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/ForceMatrixDecomposition.cpp (file contents):
Revision 1562 by gezelter, Thu May 12 17:00:14 2011 UTC vs.
Revision 1583 by gezelter, Thu Jun 16 22:00:08 2011 UTC

# Line 42 | Line 42
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 54 | Line 55 | namespace OpenMD {
55    void ForceMatrixDecomposition::distributeInitialData() {
56      snap_ = sman_->getCurrentSnapshot();
57      storageLayout_ = sman_->getStorageLayout();
58 < #ifdef IS_MPI    
59 <    int nLocal = snap_->getNumberOfAtoms();
59 <    int nGroups = snap_->getNumberOfCutoffGroups();
60 <    
61 <    AtomCommIntRow = new Communicator<Row,int>(nLocal);
62 <    AtomCommRealRow = new Communicator<Row,RealType>(nLocal);
63 <    AtomCommVectorRow = new Communicator<Row,Vector3d>(nLocal);
64 <    AtomCommMatrixRow = new Communicator<Row,Mat3x3d>(nLocal);
58 >    ff_ = info_->getForceField();
59 >    nLocal_ = snap_->getNumberOfAtoms();
60  
61 <    AtomCommIntColumn = new Communicator<Column,int>(nLocal);
62 <    AtomCommRealColumn = new Communicator<Column,RealType>(nLocal);
63 <    AtomCommVectorColumn = new Communicator<Column,Vector3d>(nLocal);
64 <    AtomCommMatrixColumn = new Communicator<Column,Mat3x3d>(nLocal);
61 >    nGroups_ = info_->getNLocalCutoffGroups();
62 >    cerr << "in dId, nGroups = " << nGroups_ << "\n";
63 >    // gather the information for atomtype IDs (atids):
64 >    idents = info_->getIdentArray();
65 >    AtomLocalToGlobal = info_->getGlobalAtomIndices();
66 >    cgLocalToGlobal = info_->getGlobalGroupIndices();
67 >    vector<int> globalGroupMembership = info_->getGlobalGroupMembership();
68 >    massFactors = info_->getMassFactors();
69 >    PairList excludes = info_->getExcludedInteractions();
70 >    PairList oneTwo = info_->getOneTwoInteractions();
71 >    PairList oneThree = info_->getOneThreeInteractions();
72 >    PairList oneFour = info_->getOneFourInteractions();
73  
74 <    cgCommIntRow = new Communicator<Row,int>(nGroups);
75 <    cgCommVectorRow = new Communicator<Row,Vector3d>(nGroups);
76 <    cgCommIntColumn = new Communicator<Column,int>(nGroups);
77 <    cgCommVectorColumn = new Communicator<Column,Vector3d>(nGroups);
74 > #ifdef IS_MPI
75 >
76 >    AtomCommIntRow = new Communicator<Row,int>(nLocal_);
77 >    AtomCommRealRow = new Communicator<Row,RealType>(nLocal_);
78 >    AtomCommVectorRow = new Communicator<Row,Vector3d>(nLocal_);
79 >    AtomCommMatrixRow = new Communicator<Row,Mat3x3d>(nLocal_);
80 >    AtomCommPotRow = new Communicator<Row,potVec>(nLocal_);
81  
82 <    int nAtomsInRow = AtomCommIntRow->getSize();
83 <    int nAtomsInCol = AtomCommIntColumn->getSize();
84 <    int nGroupsInRow = cgCommIntRow->getSize();
85 <    int nGroupsInCol = cgCommIntColumn->getSize();
82 >    AtomCommIntColumn = new Communicator<Column,int>(nLocal_);
83 >    AtomCommRealColumn = new Communicator<Column,RealType>(nLocal_);
84 >    AtomCommVectorColumn = new Communicator<Column,Vector3d>(nLocal_);
85 >    AtomCommMatrixColumn = new Communicator<Column,Mat3x3d>(nLocal_);
86 >    AtomCommPotColumn = new Communicator<Column,potVec>(nLocal_);
87  
88 +    cgCommIntRow = new Communicator<Row,int>(nGroups_);
89 +    cgCommVectorRow = new Communicator<Row,Vector3d>(nGroups_);
90 +    cgCommIntColumn = new Communicator<Column,int>(nGroups_);
91 +    cgCommVectorColumn = new Communicator<Column,Vector3d>(nGroups_);
92 +
93 +    nAtomsInRow_ = AtomCommIntRow->getSize();
94 +    nAtomsInCol_ = AtomCommIntColumn->getSize();
95 +    nGroupsInRow_ = cgCommIntRow->getSize();
96 +    nGroupsInCol_ = cgCommIntColumn->getSize();
97 +
98      // Modify the data storage objects with the correct layouts and sizes:
99 <    atomRowData.resize(nAtomsInRow);
99 >    atomRowData.resize(nAtomsInRow_);
100      atomRowData.setStorageLayout(storageLayout_);
101 <    atomColData.resize(nAtomsInCol);
101 >    atomColData.resize(nAtomsInCol_);
102      atomColData.setStorageLayout(storageLayout_);
103 <    cgRowData.resize(nGroupsInRow);
103 >    cgRowData.resize(nGroupsInRow_);
104      cgRowData.setStorageLayout(DataStorage::dslPosition);
105 <    cgColData.resize(nGroupsInCol);
105 >    cgColData.resize(nGroupsInCol_);
106      cgColData.setStorageLayout(DataStorage::dslPosition);
107 +        
108 +    identsRow.resize(nAtomsInRow_);
109 +    identsCol.resize(nAtomsInCol_);
110      
111 <    vector<vector<RealType> > pot_row(N_INTERACTION_FAMILIES,
112 <                                      vector<RealType> (nAtomsInRow, 0.0));
93 <    vector<vector<RealType> > pot_col(N_INTERACTION_FAMILIES,
94 <                                      vector<RealType> (nAtomsInCol, 0.0));
95 <
96 <
97 <    vector<RealType> pot_local(N_INTERACTION_FAMILIES, 0.0);
111 >    AtomCommIntRow->gather(idents, identsRow);
112 >    AtomCommIntColumn->gather(idents, identsCol);
113      
99 –    // gather the information for atomtype IDs (atids):
100 –    vector<int> identsLocal = info_->getIdentArray();
101 –    identsRow.reserve(nAtomsInRow);
102 –    identsCol.reserve(nAtomsInCol);
103 –    
104 –    AtomCommIntRow->gather(identsLocal, identsRow);
105 –    AtomCommIntColumn->gather(identsLocal, identsCol);
106 –    
107 –    AtomLocalToGlobal = info_->getGlobalAtomIndices();
114      AtomCommIntRow->gather(AtomLocalToGlobal, AtomRowToGlobal);
115      AtomCommIntColumn->gather(AtomLocalToGlobal, AtomColToGlobal);
116      
111 –    cgLocalToGlobal = info_->getGlobalGroupIndices();
117      cgCommIntRow->gather(cgLocalToGlobal, cgRowToGlobal);
118      cgCommIntColumn->gather(cgLocalToGlobal, cgColToGlobal);
119  
120 <    // still need:
121 <    // topoDist
122 <    // exclude
120 >    AtomCommRealRow->gather(massFactors, massFactorsRow);
121 >    AtomCommRealColumn->gather(massFactors, massFactorsCol);
122 >
123 >    groupListRow_.clear();
124 >    groupListRow_.resize(nGroupsInRow_);
125 >    for (int i = 0; i < nGroupsInRow_; i++) {
126 >      int gid = cgRowToGlobal[i];
127 >      for (int j = 0; j < nAtomsInRow_; j++) {
128 >        int aid = AtomRowToGlobal[j];
129 >        if (globalGroupMembership[aid] == gid)
130 >          groupListRow_[i].push_back(j);
131 >      }      
132 >    }
133 >
134 >    groupListCol_.clear();
135 >    groupListCol_.resize(nGroupsInCol_);
136 >    for (int i = 0; i < nGroupsInCol_; i++) {
137 >      int gid = cgColToGlobal[i];
138 >      for (int j = 0; j < nAtomsInCol_; j++) {
139 >        int aid = AtomColToGlobal[j];
140 >        if (globalGroupMembership[aid] == gid)
141 >          groupListCol_[i].push_back(j);
142 >      }      
143 >    }
144 >
145 >    skipsForAtom.clear();
146 >    skipsForAtom.resize(nAtomsInRow_);
147 >    toposForAtom.clear();
148 >    toposForAtom.resize(nAtomsInRow_);
149 >    topoDist.clear();
150 >    topoDist.resize(nAtomsInRow_);
151 >    for (int i = 0; i < nAtomsInRow_; i++) {
152 >      int iglob = AtomRowToGlobal[i];
153 >
154 >      for (int j = 0; j < nAtomsInCol_; j++) {
155 >        int jglob = AtomColToGlobal[j];
156 >
157 >        if (excludes.hasPair(iglob, jglob))
158 >          skipsForAtom[i].push_back(j);      
159 >        
160 >        if (oneTwo.hasPair(iglob, jglob)) {
161 >          toposForAtom[i].push_back(j);
162 >          topoDist[i].push_back(1);
163 >        } else {
164 >          if (oneThree.hasPair(iglob, jglob)) {
165 >            toposForAtom[i].push_back(j);
166 >            topoDist[i].push_back(2);
167 >          } else {
168 >            if (oneFour.hasPair(iglob, jglob)) {
169 >              toposForAtom[i].push_back(j);
170 >              topoDist[i].push_back(3);
171 >            }
172 >          }
173 >        }
174 >      }      
175 >    }
176 >
177   #endif
178 +
179 +    groupList_.clear();
180 +    groupList_.resize(nGroups_);
181 +    for (int i = 0; i < nGroups_; i++) {
182 +      int gid = cgLocalToGlobal[i];
183 +      for (int j = 0; j < nLocal_; j++) {
184 +        int aid = AtomLocalToGlobal[j];
185 +        if (globalGroupMembership[aid] == gid) {
186 +          groupList_[i].push_back(j);
187 +        }
188 +      }      
189 +    }
190 +
191 +    skipsForAtom.clear();
192 +    skipsForAtom.resize(nLocal_);
193 +    toposForAtom.clear();
194 +    toposForAtom.resize(nLocal_);
195 +    topoDist.clear();
196 +    topoDist.resize(nLocal_);
197 +
198 +    for (int i = 0; i < nLocal_; i++) {
199 +      int iglob = AtomLocalToGlobal[i];
200 +
201 +      for (int j = 0; j < nLocal_; j++) {
202 +        int jglob = AtomLocalToGlobal[j];
203 +
204 +        if (excludes.hasPair(iglob, jglob))
205 +          skipsForAtom[i].push_back(j);              
206 +        
207 +        if (oneTwo.hasPair(iglob, jglob)) {
208 +          toposForAtom[i].push_back(j);
209 +          topoDist[i].push_back(1);
210 +        } else {
211 +          if (oneThree.hasPair(iglob, jglob)) {
212 +            toposForAtom[i].push_back(j);
213 +            topoDist[i].push_back(2);
214 +          } else {
215 +            if (oneFour.hasPair(iglob, jglob)) {
216 +              toposForAtom[i].push_back(j);
217 +              topoDist[i].push_back(3);
218 +            }
219 +          }
220 +        }
221 +      }      
222 +    }
223 +    
224 +    createGtypeCutoffMap();
225    }
226 +  
227 +  void ForceMatrixDecomposition::createGtypeCutoffMap() {
228 +
229 +    RealType tol = 1e-6;
230 +    RealType rc;
231 +    int atid;
232 +    set<AtomType*> atypes = info_->getSimulatedAtomTypes();
233 +    vector<RealType> atypeCutoff;
234 +    atypeCutoff.resize( atypes.size() );
235 +      
236 +    for (set<AtomType*>::iterator at = atypes.begin();
237 +         at != atypes.end(); ++at){
238 +      atid = (*at)->getIdent();
239 +
240 +      if (userChoseCutoff_)
241 +        atypeCutoff[atid] = userCutoff_;
242 +      else
243 +        atypeCutoff[atid] = interactionMan_->getSuggestedCutoffRadius(*at);
244 +    }
245 +
246 +    vector<RealType> gTypeCutoffs;
247 +
248 +    // first we do a single loop over the cutoff groups to find the
249 +    // largest cutoff for any atypes present in this group.
250 + #ifdef IS_MPI
251 +    vector<RealType> groupCutoffRow(nGroupsInRow_, 0.0);
252 +    groupRowToGtype.resize(nGroupsInRow_);
253 +    for (int cg1 = 0; cg1 < nGroupsInRow_; cg1++) {
254 +      vector<int> atomListRow = getAtomsInGroupRow(cg1);
255 +      for (vector<int>::iterator ia = atomListRow.begin();
256 +           ia != atomListRow.end(); ++ia) {            
257 +        int atom1 = (*ia);
258 +        atid = identsRow[atom1];
259 +        if (atypeCutoff[atid] > groupCutoffRow[cg1]) {
260 +          groupCutoffRow[cg1] = atypeCutoff[atid];
261 +        }
262 +      }
263 +
264 +      bool gTypeFound = false;
265 +      for (int gt = 0; gt < gTypeCutoffs.size(); gt++) {
266 +        if (abs(groupCutoffRow[cg1] - gTypeCutoffs[gt]) < tol) {
267 +          groupRowToGtype[cg1] = gt;
268 +          gTypeFound = true;
269 +        }
270 +      }
271 +      if (!gTypeFound) {
272 +        gTypeCutoffs.push_back( groupCutoffRow[cg1] );
273 +        groupRowToGtype[cg1] = gTypeCutoffs.size() - 1;
274 +      }
275 +      
276 +    }
277 +    vector<RealType> groupCutoffCol(nGroupsInCol_, 0.0);
278 +    groupColToGtype.resize(nGroupsInCol_);
279 +    for (int cg2 = 0; cg2 < nGroupsInCol_; cg2++) {
280 +      vector<int> atomListCol = getAtomsInGroupColumn(cg2);
281 +      for (vector<int>::iterator jb = atomListCol.begin();
282 +           jb != atomListCol.end(); ++jb) {            
283 +        int atom2 = (*jb);
284 +        atid = identsCol[atom2];
285 +        if (atypeCutoff[atid] > groupCutoffCol[cg2]) {
286 +          groupCutoffCol[cg2] = atypeCutoff[atid];
287 +        }
288 +      }
289 +      bool gTypeFound = false;
290 +      for (int gt = 0; gt < gTypeCutoffs.size(); gt++) {
291 +        if (abs(groupCutoffCol[cg2] - gTypeCutoffs[gt]) < tol) {
292 +          groupColToGtype[cg2] = gt;
293 +          gTypeFound = true;
294 +        }
295 +      }
296 +      if (!gTypeFound) {
297 +        gTypeCutoffs.push_back( groupCutoffCol[cg2] );
298 +        groupColToGtype[cg2] = gTypeCutoffs.size() - 1;
299 +      }
300 +    }
301 + #else
302 +
303 +    vector<RealType> groupCutoff(nGroups_, 0.0);
304 +    groupToGtype.resize(nGroups_);
305 +
306 +    cerr << "nGroups = " << nGroups_ << "\n";
307 +    for (int cg1 = 0; cg1 < nGroups_; cg1++) {
308 +
309 +      groupCutoff[cg1] = 0.0;
310 +      vector<int> atomList = getAtomsInGroupRow(cg1);
311 +
312 +      for (vector<int>::iterator ia = atomList.begin();
313 +           ia != atomList.end(); ++ia) {            
314 +        int atom1 = (*ia);
315 +        atid = idents[atom1];
316 +        if (atypeCutoff[atid] > groupCutoff[cg1]) {
317 +          groupCutoff[cg1] = atypeCutoff[atid];
318 +        }
319 +      }
320 +
321 +      bool gTypeFound = false;
322 +      for (int gt = 0; gt < gTypeCutoffs.size(); gt++) {
323 +        if (abs(groupCutoff[cg1] - gTypeCutoffs[gt]) < tol) {
324 +          groupToGtype[cg1] = gt;
325 +          gTypeFound = true;
326 +        }
327 +      }
328 +      if (!gTypeFound) {
329 +        gTypeCutoffs.push_back( groupCutoff[cg1] );
330 +        groupToGtype[cg1] = gTypeCutoffs.size() - 1;
331 +      }      
332 +    }
333 + #endif
334 +
335 +    cerr << "gTypeCutoffs.size() = " << gTypeCutoffs.size() << "\n";
336 +    // Now we find the maximum group cutoff value present in the simulation
337 +
338 +    RealType groupMax = *max_element(gTypeCutoffs.begin(), gTypeCutoffs.end());
339 +
340 + #ifdef IS_MPI
341 +    MPI::COMM_WORLD.Allreduce(&groupMax, &groupMax, 1, MPI::REALTYPE, MPI::MAX);
342 + #endif
343      
344 +    RealType tradRcut = groupMax;
345  
346 +    for (int i = 0; i < gTypeCutoffs.size();  i++) {
347 +      for (int j = 0; j < gTypeCutoffs.size();  j++) {      
348 +        RealType thisRcut;
349 +        switch(cutoffPolicy_) {
350 +        case TRADITIONAL:
351 +          thisRcut = tradRcut;
352 +          break;
353 +        case MIX:
354 +          thisRcut = 0.5 * (gTypeCutoffs[i] + gTypeCutoffs[j]);
355 +          break;
356 +        case MAX:
357 +          thisRcut = max(gTypeCutoffs[i], gTypeCutoffs[j]);
358 +          break;
359 +        default:
360 +          sprintf(painCave.errMsg,
361 +                  "ForceMatrixDecomposition::createGtypeCutoffMap "
362 +                  "hit an unknown cutoff policy!\n");
363 +          painCave.severity = OPENMD_ERROR;
364 +          painCave.isFatal = 1;
365 +          simError();
366 +          break;
367 +        }
368  
369 +        pair<int,int> key = make_pair(i,j);
370 +        gTypeCutoffMap[key].first = thisRcut;
371 +
372 +        if (thisRcut > largestRcut_) largestRcut_ = thisRcut;
373 +
374 +        gTypeCutoffMap[key].second = thisRcut*thisRcut;
375 +        
376 +        gTypeCutoffMap[key].third = pow(thisRcut + skinThickness_, 2);
377 +
378 +        // sanity check
379 +        
380 +        if (userChoseCutoff_) {
381 +          if (abs(gTypeCutoffMap[key].first - userCutoff_) > 0.0001) {
382 +            sprintf(painCave.errMsg,
383 +                    "ForceMatrixDecomposition::createGtypeCutoffMap "
384 +                    "user-specified rCut (%lf) does not match computed group Cutoff\n", userCutoff_);
385 +            painCave.severity = OPENMD_ERROR;
386 +            painCave.isFatal = 1;
387 +            simError();            
388 +          }
389 +        }
390 +      }
391 +    }
392 +  }
393 +
394 +
395 +  groupCutoffs ForceMatrixDecomposition::getGroupCutoffs(int cg1, int cg2) {
396 +    int i, j;  
397 + #ifdef IS_MPI
398 +    i = groupRowToGtype[cg1];
399 +    j = groupColToGtype[cg2];
400 + #else
401 +    i = groupToGtype[cg1];
402 +    j = groupToGtype[cg2];
403 + #endif    
404 +    return gTypeCutoffMap[make_pair(i,j)];
405 +  }
406 +
407 +  int ForceMatrixDecomposition::getTopologicalDistance(int atom1, int atom2) {
408 +    for (int j = 0; j < toposForAtom[atom1].size(); j++) {
409 +      if (toposForAtom[atom1][j] == atom2)
410 +        return topoDist[atom1][j];
411 +    }
412 +    return 0;
413 +  }
414 +
415 +  void ForceMatrixDecomposition::zeroWorkArrays() {
416 +    pairwisePot = 0.0;
417 +    embeddingPot = 0.0;
418 +
419 + #ifdef IS_MPI
420 +    if (storageLayout_ & DataStorage::dslForce) {
421 +      fill(atomRowData.force.begin(), atomRowData.force.end(), V3Zero);
422 +      fill(atomColData.force.begin(), atomColData.force.end(), V3Zero);
423 +    }
424 +
425 +    if (storageLayout_ & DataStorage::dslTorque) {
426 +      fill(atomRowData.torque.begin(), atomRowData.torque.end(), V3Zero);
427 +      fill(atomColData.torque.begin(), atomColData.torque.end(), V3Zero);
428 +    }
429 +    
430 +    fill(pot_row.begin(), pot_row.end(),
431 +         Vector<RealType, N_INTERACTION_FAMILIES> (0.0));
432 +
433 +    fill(pot_col.begin(), pot_col.end(),
434 +         Vector<RealType, N_INTERACTION_FAMILIES> (0.0));  
435 +
436 +    if (storageLayout_ & DataStorage::dslParticlePot) {    
437 +      fill(atomRowData.particlePot.begin(), atomRowData.particlePot.end(), 0.0);
438 +      fill(atomColData.particlePot.begin(), atomColData.particlePot.end(), 0.0);
439 +    }
440 +
441 +    if (storageLayout_ & DataStorage::dslDensity) {      
442 +      fill(atomRowData.density.begin(), atomRowData.density.end(), 0.0);
443 +      fill(atomColData.density.begin(), atomColData.density.end(), 0.0);
444 +    }
445 +
446 +    if (storageLayout_ & DataStorage::dslFunctional) {  
447 +      fill(atomRowData.functional.begin(), atomRowData.functional.end(), 0.0);
448 +      fill(atomColData.functional.begin(), atomColData.functional.end(), 0.0);
449 +    }
450 +
451 +    if (storageLayout_ & DataStorage::dslFunctionalDerivative) {      
452 +      fill(atomRowData.functionalDerivative.begin(),
453 +           atomRowData.functionalDerivative.end(), 0.0);
454 +      fill(atomColData.functionalDerivative.begin(),
455 +           atomColData.functionalDerivative.end(), 0.0);
456 +    }
457 +
458 + #else
459 +    
460 +    if (storageLayout_ & DataStorage::dslParticlePot) {      
461 +      fill(snap_->atomData.particlePot.begin(),
462 +           snap_->atomData.particlePot.end(), 0.0);
463 +    }
464 +    
465 +    if (storageLayout_ & DataStorage::dslDensity) {      
466 +      fill(snap_->atomData.density.begin(),
467 +           snap_->atomData.density.end(), 0.0);
468 +    }
469 +    if (storageLayout_ & DataStorage::dslFunctional) {
470 +      fill(snap_->atomData.functional.begin(),
471 +           snap_->atomData.functional.end(), 0.0);
472 +    }
473 +    if (storageLayout_ & DataStorage::dslFunctionalDerivative) {      
474 +      fill(snap_->atomData.functionalDerivative.begin(),
475 +           snap_->atomData.functionalDerivative.end(), 0.0);
476 +    }
477 + #endif
478 +    
479 +  }
480 +
481 +
482    void ForceMatrixDecomposition::distributeData()  {
483      snap_ = sman_->getCurrentSnapshot();
484      storageLayout_ = sman_->getStorageLayout();
# Line 155 | Line 514 | namespace OpenMD {
514   #endif      
515    }
516    
517 +  /* collects information obtained during the pre-pair loop onto local
518 +   * data structures.
519 +   */
520    void ForceMatrixDecomposition::collectIntermediateData() {
521      snap_ = sman_->getCurrentSnapshot();
522      storageLayout_ = sman_->getStorageLayout();
# Line 166 | Line 528 | namespace OpenMD {
528                                 snap_->atomData.density);
529        
530        int n = snap_->atomData.density.size();
531 <      std::vector<RealType> rho_tmp(n, 0.0);
531 >      vector<RealType> rho_tmp(n, 0.0);
532        AtomCommRealColumn->scatter(atomColData.density, rho_tmp);
533        for (int i = 0; i < n; i++)
534          snap_->atomData.density[i] += rho_tmp[i];
535      }
536   #endif
537    }
538 <  
538 >
539 >  /*
540 >   * redistributes information obtained during the pre-pair loop out to
541 >   * row and column-indexed data structures
542 >   */
543    void ForceMatrixDecomposition::distributeIntermediateData() {
544      snap_ = sman_->getCurrentSnapshot();
545      storageLayout_ = sman_->getStorageLayout();
# Line 229 | Line 595 | namespace OpenMD {
595          snap_->atomData.torque[i] += trq_tmp[i];
596      }
597      
598 <    int nLocal = snap_->getNumberOfAtoms();
598 >    nLocal_ = snap_->getNumberOfAtoms();
599  
600 <    vector<vector<RealType> > pot_temp(N_INTERACTION_FAMILIES,
601 <                                       vector<RealType> (nLocal, 0.0));
600 >    vector<potVec> pot_temp(nLocal_,
601 >                            Vector<RealType, N_INTERACTION_FAMILIES> (0.0));
602 >
603 >    // scatter/gather pot_row into the members of my column
604 >          
605 >    AtomCommPotRow->scatter(pot_row, pot_temp);
606 >
607 >    for (int ii = 0;  ii < pot_temp.size(); ii++ )
608 >      pairwisePot += pot_temp[ii];
609      
610 <    for (int i = 0; i < N_INTERACTION_FAMILIES; i++) {
611 <      AtomCommRealRow->scatter(pot_row[i], pot_temp[i]);
612 <      for (int ii = 0;  ii < pot_temp[i].size(); ii++ ) {
613 <        pot_local[i] += pot_temp[i][ii];
614 <      }
615 <    }
610 >    fill(pot_temp.begin(), pot_temp.end(),
611 >         Vector<RealType, N_INTERACTION_FAMILIES> (0.0));
612 >      
613 >    AtomCommPotColumn->scatter(pot_col, pot_temp);    
614 >    
615 >    for (int ii = 0;  ii < pot_temp.size(); ii++ )
616 >      pairwisePot += pot_temp[ii];    
617   #endif
618 +
619    }
620  
621 +  int ForceMatrixDecomposition::getNAtomsInRow() {  
622 + #ifdef IS_MPI
623 +    return nAtomsInRow_;
624 + #else
625 +    return nLocal_;
626 + #endif
627 +  }
628 +
629 +  /**
630 +   * returns the list of atoms belonging to this group.  
631 +   */
632 +  vector<int> ForceMatrixDecomposition::getAtomsInGroupRow(int cg1){
633 + #ifdef IS_MPI
634 +    return groupListRow_[cg1];
635 + #else
636 +    return groupList_[cg1];
637 + #endif
638 +  }
639 +
640 +  vector<int> ForceMatrixDecomposition::getAtomsInGroupColumn(int cg2){
641 + #ifdef IS_MPI
642 +    return groupListCol_[cg2];
643 + #else
644 +    return groupList_[cg2];
645 + #endif
646 +  }
647    
648    Vector3d ForceMatrixDecomposition::getIntergroupVector(int cg1, int cg2){
649      Vector3d d;
# Line 284 | Line 685 | namespace OpenMD {
685      snap_->wrapVector(d);
686      return d;    
687    }
688 +
689 +  RealType ForceMatrixDecomposition::getMassFactorRow(int atom1) {
690 + #ifdef IS_MPI
691 +    return massFactorsRow[atom1];
692 + #else
693 +    return massFactors[atom1];
694 + #endif
695 +  }
696 +
697 +  RealType ForceMatrixDecomposition::getMassFactorColumn(int atom2) {
698 + #ifdef IS_MPI
699 +    return massFactorsCol[atom2];
700 + #else
701 +    return massFactors[atom2];
702 + #endif
703 +
704 +  }
705      
706    Vector3d ForceMatrixDecomposition::getInteratomicVector(int atom1, int atom2){
707      Vector3d d;
# Line 298 | Line 716 | namespace OpenMD {
716      return d;    
717    }
718  
719 +  vector<int> ForceMatrixDecomposition::getSkipsForAtom(int atom1) {
720 +    return skipsForAtom[atom1];
721 +  }
722 +
723 +  /**
724 +   * There are a number of reasons to skip a pair or a
725 +   * particle. Mostly we do this to exclude atoms who are involved in
726 +   * short range interactions (bonds, bends, torsions), but we also
727 +   * need to exclude some overcounted interactions that result from
728 +   * the parallel decomposition.
729 +   */
730 +  bool ForceMatrixDecomposition::skipAtomPair(int atom1, int atom2) {
731 +    int unique_id_1, unique_id_2;
732 +
733 + #ifdef IS_MPI
734 +    // in MPI, we have to look up the unique IDs for each atom
735 +    unique_id_1 = AtomRowToGlobal[atom1];
736 +    unique_id_2 = AtomColToGlobal[atom2];
737 +
738 +    // this situation should only arise in MPI simulations
739 +    if (unique_id_1 == unique_id_2) return true;
740 +    
741 +    // this prevents us from doing the pair on multiple processors
742 +    if (unique_id_1 < unique_id_2) {
743 +      if ((unique_id_1 + unique_id_2) % 2 == 0) return true;
744 +    } else {
745 +      if ((unique_id_1 + unique_id_2) % 2 == 1) return true;
746 +    }
747 + #else
748 +    // in the normal loop, the atom numbers are unique
749 +    unique_id_1 = atom1;
750 +    unique_id_2 = atom2;
751 + #endif
752 +    
753 +    for (vector<int>::iterator i = skipsForAtom[atom1].begin();
754 +         i != skipsForAtom[atom1].end(); ++i) {
755 +      if ( (*i) == unique_id_2 ) return true;
756 +    }
757 +
758 +    return false;
759 +  }
760 +
761 +
762    void ForceMatrixDecomposition::addForceToAtomRow(int atom1, Vector3d fg){
763   #ifdef IS_MPI
764      atomRowData.force[atom1] += fg;
# Line 312 | Line 773 | namespace OpenMD {
773   #else
774      snap_->atomData.force[atom2] += fg;
775   #endif
315 –
776    }
777  
778      // filling interaction blocks with pointers
779 <  InteractionData ForceMatrixDecomposition::fillInteractionData(int atom1, int atom2) {    
780 <
321 <    InteractionData idat;
779 >  void ForceMatrixDecomposition::fillInteractionData(InteractionData &idat,
780 >                                                     int atom1, int atom2) {    
781   #ifdef IS_MPI
782 +    
783 +    idat.atypes = make_pair( ff_->getAtomType(identsRow[atom1]),
784 +                             ff_->getAtomType(identsCol[atom2]) );
785 +    
786      if (storageLayout_ & DataStorage::dslAmat) {
787        idat.A1 = &(atomRowData.aMat[atom1]);
788        idat.A2 = &(atomColData.aMat[atom2]);
789      }
790 <
790 >    
791      if (storageLayout_ & DataStorage::dslElectroFrame) {
792        idat.eFrame1 = &(atomRowData.electroFrame[atom1]);
793        idat.eFrame2 = &(atomColData.electroFrame[atom2]);
# Line 338 | Line 801 | namespace OpenMD {
801      if (storageLayout_ & DataStorage::dslDensity) {
802        idat.rho1 = &(atomRowData.density[atom1]);
803        idat.rho2 = &(atomColData.density[atom2]);
804 +    }
805 +
806 +    if (storageLayout_ & DataStorage::dslFunctional) {
807 +      idat.frho1 = &(atomRowData.functional[atom1]);
808 +      idat.frho2 = &(atomColData.functional[atom2]);
809      }
810  
811      if (storageLayout_ & DataStorage::dslFunctionalDerivative) {
812        idat.dfrho1 = &(atomRowData.functionalDerivative[atom1]);
813        idat.dfrho2 = &(atomColData.functionalDerivative[atom2]);
814      }
815 +
816 +    if (storageLayout_ & DataStorage::dslParticlePot) {
817 +      idat.particlePot1 = &(atomRowData.particlePot[atom1]);
818 +      idat.particlePot2 = &(atomColData.particlePot[atom2]);
819 +    }
820 +
821   #else
822 +
823 +    idat.atypes = make_pair( ff_->getAtomType(idents[atom1]),
824 +                             ff_->getAtomType(idents[atom2]) );
825 +
826      if (storageLayout_ & DataStorage::dslAmat) {
827        idat.A1 = &(snap_->atomData.aMat[atom1]);
828        idat.A2 = &(snap_->atomData.aMat[atom2]);
# Line 360 | Line 838 | namespace OpenMD {
838        idat.t2 = &(snap_->atomData.torque[atom2]);
839      }
840  
841 <    if (storageLayout_ & DataStorage::dslDensity) {
841 >    if (storageLayout_ & DataStorage::dslDensity) {    
842        idat.rho1 = &(snap_->atomData.density[atom1]);
843        idat.rho2 = &(snap_->atomData.density[atom2]);
844      }
845  
846 +    if (storageLayout_ & DataStorage::dslFunctional) {
847 +      idat.frho1 = &(snap_->atomData.functional[atom1]);
848 +      idat.frho2 = &(snap_->atomData.functional[atom2]);
849 +    }
850 +
851      if (storageLayout_ & DataStorage::dslFunctionalDerivative) {
852        idat.dfrho1 = &(snap_->atomData.functionalDerivative[atom1]);
853        idat.dfrho2 = &(snap_->atomData.functionalDerivative[atom2]);
854      }
855 +
856 +    if (storageLayout_ & DataStorage::dslParticlePot) {
857 +      idat.particlePot1 = &(snap_->atomData.particlePot[atom1]);
858 +      idat.particlePot2 = &(snap_->atomData.particlePot[atom2]);
859 +    }
860 +
861   #endif
373 –    
862    }
863 <  InteractionData ForceMatrixDecomposition::fillSkipData(int atom1, int atom2){
864 <    InteractionData idat;
865 <    skippedCharge1
378 <      skippedCharge2
379 <      rij
380 <      d
381 <    electroMult
382 <    sw
383 <    f
863 >
864 >  
865 >  void ForceMatrixDecomposition::unpackInteractionData(InteractionData &idat, int atom1, int atom2) {    
866   #ifdef IS_MPI
867 +    pot_row[atom1] += 0.5 *  *(idat.pot);
868 +    pot_col[atom2] += 0.5 *  *(idat.pot);
869  
870 +    atomRowData.force[atom1] += *(idat.f1);
871 +    atomColData.force[atom2] -= *(idat.f1);
872 + #else
873 +    pairwisePot += *(idat.pot);
874 +
875 +    snap_->atomData.force[atom1] += *(idat.f1);
876 +    snap_->atomData.force[atom2] -= *(idat.f1);
877 + #endif
878 +
879 +  }
880 +
881 +
882 +  void ForceMatrixDecomposition::fillSkipData(InteractionData &idat,
883 +                                              int atom1, int atom2) {
884 +    // Still Missing:: skippedCharge fill must be added to DataStorage
885 + #ifdef IS_MPI
886 +    idat.atypes = make_pair( ff_->getAtomType(identsRow[atom1]),
887 +                             ff_->getAtomType(identsCol[atom2]) );
888 +
889      if (storageLayout_ & DataStorage::dslElectroFrame) {
890        idat.eFrame1 = &(atomRowData.electroFrame[atom1]);
891        idat.eFrame2 = &(atomColData.electroFrame[atom2]);
# Line 391 | Line 894 | namespace OpenMD {
894        idat.t1 = &(atomRowData.torque[atom1]);
895        idat.t2 = &(atomColData.torque[atom2]);
896      }
897 + #else
898 +    idat.atypes = make_pair( ff_->getAtomType(idents[atom1]),
899 +                             ff_->getAtomType(idents[atom2]) );
900  
901 <    
901 >    if (storageLayout_ & DataStorage::dslElectroFrame) {
902 >      idat.eFrame1 = &(snap_->atomData.electroFrame[atom1]);
903 >      idat.eFrame2 = &(snap_->atomData.electroFrame[atom2]);
904 >    }
905 >    if (storageLayout_ & DataStorage::dslTorque) {
906 >      idat.t1 = &(snap_->atomData.torque[atom1]);
907 >      idat.t2 = &(snap_->atomData.torque[atom2]);
908 >    }
909 > #endif    
910    }
911 <  SelfData ForceMatrixDecomposition::fillSelfData(int atom1) {
911 >
912 >
913 >  void ForceMatrixDecomposition::unpackSkipData(InteractionData &idat, int atom1, int atom2) {    
914 > #ifdef IS_MPI
915 >    pot_row[atom1] += 0.5 *  *(idat.pot);
916 >    pot_col[atom2] += 0.5 *  *(idat.pot);
917 > #else
918 >    pairwisePot += *(idat.pot);  
919 > #endif
920 >
921    }
922  
923  
# Line 404 | Line 927 | namespace OpenMD {
927     * first element of pair is row-indexed CutoffGroup
928     * second element of pair is column-indexed CutoffGroup
929     */
930 <  vector<pair<int, int> >  buildNeighborList() {
931 <    Vector3d dr, invWid, rs, shift;
932 <    Vector3i cc, m1v, m2s;
933 <    RealType rrNebr;
934 <    int c, j1, j2, m1, m1x, m1y, m1z, m2, n, offset;
930 >  vector<pair<int, int> > ForceMatrixDecomposition::buildNeighborList() {
931 >      
932 >    vector<pair<int, int> > neighborList;
933 >    groupCutoffs cuts;
934 > #ifdef IS_MPI
935 >    cellListRow_.clear();
936 >    cellListCol_.clear();
937 > #else
938 >    cellList_.clear();
939 > #endif
940  
941 <
942 <    vector<pair<int, int> > neighborList;  
943 <    Vector3i nCells;
416 <    Vector3d invWid, r;
417 <
418 <    rList_ = (rCut_ + skinThickness_);
419 <    rl2 = rList_ * rList_;
420 <
421 <    snap_ = sman_->getCurrentSnapshot();
941 >    RealType rList_ = (largestRcut_ + skinThickness_);
942 >    RealType rl2 = rList_ * rList_;
943 >    Snapshot* snap_ = sman_->getCurrentSnapshot();
944      Mat3x3d Hmat = snap_->getHmat();
945      Vector3d Hx = Hmat.getColumn(0);
946      Vector3d Hy = Hmat.getColumn(1);
947      Vector3d Hz = Hmat.getColumn(2);
948  
949 <    nCells.x() = (int) ( Hx.length() )/ rList_;
950 <    nCells.y() = (int) ( Hy.length() )/ rList_;
951 <    nCells.z() = (int) ( Hz.length() )/ rList_;
949 >    nCells_.x() = (int) ( Hx.length() )/ rList_;
950 >    nCells_.y() = (int) ( Hy.length() )/ rList_;
951 >    nCells_.z() = (int) ( Hz.length() )/ rList_;
952  
953 <    for (i = 0; i < nGroupsInRow; i++) {
953 >    Mat3x3d invHmat = snap_->getInvHmat();
954 >    Vector3d rs, scaled, dr;
955 >    Vector3i whichCell;
956 >    int cellIndex;
957 >    int nCtot = nCells_.x() * nCells_.y() * nCells_.z();
958 >
959 > #ifdef IS_MPI
960 >    cellListRow_.resize(nCtot);
961 >    cellListCol_.resize(nCtot);
962 > #else
963 >    cellList_.resize(nCtot);
964 > #endif
965 >
966 > #ifdef IS_MPI
967 >    for (int i = 0; i < nGroupsInRow_; i++) {
968        rs = cgRowData.position[i];
969 <      snap_->scaleVector(rs);    
969 >
970 >      // scaled positions relative to the box vectors
971 >      scaled = invHmat * rs;
972 >
973 >      // wrap the vector back into the unit box by subtracting integer box
974 >      // numbers
975 >      for (int j = 0; j < 3; j++) {
976 >        scaled[j] -= roundMe(scaled[j]);
977 >        scaled[j] += 0.5;
978 >      }
979 >    
980 >      // find xyz-indices of cell that cutoffGroup is in.
981 >      whichCell.x() = nCells_.x() * scaled.x();
982 >      whichCell.y() = nCells_.y() * scaled.y();
983 >      whichCell.z() = nCells_.z() * scaled.z();
984 >
985 >      // find single index of this cell:
986 >      cellIndex = Vlinear(whichCell, nCells_);
987 >
988 >      // add this cutoff group to the list of groups in this cell;
989 >      cellListRow_[cellIndex].push_back(i);
990      }
435 –    
991  
992 <    VDiv (invWid, cells, region);
993 <    for (n = nMol; n < nMol + cells.componentProduct(); n ++) cellList[n] = -1;
994 <    for (n = 0; n < nMol; n ++) {
995 <      VSAdd (rs, mol[n].r, 0.5, region);
996 <      VMul (cc, rs, invWid);
997 <      c = VLinear (cc, cells) + nMol;
998 <      cellList[n] = cellList[c];
999 <      cellList[c] = n;
992 >    for (int i = 0; i < nGroupsInCol_; i++) {
993 >      rs = cgColData.position[i];
994 >
995 >      // scaled positions relative to the box vectors
996 >      scaled = invHmat * rs;
997 >
998 >      // wrap the vector back into the unit box by subtracting integer box
999 >      // numbers
1000 >      for (int j = 0; j < 3; j++) {
1001 >        scaled[j] -= roundMe(scaled[j]);
1002 >        scaled[j] += 0.5;
1003 >      }
1004 >
1005 >      // find xyz-indices of cell that cutoffGroup is in.
1006 >      whichCell.x() = nCells_.x() * scaled.x();
1007 >      whichCell.y() = nCells_.y() * scaled.y();
1008 >      whichCell.z() = nCells_.z() * scaled.z();
1009 >
1010 >      // find single index of this cell:
1011 >      cellIndex = Vlinear(whichCell, nCells_);
1012 >
1013 >      // add this cutoff group to the list of groups in this cell;
1014 >      cellListCol_[cellIndex].push_back(i);
1015      }
1016 <    nebrTabLen = 0;
1017 <    for (m1z = 0; m1z < cells.z(); m1z++) {
1018 <      for (m1y = 0; m1y < cells.y(); m1y++) {
1019 <        for (m1x = 0; m1x < cells.x(); m1x++) {
1016 > #else
1017 >    for (int i = 0; i < nGroups_; i++) {
1018 >      rs = snap_->cgData.position[i];
1019 >
1020 >      // scaled positions relative to the box vectors
1021 >      scaled = invHmat * rs;
1022 >
1023 >      // wrap the vector back into the unit box by subtracting integer box
1024 >      // numbers
1025 >      for (int j = 0; j < 3; j++) {
1026 >        scaled[j] -= roundMe(scaled[j]);
1027 >        scaled[j] += 0.5;
1028 >      }
1029 >
1030 >      // find xyz-indices of cell that cutoffGroup is in.
1031 >      whichCell.x() = nCells_.x() * scaled.x();
1032 >      whichCell.y() = nCells_.y() * scaled.y();
1033 >      whichCell.z() = nCells_.z() * scaled.z();
1034 >
1035 >      // find single index of this cell:
1036 >      cellIndex = Vlinear(whichCell, nCells_);      
1037 >
1038 >      // add this cutoff group to the list of groups in this cell;
1039 >      cellList_[cellIndex].push_back(i);
1040 >    }
1041 > #endif
1042 >
1043 >    for (int m1z = 0; m1z < nCells_.z(); m1z++) {
1044 >      for (int m1y = 0; m1y < nCells_.y(); m1y++) {
1045 >        for (int m1x = 0; m1x < nCells_.x(); m1x++) {
1046            Vector3i m1v(m1x, m1y, m1z);
1047 <          m1 = VLinear(m1v, cells) + nMol;
452 <          for (offset = 0; offset < nOffset_; offset++) {
453 <            m2v = m1v + cellOffsets_[offset];
454 <            shift = V3Zero();
1047 >          int m1 = Vlinear(m1v, nCells_);
1048  
1049 <            if (m2v.x() >= cells.x) {
1049 >          for (vector<Vector3i>::iterator os = cellOffsets_.begin();
1050 >               os != cellOffsets_.end(); ++os) {
1051 >            
1052 >            Vector3i m2v = m1v + (*os);
1053 >            
1054 >            if (m2v.x() >= nCells_.x()) {
1055                m2v.x() = 0;          
458 –              shift.x() = region.x();  
1056              } else if (m2v.x() < 0) {
1057 <              m2v.x() = cells.x() - 1;
461 <              shift.x() = - region.x();
1057 >              m2v.x() = nCells_.x() - 1;
1058              }
1059 <
1060 <            if (m2v.y() >= cells.y()) {
1059 >            
1060 >            if (m2v.y() >= nCells_.y()) {
1061                m2v.y() = 0;          
466 –              shift.y() = region.y();  
1062              } else if (m2v.y() < 0) {
1063 <              m2v.y() = cells.y() - 1;
469 <              shift.y() = - region.y();
1063 >              m2v.y() = nCells_.y() - 1;
1064              }
1065 +            
1066 +            if (m2v.z() >= nCells_.z()) {
1067 +              m2v.z() = 0;          
1068 +            } else if (m2v.z() < 0) {
1069 +              m2v.z() = nCells_.z() - 1;
1070 +            }
1071 +            
1072 +            int m2 = Vlinear (m2v, nCells_);
1073  
1074 <            m2 = VLinear (m2v, cells) + nMol;
1075 <            for (j1 = cellList[m1]; j1 >= 0; j1 = cellList[j1]) {
1076 <              for (j2 = cellList[m2]; j2 >= 0; j2 = cellList[j2]) {
1077 <                if (m1 != m2 || j2 < j1) {
1078 <                  dr = mol[j1].r - mol[j2].r;
1079 <                  VSub (dr, mol[j1].r, mol[j2].r);
1080 <                  VVSub (dr, shift);
1081 <                  if (VLenSq (dr) < rrNebr) {
1082 <                    neighborList.push_back(make_pair(j1, j2));
1074 > #ifdef IS_MPI
1075 >            for (vector<int>::iterator j1 = cellListRow_[m1].begin();
1076 >                 j1 != cellListRow_[m1].end(); ++j1) {
1077 >              for (vector<int>::iterator j2 = cellListCol_[m2].begin();
1078 >                   j2 != cellListCol_[m2].end(); ++j2) {
1079 >                              
1080 >                // Always do this if we're in different cells or if
1081 >                // we're in the same cell and the global index of the
1082 >                // j2 cutoff group is less than the j1 cutoff group
1083 >
1084 >                if (m2 != m1 || cgColToGlobal[(*j2)] < cgRowToGlobal[(*j1)]) {
1085 >                  dr = cgColData.position[(*j2)] - cgRowData.position[(*j1)];
1086 >                  snap_->wrapVector(dr);
1087 >                  cuts = getGroupCutoffs( (*j1), (*j2) );
1088 >                  if (dr.lengthSquare() < cuts.third) {
1089 >                    neighborList.push_back(make_pair((*j1), (*j2)));
1090                    }
1091                  }
1092                }
1093              }
1094 + #else
1095 +
1096 +            for (vector<int>::iterator j1 = cellList_[m1].begin();
1097 +                 j1 != cellList_[m1].end(); ++j1) {
1098 +              for (vector<int>::iterator j2 = cellList_[m2].begin();
1099 +                   j2 != cellList_[m2].end(); ++j2) {
1100 +
1101 +                // Always do this if we're in different cells or if
1102 +                // we're in the same cell and the global index of the
1103 +                // j2 cutoff group is less than the j1 cutoff group
1104 +
1105 +                if (m2 != m1 || (*j2) < (*j1)) {
1106 +                  dr = snap_->cgData.position[(*j2)] - snap_->cgData.position[(*j1)];
1107 +                  snap_->wrapVector(dr);
1108 +                  cuts = getGroupCutoffs( (*j1), (*j2) );
1109 +                  if (dr.lengthSquare() < cuts.third) {
1110 +                    neighborList.push_back(make_pair((*j1), (*j2)));
1111 +                  }
1112 +                }
1113 +              }
1114 +            }
1115 + #endif
1116            }
1117          }
1118        }
1119      }
1120 +    
1121 +    // save the local cutoff group positions for the check that is
1122 +    // done on each loop:
1123 +    saved_CG_positions_.clear();
1124 +    for (int i = 0; i < nGroups_; i++)
1125 +      saved_CG_positions_.push_back(snap_->cgData.position[i]);
1126 +  
1127 +    return neighborList;
1128    }
490 –
491 –  
1129   } //end namespace OpenMD

Diff Legend

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