);
547 <    AtomCommVectorJ->gather(snap->atomData.position,
548 <                            snap->atomJData.position);
545 >    AtomPlanVectorRow->gather(snap_->atomData.position,
546 >                              atomRowData.position);
547 >    AtomPlanVectorColumn->gather(snap_->atomData.position,
548 >                                 atomColData.position);
549      
550      // gather up the cutoff group positions
551 <    cgCommVectorI->gather(snap->cgData.position,
552 <                          snap->cgIData.position);
553 <    cgCommVectorJ->gather(snap->cgData.position,
554 <                          snap->cgJData.position);
551 >
552 >    cgPlanVectorRow->gather(snap_->cgData.position,
553 >                            cgRowData.position);
554 >
555 >    cgPlanVectorColumn->gather(snap_->cgData.position,
556 >                               cgColData.position);
557 >
558      
559      // if needed, gather the atomic rotation matrices
560 <    if (snap->atomData.getStorageLayout() & DataStorage::dslAmat) {
561 <      AtomCommMatrixI->gather(snap->atomData.aMat,
562 <                              snap->atomIData.aMat);
563 <      AtomCommMatrixJ->gather(snap->atomData.aMat,
564 <                              snap->atomJData.aMat);
560 >    if (storageLayout_ & DataStorage::dslAmat) {
561 >      AtomPlanMatrixRow->gather(snap_->atomData.aMat,
562 >                                atomRowData.aMat);
563 >      AtomPlanMatrixColumn->gather(snap_->atomData.aMat,
564 >                                   atomColData.aMat);
565      }
566      
567      // if needed, gather the atomic eletrostatic frames
568 <    if (snap->atomData.getStorageLayout() & DataStorage::dslElectroFrame) {
569 <      AtomCommMatrixI->gather(snap->atomData.electroFrame,
570 <                              snap->atomIData.electroFrame);
571 <      AtomCommMatrixJ->gather(snap->atomData.electroFrame,
572 <                              snap->atomJData.electroFrame);
568 >    if (storageLayout_ & DataStorage::dslElectroFrame) {
569 >      AtomPlanMatrixRow->gather(snap_->atomData.electroFrame,
570 >                                atomRowData.electroFrame);
571 >      AtomPlanMatrixColumn->gather(snap_->atomData.electroFrame,
572 >                                   atomColData.electroFrame);
573      }
574 +
575   #endif      
576    }
577    
578 <  void ForceDecomposition::collectIntermediateData() {
578 >  /* collects information obtained during the pre-pair loop onto local
579 >   * data structures.
580 >   */
581 >  void ForceMatrixDecomposition::collectIntermediateData() {
582 >    snap_ = sman_->getCurrentSnapshot();
583 >    storageLayout_ = sman_->getStorageLayout();
584   #ifdef IS_MPI
149 –    Snapshot* snap = sman_->getCurrentSnapshot();
585      
586 <    if (snap->atomData.getStorageLayout() & DataStorage::dslDensity) {
587 <
588 <      AtomCommRealI->scatter(snap->atomIData.density,
589 <                             snap->atomData.density);
590 <
591 <      int n = snap->atomData.density.size();
592 <      std::vector<RealType> rho_tmp(n, 0.0);
593 <      AtomCommRealJ->scatter(snap->atomJData.density, rho_tmp);
586 >    if (storageLayout_ & DataStorage::dslDensity) {
587 >      
588 >      AtomPlanRealRow->scatter(atomRowData.density,
589 >                               snap_->atomData.density);
590 >      
591 >      int n = snap_->atomData.density.size();
592 >      vector<RealType> rho_tmp(n, 0.0);
593 >      AtomPlanRealColumn->scatter(atomColData.density, rho_tmp);
594        for (int i = 0; i < n; i++)
595 <        snap->atomData.density[i] += rho_tmp[i];
595 >        snap_->atomData.density[i] += rho_tmp[i];
596      }
597   #endif
598    }
599 <  
600 <  void ForceDecomposition::distributeIntermediateData() {
599 >
600 >  /*
601 >   * redistributes information obtained during the pre-pair loop out to
602 >   * row and column-indexed data structures
603 >   */
604 >  void ForceMatrixDecomposition::distributeIntermediateData() {
605 >    snap_ = sman_->getCurrentSnapshot();
606 >    storageLayout_ = sman_->getStorageLayout();
607   #ifdef IS_MPI
608 <    Snapshot* snap = sman_->getCurrentSnapshot();
609 <    if (snap->atomData.getStorageLayout() & DataStorage::dslFunctional) {
610 <      AtomCommRealI->gather(snap->atomData.functional,
611 <                            snap->atomIData.functional);
612 <      AtomCommRealJ->gather(snap->atomData.functional,
172 <                            snap->atomJData.functional);
608 >    if (storageLayout_ & DataStorage::dslFunctional) {
609 >      AtomPlanRealRow->gather(snap_->atomData.functional,
610 >                              atomRowData.functional);
611 >      AtomPlanRealColumn->gather(snap_->atomData.functional,
612 >                                 atomColData.functional);
613      }
614      
615 <    if (snap->atomData.getStorageLayout() & DataStorage::dslFunctionalDerivative) {
616 <      AtomCommRealI->gather(snap->atomData.functionalDerivative,
617 <                            snap->atomIData.functionalDerivative);
618 <      AtomCommRealJ->gather(snap->atomData.functionalDerivative,
619 <                            snap->atomJData.functionalDerivative);
615 >    if (storageLayout_ & DataStorage::dslFunctionalDerivative) {
616 >      AtomPlanRealRow->gather(snap_->atomData.functionalDerivative,
617 >                              atomRowData.functionalDerivative);
618 >      AtomPlanRealColumn->gather(snap_->atomData.functionalDerivative,
619 >                                 atomColData.functionalDerivative);
620      }
621   #endif
622    }
623    
624    
625 <  void ForceDecomposition::collectData() {
626 < #ifdef IS_MPI
627 <    Snapshot* snap = sman_->getCurrentSnapshot();
628 <    
629 <    int n = snap->atomData.force.size();
625 >  void ForceMatrixDecomposition::collectData() {
626 >    snap_ = sman_->getCurrentSnapshot();
627 >    storageLayout_ = sman_->getStorageLayout();
628 > #ifdef IS_MPI    
629 >    int n = snap_->atomData.force.size();
630      vector<Vector3d> frc_tmp(n, V3Zero);
631      
632 <    AtomCommVectorI->scatter(snap->atomIData.force, frc_tmp);
632 >    AtomPlanVectorRow->scatter(atomRowData.force, frc_tmp);
633      for (int i = 0; i < n; i++) {
634 <      snap->atomData.force[i] += frc_tmp[i];
634 >      snap_->atomData.force[i] += frc_tmp[i];
635        frc_tmp[i] = 0.0;
636      }
637      
638 <    AtomCommVectorJ->scatter(snap->atomJData.force, frc_tmp);
639 <    for (int i = 0; i < n; i++)
640 <      snap->atomData.force[i] += frc_tmp[i];
641 <    
642 <    
643 <    if (snap->atomData.getStorageLayout() & DataStorage::dslTorque) {
638 >    AtomPlanVectorColumn->scatter(atomColData.force, frc_tmp);
639 >    for (int i = 0; i < n; i++) {
640 >      snap_->atomData.force[i] += frc_tmp[i];
641 >    }
642 >        
643 >    if (storageLayout_ & DataStorage::dslTorque) {
644  
645 <      int nt = snap->atomData.force.size();
645 >      int nt = snap_->atomData.torque.size();
646        vector<Vector3d> trq_tmp(nt, V3Zero);
647  
648 <      AtomCommVectorI->scatter(snap->atomIData.torque, trq_tmp);
649 <      for (int i = 0; i < n; i++) {
650 <        snap->atomData.torque[i] += trq_tmp[i];
648 >      AtomPlanVectorRow->scatter(atomRowData.torque, trq_tmp);
649 >      for (int i = 0; i < nt; i++) {
650 >        snap_->atomData.torque[i] += trq_tmp[i];
651          trq_tmp[i] = 0.0;
652        }
653        
654 <      AtomCommVectorJ->scatter(snap->atomJData.torque, trq_tmp);
655 <      for (int i = 0; i < n; i++)
656 <        snap->atomData.torque[i] += trq_tmp[i];
654 >      AtomPlanVectorColumn->scatter(atomColData.torque, trq_tmp);
655 >      for (int i = 0; i < nt; i++)
656 >        snap_->atomData.torque[i] += trq_tmp[i];
657      }
658 +
659 +    if (storageLayout_ & DataStorage::dslSkippedCharge) {
660 +
661 +      int ns = snap_->atomData.skippedCharge.size();
662 +      vector<RealType> skch_tmp(ns, 0.0);
663 +
664 +      AtomPlanRealRow->scatter(atomRowData.skippedCharge, skch_tmp);
665 +      for (int i = 0; i < ns; i++) {
666 +        snap_->atomData.skippedCharge[i] += skch_tmp[i];
667 +        skch_tmp[i] = 0.0;
668 +      }
669 +      
670 +      AtomPlanRealColumn->scatter(atomColData.skippedCharge, skch_tmp);
671 +      for (int i = 0; i < ns; i++)
672 +        snap_->atomData.skippedCharge[i] += skch_tmp[i];
673 +    }
674      
675 <    int nLocal = snap->getNumberOfAtoms();
675 >    nLocal_ = snap_->getNumberOfAtoms();
676  
677 <    vector<vector<RealType> > pot_temp(N_INTERACTION_FAMILIES,
678 <                                       vector<RealType> (nLocal, 0.0));
677 >    vector<potVec> pot_temp(nLocal_,
678 >                            Vector<RealType, N_INTERACTION_FAMILIES> (0.0));
679 >
680 >    // scatter/gather pot_row into the members of my column
681 >          
682 >    AtomPlanPotRow->scatter(pot_row, pot_temp);
683 >
684 >    for (int ii = 0;  ii < pot_temp.size(); ii++ )
685 >      pairwisePot += pot_temp[ii];
686      
687 <    for (int i = 0; i < N_INTERACTION_FAMILIES; i++) {
688 <      AtomCommRealI->scatter(pot_row[i], pot_temp[i]);
689 <      for (int ii = 0;  ii < pot_temp[i].size(); ii++ ) {
690 <        pot_local[i] += pot_temp[i][ii];
691 <      }
687 >    fill(pot_temp.begin(), pot_temp.end(),
688 >         Vector<RealType, N_INTERACTION_FAMILIES> (0.0));
689 >      
690 >    AtomPlanPotColumn->scatter(pot_col, pot_temp);    
691 >    
692 >    for (int ii = 0;  ii < pot_temp.size(); ii++ )
693 >      pairwisePot += pot_temp[ii];    
694 >    
695 >    for (int ii = 0; ii < N_INTERACTION_FAMILIES; ii++) {
696 >      RealType ploc1 = pairwisePot[ii];
697 >      RealType ploc2 = 0.0;
698 >      MPI::COMM_WORLD.Allreduce(&ploc1, &ploc2, 1, MPI::REALTYPE, MPI::SUM);
699 >      pairwisePot[ii] = ploc2;
700 >    }
701 >
702 > #endif
703 >
704 >  }
705 >
706 >  int ForceMatrixDecomposition::getNAtomsInRow() {  
707 > #ifdef IS_MPI
708 >    return nAtomsInRow_;
709 > #else
710 >    return nLocal_;
711 > #endif
712 >  }
713 >
714 >  /**
715 >   * returns the list of atoms belonging to this group.  
716 >   */
717 >  vector<int> ForceMatrixDecomposition::getAtomsInGroupRow(int cg1){
718 > #ifdef IS_MPI
719 >    return groupListRow_[cg1];
720 > #else
721 >    return groupList_[cg1];
722 > #endif
723 >  }
724 >
725 >  vector<int> ForceMatrixDecomposition::getAtomsInGroupColumn(int cg2){
726 > #ifdef IS_MPI
727 >    return groupListCol_[cg2];
728 > #else
729 >    return groupList_[cg2];
730 > #endif
731 >  }
732 >  
733 >  Vector3d ForceMatrixDecomposition::getIntergroupVector(int cg1, int cg2){
734 >    Vector3d d;
735 >    
736 > #ifdef IS_MPI
737 >    d = cgColData.position[cg2] - cgRowData.position[cg1];
738 > #else
739 >    d = snap_->cgData.position[cg2] - snap_->cgData.position[cg1];
740 > #endif
741 >    
742 >    snap_->wrapVector(d);
743 >    return d;    
744 >  }
745 >
746 >
747 >  Vector3d ForceMatrixDecomposition::getAtomToGroupVectorRow(int atom1, int cg1){
748 >
749 >    Vector3d d;
750 >    
751 > #ifdef IS_MPI
752 >    d = cgRowData.position[cg1] - atomRowData.position[atom1];
753 > #else
754 >    d = snap_->cgData.position[cg1] - snap_->atomData.position[atom1];
755 > #endif
756 >
757 >    snap_->wrapVector(d);
758 >    return d;    
759 >  }
760 >  
761 >  Vector3d ForceMatrixDecomposition::getAtomToGroupVectorColumn(int atom2, int cg2){
762 >    Vector3d d;
763 >    
764 > #ifdef IS_MPI
765 >    d = cgColData.position[cg2] - atomColData.position[atom2];
766 > #else
767 >    d = snap_->cgData.position[cg2] - snap_->atomData.position[atom2];
768 > #endif
769 >    
770 >    snap_->wrapVector(d);
771 >    return d;    
772 >  }
773 >
774 >  RealType ForceMatrixDecomposition::getMassFactorRow(int atom1) {
775 > #ifdef IS_MPI
776 >    return massFactorsRow[atom1];
777 > #else
778 >    return massFactors[atom1];
779 > #endif
780 >  }
781 >
782 >  RealType ForceMatrixDecomposition::getMassFactorColumn(int atom2) {
783 > #ifdef IS_MPI
784 >    return massFactorsCol[atom2];
785 > #else
786 >    return massFactors[atom2];
787 > #endif
788 >
789 >  }
790 >    
791 >  Vector3d ForceMatrixDecomposition::getInteratomicVector(int atom1, int atom2){
792 >    Vector3d d;
793 >    
794 > #ifdef IS_MPI
795 >    d = atomColData.position[atom2] - atomRowData.position[atom1];
796 > #else
797 >    d = snap_->atomData.position[atom2] - snap_->atomData.position[atom1];
798 > #endif
799 >
800 >    snap_->wrapVector(d);
801 >    return d;    
802 >  }
803 >
804 >  vector<int> ForceMatrixDecomposition::getExcludesForAtom(int atom1) {
805 >    return excludesForAtom[atom1];
806 >  }
807 >
808 >  /**
809 >   * We need to exclude some overcounted interactions that result from
810 >   * the parallel decomposition.
811 >   */
812 >  bool ForceMatrixDecomposition::skipAtomPair(int atom1, int atom2) {
813 >    int unique_id_1, unique_id_2;
814 >    
815 > #ifdef IS_MPI
816 >    // in MPI, we have to look up the unique IDs for each atom
817 >    unique_id_1 = AtomRowToGlobal[atom1];
818 >    unique_id_2 = AtomColToGlobal[atom2];
819 >
820 >    // this situation should only arise in MPI simulations
821 >    if (unique_id_1 == unique_id_2) return true;
822 >    
823 >    // this prevents us from doing the pair on multiple processors
824 >    if (unique_id_1 < unique_id_2) {
825 >      if ((unique_id_1 + unique_id_2) % 2 == 0) return true;
826 >    } else {
827 >      if ((unique_id_1 + unique_id_2) % 2 == 1) return true;
828 >    }
829 > #endif
830 >    return false;
831 >  }
832 >
833 >  /**
834 >   * We need to handle the interactions for atoms who are involved in
835 >   * the same rigid body as well as some short range interactions
836 >   * (bonds, bends, torsions) differently from other interactions.
837 >   * We'll still visit the pairwise routines, but with a flag that
838 >   * tells those routines to exclude the pair from direct long range
839 >   * interactions.  Some indirect interactions (notably reaction
840 >   * field) must still be handled for these pairs.
841 >   */
842 >  bool ForceMatrixDecomposition::excludeAtomPair(int atom1, int atom2) {
843 >    int unique_id_2;
844 > #ifdef IS_MPI
845 >    // in MPI, we have to look up the unique IDs for the row atom.
846 >    unique_id_2 = AtomColToGlobal[atom2];
847 > #else
848 >    // in the normal loop, the atom numbers are unique
849 >    unique_id_2 = atom2;
850 > #endif
851 >    
852 >    for (vector<int>::iterator i = excludesForAtom[atom1].begin();
853 >         i != excludesForAtom[atom1].end(); ++i) {
854 >      if ( (*i) == unique_id_2 ) return true;
855 >    }
856 >
857 >    return false;
858 >  }
859 >
860 >
861 >  void ForceMatrixDecomposition::addForceToAtomRow(int atom1, Vector3d fg){
862 > #ifdef IS_MPI
863 >    atomRowData.force[atom1] += fg;
864 > #else
865 >    snap_->atomData.force[atom1] += fg;
866 > #endif
867 >  }
868 >
869 >  void ForceMatrixDecomposition::addForceToAtomColumn(int atom2, Vector3d fg){
870 > #ifdef IS_MPI
871 >    atomColData.force[atom2] += fg;
872 > #else
873 >    snap_->atomData.force[atom2] += fg;
874 > #endif
875 >  }
876 >
877 >    // filling interaction blocks with pointers
878 >  void ForceMatrixDecomposition::fillInteractionData(InteractionData &idat,
879 >                                                     int atom1, int atom2) {
880 >
881 >    idat.excluded = excludeAtomPair(atom1, atom2);
882 >  
883 > #ifdef IS_MPI
884 >    idat.atypes = make_pair( atypesRow[atom1], atypesCol[atom2]);
885 >    //idat.atypes = make_pair( ff_->getAtomType(identsRow[atom1]),
886 >    //                         ff_->getAtomType(identsCol[atom2]) );
887 >    
888 >    if (storageLayout_ & DataStorage::dslAmat) {
889 >      idat.A1 = &(atomRowData.aMat[atom1]);
890 >      idat.A2 = &(atomColData.aMat[atom2]);
891 >    }
892 >    
893 >    if (storageLayout_ & DataStorage::dslElectroFrame) {
894 >      idat.eFrame1 = &(atomRowData.electroFrame[atom1]);
895 >      idat.eFrame2 = &(atomColData.electroFrame[atom2]);
896 >    }
897 >
898 >    if (storageLayout_ & DataStorage::dslTorque) {
899 >      idat.t1 = &(atomRowData.torque[atom1]);
900 >      idat.t2 = &(atomColData.torque[atom2]);
901 >    }
902 >
903 >    if (storageLayout_ & DataStorage::dslDensity) {
904 >      idat.rho1 = &(atomRowData.density[atom1]);
905 >      idat.rho2 = &(atomColData.density[atom2]);
906 >    }
907 >
908 >    if (storageLayout_ & DataStorage::dslFunctional) {
909 >      idat.frho1 = &(atomRowData.functional[atom1]);
910 >      idat.frho2 = &(atomColData.functional[atom2]);
911 >    }
912 >
913 >    if (storageLayout_ & DataStorage::dslFunctionalDerivative) {
914 >      idat.dfrho1 = &(atomRowData.functionalDerivative[atom1]);
915 >      idat.dfrho2 = &(atomColData.functionalDerivative[atom2]);
916 >    }
917 >
918 >    if (storageLayout_ & DataStorage::dslParticlePot) {
919 >      idat.particlePot1 = &(atomRowData.particlePot[atom1]);
920 >      idat.particlePot2 = &(atomColData.particlePot[atom2]);
921 >    }
922 >
923 >    if (storageLayout_ & DataStorage::dslSkippedCharge) {              
924 >      idat.skippedCharge1 = &(atomRowData.skippedCharge[atom1]);
925 >      idat.skippedCharge2 = &(atomColData.skippedCharge[atom2]);
926      }
927 +
928 + #else
929 +
930 +    idat.atypes = make_pair( atypesLocal[atom1], atypesLocal[atom2]);
931 +    //idat.atypes = make_pair( ff_->getAtomType(idents[atom1]),
932 +    //                         ff_->getAtomType(idents[atom2]) );
933 +
934 +    if (storageLayout_ & DataStorage::dslAmat) {
935 +      idat.A1 = &(snap_->atomData.aMat[atom1]);
936 +      idat.A2 = &(snap_->atomData.aMat[atom2]);
937 +    }
938 +
939 +    if (storageLayout_ & DataStorage::dslElectroFrame) {
940 +      idat.eFrame1 = &(snap_->atomData.electroFrame[atom1]);
941 +      idat.eFrame2 = &(snap_->atomData.electroFrame[atom2]);
942 +    }
943 +
944 +    if (storageLayout_ & DataStorage::dslTorque) {
945 +      idat.t1 = &(snap_->atomData.torque[atom1]);
946 +      idat.t2 = &(snap_->atomData.torque[atom2]);
947 +    }
948 +
949 +    if (storageLayout_ & DataStorage::dslDensity) {    
950 +      idat.rho1 = &(snap_->atomData.density[atom1]);
951 +      idat.rho2 = &(snap_->atomData.density[atom2]);
952 +    }
953 +
954 +    if (storageLayout_ & DataStorage::dslFunctional) {
955 +      idat.frho1 = &(snap_->atomData.functional[atom1]);
956 +      idat.frho2 = &(snap_->atomData.functional[atom2]);
957 +    }
958 +
959 +    if (storageLayout_ & DataStorage::dslFunctionalDerivative) {
960 +      idat.dfrho1 = &(snap_->atomData.functionalDerivative[atom1]);
961 +      idat.dfrho2 = &(snap_->atomData.functionalDerivative[atom2]);
962 +    }
963 +
964 +    if (storageLayout_ & DataStorage::dslParticlePot) {
965 +      idat.particlePot1 = &(snap_->atomData.particlePot[atom1]);
966 +      idat.particlePot2 = &(snap_->atomData.particlePot[atom2]);
967 +    }
968 +
969 +    if (storageLayout_ & DataStorage::dslSkippedCharge) {
970 +      idat.skippedCharge1 = &(snap_->atomData.skippedCharge[atom1]);
971 +      idat.skippedCharge2 = &(snap_->atomData.skippedCharge[atom2]);
972 +    }
973   #endif
974    }
975 +
976    
977 +  void ForceMatrixDecomposition::unpackInteractionData(InteractionData &idat, int atom1, int atom2) {    
978 + #ifdef IS_MPI
979 +    pot_row[atom1] += 0.5 *  *(idat.pot);
980 +    pot_col[atom2] += 0.5 *  *(idat.pot);
981 +
982 +    atomRowData.force[atom1] += *(idat.f1);
983 +    atomColData.force[atom2] -= *(idat.f1);
984 + #else
985 +    pairwisePot += *(idat.pot);
986 +
987 +    snap_->atomData.force[atom1] += *(idat.f1);
988 +    snap_->atomData.force[atom2] -= *(idat.f1);
989 + #endif
990 +    
991 +  }
992 +
993 +  /*
994 +   * buildNeighborList
995 +   *
996 +   * first element of pair is row-indexed CutoffGroup
997 +   * second element of pair is column-indexed CutoffGroup
998 +   */
999 +  vector<pair<int, int> > ForceMatrixDecomposition::buildNeighborList() {
1000 +      
1001 +    vector<pair<int, int> > neighborList;
1002 +    groupCutoffs cuts;
1003 +    bool doAllPairs = false;
1004 +
1005 + #ifdef IS_MPI
1006 +    cellListRow_.clear();
1007 +    cellListCol_.clear();
1008 + #else
1009 +    cellList_.clear();
1010 + #endif
1011 +
1012 +    RealType rList_ = (largestRcut_ + skinThickness_);
1013 +    RealType rl2 = rList_ * rList_;
1014 +    Snapshot* snap_ = sman_->getCurrentSnapshot();
1015 +    Mat3x3d Hmat = snap_->getHmat();
1016 +    Vector3d Hx = Hmat.getColumn(0);
1017 +    Vector3d Hy = Hmat.getColumn(1);
1018 +    Vector3d Hz = Hmat.getColumn(2);
1019 +
1020 +    nCells_.x() = (int) ( Hx.length() )/ rList_;
1021 +    nCells_.y() = (int) ( Hy.length() )/ rList_;
1022 +    nCells_.z() = (int) ( Hz.length() )/ rList_;
1023 +
1024 +    // handle small boxes where the cell offsets can end up repeating cells
1025 +    
1026 +    if (nCells_.x() < 3) doAllPairs = true;
1027 +    if (nCells_.y() < 3) doAllPairs = true;
1028 +    if (nCells_.z() < 3) doAllPairs = true;
1029 +
1030 +    Mat3x3d invHmat = snap_->getInvHmat();
1031 +    Vector3d rs, scaled, dr;
1032 +    Vector3i whichCell;
1033 +    int cellIndex;
1034 +    int nCtot = nCells_.x() * nCells_.y() * nCells_.z();
1035 +
1036 + #ifdef IS_MPI
1037 +    cellListRow_.resize(nCtot);
1038 +    cellListCol_.resize(nCtot);
1039 + #else
1040 +    cellList_.resize(nCtot);
1041 + #endif
1042 +
1043 +    if (!doAllPairs) {
1044 + #ifdef IS_MPI
1045 +
1046 +      for (int i = 0; i < nGroupsInRow_; i++) {
1047 +        rs = cgRowData.position[i];
1048 +        
1049 +        // scaled positions relative to the box vectors
1050 +        scaled = invHmat * rs;
1051 +        
1052 +        // wrap the vector back into the unit box by subtracting integer box
1053 +        // numbers
1054 +        for (int j = 0; j < 3; j++) {
1055 +          scaled[j] -= roundMe(scaled[j]);
1056 +          scaled[j] += 0.5;
1057 +        }
1058 +        
1059 +        // find xyz-indices of cell that cutoffGroup is in.
1060 +        whichCell.x() = nCells_.x() * scaled.x();
1061 +        whichCell.y() = nCells_.y() * scaled.y();
1062 +        whichCell.z() = nCells_.z() * scaled.z();
1063 +        
1064 +        // find single index of this cell:
1065 +        cellIndex = Vlinear(whichCell, nCells_);
1066 +        
1067 +        // add this cutoff group to the list of groups in this cell;
1068 +        cellListRow_[cellIndex].push_back(i);
1069 +      }
1070 +      for (int i = 0; i < nGroupsInCol_; i++) {
1071 +        rs = cgColData.position[i];
1072 +        
1073 +        // scaled positions relative to the box vectors
1074 +        scaled = invHmat * rs;
1075 +        
1076 +        // wrap the vector back into the unit box by subtracting integer box
1077 +        // numbers
1078 +        for (int j = 0; j < 3; j++) {
1079 +          scaled[j] -= roundMe(scaled[j]);
1080 +          scaled[j] += 0.5;
1081 +        }
1082 +        
1083 +        // find xyz-indices of cell that cutoffGroup is in.
1084 +        whichCell.x() = nCells_.x() * scaled.x();
1085 +        whichCell.y() = nCells_.y() * scaled.y();
1086 +        whichCell.z() = nCells_.z() * scaled.z();
1087 +        
1088 +        // find single index of this cell:
1089 +        cellIndex = Vlinear(whichCell, nCells_);
1090 +        
1091 +        // add this cutoff group to the list of groups in this cell;
1092 +        cellListCol_[cellIndex].push_back(i);
1093 +      }
1094 + #else
1095 +      for (int i = 0; i < nGroups_; i++) {
1096 +        rs = snap_->cgData.position[i];
1097 +        
1098 +        // scaled positions relative to the box vectors
1099 +        scaled = invHmat * rs;
1100 +        
1101 +        // wrap the vector back into the unit box by subtracting integer box
1102 +        // numbers
1103 +        for (int j = 0; j < 3; j++) {
1104 +          scaled[j] -= roundMe(scaled[j]);
1105 +          scaled[j] += 0.5;
1106 +        }
1107 +        
1108 +        // find xyz-indices of cell that cutoffGroup is in.
1109 +        whichCell.x() = nCells_.x() * scaled.x();
1110 +        whichCell.y() = nCells_.y() * scaled.y();
1111 +        whichCell.z() = nCells_.z() * scaled.z();
1112 +        
1113 +        // find single index of this cell:
1114 +        cellIndex = Vlinear(whichCell, nCells_);
1115 +        
1116 +        // add this cutoff group to the list of groups in this cell;
1117 +        cellList_[cellIndex].push_back(i);
1118 +      }
1119 + #endif
1120 +
1121 +      for (int m1z = 0; m1z < nCells_.z(); m1z++) {
1122 +        for (int m1y = 0; m1y < nCells_.y(); m1y++) {
1123 +          for (int m1x = 0; m1x < nCells_.x(); m1x++) {
1124 +            Vector3i m1v(m1x, m1y, m1z);
1125 +            int m1 = Vlinear(m1v, nCells_);
1126 +            
1127 +            for (vector<Vector3i>::iterator os = cellOffsets_.begin();
1128 +                 os != cellOffsets_.end(); ++os) {
1129 +              
1130 +              Vector3i m2v = m1v + (*os);
1131 +              
1132 +              if (m2v.x() >= nCells_.x()) {
1133 +                m2v.x() = 0;          
1134 +              } else if (m2v.x() < 0) {
1135 +                m2v.x() = nCells_.x() - 1;
1136 +              }
1137 +              
1138 +              if (m2v.y() >= nCells_.y()) {
1139 +                m2v.y() = 0;          
1140 +              } else if (m2v.y() < 0) {
1141 +                m2v.y() = nCells_.y() - 1;
1142 +              }
1143 +              
1144 +              if (m2v.z() >= nCells_.z()) {
1145 +                m2v.z() = 0;          
1146 +              } else if (m2v.z() < 0) {
1147 +                m2v.z() = nCells_.z() - 1;
1148 +              }
1149 +              
1150 +              int m2 = Vlinear (m2v, nCells_);
1151 +              
1152 + #ifdef IS_MPI
1153 +              for (vector<int>::iterator j1 = cellListRow_[m1].begin();
1154 +                   j1 != cellListRow_[m1].end(); ++j1) {
1155 +                for (vector<int>::iterator j2 = cellListCol_[m2].begin();
1156 +                     j2 != cellListCol_[m2].end(); ++j2) {
1157 +                  
1158 +                  // In parallel, we need to visit *all* pairs of row &
1159 +                  // column indicies and will truncate later on.
1160 +                  dr = cgColData.position[(*j2)] - cgRowData.position[(*j1)];
1161 +                  snap_->wrapVector(dr);
1162 +                  cuts = getGroupCutoffs( (*j1), (*j2) );
1163 +                  if (dr.lengthSquare() < cuts.third) {
1164 +                    neighborList.push_back(make_pair((*j1), (*j2)));
1165 +                  }                  
1166 +                }
1167 +              }
1168 + #else
1169 +              
1170 +              for (vector<int>::iterator j1 = cellList_[m1].begin();
1171 +                   j1 != cellList_[m1].end(); ++j1) {
1172 +                for (vector<int>::iterator j2 = cellList_[m2].begin();
1173 +                     j2 != cellList_[m2].end(); ++j2) {
1174 +                  
1175 +                  // Always do this if we're in different cells or if
1176 +                  // we're in the same cell and the global index of the
1177 +                  // j2 cutoff group is less than the j1 cutoff group
1178 +                  
1179 +                  if (m2 != m1 || (*j2) < (*j1)) {
1180 +                    dr = snap_->cgData.position[(*j2)] - snap_->cgData.position[(*j1)];
1181 +                    snap_->wrapVector(dr);
1182 +                    cuts = getGroupCutoffs( (*j1), (*j2) );
1183 +                    if (dr.lengthSquare() < cuts.third) {
1184 +                      neighborList.push_back(make_pair((*j1), (*j2)));
1185 +                    }
1186 +                  }
1187 +                }
1188 +              }
1189 + #endif
1190 +            }
1191 +          }
1192 +        }
1193 +      }
1194 +    } else {
1195 +      // branch to do all cutoff group pairs
1196 + #ifdef IS_MPI
1197 +      for (int j1 = 0; j1 < nGroupsInRow_; j1++) {
1198 +        for (int j2 = 0; j2 < nGroupsInCol_; j2++) {      
1199 +          dr = cgColData.position[j2] - cgRowData.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 + #else
1208 +      for (int j1 = 0; j1 < nGroups_ - 1; j1++) {
1209 +        for (int j2 = j1 + 1; j2 < nGroups_; j2++) {
1210 +          dr = snap_->cgData.position[j2] - snap_->cgData.position[j1];
1211 +          snap_->wrapVector(dr);
1212 +          cuts = getGroupCutoffs( j1, j2 );
1213 +          if (dr.lengthSquare() < cuts.third) {
1214 +            neighborList.push_back(make_pair(j1, j2));
1215 +          }
1216 +        }
1217 +      }        
1218 + #endif
1219 +    }
1220 +      
1221 +    // save the local cutoff group positions for the check that is
1222 +    // done on each loop:
1223 +    saved_CG_positions_.clear();
1224 +    for (int i = 0; i < nGroups_; i++)
1225 +      saved_CG_positions_.push_back(snap_->cgData.position[i]);
1226 +    
1227 +    return neighborList;
1228 +  }
1229   } //end namespace OpenMD

Diff Legend

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