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 1591 by gezelter, Tue Jul 12 15:25:07 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 {
# Line 51 | Line 52 | namespace OpenMD {
52     * SimulationSetup
53     */
54    
55 <  void ForceDecomposition::distributeInitialData() {
56 < #ifdef IS_MPI    
57 <    Snapshot* snap = sman_->getCurrentSnapshot();
58 <    int nLocal = snap->getNumberOfAtoms();
59 <    int nGroups = snap->getNumberOfCutoffGroups();
55 >  void ForceMatrixDecomposition::distributeInitialData() {
56 >    snap_ = sman_->getCurrentSnapshot();
57 >    storageLayout_ = sman_->getStorageLayout();
58 >    ff_ = info_->getForceField();
59 >    nLocal_ = snap_->getNumberOfAtoms();
60 >    
61 >    nGroups_ = info_->getNLocalCutoffGroups();
62 >    // gather the information for atomtype IDs (atids):
63 >    idents = info_->getIdentArray();
64 >    AtomLocalToGlobal = info_->getGlobalAtomIndices();
65 >    cgLocalToGlobal = info_->getGlobalGroupIndices();
66 >    vector<int> globalGroupMembership = info_->getGlobalGroupMembership();
67  
68 <    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);
68 >    massFactors = info_->getMassFactors();
69  
70 <    AtomCommIntJ = new Communicator<Column,int>(nLocal);
71 <    AtomCommRealJ = new Communicator<Column,RealType>(nLocal);
72 <    AtomCommVectorJ = new Communicator<Column,Vector3d>(nLocal);
73 <    AtomCommMatrixJ = new Communicator<Column,Mat3x3d>(nLocal);
70 >    PairList* excludes = info_->getExcludedInteractions();
71 >    PairList* oneTwo = info_->getOneTwoInteractions();
72 >    PairList* oneThree = info_->getOneThreeInteractions();
73 >    PairList* oneFour = info_->getOneFourInteractions();
74  
75 <    cgCommIntI = new Communicator<Row,int>(nGroups);
76 <    cgCommVectorI = new Communicator<Row,Vector3d>(nGroups);
77 <    cgCommIntJ = new Communicator<Column,int>(nGroups);
78 <    cgCommVectorJ = new Communicator<Column,Vector3d>(nGroups);
75 > #ifdef IS_MPI
76 >
77 >    AtomCommIntRow = new Communicator<Row,int>(nLocal_);
78 >    AtomCommRealRow = new Communicator<Row,RealType>(nLocal_);
79 >    AtomCommVectorRow = new Communicator<Row,Vector3d>(nLocal_);
80 >    AtomCommMatrixRow = new Communicator<Row,Mat3x3d>(nLocal_);
81 >    AtomCommPotRow = new Communicator<Row,potVec>(nLocal_);
82  
83 <    int nAtomsInRow = AtomCommIntI->getSize();
84 <    int nAtomsInCol = AtomCommIntJ->getSize();
85 <    int nGroupsInRow = cgCommIntI->getSize();
86 <    int nGroupsInCol = cgCommIntJ->getSize();
83 >    AtomCommIntColumn = new Communicator<Column,int>(nLocal_);
84 >    AtomCommRealColumn = new Communicator<Column,RealType>(nLocal_);
85 >    AtomCommVectorColumn = new Communicator<Column,Vector3d>(nLocal_);
86 >    AtomCommMatrixColumn = new Communicator<Column,Mat3x3d>(nLocal_);
87 >    AtomCommPotColumn = new Communicator<Column,potVec>(nLocal_);
88  
89 <    vector<vector<RealType> > pot_row(N_INTERACTION_FAMILIES,
90 <                                      vector<RealType> (nAtomsInRow, 0.0));
91 <    vector<vector<RealType> > pot_col(N_INTERACTION_FAMILIES,
92 <                                      vector<RealType> (nAtomsInCol, 0.0));
89 >    cgCommIntRow = new Communicator<Row,int>(nGroups_);
90 >    cgCommVectorRow = new Communicator<Row,Vector3d>(nGroups_);
91 >    cgCommIntColumn = new Communicator<Column,int>(nGroups_);
92 >    cgCommVectorColumn = new Communicator<Column,Vector3d>(nGroups_);
93 >
94 >    nAtomsInRow_ = AtomCommIntRow->getSize();
95 >    nAtomsInCol_ = AtomCommIntColumn->getSize();
96 >    nGroupsInRow_ = cgCommIntRow->getSize();
97 >    nGroupsInCol_ = cgCommIntColumn->getSize();
98 >
99 >    // Modify the data storage objects with the correct layouts and sizes:
100 >    atomRowData.resize(nAtomsInRow_);
101 >    atomRowData.setStorageLayout(storageLayout_);
102 >    atomColData.resize(nAtomsInCol_);
103 >    atomColData.setStorageLayout(storageLayout_);
104 >    cgRowData.resize(nGroupsInRow_);
105 >    cgRowData.setStorageLayout(DataStorage::dslPosition);
106 >    cgColData.resize(nGroupsInCol_);
107 >    cgColData.setStorageLayout(DataStorage::dslPosition);
108 >        
109 >    identsRow.resize(nAtomsInRow_);
110 >    identsCol.resize(nAtomsInCol_);
111      
112 <    vector<RealType> pot_local(N_INTERACTION_FAMILIES, 0.0);
112 >    AtomCommIntRow->gather(idents, identsRow);
113 >    AtomCommIntColumn->gather(idents, identsCol);
114 >    
115 >    // allocate memory for the parallel objects
116 >    atypesRow.resize(nAtomsInRow_);
117 >    atypesCol.resize(nAtomsInCol_);
118  
119 <    // gather the information for atomtype IDs (atids):
120 <    vector<int> identsLocal = info_->getIdentArray();
121 <    identsRow.reserve(nAtomsInRow);
122 <    identsCol.reserve(nAtomsInCol);
119 >    for (int i = 0; i < nAtomsInRow_; i++)
120 >      atypesRow[i] = ff_->getAtomType(identsRow[i]);
121 >    for (int i = 0; i < nAtomsInCol_; i++)
122 >      atypesCol[i] = ff_->getAtomType(identsCol[i]);        
123  
124 <    AtomCommIntI->gather(identsLocal, identsRow);
125 <    AtomCommIntJ->gather(identsLocal, identsCol);
124 >    pot_row.resize(nAtomsInRow_);
125 >    pot_col.resize(nAtomsInCol_);
126  
127 <    AtomLocalToGlobal = info_->getLocalToGlobalAtomIndex();
128 <    AtomCommIntI->gather(AtomLocalToGlobal, AtomRowToGlobal);
129 <    AtomCommIntJ->gather(AtomLocalToGlobal, AtomColToGlobal);
127 >    AtomRowToGlobal.resize(nAtomsInRow_);
128 >    AtomColToGlobal.resize(nAtomsInCol_);
129 >    AtomCommIntRow->gather(AtomLocalToGlobal, AtomRowToGlobal);
130 >    AtomCommIntColumn->gather(AtomLocalToGlobal, AtomColToGlobal);
131 >    
132 >    cgRowToGlobal.resize(nGroupsInRow_);
133 >    cgColToGlobal.resize(nGroupsInCol_);
134 >    cgCommIntRow->gather(cgLocalToGlobal, cgRowToGlobal);
135 >    cgCommIntColumn->gather(cgLocalToGlobal, cgColToGlobal);
136  
137 <    cgLocalToGlobal = info_->getLocalToGlobalCutoffGroupIndex();
138 <    cgCommIntI->gather(cgLocalToGlobal, cgRowToGlobal);
139 <    cgCommIntJ->gather(cgLocalToGlobal, cgColToGlobal);
137 >    massFactorsRow.resize(nAtomsInRow_);
138 >    massFactorsCol.resize(nAtomsInCol_);
139 >    AtomCommRealRow->gather(massFactors, massFactorsRow);
140 >    AtomCommRealColumn->gather(massFactors, massFactorsCol);
141  
142 <    
142 >    groupListRow_.clear();
143 >    groupListRow_.resize(nGroupsInRow_);
144 >    for (int i = 0; i < nGroupsInRow_; i++) {
145 >      int gid = cgRowToGlobal[i];
146 >      for (int j = 0; j < nAtomsInRow_; j++) {
147 >        int aid = AtomRowToGlobal[j];
148 >        if (globalGroupMembership[aid] == gid)
149 >          groupListRow_[i].push_back(j);
150 >      }      
151 >    }
152  
153 <    // still need:
154 <    // topoDist
155 <    // exclude
153 >    groupListCol_.clear();
154 >    groupListCol_.resize(nGroupsInCol_);
155 >    for (int i = 0; i < nGroupsInCol_; i++) {
156 >      int gid = cgColToGlobal[i];
157 >      for (int j = 0; j < nAtomsInCol_; j++) {
158 >        int aid = AtomColToGlobal[j];
159 >        if (globalGroupMembership[aid] == gid)
160 >          groupListCol_[i].push_back(j);
161 >      }      
162 >    }
163 >
164 >    excludesForAtom.clear();
165 >    excludesForAtom.resize(nAtomsInRow_);
166 >    toposForAtom.clear();
167 >    toposForAtom.resize(nAtomsInRow_);
168 >    topoDist.clear();
169 >    topoDist.resize(nAtomsInRow_);
170 >    for (int i = 0; i < nAtomsInRow_; i++) {
171 >      int iglob = AtomRowToGlobal[i];
172 >
173 >      for (int j = 0; j < nAtomsInCol_; j++) {
174 >        int jglob = AtomColToGlobal[j];
175 >
176 >        if (excludes->hasPair(iglob, jglob))
177 >          excludesForAtom[i].push_back(j);      
178 >        
179 >        if (oneTwo->hasPair(iglob, jglob)) {
180 >          toposForAtom[i].push_back(j);
181 >          topoDist[i].push_back(1);
182 >        } else {
183 >          if (oneThree->hasPair(iglob, jglob)) {
184 >            toposForAtom[i].push_back(j);
185 >            topoDist[i].push_back(2);
186 >          } else {
187 >            if (oneFour->hasPair(iglob, jglob)) {
188 >              toposForAtom[i].push_back(j);
189 >              topoDist[i].push_back(3);
190 >            }
191 >          }
192 >        }
193 >      }      
194 >    }
195 >
196   #endif
197 +
198 +    // allocate memory for the parallel objects
199 +    atypesLocal.resize(nLocal_);
200 +
201 +    for (int i = 0; i < nLocal_; i++)
202 +      atypesLocal[i] = ff_->getAtomType(idents[i]);
203 +
204 +    groupList_.clear();
205 +    groupList_.resize(nGroups_);
206 +    for (int i = 0; i < nGroups_; i++) {
207 +      int gid = cgLocalToGlobal[i];
208 +      for (int j = 0; j < nLocal_; j++) {
209 +        int aid = AtomLocalToGlobal[j];
210 +        if (globalGroupMembership[aid] == gid) {
211 +          groupList_[i].push_back(j);
212 +        }
213 +      }      
214 +    }
215 +
216 +    excludesForAtom.clear();
217 +    excludesForAtom.resize(nLocal_);
218 +    toposForAtom.clear();
219 +    toposForAtom.resize(nLocal_);
220 +    topoDist.clear();
221 +    topoDist.resize(nLocal_);
222 +
223 +    for (int i = 0; i < nLocal_; i++) {
224 +      int iglob = AtomLocalToGlobal[i];
225 +
226 +      for (int j = 0; j < nLocal_; j++) {
227 +        int jglob = AtomLocalToGlobal[j];
228 +
229 +        if (excludes->hasPair(iglob, jglob))
230 +          excludesForAtom[i].push_back(j);              
231 +        
232 +        if (oneTwo->hasPair(iglob, jglob)) {
233 +          toposForAtom[i].push_back(j);
234 +          topoDist[i].push_back(1);
235 +        } else {
236 +          if (oneThree->hasPair(iglob, jglob)) {
237 +            toposForAtom[i].push_back(j);
238 +            topoDist[i].push_back(2);
239 +          } else {
240 +            if (oneFour->hasPair(iglob, jglob)) {
241 +              toposForAtom[i].push_back(j);
242 +              topoDist[i].push_back(3);
243 +            }
244 +          }
245 +        }
246 +      }      
247 +    }
248 +    
249 +    createGtypeCutoffMap();
250 +
251    }
252 +  
253 +  void ForceMatrixDecomposition::createGtypeCutoffMap() {
254      
255 +    RealType tol = 1e-6;
256 +    RealType rc;
257 +    int atid;
258 +    set<AtomType*> atypes = info_->getSimulatedAtomTypes();
259 +    map<int, RealType> atypeCutoff;
260 +      
261 +    for (set<AtomType*>::iterator at = atypes.begin();
262 +         at != atypes.end(); ++at){
263 +      atid = (*at)->getIdent();
264 +      if (userChoseCutoff_)
265 +        atypeCutoff[atid] = userCutoff_;
266 +      else
267 +        atypeCutoff[atid] = interactionMan_->getSuggestedCutoffRadius(*at);
268 +    }
269  
270 +    vector<RealType> gTypeCutoffs;
271 +    // first we do a single loop over the cutoff groups to find the
272 +    // largest cutoff for any atypes present in this group.
273 + #ifdef IS_MPI
274 +    vector<RealType> groupCutoffRow(nGroupsInRow_, 0.0);
275 +    groupRowToGtype.resize(nGroupsInRow_);
276 +    for (int cg1 = 0; cg1 < nGroupsInRow_; cg1++) {
277 +      vector<int> atomListRow = getAtomsInGroupRow(cg1);
278 +      for (vector<int>::iterator ia = atomListRow.begin();
279 +           ia != atomListRow.end(); ++ia) {            
280 +        int atom1 = (*ia);
281 +        atid = identsRow[atom1];
282 +        if (atypeCutoff[atid] > groupCutoffRow[cg1]) {
283 +          groupCutoffRow[cg1] = atypeCutoff[atid];
284 +        }
285 +      }
286  
287 <  void ForceDecomposition::distributeData()  {
287 >      bool gTypeFound = false;
288 >      for (int gt = 0; gt < gTypeCutoffs.size(); gt++) {
289 >        if (abs(groupCutoffRow[cg1] - gTypeCutoffs[gt]) < tol) {
290 >          groupRowToGtype[cg1] = gt;
291 >          gTypeFound = true;
292 >        }
293 >      }
294 >      if (!gTypeFound) {
295 >        gTypeCutoffs.push_back( groupCutoffRow[cg1] );
296 >        groupRowToGtype[cg1] = gTypeCutoffs.size() - 1;
297 >      }
298 >      
299 >    }
300 >    vector<RealType> groupCutoffCol(nGroupsInCol_, 0.0);
301 >    groupColToGtype.resize(nGroupsInCol_);
302 >    for (int cg2 = 0; cg2 < nGroupsInCol_; cg2++) {
303 >      vector<int> atomListCol = getAtomsInGroupColumn(cg2);
304 >      for (vector<int>::iterator jb = atomListCol.begin();
305 >           jb != atomListCol.end(); ++jb) {            
306 >        int atom2 = (*jb);
307 >        atid = identsCol[atom2];
308 >        if (atypeCutoff[atid] > groupCutoffCol[cg2]) {
309 >          groupCutoffCol[cg2] = atypeCutoff[atid];
310 >        }
311 >      }
312 >      bool gTypeFound = false;
313 >      for (int gt = 0; gt < gTypeCutoffs.size(); gt++) {
314 >        if (abs(groupCutoffCol[cg2] - gTypeCutoffs[gt]) < tol) {
315 >          groupColToGtype[cg2] = gt;
316 >          gTypeFound = true;
317 >        }
318 >      }
319 >      if (!gTypeFound) {
320 >        gTypeCutoffs.push_back( groupCutoffCol[cg2] );
321 >        groupColToGtype[cg2] = gTypeCutoffs.size() - 1;
322 >      }
323 >    }
324 > #else
325 >
326 >    vector<RealType> groupCutoff(nGroups_, 0.0);
327 >    groupToGtype.resize(nGroups_);
328 >    for (int cg1 = 0; cg1 < nGroups_; cg1++) {
329 >
330 >      groupCutoff[cg1] = 0.0;
331 >      vector<int> atomList = getAtomsInGroupRow(cg1);
332 >
333 >      for (vector<int>::iterator ia = atomList.begin();
334 >           ia != atomList.end(); ++ia) {            
335 >        int atom1 = (*ia);
336 >        atid = idents[atom1];
337 >        if (atypeCutoff[atid] > groupCutoff[cg1]) {
338 >          groupCutoff[cg1] = atypeCutoff[atid];
339 >        }
340 >      }
341 >
342 >      bool gTypeFound = false;
343 >      for (int gt = 0; gt < gTypeCutoffs.size(); gt++) {
344 >        if (abs(groupCutoff[cg1] - gTypeCutoffs[gt]) < tol) {
345 >          groupToGtype[cg1] = gt;
346 >          gTypeFound = true;
347 >        }
348 >      }
349 >      if (!gTypeFound) {
350 >        gTypeCutoffs.push_back( groupCutoff[cg1] );
351 >        groupToGtype[cg1] = gTypeCutoffs.size() - 1;
352 >      }      
353 >    }
354 > #endif
355 >
356 >    // Now we find the maximum group cutoff value present in the simulation
357 >
358 >    RealType groupMax = *max_element(gTypeCutoffs.begin(),
359 >                                     gTypeCutoffs.end());
360 >
361   #ifdef IS_MPI
362 <    Snapshot* snap = sman_->getCurrentSnapshot();
362 >    MPI::COMM_WORLD.Allreduce(&groupMax, &groupMax, 1, MPI::REALTYPE,
363 >                              MPI::MAX);
364 > #endif
365      
366 +    RealType tradRcut = groupMax;
367 +
368 +    for (int i = 0; i < gTypeCutoffs.size();  i++) {
369 +      for (int j = 0; j < gTypeCutoffs.size();  j++) {      
370 +        RealType thisRcut;
371 +        switch(cutoffPolicy_) {
372 +        case TRADITIONAL:
373 +          thisRcut = tradRcut;
374 +          break;
375 +        case MIX:
376 +          thisRcut = 0.5 * (gTypeCutoffs[i] + gTypeCutoffs[j]);
377 +          break;
378 +        case MAX:
379 +          thisRcut = max(gTypeCutoffs[i], gTypeCutoffs[j]);
380 +          break;
381 +        default:
382 +          sprintf(painCave.errMsg,
383 +                  "ForceMatrixDecomposition::createGtypeCutoffMap "
384 +                  "hit an unknown cutoff policy!\n");
385 +          painCave.severity = OPENMD_ERROR;
386 +          painCave.isFatal = 1;
387 +          simError();
388 +          break;
389 +        }
390 +
391 +        pair<int,int> key = make_pair(i,j);
392 +        gTypeCutoffMap[key].first = thisRcut;
393 +
394 +        if (thisRcut > largestRcut_) largestRcut_ = thisRcut;
395 +
396 +        gTypeCutoffMap[key].second = thisRcut*thisRcut;
397 +        
398 +        gTypeCutoffMap[key].third = pow(thisRcut + skinThickness_, 2);
399 +
400 +        // sanity check
401 +        
402 +        if (userChoseCutoff_) {
403 +          if (abs(gTypeCutoffMap[key].first - userCutoff_) > 0.0001) {
404 +            sprintf(painCave.errMsg,
405 +                    "ForceMatrixDecomposition::createGtypeCutoffMap "
406 +                    "user-specified rCut (%lf) does not match computed group Cutoff\n", userCutoff_);
407 +            painCave.severity = OPENMD_ERROR;
408 +            painCave.isFatal = 1;
409 +            simError();            
410 +          }
411 +        }
412 +      }
413 +    }
414 +  }
415 +
416 +
417 +  groupCutoffs ForceMatrixDecomposition::getGroupCutoffs(int cg1, int cg2) {
418 +    int i, j;  
419 + #ifdef IS_MPI
420 +    i = groupRowToGtype[cg1];
421 +    j = groupColToGtype[cg2];
422 + #else
423 +    i = groupToGtype[cg1];
424 +    j = groupToGtype[cg2];
425 + #endif    
426 +    return gTypeCutoffMap[make_pair(i,j)];
427 +  }
428 +
429 +  int ForceMatrixDecomposition::getTopologicalDistance(int atom1, int atom2) {
430 +    for (int j = 0; j < toposForAtom[atom1].size(); j++) {
431 +      if (toposForAtom[atom1][j] == atom2)
432 +        return topoDist[atom1][j];
433 +    }
434 +    return 0;
435 +  }
436 +
437 +  void ForceMatrixDecomposition::zeroWorkArrays() {
438 +    pairwisePot = 0.0;
439 +    embeddingPot = 0.0;
440 +
441 + #ifdef IS_MPI
442 +    if (storageLayout_ & DataStorage::dslForce) {
443 +      fill(atomRowData.force.begin(), atomRowData.force.end(), V3Zero);
444 +      fill(atomColData.force.begin(), atomColData.force.end(), V3Zero);
445 +    }
446 +
447 +    if (storageLayout_ & DataStorage::dslTorque) {
448 +      fill(atomRowData.torque.begin(), atomRowData.torque.end(), V3Zero);
449 +      fill(atomColData.torque.begin(), atomColData.torque.end(), V3Zero);
450 +    }
451 +    
452 +    fill(pot_row.begin(), pot_row.end(),
453 +         Vector<RealType, N_INTERACTION_FAMILIES> (0.0));
454 +
455 +    fill(pot_col.begin(), pot_col.end(),
456 +         Vector<RealType, N_INTERACTION_FAMILIES> (0.0));  
457 +
458 +    if (storageLayout_ & DataStorage::dslParticlePot) {    
459 +      fill(atomRowData.particlePot.begin(), atomRowData.particlePot.end(),
460 +           0.0);
461 +      fill(atomColData.particlePot.begin(), atomColData.particlePot.end(),
462 +           0.0);
463 +    }
464 +
465 +    if (storageLayout_ & DataStorage::dslDensity) {      
466 +      fill(atomRowData.density.begin(), atomRowData.density.end(), 0.0);
467 +      fill(atomColData.density.begin(), atomColData.density.end(), 0.0);
468 +    }
469 +
470 +    if (storageLayout_ & DataStorage::dslFunctional) {  
471 +      fill(atomRowData.functional.begin(), atomRowData.functional.end(),
472 +           0.0);
473 +      fill(atomColData.functional.begin(), atomColData.functional.end(),
474 +           0.0);
475 +    }
476 +
477 +    if (storageLayout_ & DataStorage::dslFunctionalDerivative) {      
478 +      fill(atomRowData.functionalDerivative.begin(),
479 +           atomRowData.functionalDerivative.end(), 0.0);
480 +      fill(atomColData.functionalDerivative.begin(),
481 +           atomColData.functionalDerivative.end(), 0.0);
482 +    }
483 +
484 +    if (storageLayout_ & DataStorage::dslSkippedCharge) {      
485 +      fill(atomRowData.skippedCharge.begin(),
486 +           atomRowData.skippedCharge.end(), 0.0);
487 +      fill(atomColData.skippedCharge.begin(),
488 +           atomColData.skippedCharge.end(), 0.0);
489 +    }
490 +
491 + #endif
492 +    // even in parallel, we need to zero out the local arrays:
493 +
494 +    if (storageLayout_ & DataStorage::dslParticlePot) {      
495 +      fill(snap_->atomData.particlePot.begin(),
496 +           snap_->atomData.particlePot.end(), 0.0);
497 +    }
498 +    
499 +    if (storageLayout_ & DataStorage::dslDensity) {      
500 +      fill(snap_->atomData.density.begin(),
501 +           snap_->atomData.density.end(), 0.0);
502 +    }
503 +    if (storageLayout_ & DataStorage::dslFunctional) {
504 +      fill(snap_->atomData.functional.begin(),
505 +           snap_->atomData.functional.end(), 0.0);
506 +    }
507 +    if (storageLayout_ & DataStorage::dslFunctionalDerivative) {      
508 +      fill(snap_->atomData.functionalDerivative.begin(),
509 +           snap_->atomData.functionalDerivative.end(), 0.0);
510 +    }
511 +    if (storageLayout_ & DataStorage::dslSkippedCharge) {      
512 +      fill(snap_->atomData.skippedCharge.begin(),
513 +           snap_->atomData.skippedCharge.end(), 0.0);
514 +    }
515 +    
516 +  }
517 +
518 +
519 +  void ForceMatrixDecomposition::distributeData()  {
520 +    snap_ = sman_->getCurrentSnapshot();
521 +    storageLayout_ = sman_->getStorageLayout();
522 + #ifdef IS_MPI
523 +    
524      // gather up the atomic positions
525 <    AtomCommVectorI->gather(snap->atomData.position,
526 <                            snap->atomIData.position);
527 <    AtomCommVectorJ->gather(snap->atomData.position,
528 <                            snap->atomJData.position);
525 >    AtomCommVectorRow->gather(snap_->atomData.position,
526 >                              atomRowData.position);
527 >    AtomCommVectorColumn->gather(snap_->atomData.position,
528 >                                 atomColData.position);
529      
530      // gather up the cutoff group positions
531 <    cgCommVectorI->gather(snap->cgData.position,
532 <                          snap->cgIData.position);
533 <    cgCommVectorJ->gather(snap->cgData.position,
534 <                          snap->cgJData.position);
531 >    cgCommVectorRow->gather(snap_->cgData.position,
532 >                            cgRowData.position);
533 >    cgCommVectorColumn->gather(snap_->cgData.position,
534 >                               cgColData.position);
535      
536      // if needed, gather the atomic rotation matrices
537 <    if (snap->atomData.getStorageLayout() & DataStorage::dslAmat) {
538 <      AtomCommMatrixI->gather(snap->atomData.aMat,
539 <                              snap->atomIData.aMat);
540 <      AtomCommMatrixJ->gather(snap->atomData.aMat,
541 <                              snap->atomJData.aMat);
537 >    if (storageLayout_ & DataStorage::dslAmat) {
538 >      AtomCommMatrixRow->gather(snap_->atomData.aMat,
539 >                                atomRowData.aMat);
540 >      AtomCommMatrixColumn->gather(snap_->atomData.aMat,
541 >                                   atomColData.aMat);
542      }
543      
544      // if needed, gather the atomic eletrostatic frames
545 <    if (snap->atomData.getStorageLayout() & DataStorage::dslElectroFrame) {
546 <      AtomCommMatrixI->gather(snap->atomData.electroFrame,
547 <                              snap->atomIData.electroFrame);
548 <      AtomCommMatrixJ->gather(snap->atomData.electroFrame,
549 <                              snap->atomJData.electroFrame);
545 >    if (storageLayout_ & DataStorage::dslElectroFrame) {
546 >      AtomCommMatrixRow->gather(snap_->atomData.electroFrame,
547 >                                atomRowData.electroFrame);
548 >      AtomCommMatrixColumn->gather(snap_->atomData.electroFrame,
549 >                                   atomColData.electroFrame);
550      }
551 +
552   #endif      
553    }
554    
555 <  void ForceDecomposition::collectIntermediateData() {
555 >  /* collects information obtained during the pre-pair loop onto local
556 >   * data structures.
557 >   */
558 >  void ForceMatrixDecomposition::collectIntermediateData() {
559 >    snap_ = sman_->getCurrentSnapshot();
560 >    storageLayout_ = sman_->getStorageLayout();
561   #ifdef IS_MPI
149    Snapshot* snap = sman_->getCurrentSnapshot();
562      
563 <    if (snap->atomData.getStorageLayout() & DataStorage::dslDensity) {
564 <
565 <      AtomCommRealI->scatter(snap->atomIData.density,
566 <                             snap->atomData.density);
567 <
568 <      int n = snap->atomData.density.size();
569 <      std::vector<RealType> rho_tmp(n, 0.0);
570 <      AtomCommRealJ->scatter(snap->atomJData.density, rho_tmp);
563 >    if (storageLayout_ & DataStorage::dslDensity) {
564 >      
565 >      AtomCommRealRow->scatter(atomRowData.density,
566 >                               snap_->atomData.density);
567 >      
568 >      int n = snap_->atomData.density.size();
569 >      vector<RealType> rho_tmp(n, 0.0);
570 >      AtomCommRealColumn->scatter(atomColData.density, rho_tmp);
571        for (int i = 0; i < n; i++)
572 <        snap->atomData.density[i] += rho_tmp[i];
572 >        snap_->atomData.density[i] += rho_tmp[i];
573      }
574   #endif
575    }
576 <  
577 <  void ForceDecomposition::distributeIntermediateData() {
576 >
577 >  /*
578 >   * redistributes information obtained during the pre-pair loop out to
579 >   * row and column-indexed data structures
580 >   */
581 >  void ForceMatrixDecomposition::distributeIntermediateData() {
582 >    snap_ = sman_->getCurrentSnapshot();
583 >    storageLayout_ = sman_->getStorageLayout();
584   #ifdef IS_MPI
585 <    Snapshot* snap = sman_->getCurrentSnapshot();
586 <    if (snap->atomData.getStorageLayout() & DataStorage::dslFunctional) {
587 <      AtomCommRealI->gather(snap->atomData.functional,
588 <                            snap->atomIData.functional);
589 <      AtomCommRealJ->gather(snap->atomData.functional,
172 <                            snap->atomJData.functional);
585 >    if (storageLayout_ & DataStorage::dslFunctional) {
586 >      AtomCommRealRow->gather(snap_->atomData.functional,
587 >                              atomRowData.functional);
588 >      AtomCommRealColumn->gather(snap_->atomData.functional,
589 >                                 atomColData.functional);
590      }
591      
592 <    if (snap->atomData.getStorageLayout() & DataStorage::dslFunctionalDerivative) {
593 <      AtomCommRealI->gather(snap->atomData.functionalDerivative,
594 <                            snap->atomIData.functionalDerivative);
595 <      AtomCommRealJ->gather(snap->atomData.functionalDerivative,
596 <                            snap->atomJData.functionalDerivative);
592 >    if (storageLayout_ & DataStorage::dslFunctionalDerivative) {
593 >      AtomCommRealRow->gather(snap_->atomData.functionalDerivative,
594 >                              atomRowData.functionalDerivative);
595 >      AtomCommRealColumn->gather(snap_->atomData.functionalDerivative,
596 >                                 atomColData.functionalDerivative);
597      }
598   #endif
599    }
600    
601    
602 <  void ForceDecomposition::collectData() {
603 < #ifdef IS_MPI
604 <    Snapshot* snap = sman_->getCurrentSnapshot();
605 <    
606 <    int n = snap->atomData.force.size();
602 >  void ForceMatrixDecomposition::collectData() {
603 >    snap_ = sman_->getCurrentSnapshot();
604 >    storageLayout_ = sman_->getStorageLayout();
605 > #ifdef IS_MPI    
606 >    int n = snap_->atomData.force.size();
607      vector<Vector3d> frc_tmp(n, V3Zero);
608      
609 <    AtomCommVectorI->scatter(snap->atomIData.force, frc_tmp);
609 >    AtomCommVectorRow->scatter(atomRowData.force, frc_tmp);
610      for (int i = 0; i < n; i++) {
611 <      snap->atomData.force[i] += frc_tmp[i];
611 >      snap_->atomData.force[i] += frc_tmp[i];
612        frc_tmp[i] = 0.0;
613      }
614      
615 <    AtomCommVectorJ->scatter(snap->atomJData.force, frc_tmp);
615 >    AtomCommVectorColumn->scatter(atomColData.force, frc_tmp);
616      for (int i = 0; i < n; i++)
617 <      snap->atomData.force[i] += frc_tmp[i];
618 <    
619 <    
203 <    if (snap->atomData.getStorageLayout() & DataStorage::dslTorque) {
617 >      snap_->atomData.force[i] += frc_tmp[i];
618 >        
619 >    if (storageLayout_ & DataStorage::dslTorque) {
620  
621 <      int nt = snap->atomData.force.size();
621 >      int nt = snap_->atomData.torque.size();
622        vector<Vector3d> trq_tmp(nt, V3Zero);
623  
624 <      AtomCommVectorI->scatter(snap->atomIData.torque, trq_tmp);
625 <      for (int i = 0; i < n; i++) {
626 <        snap->atomData.torque[i] += trq_tmp[i];
624 >      AtomCommVectorRow->scatter(atomRowData.torque, trq_tmp);
625 >      for (int i = 0; i < nt; i++) {
626 >        snap_->atomData.torque[i] += trq_tmp[i];
627          trq_tmp[i] = 0.0;
628        }
629        
630 <      AtomCommVectorJ->scatter(snap->atomJData.torque, trq_tmp);
631 <      for (int i = 0; i < n; i++)
632 <        snap->atomData.torque[i] += trq_tmp[i];
630 >      AtomCommVectorColumn->scatter(atomColData.torque, trq_tmp);
631 >      for (int i = 0; i < nt; i++)
632 >        snap_->atomData.torque[i] += trq_tmp[i];
633      }
634 +
635 +    if (storageLayout_ & DataStorage::dslSkippedCharge) {
636 +
637 +      int ns = snap_->atomData.skippedCharge.size();
638 +      vector<RealType> skch_tmp(ns, 0.0);
639 +
640 +      AtomCommRealRow->scatter(atomRowData.skippedCharge, skch_tmp);
641 +      for (int i = 0; i < ns; i++) {
642 +        snap_->atomData.skippedCharge[i] += skch_tmp[i];
643 +        skch_tmp[i] = 0.0;
644 +      }
645 +      
646 +      AtomCommRealColumn->scatter(atomColData.skippedCharge, skch_tmp);
647 +      for (int i = 0; i < ns; i++)
648 +        snap_->atomData.skippedCharge[i] += skch_tmp[i];
649 +    }
650      
651 <    int nLocal = snap->getNumberOfAtoms();
651 >    nLocal_ = snap_->getNumberOfAtoms();
652  
653 <    vector<vector<RealType> > pot_temp(N_INTERACTION_FAMILIES,
654 <                                       vector<RealType> (nLocal, 0.0));
653 >    vector<potVec> pot_temp(nLocal_,
654 >                            Vector<RealType, N_INTERACTION_FAMILIES> (0.0));
655 >
656 >    // scatter/gather pot_row into the members of my column
657 >          
658 >    AtomCommPotRow->scatter(pot_row, pot_temp);
659 >
660 >    for (int ii = 0;  ii < pot_temp.size(); ii++ )
661 >      pairwisePot += pot_temp[ii];
662      
663 <    for (int i = 0; i < N_INTERACTION_FAMILIES; i++) {
664 <      AtomCommRealI->scatter(pot_row[i], pot_temp[i]);
665 <      for (int ii = 0;  ii < pot_temp[i].size(); ii++ ) {
666 <        pot_local[i] += pot_temp[i][ii];
667 <      }
663 >    fill(pot_temp.begin(), pot_temp.end(),
664 >         Vector<RealType, N_INTERACTION_FAMILIES> (0.0));
665 >      
666 >    AtomCommPotColumn->scatter(pot_col, pot_temp);    
667 >    
668 >    for (int ii = 0;  ii < pot_temp.size(); ii++ )
669 >      pairwisePot += pot_temp[ii];    
670 > #endif
671 >
672 >  }
673 >
674 >  int ForceMatrixDecomposition::getNAtomsInRow() {  
675 > #ifdef IS_MPI
676 >    return nAtomsInRow_;
677 > #else
678 >    return nLocal_;
679 > #endif
680 >  }
681 >
682 >  /**
683 >   * returns the list of atoms belonging to this group.  
684 >   */
685 >  vector<int> ForceMatrixDecomposition::getAtomsInGroupRow(int cg1){
686 > #ifdef IS_MPI
687 >    return groupListRow_[cg1];
688 > #else
689 >    return groupList_[cg1];
690 > #endif
691 >  }
692 >
693 >  vector<int> ForceMatrixDecomposition::getAtomsInGroupColumn(int cg2){
694 > #ifdef IS_MPI
695 >    return groupListCol_[cg2];
696 > #else
697 >    return groupList_[cg2];
698 > #endif
699 >  }
700 >  
701 >  Vector3d ForceMatrixDecomposition::getIntergroupVector(int cg1, int cg2){
702 >    Vector3d d;
703 >    
704 > #ifdef IS_MPI
705 >    d = cgColData.position[cg2] - cgRowData.position[cg1];
706 > #else
707 >    d = snap_->cgData.position[cg2] - snap_->cgData.position[cg1];
708 > #endif
709 >    
710 >    snap_->wrapVector(d);
711 >    return d;    
712 >  }
713 >
714 >
715 >  Vector3d ForceMatrixDecomposition::getAtomToGroupVectorRow(int atom1, int cg1){
716 >
717 >    Vector3d d;
718 >    
719 > #ifdef IS_MPI
720 >    d = cgRowData.position[cg1] - atomRowData.position[atom1];
721 > #else
722 >    d = snap_->cgData.position[cg1] - snap_->atomData.position[atom1];
723 > #endif
724 >
725 >    snap_->wrapVector(d);
726 >    return d;    
727 >  }
728 >  
729 >  Vector3d ForceMatrixDecomposition::getAtomToGroupVectorColumn(int atom2, int cg2){
730 >    Vector3d d;
731 >    
732 > #ifdef IS_MPI
733 >    d = cgColData.position[cg2] - atomColData.position[atom2];
734 > #else
735 >    d = snap_->cgData.position[cg2] - snap_->atomData.position[atom2];
736 > #endif
737 >    
738 >    snap_->wrapVector(d);
739 >    return d;    
740 >  }
741 >
742 >  RealType ForceMatrixDecomposition::getMassFactorRow(int atom1) {
743 > #ifdef IS_MPI
744 >    return massFactorsRow[atom1];
745 > #else
746 >    return massFactors[atom1];
747 > #endif
748 >  }
749 >
750 >  RealType ForceMatrixDecomposition::getMassFactorColumn(int atom2) {
751 > #ifdef IS_MPI
752 >    return massFactorsCol[atom2];
753 > #else
754 >    return massFactors[atom2];
755 > #endif
756 >
757 >  }
758 >    
759 >  Vector3d ForceMatrixDecomposition::getInteratomicVector(int atom1, int atom2){
760 >    Vector3d d;
761 >    
762 > #ifdef IS_MPI
763 >    d = atomColData.position[atom2] - atomRowData.position[atom1];
764 > #else
765 >    d = snap_->atomData.position[atom2] - snap_->atomData.position[atom1];
766 > #endif
767 >
768 >    snap_->wrapVector(d);
769 >    return d;    
770 >  }
771 >
772 >  vector<int> ForceMatrixDecomposition::getExcludesForAtom(int atom1) {
773 >    return excludesForAtom[atom1];
774 >  }
775 >
776 >  /**
777 >   * We need to exclude some overcounted interactions that result from
778 >   * the parallel decomposition.
779 >   */
780 >  bool ForceMatrixDecomposition::skipAtomPair(int atom1, int atom2) {
781 >    int unique_id_1, unique_id_2;
782 >
783 > #ifdef IS_MPI
784 >    // in MPI, we have to look up the unique IDs for each atom
785 >    unique_id_1 = AtomRowToGlobal[atom1];
786 >    unique_id_2 = AtomColToGlobal[atom2];
787 >
788 >    // this situation should only arise in MPI simulations
789 >    if (unique_id_1 == unique_id_2) return true;
790 >    
791 >    // this prevents us from doing the pair on multiple processors
792 >    if (unique_id_1 < unique_id_2) {
793 >      if ((unique_id_1 + unique_id_2) % 2 == 0) return true;
794 >    } else {
795 >      if ((unique_id_1 + unique_id_2) % 2 == 1) return true;
796      }
797   #endif
798 +    return false;
799    }
800 +
801 +  /**
802 +   * We need to handle the interactions for atoms who are involved in
803 +   * the same rigid body as well as some short range interactions
804 +   * (bonds, bends, torsions) differently from other interactions.
805 +   * We'll still visit the pairwise routines, but with a flag that
806 +   * tells those routines to exclude the pair from direct long range
807 +   * interactions.  Some indirect interactions (notably reaction
808 +   * field) must still be handled for these pairs.
809 +   */
810 +  bool ForceMatrixDecomposition::excludeAtomPair(int atom1, int atom2) {
811 +    int unique_id_2;
812 +    
813 + #ifdef IS_MPI
814 +    // in MPI, we have to look up the unique IDs for the row atom.
815 +    unique_id_2 = AtomColToGlobal[atom2];
816 + #else
817 +    // in the normal loop, the atom numbers are unique
818 +    unique_id_2 = atom2;
819 + #endif
820 +    
821 +    for (vector<int>::iterator i = excludesForAtom[atom1].begin();
822 +         i != excludesForAtom[atom1].end(); ++i) {
823 +      if ( (*i) == unique_id_2 ) return true;
824 +    }
825 +
826 +    return false;
827 +  }
828 +
829 +
830 +  void ForceMatrixDecomposition::addForceToAtomRow(int atom1, Vector3d fg){
831 + #ifdef IS_MPI
832 +    atomRowData.force[atom1] += fg;
833 + #else
834 +    snap_->atomData.force[atom1] += fg;
835 + #endif
836 +  }
837 +
838 +  void ForceMatrixDecomposition::addForceToAtomColumn(int atom2, Vector3d fg){
839 + #ifdef IS_MPI
840 +    atomColData.force[atom2] += fg;
841 + #else
842 +    snap_->atomData.force[atom2] += fg;
843 + #endif
844 +  }
845 +
846 +    // filling interaction blocks with pointers
847 +  void ForceMatrixDecomposition::fillInteractionData(InteractionData &idat,
848 +                                                     int atom1, int atom2) {
849 +
850 +    idat.excluded = excludeAtomPair(atom1, atom2);
851 +  
852 + #ifdef IS_MPI
853 +    idat.atypes = make_pair( atypesRow[atom1], atypesCol[atom2]);
854 +    //idat.atypes = make_pair( ff_->getAtomType(identsRow[atom1]),
855 +    //                         ff_->getAtomType(identsCol[atom2]) );
856 +    
857 +    if (storageLayout_ & DataStorage::dslAmat) {
858 +      idat.A1 = &(atomRowData.aMat[atom1]);
859 +      idat.A2 = &(atomColData.aMat[atom2]);
860 +    }
861 +    
862 +    if (storageLayout_ & DataStorage::dslElectroFrame) {
863 +      idat.eFrame1 = &(atomRowData.electroFrame[atom1]);
864 +      idat.eFrame2 = &(atomColData.electroFrame[atom2]);
865 +    }
866 +
867 +    if (storageLayout_ & DataStorage::dslTorque) {
868 +      idat.t1 = &(atomRowData.torque[atom1]);
869 +      idat.t2 = &(atomColData.torque[atom2]);
870 +    }
871 +
872 +    if (storageLayout_ & DataStorage::dslDensity) {
873 +      idat.rho1 = &(atomRowData.density[atom1]);
874 +      idat.rho2 = &(atomColData.density[atom2]);
875 +    }
876 +
877 +    if (storageLayout_ & DataStorage::dslFunctional) {
878 +      idat.frho1 = &(atomRowData.functional[atom1]);
879 +      idat.frho2 = &(atomColData.functional[atom2]);
880 +    }
881 +
882 +    if (storageLayout_ & DataStorage::dslFunctionalDerivative) {
883 +      idat.dfrho1 = &(atomRowData.functionalDerivative[atom1]);
884 +      idat.dfrho2 = &(atomColData.functionalDerivative[atom2]);
885 +    }
886 +
887 +    if (storageLayout_ & DataStorage::dslParticlePot) {
888 +      idat.particlePot1 = &(atomRowData.particlePot[atom1]);
889 +      idat.particlePot2 = &(atomColData.particlePot[atom2]);
890 +    }
891 +
892 +    if (storageLayout_ & DataStorage::dslSkippedCharge) {              
893 +      idat.skippedCharge1 = &(atomRowData.skippedCharge[atom1]);
894 +      idat.skippedCharge2 = &(atomColData.skippedCharge[atom2]);
895 +    }
896 +
897 + #else
898 +
899 +    idat.atypes = make_pair( atypesLocal[atom1], atypesLocal[atom2]);
900 +    //idat.atypes = make_pair( ff_->getAtomType(idents[atom1]),
901 +    //                         ff_->getAtomType(idents[atom2]) );
902 +
903 +    if (storageLayout_ & DataStorage::dslAmat) {
904 +      idat.A1 = &(snap_->atomData.aMat[atom1]);
905 +      idat.A2 = &(snap_->atomData.aMat[atom2]);
906 +    }
907 +
908 +    if (storageLayout_ & DataStorage::dslElectroFrame) {
909 +      idat.eFrame1 = &(snap_->atomData.electroFrame[atom1]);
910 +      idat.eFrame2 = &(snap_->atomData.electroFrame[atom2]);
911 +    }
912 +
913 +    if (storageLayout_ & DataStorage::dslTorque) {
914 +      idat.t1 = &(snap_->atomData.torque[atom1]);
915 +      idat.t2 = &(snap_->atomData.torque[atom2]);
916 +    }
917 +
918 +    if (storageLayout_ & DataStorage::dslDensity) {    
919 +      idat.rho1 = &(snap_->atomData.density[atom1]);
920 +      idat.rho2 = &(snap_->atomData.density[atom2]);
921 +    }
922 +
923 +    if (storageLayout_ & DataStorage::dslFunctional) {
924 +      idat.frho1 = &(snap_->atomData.functional[atom1]);
925 +      idat.frho2 = &(snap_->atomData.functional[atom2]);
926 +    }
927 +
928 +    if (storageLayout_ & DataStorage::dslFunctionalDerivative) {
929 +      idat.dfrho1 = &(snap_->atomData.functionalDerivative[atom1]);
930 +      idat.dfrho2 = &(snap_->atomData.functionalDerivative[atom2]);
931 +    }
932 +
933 +    if (storageLayout_ & DataStorage::dslParticlePot) {
934 +      idat.particlePot1 = &(snap_->atomData.particlePot[atom1]);
935 +      idat.particlePot2 = &(snap_->atomData.particlePot[atom2]);
936 +    }
937 +
938 +    if (storageLayout_ & DataStorage::dslSkippedCharge) {
939 +      idat.skippedCharge1 = &(snap_->atomData.skippedCharge[atom1]);
940 +      idat.skippedCharge2 = &(snap_->atomData.skippedCharge[atom2]);
941 +    }
942 + #endif
943 +  }
944 +
945    
946 +  void ForceMatrixDecomposition::unpackInteractionData(InteractionData &idat, int atom1, int atom2) {    
947 + #ifdef IS_MPI
948 +    pot_row[atom1] += 0.5 *  *(idat.pot);
949 +    pot_col[atom2] += 0.5 *  *(idat.pot);
950 +
951 +    atomRowData.force[atom1] += *(idat.f1);
952 +    atomColData.force[atom2] -= *(idat.f1);
953 + #else
954 +    pairwisePot += *(idat.pot);
955 +
956 +    snap_->atomData.force[atom1] += *(idat.f1);
957 +    snap_->atomData.force[atom2] -= *(idat.f1);
958 + #endif
959 +    
960 +  }
961 +
962 +  /*
963 +   * buildNeighborList
964 +   *
965 +   * first element of pair is row-indexed CutoffGroup
966 +   * second element of pair is column-indexed CutoffGroup
967 +   */
968 +  vector<pair<int, int> > ForceMatrixDecomposition::buildNeighborList() {
969 +      
970 +    vector<pair<int, int> > neighborList;
971 +    groupCutoffs cuts;
972 +    bool doAllPairs = false;
973 +
974 + #ifdef IS_MPI
975 +    cellListRow_.clear();
976 +    cellListCol_.clear();
977 + #else
978 +    cellList_.clear();
979 + #endif
980 +
981 +    RealType rList_ = (largestRcut_ + skinThickness_);
982 +    RealType rl2 = rList_ * rList_;
983 +    Snapshot* snap_ = sman_->getCurrentSnapshot();
984 +    Mat3x3d Hmat = snap_->getHmat();
985 +    Vector3d Hx = Hmat.getColumn(0);
986 +    Vector3d Hy = Hmat.getColumn(1);
987 +    Vector3d Hz = Hmat.getColumn(2);
988 +
989 +    nCells_.x() = (int) ( Hx.length() )/ rList_;
990 +    nCells_.y() = (int) ( Hy.length() )/ rList_;
991 +    nCells_.z() = (int) ( Hz.length() )/ rList_;
992 +
993 +    // handle small boxes where the cell offsets can end up repeating cells
994 +    
995 +    if (nCells_.x() < 3) doAllPairs = true;
996 +    if (nCells_.y() < 3) doAllPairs = true;
997 +    if (nCells_.z() < 3) doAllPairs = true;
998 +
999 +    Mat3x3d invHmat = snap_->getInvHmat();
1000 +    Vector3d rs, scaled, dr;
1001 +    Vector3i whichCell;
1002 +    int cellIndex;
1003 +    int nCtot = nCells_.x() * nCells_.y() * nCells_.z();
1004 +
1005 + #ifdef IS_MPI
1006 +    cellListRow_.resize(nCtot);
1007 +    cellListCol_.resize(nCtot);
1008 + #else
1009 +    cellList_.resize(nCtot);
1010 + #endif
1011 +
1012 +    if (!doAllPairs) {
1013 + #ifdef IS_MPI
1014 +
1015 +      for (int i = 0; i < nGroupsInRow_; i++) {
1016 +        rs = cgRowData.position[i];
1017 +        
1018 +        // scaled positions relative to the box vectors
1019 +        scaled = invHmat * rs;
1020 +        
1021 +        // wrap the vector back into the unit box by subtracting integer box
1022 +        // numbers
1023 +        for (int j = 0; j < 3; j++) {
1024 +          scaled[j] -= roundMe(scaled[j]);
1025 +          scaled[j] += 0.5;
1026 +        }
1027 +        
1028 +        // find xyz-indices of cell that cutoffGroup is in.
1029 +        whichCell.x() = nCells_.x() * scaled.x();
1030 +        whichCell.y() = nCells_.y() * scaled.y();
1031 +        whichCell.z() = nCells_.z() * scaled.z();
1032 +        
1033 +        // find single index of this cell:
1034 +        cellIndex = Vlinear(whichCell, nCells_);
1035 +        
1036 +        // add this cutoff group to the list of groups in this cell;
1037 +        cellListRow_[cellIndex].push_back(i);
1038 +      }
1039 +      
1040 +      for (int i = 0; i < nGroupsInCol_; i++) {
1041 +        rs = cgColData.position[i];
1042 +        
1043 +        // scaled positions relative to the box vectors
1044 +        scaled = invHmat * rs;
1045 +        
1046 +        // wrap the vector back into the unit box by subtracting integer box
1047 +        // numbers
1048 +        for (int j = 0; j < 3; j++) {
1049 +          scaled[j] -= roundMe(scaled[j]);
1050 +          scaled[j] += 0.5;
1051 +        }
1052 +        
1053 +        // find xyz-indices of cell that cutoffGroup is in.
1054 +        whichCell.x() = nCells_.x() * scaled.x();
1055 +        whichCell.y() = nCells_.y() * scaled.y();
1056 +        whichCell.z() = nCells_.z() * scaled.z();
1057 +        
1058 +        // find single index of this cell:
1059 +        cellIndex = Vlinear(whichCell, nCells_);
1060 +        
1061 +        // add this cutoff group to the list of groups in this cell;
1062 +        cellListCol_[cellIndex].push_back(i);
1063 +      }
1064 + #else
1065 +      for (int i = 0; i < nGroups_; i++) {
1066 +        rs = snap_->cgData.position[i];
1067 +        
1068 +        // scaled positions relative to the box vectors
1069 +        scaled = invHmat * rs;
1070 +        
1071 +        // wrap the vector back into the unit box by subtracting integer box
1072 +        // numbers
1073 +        for (int j = 0; j < 3; j++) {
1074 +          scaled[j] -= roundMe(scaled[j]);
1075 +          scaled[j] += 0.5;
1076 +        }
1077 +        
1078 +        // find xyz-indices of cell that cutoffGroup is in.
1079 +        whichCell.x() = nCells_.x() * scaled.x();
1080 +        whichCell.y() = nCells_.y() * scaled.y();
1081 +        whichCell.z() = nCells_.z() * scaled.z();
1082 +        
1083 +        // find single index of this cell:
1084 +        cellIndex = Vlinear(whichCell, nCells_);      
1085 +        
1086 +        // add this cutoff group to the list of groups in this cell;
1087 +        cellList_[cellIndex].push_back(i);
1088 +      }
1089 + #endif
1090 +
1091 +      for (int m1z = 0; m1z < nCells_.z(); m1z++) {
1092 +        for (int m1y = 0; m1y < nCells_.y(); m1y++) {
1093 +          for (int m1x = 0; m1x < nCells_.x(); m1x++) {
1094 +            Vector3i m1v(m1x, m1y, m1z);
1095 +            int m1 = Vlinear(m1v, nCells_);
1096 +            
1097 +            for (vector<Vector3i>::iterator os = cellOffsets_.begin();
1098 +                 os != cellOffsets_.end(); ++os) {
1099 +              
1100 +              Vector3i m2v = m1v + (*os);
1101 +              
1102 +              if (m2v.x() >= nCells_.x()) {
1103 +                m2v.x() = 0;          
1104 +              } else if (m2v.x() < 0) {
1105 +                m2v.x() = nCells_.x() - 1;
1106 +              }
1107 +              
1108 +              if (m2v.y() >= nCells_.y()) {
1109 +                m2v.y() = 0;          
1110 +              } else if (m2v.y() < 0) {
1111 +                m2v.y() = nCells_.y() - 1;
1112 +              }
1113 +              
1114 +              if (m2v.z() >= nCells_.z()) {
1115 +                m2v.z() = 0;          
1116 +              } else if (m2v.z() < 0) {
1117 +                m2v.z() = nCells_.z() - 1;
1118 +              }
1119 +              
1120 +              int m2 = Vlinear (m2v, nCells_);
1121 +              
1122 + #ifdef IS_MPI
1123 +              for (vector<int>::iterator j1 = cellListRow_[m1].begin();
1124 +                   j1 != cellListRow_[m1].end(); ++j1) {
1125 +                for (vector<int>::iterator j2 = cellListCol_[m2].begin();
1126 +                     j2 != cellListCol_[m2].end(); ++j2) {
1127 +                  
1128 +                  // Always do this if we're in different cells or if
1129 +                  // we're in the same cell and the global index of the
1130 +                  // j2 cutoff group is less than the j1 cutoff group
1131 +                  
1132 +                  if (m2 != m1 || cgColToGlobal[(*j2)] < cgRowToGlobal[(*j1)]) {
1133 +                    dr = cgColData.position[(*j2)] - cgRowData.position[(*j1)];
1134 +                    snap_->wrapVector(dr);
1135 +                    cuts = getGroupCutoffs( (*j1), (*j2) );
1136 +                    if (dr.lengthSquare() < cuts.third) {
1137 +                      neighborList.push_back(make_pair((*j1), (*j2)));
1138 +                    }
1139 +                  }
1140 +                }
1141 +              }
1142 + #else
1143 +              
1144 +              for (vector<int>::iterator j1 = cellList_[m1].begin();
1145 +                   j1 != cellList_[m1].end(); ++j1) {
1146 +                for (vector<int>::iterator j2 = cellList_[m2].begin();
1147 +                     j2 != cellList_[m2].end(); ++j2) {
1148 +                  
1149 +                  // Always do this if we're in different cells or if
1150 +                  // we're in the same cell and the global index of the
1151 +                  // j2 cutoff group is less than the j1 cutoff group
1152 +                  
1153 +                  if (m2 != m1 || (*j2) < (*j1)) {
1154 +                    dr = snap_->cgData.position[(*j2)] - snap_->cgData.position[(*j1)];
1155 +                    snap_->wrapVector(dr);
1156 +                    cuts = getGroupCutoffs( (*j1), (*j2) );
1157 +                    if (dr.lengthSquare() < cuts.third) {
1158 +                      neighborList.push_back(make_pair((*j1), (*j2)));
1159 +                    }
1160 +                  }
1161 +                }
1162 +              }
1163 + #endif
1164 +            }
1165 +          }
1166 +        }
1167 +      }
1168 +    } else {
1169 +      // branch to do all cutoff group pairs
1170 + #ifdef IS_MPI
1171 +      for (int j1 = 0; j1 < nGroupsInRow_; j1++) {
1172 +        for (int j2 = 0; j2 < nGroupsInCol_; j2++) {      
1173 +          dr = cgColData.position[j2] - cgRowData.position[j1];
1174 +          snap_->wrapVector(dr);
1175 +          cuts = getGroupCutoffs( j1, j2 );
1176 +          if (dr.lengthSquare() < cuts.third) {
1177 +            neighborList.push_back(make_pair(j1, j2));
1178 +          }
1179 +        }
1180 +      }
1181 + #else
1182 +      for (int j1 = 0; j1 < nGroups_ - 1; j1++) {
1183 +        for (int j2 = j1 + 1; j2 < nGroups_; j2++) {
1184 +          dr = snap_->cgData.position[j2] - snap_->cgData.position[j1];
1185 +          snap_->wrapVector(dr);
1186 +          cuts = getGroupCutoffs( j1, j2 );
1187 +          if (dr.lengthSquare() < cuts.third) {
1188 +            neighborList.push_back(make_pair(j1, j2));
1189 +          }
1190 +        }
1191 +      }        
1192 + #endif
1193 +    }
1194 +      
1195 +    // save the local cutoff group positions for the check that is
1196 +    // done on each loop:
1197 +    saved_CG_positions_.clear();
1198 +    for (int i = 0; i < nGroups_; i++)
1199 +      saved_CG_positions_.push_back(snap_->cgData.position[i]);
1200 +    
1201 +    return neighborList;
1202 +  }
1203   } //end namespace OpenMD

Diff Legend

Removed lines
+ Added lines
< Changed lines
> Changed lines