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

Comparing:
branches/development/src/parallel/ForceDecomposition.cpp (file contents), Revision 1538 by chuckv, Tue Jan 11 18:58:12 2011 UTC vs.
branches/development/src/parallel/ForceMatrixDecomposition.cpp (file contents), Revision 1723 by gezelter, Thu May 24 20:59:54 2012 UTC

# Line 1 | Line 1
1 < /**
2 < * @file ForceDecomposition.cpp
3 < * @author Charles Vardeman <cvardema.at.nd.edu>
4 < * @date 08/18/2010
5 < * @time 11:56am
6 < * @version 1.0
1 > /*
2 > * Copyright (c) 2005 The University of Notre Dame. All Rights Reserved.
3   *
8 * @section LICENSE
9 * Copyright (c) 2010 The University of Notre Dame. All Rights Reserved.
10 *
4   * The University of Notre Dame grants you ("Licensee") a
5   * non-exclusive, royalty free, license to use, modify and
6   * redistribute this software in source and binary code form, provided
# Line 43 | Line 36
36   * [1]  Meineke, et al., J. Comp. Chem. 26, 252-271 (2005).            
37   * [2]  Fennell & Gezelter, J. Chem. Phys. 124, 234104 (2006).          
38   * [3]  Sun, Lin & Gezelter, J. Chem. Phys. 128, 24107 (2008).          
39 < * [4]  Vardeman & Gezelter, in progress (2009).                        
39 > * [4]  Kuang & Gezelter,  J. Chem. Phys. 133, 164101 (2010).
40 > * [5]  Vardeman, Stocker & Gezelter, J. Chem. Theory Comput. 7, 834 (2011).
41   */
42 + #include "parallel/ForceMatrixDecomposition.hpp"
43 + #include "math/SquareMatrix3.hpp"
44 + #include "nonbonded/NonBondedInteraction.hpp"
45 + #include "brains/SnapshotManager.hpp"
46 + #include "brains/PairList.hpp"
47  
48 + using namespace std;
49 + namespace OpenMD {
50  
51 +  ForceMatrixDecomposition::ForceMatrixDecomposition(SimInfo* info, InteractionManager* iMan) : ForceDecomposition(info, iMan) {
52  
53 < /*  -*- c++ -*-  */
54 < #include "config.h"
55 < #include <stdlib.h>
53 >    // In a parallel computation, row and colum scans must visit all
54 >    // surrounding cells (not just the 14 upper triangular blocks that
55 >    // are used when the processor can see all pairs)
56   #ifdef IS_MPI
57 < #include <mpi.h>
58 < #endif
57 >    cellOffsets_.clear();
58 >    cellOffsets_.push_back( Vector3i(-1,-1,-1) );
59 >    cellOffsets_.push_back( Vector3i( 0,-1,-1) );
60 >    cellOffsets_.push_back( Vector3i( 1,-1,-1) );                          
61 >    cellOffsets_.push_back( Vector3i(-1, 0,-1) );
62 >    cellOffsets_.push_back( Vector3i( 0, 0,-1) );
63 >    cellOffsets_.push_back( Vector3i( 1, 0,-1) );
64 >    cellOffsets_.push_back( Vector3i(-1, 1,-1) );
65 >    cellOffsets_.push_back( Vector3i( 0, 1,-1) );      
66 >    cellOffsets_.push_back( Vector3i( 1, 1,-1) );
67 >    cellOffsets_.push_back( Vector3i(-1,-1, 0) );
68 >    cellOffsets_.push_back( Vector3i( 0,-1, 0) );
69 >    cellOffsets_.push_back( Vector3i( 1,-1, 0) );
70 >    cellOffsets_.push_back( Vector3i(-1, 0, 0) );      
71 >    cellOffsets_.push_back( Vector3i( 0, 0, 0) );
72 >    cellOffsets_.push_back( Vector3i( 1, 0, 0) );
73 >    cellOffsets_.push_back( Vector3i(-1, 1, 0) );
74 >    cellOffsets_.push_back( Vector3i( 0, 1, 0) );
75 >    cellOffsets_.push_back( Vector3i( 1, 1, 0) );
76 >    cellOffsets_.push_back( Vector3i(-1,-1, 1) );
77 >    cellOffsets_.push_back( Vector3i( 0,-1, 1) );
78 >    cellOffsets_.push_back( Vector3i( 1,-1, 1) );
79 >    cellOffsets_.push_back( Vector3i(-1, 0, 1) );
80 >    cellOffsets_.push_back( Vector3i( 0, 0, 1) );
81 >    cellOffsets_.push_back( Vector3i( 1, 0, 1) );
82 >    cellOffsets_.push_back( Vector3i(-1, 1, 1) );
83 >    cellOffsets_.push_back( Vector3i( 0, 1, 1) );
84 >    cellOffsets_.push_back( Vector3i( 1, 1, 1) );
85 > #endif    
86 >  }
87  
58 #include <iostream>
59 #include <vector>
60 #include <algorithm>
61 #include <cmath>
62 #include "parallel/ForceDecomposition.hpp"
88  
89 +  /**
90 +   * distributeInitialData is essentially a copy of the older fortran
91 +   * SimulationSetup
92 +   */
93 +  void ForceMatrixDecomposition::distributeInitialData() {
94 +    snap_ = sman_->getCurrentSnapshot();
95 +    storageLayout_ = sman_->getStorageLayout();
96 +    ff_ = info_->getForceField();
97 +    nLocal_ = snap_->getNumberOfAtoms();
98 +  
99 +    nGroups_ = info_->getNLocalCutoffGroups();
100 +    // gather the information for atomtype IDs (atids):
101 +    idents = info_->getIdentArray();
102 +    AtomLocalToGlobal = info_->getGlobalAtomIndices();
103 +    cgLocalToGlobal = info_->getGlobalGroupIndices();
104 +    vector<int> globalGroupMembership = info_->getGlobalGroupMembership();
105  
106 < using namespace std;
66 < using namespace OpenMD;
106 >    massFactors = info_->getMassFactors();
107  
108 < //__static
108 >    PairList* excludes = info_->getExcludedInteractions();
109 >    PairList* oneTwo = info_->getOneTwoInteractions();
110 >    PairList* oneThree = info_->getOneThreeInteractions();
111 >    PairList* oneFour = info_->getOneFourInteractions();
112 >    
113 >    if (needVelocities_)
114 >      snap_->cgData.setStorageLayout(DataStorage::dslPosition |
115 >                                     DataStorage::dslVelocity);
116 >    else
117 >      snap_->cgData.setStorageLayout(DataStorage::dslPosition);
118 >    
119   #ifdef IS_MPI
120 < static vector<MPI:Comm> communictors;
121 < #endif
120 >
121 >    MPI::Intracomm row = rowComm.getComm();
122 >    MPI::Intracomm col = colComm.getComm();
123  
124 < //____ MPITypeTraits
125 < template<typename T>
126 < struct MPITypeTraits;
124 >    AtomPlanIntRow = new Plan<int>(row, nLocal_);
125 >    AtomPlanRealRow = new Plan<RealType>(row, nLocal_);
126 >    AtomPlanVectorRow = new Plan<Vector3d>(row, nLocal_);
127 >    AtomPlanMatrixRow = new Plan<Mat3x3d>(row, nLocal_);
128 >    AtomPlanPotRow = new Plan<potVec>(row, nLocal_);
129  
130 < #ifdef IS_MPI
131 < template<>
132 < struct MPITypeTraits<RealType> {
133 <  static const MPI::Datatype datatype;
134 < };
82 < const MPI_Datatype MPITypeTraits<RealType>::datatype = MY_MPI_REAL;
130 >    AtomPlanIntColumn = new Plan<int>(col, nLocal_);
131 >    AtomPlanRealColumn = new Plan<RealType>(col, nLocal_);
132 >    AtomPlanVectorColumn = new Plan<Vector3d>(col, nLocal_);
133 >    AtomPlanMatrixColumn = new Plan<Mat3x3d>(col, nLocal_);
134 >    AtomPlanPotColumn = new Plan<potVec>(col, nLocal_);
135  
136 < template<>
137 < struct MPITypeTraits<int> {
138 <  static const MPI::Datatype datatype;
139 < };
88 < const MPI::Datatype MPITypeTraits<int>::datatype = MPI_INT;
89 < #endif
136 >    cgPlanIntRow = new Plan<int>(row, nGroups_);
137 >    cgPlanVectorRow = new Plan<Vector3d>(row, nGroups_);
138 >    cgPlanIntColumn = new Plan<int>(col, nGroups_);
139 >    cgPlanVectorColumn = new Plan<Vector3d>(col, nGroups_);
140  
141 < /**
142 < * Constructor for ForceDecomposition Parallel Decomposition Method
143 < * Will try to construct a symmetric grid of processors. Ideally, the
144 < * number of processors will be a square ex: 4, 9, 16, 25.
95 < *
96 < */
141 >    nAtomsInRow_ = AtomPlanIntRow->getSize();
142 >    nAtomsInCol_ = AtomPlanIntColumn->getSize();
143 >    nGroupsInRow_ = cgPlanIntRow->getSize();
144 >    nGroupsInCol_ = cgPlanIntColumn->getSize();
145  
146 < ForceDecomposition::ForceDecomposition() {
146 >    // Modify the data storage objects with the correct layouts and sizes:
147 >    atomRowData.resize(nAtomsInRow_);
148 >    atomRowData.setStorageLayout(storageLayout_);
149 >    atomColData.resize(nAtomsInCol_);
150 >    atomColData.setStorageLayout(storageLayout_);
151 >    cgRowData.resize(nGroupsInRow_);
152 >    cgRowData.setStorageLayout(DataStorage::dslPosition);
153 >    cgColData.resize(nGroupsInCol_);
154 >    if (needVelocities_)
155 >      // we only need column velocities if we need them.
156 >      cgColData.setStorageLayout(DataStorage::dslPosition |
157 >                                 DataStorage::dslVelocity);
158 >    else    
159 >      cgColData.setStorageLayout(DataStorage::dslPosition);
160 >      
161 >    identsRow.resize(nAtomsInRow_);
162 >    identsCol.resize(nAtomsInCol_);
163 >    
164 >    AtomPlanIntRow->gather(idents, identsRow);
165 >    AtomPlanIntColumn->gather(idents, identsCol);
166 >    
167 >    // allocate memory for the parallel objects
168 >    atypesRow.resize(nAtomsInRow_);
169 >    atypesCol.resize(nAtomsInCol_);
170  
171 < #ifdef IS_MPI
172 <  int nProcs = MPI::COMM_WORLD.Get_size();
173 <  int worldRank = MPI::COMM_WORLD.Get_rank();
171 >    for (int i = 0; i < nAtomsInRow_; i++)
172 >      atypesRow[i] = ff_->getAtomType(identsRow[i]);
173 >    for (int i = 0; i < nAtomsInCol_; i++)
174 >      atypesCol[i] = ff_->getAtomType(identsCol[i]);        
175 >
176 >    pot_row.resize(nAtomsInRow_);
177 >    pot_col.resize(nAtomsInCol_);
178 >
179 >    AtomRowToGlobal.resize(nAtomsInRow_);
180 >    AtomColToGlobal.resize(nAtomsInCol_);
181 >    AtomPlanIntRow->gather(AtomLocalToGlobal, AtomRowToGlobal);
182 >    AtomPlanIntColumn->gather(AtomLocalToGlobal, AtomColToGlobal);
183 >
184 >    cgRowToGlobal.resize(nGroupsInRow_);
185 >    cgColToGlobal.resize(nGroupsInCol_);
186 >    cgPlanIntRow->gather(cgLocalToGlobal, cgRowToGlobal);
187 >    cgPlanIntColumn->gather(cgLocalToGlobal, cgColToGlobal);
188 >
189 >    massFactorsRow.resize(nAtomsInRow_);
190 >    massFactorsCol.resize(nAtomsInCol_);
191 >    AtomPlanRealRow->gather(massFactors, massFactorsRow);
192 >    AtomPlanRealColumn->gather(massFactors, massFactorsCol);
193 >
194 >    groupListRow_.clear();
195 >    groupListRow_.resize(nGroupsInRow_);
196 >    for (int i = 0; i < nGroupsInRow_; i++) {
197 >      int gid = cgRowToGlobal[i];
198 >      for (int j = 0; j < nAtomsInRow_; j++) {
199 >        int aid = AtomRowToGlobal[j];
200 >        if (globalGroupMembership[aid] == gid)
201 >          groupListRow_[i].push_back(j);
202 >      }      
203 >    }
204 >
205 >    groupListCol_.clear();
206 >    groupListCol_.resize(nGroupsInCol_);
207 >    for (int i = 0; i < nGroupsInCol_; i++) {
208 >      int gid = cgColToGlobal[i];
209 >      for (int j = 0; j < nAtomsInCol_; j++) {
210 >        int aid = AtomColToGlobal[j];
211 >        if (globalGroupMembership[aid] == gid)
212 >          groupListCol_[i].push_back(j);
213 >      }      
214 >    }
215 >
216 >    excludesForAtom.clear();
217 >    excludesForAtom.resize(nAtomsInRow_);
218 >    toposForAtom.clear();
219 >    toposForAtom.resize(nAtomsInRow_);
220 >    topoDist.clear();
221 >    topoDist.resize(nAtomsInRow_);
222 >    for (int i = 0; i < nAtomsInRow_; i++) {
223 >      int iglob = AtomRowToGlobal[i];
224 >
225 >      for (int j = 0; j < nAtomsInCol_; j++) {
226 >        int jglob = AtomColToGlobal[j];
227 >
228 >        if (excludes->hasPair(iglob, jglob))
229 >          excludesForAtom[i].push_back(j);      
230 >        
231 >        if (oneTwo->hasPair(iglob, jglob)) {
232 >          toposForAtom[i].push_back(j);
233 >          topoDist[i].push_back(1);
234 >        } else {
235 >          if (oneThree->hasPair(iglob, jglob)) {
236 >            toposForAtom[i].push_back(j);
237 >            topoDist[i].push_back(2);
238 >          } else {
239 >            if (oneFour->hasPair(iglob, jglob)) {
240 >              toposForAtom[i].push_back(j);
241 >              topoDist[i].push_back(3);
242 >            }
243 >          }
244 >        }
245 >      }      
246 >    }
247 >
248 > #else
249 >    excludesForAtom.clear();
250 >    excludesForAtom.resize(nLocal_);
251 >    toposForAtom.clear();
252 >    toposForAtom.resize(nLocal_);
253 >    topoDist.clear();
254 >    topoDist.resize(nLocal_);
255 >
256 >    for (int i = 0; i < nLocal_; i++) {
257 >      int iglob = AtomLocalToGlobal[i];
258 >
259 >      for (int j = 0; j < nLocal_; j++) {
260 >        int jglob = AtomLocalToGlobal[j];
261 >
262 >        if (excludes->hasPair(iglob, jglob))
263 >          excludesForAtom[i].push_back(j);              
264 >        
265 >        if (oneTwo->hasPair(iglob, jglob)) {
266 >          toposForAtom[i].push_back(j);
267 >          topoDist[i].push_back(1);
268 >        } else {
269 >          if (oneThree->hasPair(iglob, jglob)) {
270 >            toposForAtom[i].push_back(j);
271 >            topoDist[i].push_back(2);
272 >          } else {
273 >            if (oneFour->hasPair(iglob, jglob)) {
274 >              toposForAtom[i].push_back(j);
275 >              topoDist[i].push_back(3);
276 >            }
277 >          }
278 >        }
279 >      }      
280 >    }
281   #endif
282  
283 <  // First time through, construct column stride.
284 <  if (communicators.size() == 0)
285 <  {
286 <    int nColumnsMax = (int) round(sqrt((float) nProcs));
287 <    for (int i = 0; i < nProcs; ++i)
288 <    {
289 <      if (nProcs%i==0) nColumns=i;
283 >    // allocate memory for the parallel objects
284 >    atypesLocal.resize(nLocal_);
285 >
286 >    for (int i = 0; i < nLocal_; i++)
287 >      atypesLocal[i] = ff_->getAtomType(idents[i]);
288 >
289 >    groupList_.clear();
290 >    groupList_.resize(nGroups_);
291 >    for (int i = 0; i < nGroups_; i++) {
292 >      int gid = cgLocalToGlobal[i];
293 >      for (int j = 0; j < nLocal_; j++) {
294 >        int aid = AtomLocalToGlobal[j];
295 >        if (globalGroupMembership[aid] == gid) {
296 >          groupList_[i].push_back(j);
297 >        }
298 >      }      
299      }
300  
301 <    int nRows = nProcs/nColumns;    
302 <    myRank_ = (int) worldRank%nColumns;
301 >
302 >    createGtypeCutoffMap();
303 >
304    }
305 <  else
306 <  {
307 <    myRank_ = myRank/nColumns;
308 <  }
309 <  MPI::Comm newComm = MPI:COMM_WORLD.Split(myRank_,0);
310 <  
311 <  isColumn_ = false;
312 <  
313 < }
305 >  
306 >  void ForceMatrixDecomposition::createGtypeCutoffMap() {
307 >    
308 >    RealType tol = 1e-6;
309 >    largestRcut_ = 0.0;
310 >    RealType rc;
311 >    int atid;
312 >    set<AtomType*> atypes = info_->getSimulatedAtomTypes();
313 >    
314 >    map<int, RealType> atypeCutoff;
315 >      
316 >    for (set<AtomType*>::iterator at = atypes.begin();
317 >         at != atypes.end(); ++at){
318 >      atid = (*at)->getIdent();
319 >      if (userChoseCutoff_)
320 >        atypeCutoff[atid] = userCutoff_;
321 >      else
322 >        atypeCutoff[atid] = interactionMan_->getSuggestedCutoffRadius(*at);
323 >    }
324 >    
325 >    vector<RealType> gTypeCutoffs;
326 >    // first we do a single loop over the cutoff groups to find the
327 >    // largest cutoff for any atypes present in this group.
328 > #ifdef IS_MPI
329 >    vector<RealType> groupCutoffRow(nGroupsInRow_, 0.0);
330 >    groupRowToGtype.resize(nGroupsInRow_);
331 >    for (int cg1 = 0; cg1 < nGroupsInRow_; cg1++) {
332 >      vector<int> atomListRow = getAtomsInGroupRow(cg1);
333 >      for (vector<int>::iterator ia = atomListRow.begin();
334 >           ia != atomListRow.end(); ++ia) {            
335 >        int atom1 = (*ia);
336 >        atid = identsRow[atom1];
337 >        if (atypeCutoff[atid] > groupCutoffRow[cg1]) {
338 >          groupCutoffRow[cg1] = atypeCutoff[atid];
339 >        }
340 >      }
341  
342 < ForceDecomposition::gather(sendbuf, receivebuf){
343 <  communicators(myIndex_).Allgatherv();
344 < }
342 >      bool gTypeFound = false;
343 >      for (int gt = 0; gt < gTypeCutoffs.size(); gt++) {
344 >        if (abs(groupCutoffRow[cg1] - gTypeCutoffs[gt]) < tol) {
345 >          groupRowToGtype[cg1] = gt;
346 >          gTypeFound = true;
347 >        }
348 >      }
349 >      if (!gTypeFound) {
350 >        gTypeCutoffs.push_back( groupCutoffRow[cg1] );
351 >        groupRowToGtype[cg1] = gTypeCutoffs.size() - 1;
352 >      }
353 >      
354 >    }
355 >    vector<RealType> groupCutoffCol(nGroupsInCol_, 0.0);
356 >    groupColToGtype.resize(nGroupsInCol_);
357 >    for (int cg2 = 0; cg2 < nGroupsInCol_; cg2++) {
358 >      vector<int> atomListCol = getAtomsInGroupColumn(cg2);
359 >      for (vector<int>::iterator jb = atomListCol.begin();
360 >           jb != atomListCol.end(); ++jb) {            
361 >        int atom2 = (*jb);
362 >        atid = identsCol[atom2];
363 >        if (atypeCutoff[atid] > groupCutoffCol[cg2]) {
364 >          groupCutoffCol[cg2] = atypeCutoff[atid];
365 >        }
366 >      }
367 >      bool gTypeFound = false;
368 >      for (int gt = 0; gt < gTypeCutoffs.size(); gt++) {
369 >        if (abs(groupCutoffCol[cg2] - gTypeCutoffs[gt]) < tol) {
370 >          groupColToGtype[cg2] = gt;
371 >          gTypeFound = true;
372 >        }
373 >      }
374 >      if (!gTypeFound) {
375 >        gTypeCutoffs.push_back( groupCutoffCol[cg2] );
376 >        groupColToGtype[cg2] = gTypeCutoffs.size() - 1;
377 >      }
378 >    }
379 > #else
380  
381 +    vector<RealType> groupCutoff(nGroups_, 0.0);
382 +    groupToGtype.resize(nGroups_);
383 +    for (int cg1 = 0; cg1 < nGroups_; cg1++) {
384 +      groupCutoff[cg1] = 0.0;
385 +      vector<int> atomList = getAtomsInGroupRow(cg1);
386 +      for (vector<int>::iterator ia = atomList.begin();
387 +           ia != atomList.end(); ++ia) {            
388 +        int atom1 = (*ia);
389 +        atid = idents[atom1];
390 +        if (atypeCutoff[atid] > groupCutoff[cg1])
391 +          groupCutoff[cg1] = atypeCutoff[atid];
392 +      }
393 +      
394 +      bool gTypeFound = false;
395 +      for (int gt = 0; gt < gTypeCutoffs.size(); gt++) {
396 +        if (abs(groupCutoff[cg1] - gTypeCutoffs[gt]) < tol) {
397 +          groupToGtype[cg1] = gt;
398 +          gTypeFound = true;
399 +        }
400 +      }
401 +      if (!gTypeFound) {      
402 +        gTypeCutoffs.push_back( groupCutoff[cg1] );
403 +        groupToGtype[cg1] = gTypeCutoffs.size() - 1;
404 +      }      
405 +    }
406 + #endif
407  
408 +    // Now we find the maximum group cutoff value present in the simulation
409  
410 < ForceDecomposition::scatter(sbuffer, rbuffer){
411 <  communicators(myIndex_).Reduce_scatter(sbuffer, recevbuf. recvcounts, MPI::DOUBLE, MPI::SUM);
412 < }
410 >    RealType groupMax = *max_element(gTypeCutoffs.begin(),
411 >                                     gTypeCutoffs.end());
412 >
413 > #ifdef IS_MPI
414 >    MPI::COMM_WORLD.Allreduce(&groupMax, &groupMax, 1, MPI::REALTYPE,
415 >                              MPI::MAX);
416 > #endif
417 >    
418 >    RealType tradRcut = groupMax;
419 >
420 >    for (int i = 0; i < gTypeCutoffs.size();  i++) {
421 >      for (int j = 0; j < gTypeCutoffs.size();  j++) {      
422 >        RealType thisRcut;
423 >        switch(cutoffPolicy_) {
424 >        case TRADITIONAL:
425 >          thisRcut = tradRcut;
426 >          break;
427 >        case MIX:
428 >          thisRcut = 0.5 * (gTypeCutoffs[i] + gTypeCutoffs[j]);
429 >          break;
430 >        case MAX:
431 >          thisRcut = max(gTypeCutoffs[i], gTypeCutoffs[j]);
432 >          break;
433 >        default:
434 >          sprintf(painCave.errMsg,
435 >                  "ForceMatrixDecomposition::createGtypeCutoffMap "
436 >                  "hit an unknown cutoff policy!\n");
437 >          painCave.severity = OPENMD_ERROR;
438 >          painCave.isFatal = 1;
439 >          simError();
440 >          break;
441 >        }
442 >
443 >        pair<int,int> key = make_pair(i,j);
444 >        gTypeCutoffMap[key].first = thisRcut;
445 >        if (thisRcut > largestRcut_) largestRcut_ = thisRcut;
446 >        gTypeCutoffMap[key].second = thisRcut*thisRcut;
447 >        gTypeCutoffMap[key].third = pow(thisRcut + skinThickness_, 2);
448 >        // sanity check
449 >        
450 >        if (userChoseCutoff_) {
451 >          if (abs(gTypeCutoffMap[key].first - userCutoff_) > 0.0001) {
452 >            sprintf(painCave.errMsg,
453 >                    "ForceMatrixDecomposition::createGtypeCutoffMap "
454 >                    "user-specified rCut (%lf) does not match computed group Cutoff\n", userCutoff_);
455 >            painCave.severity = OPENMD_ERROR;
456 >            painCave.isFatal = 1;
457 >            simError();            
458 >          }
459 >        }
460 >      }
461 >    }
462 >  }
463 >
464 >
465 >  groupCutoffs ForceMatrixDecomposition::getGroupCutoffs(int cg1, int cg2) {
466 >    int i, j;  
467 > #ifdef IS_MPI
468 >    i = groupRowToGtype[cg1];
469 >    j = groupColToGtype[cg2];
470 > #else
471 >    i = groupToGtype[cg1];
472 >    j = groupToGtype[cg2];
473 > #endif    
474 >    return gTypeCutoffMap[make_pair(i,j)];
475 >  }
476 >
477 >  int ForceMatrixDecomposition::getTopologicalDistance(int atom1, int atom2) {
478 >    for (int j = 0; j < toposForAtom[atom1].size(); j++) {
479 >      if (toposForAtom[atom1][j] == atom2)
480 >        return topoDist[atom1][j];
481 >    }
482 >    return 0;
483 >  }
484 >
485 >  void ForceMatrixDecomposition::zeroWorkArrays() {
486 >    pairwisePot = 0.0;
487 >    embeddingPot = 0.0;
488 >
489 > #ifdef IS_MPI
490 >    if (storageLayout_ & DataStorage::dslForce) {
491 >      fill(atomRowData.force.begin(), atomRowData.force.end(), V3Zero);
492 >      fill(atomColData.force.begin(), atomColData.force.end(), V3Zero);
493 >    }
494 >
495 >    if (storageLayout_ & DataStorage::dslTorque) {
496 >      fill(atomRowData.torque.begin(), atomRowData.torque.end(), V3Zero);
497 >      fill(atomColData.torque.begin(), atomColData.torque.end(), V3Zero);
498 >    }
499 >    
500 >    fill(pot_row.begin(), pot_row.end(),
501 >         Vector<RealType, N_INTERACTION_FAMILIES> (0.0));
502 >
503 >    fill(pot_col.begin(), pot_col.end(),
504 >         Vector<RealType, N_INTERACTION_FAMILIES> (0.0));  
505 >
506 >    if (storageLayout_ & DataStorage::dslParticlePot) {    
507 >      fill(atomRowData.particlePot.begin(), atomRowData.particlePot.end(),
508 >           0.0);
509 >      fill(atomColData.particlePot.begin(), atomColData.particlePot.end(),
510 >           0.0);
511 >    }
512 >
513 >    if (storageLayout_ & DataStorage::dslDensity) {      
514 >      fill(atomRowData.density.begin(), atomRowData.density.end(), 0.0);
515 >      fill(atomColData.density.begin(), atomColData.density.end(), 0.0);
516 >    }
517 >
518 >    if (storageLayout_ & DataStorage::dslFunctional) {  
519 >      fill(atomRowData.functional.begin(), atomRowData.functional.end(),
520 >           0.0);
521 >      fill(atomColData.functional.begin(), atomColData.functional.end(),
522 >           0.0);
523 >    }
524 >
525 >    if (storageLayout_ & DataStorage::dslFunctionalDerivative) {      
526 >      fill(atomRowData.functionalDerivative.begin(),
527 >           atomRowData.functionalDerivative.end(), 0.0);
528 >      fill(atomColData.functionalDerivative.begin(),
529 >           atomColData.functionalDerivative.end(), 0.0);
530 >    }
531 >
532 >    if (storageLayout_ & DataStorage::dslSkippedCharge) {      
533 >      fill(atomRowData.skippedCharge.begin(),
534 >           atomRowData.skippedCharge.end(), 0.0);
535 >      fill(atomColData.skippedCharge.begin(),
536 >           atomColData.skippedCharge.end(), 0.0);
537 >    }
538 >
539 >    if (storageLayout_ & DataStorage::dslFlucQForce) {      
540 >      fill(atomRowData.flucQFrc.begin(),
541 >           atomRowData.flucQFrc.end(), 0.0);
542 >      fill(atomColData.flucQFrc.begin(),
543 >           atomColData.flucQFrc.end(), 0.0);
544 >    }
545 >
546 >    if (storageLayout_ & DataStorage::dslElectricField) {    
547 >      fill(atomRowData.electricField.begin(),
548 >           atomRowData.electricField.end(), V3Zero);
549 >      fill(atomColData.electricField.begin(),
550 >           atomColData.electricField.end(), V3Zero);
551 >    }
552 >
553 >    if (storageLayout_ & DataStorage::dslFlucQForce) {    
554 >      fill(atomRowData.flucQFrc.begin(), atomRowData.flucQFrc.end(),
555 >           0.0);
556 >      fill(atomColData.flucQFrc.begin(), atomColData.flucQFrc.end(),
557 >           0.0);
558 >    }
559 >
560 > #endif
561 >    // even in parallel, we need to zero out the local arrays:
562 >
563 >    if (storageLayout_ & DataStorage::dslParticlePot) {      
564 >      fill(snap_->atomData.particlePot.begin(),
565 >           snap_->atomData.particlePot.end(), 0.0);
566 >    }
567 >    
568 >    if (storageLayout_ & DataStorage::dslDensity) {      
569 >      fill(snap_->atomData.density.begin(),
570 >           snap_->atomData.density.end(), 0.0);
571 >    }
572 >
573 >    if (storageLayout_ & DataStorage::dslFunctional) {
574 >      fill(snap_->atomData.functional.begin(),
575 >           snap_->atomData.functional.end(), 0.0);
576 >    }
577 >
578 >    if (storageLayout_ & DataStorage::dslFunctionalDerivative) {      
579 >      fill(snap_->atomData.functionalDerivative.begin(),
580 >           snap_->atomData.functionalDerivative.end(), 0.0);
581 >    }
582 >
583 >    if (storageLayout_ & DataStorage::dslSkippedCharge) {      
584 >      fill(snap_->atomData.skippedCharge.begin(),
585 >           snap_->atomData.skippedCharge.end(), 0.0);
586 >    }
587 >
588 >    if (storageLayout_ & DataStorage::dslElectricField) {      
589 >      fill(snap_->atomData.electricField.begin(),
590 >           snap_->atomData.electricField.end(), V3Zero);
591 >    }
592 >  }
593 >
594 >
595 >  void ForceMatrixDecomposition::distributeData()  {
596 >    snap_ = sman_->getCurrentSnapshot();
597 >    storageLayout_ = sman_->getStorageLayout();
598 > #ifdef IS_MPI
599 >    
600 >    // gather up the atomic positions
601 >    AtomPlanVectorRow->gather(snap_->atomData.position,
602 >                              atomRowData.position);
603 >    AtomPlanVectorColumn->gather(snap_->atomData.position,
604 >                                 atomColData.position);
605 >    
606 >    // gather up the cutoff group positions
607 >
608 >    cgPlanVectorRow->gather(snap_->cgData.position,
609 >                            cgRowData.position);
610 >
611 >    cgPlanVectorColumn->gather(snap_->cgData.position,
612 >                               cgColData.position);
613 >
614 >
615 >
616 >    if (needVelocities_) {
617 >      // gather up the atomic velocities
618 >      AtomPlanVectorColumn->gather(snap_->atomData.velocity,
619 >                                   atomColData.velocity);
620 >      
621 >      cgPlanVectorColumn->gather(snap_->cgData.velocity,
622 >                                 cgColData.velocity);
623 >    }
624 >
625 >    
626 >    // if needed, gather the atomic rotation matrices
627 >    if (storageLayout_ & DataStorage::dslAmat) {
628 >      AtomPlanMatrixRow->gather(snap_->atomData.aMat,
629 >                                atomRowData.aMat);
630 >      AtomPlanMatrixColumn->gather(snap_->atomData.aMat,
631 >                                   atomColData.aMat);
632 >    }
633 >    
634 >    // if needed, gather the atomic eletrostatic frames
635 >    if (storageLayout_ & DataStorage::dslElectroFrame) {
636 >      AtomPlanMatrixRow->gather(snap_->atomData.electroFrame,
637 >                                atomRowData.electroFrame);
638 >      AtomPlanMatrixColumn->gather(snap_->atomData.electroFrame,
639 >                                   atomColData.electroFrame);
640 >    }
641 >
642 >    // if needed, gather the atomic fluctuating charge values
643 >    if (storageLayout_ & DataStorage::dslFlucQPosition) {
644 >      AtomPlanRealRow->gather(snap_->atomData.flucQPos,
645 >                              atomRowData.flucQPos);
646 >      AtomPlanRealColumn->gather(snap_->atomData.flucQPos,
647 >                                 atomColData.flucQPos);
648 >    }
649 >
650 > #endif      
651 >  }
652 >  
653 >  /* collects information obtained during the pre-pair loop onto local
654 >   * data structures.
655 >   */
656 >  void ForceMatrixDecomposition::collectIntermediateData() {
657 >    snap_ = sman_->getCurrentSnapshot();
658 >    storageLayout_ = sman_->getStorageLayout();
659 > #ifdef IS_MPI
660 >    
661 >    if (storageLayout_ & DataStorage::dslDensity) {
662 >      
663 >      AtomPlanRealRow->scatter(atomRowData.density,
664 >                               snap_->atomData.density);
665 >      
666 >      int n = snap_->atomData.density.size();
667 >      vector<RealType> rho_tmp(n, 0.0);
668 >      AtomPlanRealColumn->scatter(atomColData.density, rho_tmp);
669 >      for (int i = 0; i < n; i++)
670 >        snap_->atomData.density[i] += rho_tmp[i];
671 >    }
672 >
673 >    if (storageLayout_ & DataStorage::dslElectricField) {
674 >      
675 >      AtomPlanVectorRow->scatter(atomRowData.electricField,
676 >                                 snap_->atomData.electricField);
677 >      
678 >      int n = snap_->atomData.electricField.size();
679 >      vector<Vector3d> field_tmp(n, V3Zero);
680 >      AtomPlanVectorColumn->scatter(atomColData.electricField, field_tmp);
681 >      for (int i = 0; i < n; i++)
682 >        snap_->atomData.electricField[i] += field_tmp[i];
683 >    }
684 > #endif
685 >  }
686 >
687 >  /*
688 >   * redistributes information obtained during the pre-pair loop out to
689 >   * row and column-indexed data structures
690 >   */
691 >  void ForceMatrixDecomposition::distributeIntermediateData() {
692 >    snap_ = sman_->getCurrentSnapshot();
693 >    storageLayout_ = sman_->getStorageLayout();
694 > #ifdef IS_MPI
695 >    if (storageLayout_ & DataStorage::dslFunctional) {
696 >      AtomPlanRealRow->gather(snap_->atomData.functional,
697 >                              atomRowData.functional);
698 >      AtomPlanRealColumn->gather(snap_->atomData.functional,
699 >                                 atomColData.functional);
700 >    }
701 >    
702 >    if (storageLayout_ & DataStorage::dslFunctionalDerivative) {
703 >      AtomPlanRealRow->gather(snap_->atomData.functionalDerivative,
704 >                              atomRowData.functionalDerivative);
705 >      AtomPlanRealColumn->gather(snap_->atomData.functionalDerivative,
706 >                                 atomColData.functionalDerivative);
707 >    }
708 > #endif
709 >  }
710 >  
711 >  
712 >  void ForceMatrixDecomposition::collectData() {
713 >    snap_ = sman_->getCurrentSnapshot();
714 >    storageLayout_ = sman_->getStorageLayout();
715 > #ifdef IS_MPI    
716 >    int n = snap_->atomData.force.size();
717 >    vector<Vector3d> frc_tmp(n, V3Zero);
718 >    
719 >    AtomPlanVectorRow->scatter(atomRowData.force, frc_tmp);
720 >    for (int i = 0; i < n; i++) {
721 >      snap_->atomData.force[i] += frc_tmp[i];
722 >      frc_tmp[i] = 0.0;
723 >    }
724 >    
725 >    AtomPlanVectorColumn->scatter(atomColData.force, frc_tmp);
726 >    for (int i = 0; i < n; i++) {
727 >      snap_->atomData.force[i] += frc_tmp[i];
728 >    }
729 >        
730 >    if (storageLayout_ & DataStorage::dslTorque) {
731 >
732 >      int nt = snap_->atomData.torque.size();
733 >      vector<Vector3d> trq_tmp(nt, V3Zero);
734 >
735 >      AtomPlanVectorRow->scatter(atomRowData.torque, trq_tmp);
736 >      for (int i = 0; i < nt; i++) {
737 >        snap_->atomData.torque[i] += trq_tmp[i];
738 >        trq_tmp[i] = 0.0;
739 >      }
740 >      
741 >      AtomPlanVectorColumn->scatter(atomColData.torque, trq_tmp);
742 >      for (int i = 0; i < nt; i++)
743 >        snap_->atomData.torque[i] += trq_tmp[i];
744 >    }
745  
746 +    if (storageLayout_ & DataStorage::dslSkippedCharge) {
747  
748 +      int ns = snap_->atomData.skippedCharge.size();
749 +      vector<RealType> skch_tmp(ns, 0.0);
750 +
751 +      AtomPlanRealRow->scatter(atomRowData.skippedCharge, skch_tmp);
752 +      for (int i = 0; i < ns; i++) {
753 +        snap_->atomData.skippedCharge[i] += skch_tmp[i];
754 +        skch_tmp[i] = 0.0;
755 +      }
756 +      
757 +      AtomPlanRealColumn->scatter(atomColData.skippedCharge, skch_tmp);
758 +      for (int i = 0; i < ns; i++)
759 +        snap_->atomData.skippedCharge[i] += skch_tmp[i];
760 +            
761 +    }
762 +    
763 +    if (storageLayout_ & DataStorage::dslFlucQForce) {
764 +
765 +      int nq = snap_->atomData.flucQFrc.size();
766 +      vector<RealType> fqfrc_tmp(nq, 0.0);
767 +
768 +      AtomPlanRealRow->scatter(atomRowData.flucQFrc, fqfrc_tmp);
769 +      for (int i = 0; i < nq; i++) {
770 +        snap_->atomData.flucQFrc[i] += fqfrc_tmp[i];
771 +        fqfrc_tmp[i] = 0.0;
772 +      }
773 +      
774 +      AtomPlanRealColumn->scatter(atomColData.flucQFrc, fqfrc_tmp);
775 +      for (int i = 0; i < nq; i++)
776 +        snap_->atomData.flucQFrc[i] += fqfrc_tmp[i];
777 +            
778 +    }
779 +
780 +    nLocal_ = snap_->getNumberOfAtoms();
781 +
782 +    vector<potVec> pot_temp(nLocal_,
783 +                            Vector<RealType, N_INTERACTION_FAMILIES> (0.0));
784 +
785 +    // scatter/gather pot_row into the members of my column
786 +          
787 +    AtomPlanPotRow->scatter(pot_row, pot_temp);
788 +
789 +    for (int ii = 0;  ii < pot_temp.size(); ii++ )
790 +      pairwisePot += pot_temp[ii];
791 +        
792 +    if (storageLayout_ & DataStorage::dslParticlePot) {
793 +      // This is the pairwise contribution to the particle pot.  The
794 +      // embedding contribution is added in each of the low level
795 +      // non-bonded routines.  In single processor, this is done in
796 +      // unpackInteractionData, not in collectData.
797 +      for (int ii = 0; ii < N_INTERACTION_FAMILIES; ii++) {
798 +        for (int i = 0; i < nLocal_; i++) {
799 +          // factor of two is because the total potential terms are divided
800 +          // by 2 in parallel due to row/ column scatter      
801 +          snap_->atomData.particlePot[i] += 2.0 * pot_temp[i](ii);
802 +        }
803 +      }
804 +    }
805 +
806 +    fill(pot_temp.begin(), pot_temp.end(),
807 +         Vector<RealType, N_INTERACTION_FAMILIES> (0.0));
808 +      
809 +    AtomPlanPotColumn->scatter(pot_col, pot_temp);    
810 +    
811 +    for (int ii = 0;  ii < pot_temp.size(); ii++ )
812 +      pairwisePot += pot_temp[ii];    
813 +
814 +    if (storageLayout_ & DataStorage::dslParticlePot) {
815 +      // This is the pairwise contribution to the particle pot.  The
816 +      // embedding contribution is added in each of the low level
817 +      // non-bonded routines.  In single processor, this is done in
818 +      // unpackInteractionData, not in collectData.
819 +      for (int ii = 0; ii < N_INTERACTION_FAMILIES; ii++) {
820 +        for (int i = 0; i < nLocal_; i++) {
821 +          // factor of two is because the total potential terms are divided
822 +          // by 2 in parallel due to row/ column scatter      
823 +          snap_->atomData.particlePot[i] += 2.0 * pot_temp[i](ii);
824 +        }
825 +      }
826 +    }
827 +    
828 +    if (storageLayout_ & DataStorage::dslParticlePot) {
829 +      int npp = snap_->atomData.particlePot.size();
830 +      vector<RealType> ppot_temp(npp, 0.0);
831 +
832 +      // This is the direct or embedding contribution to the particle
833 +      // pot.
834 +      
835 +      AtomPlanRealRow->scatter(atomRowData.particlePot, ppot_temp);
836 +      for (int i = 0; i < npp; i++) {
837 +        snap_->atomData.particlePot[i] += ppot_temp[i];
838 +      }
839 +
840 +      fill(ppot_temp.begin(), ppot_temp.end(), 0.0);
841 +      
842 +      AtomPlanRealColumn->scatter(atomColData.particlePot, ppot_temp);
843 +      for (int i = 0; i < npp; i++) {
844 +        snap_->atomData.particlePot[i] += ppot_temp[i];
845 +      }
846 +    }
847 +
848 +    for (int ii = 0; ii < N_INTERACTION_FAMILIES; ii++) {
849 +      RealType ploc1 = pairwisePot[ii];
850 +      RealType ploc2 = 0.0;
851 +      MPI::COMM_WORLD.Allreduce(&ploc1, &ploc2, 1, MPI::REALTYPE, MPI::SUM);
852 +      pairwisePot[ii] = ploc2;
853 +    }
854 +
855 +    for (int ii = 0; ii < N_INTERACTION_FAMILIES; ii++) {
856 +      RealType ploc1 = embeddingPot[ii];
857 +      RealType ploc2 = 0.0;
858 +      MPI::COMM_WORLD.Allreduce(&ploc1, &ploc2, 1, MPI::REALTYPE, MPI::SUM);
859 +      embeddingPot[ii] = ploc2;
860 +    }
861 +    
862 +    // Here be dragons.
863 +    MPI::Intracomm col = colComm.getComm();
864 +
865 +    col.Allreduce(MPI::IN_PLACE,
866 +                  &snap_->frameData.conductiveHeatFlux[0], 3,
867 +                  MPI::REALTYPE, MPI::SUM);
868 +
869 +
870 + #endif
871 +
872 +  }
873 +
874 +  int ForceMatrixDecomposition::getNAtomsInRow() {  
875 + #ifdef IS_MPI
876 +    return nAtomsInRow_;
877 + #else
878 +    return nLocal_;
879 + #endif
880 +  }
881 +
882 +  /**
883 +   * returns the list of atoms belonging to this group.  
884 +   */
885 +  vector<int> ForceMatrixDecomposition::getAtomsInGroupRow(int cg1){
886 + #ifdef IS_MPI
887 +    return groupListRow_[cg1];
888 + #else
889 +    return groupList_[cg1];
890 + #endif
891 +  }
892 +
893 +  vector<int> ForceMatrixDecomposition::getAtomsInGroupColumn(int cg2){
894 + #ifdef IS_MPI
895 +    return groupListCol_[cg2];
896 + #else
897 +    return groupList_[cg2];
898 + #endif
899 +  }
900 +  
901 +  Vector3d ForceMatrixDecomposition::getIntergroupVector(int cg1, int cg2){
902 +    Vector3d d;
903 +    
904 + #ifdef IS_MPI
905 +    d = cgColData.position[cg2] - cgRowData.position[cg1];
906 + #else
907 +    d = snap_->cgData.position[cg2] - snap_->cgData.position[cg1];
908 + #endif
909 +    
910 +    snap_->wrapVector(d);
911 +    return d;    
912 +  }
913 +
914 +  Vector3d ForceMatrixDecomposition::getGroupVelocityColumn(int cg2){
915 + #ifdef IS_MPI
916 +    return cgColData.velocity[cg2];
917 + #else
918 +    return snap_->cgData.velocity[cg2];
919 + #endif
920 +  }
921 +
922 +  Vector3d ForceMatrixDecomposition::getAtomVelocityColumn(int atom2){
923 + #ifdef IS_MPI
924 +    return atomColData.velocity[atom2];
925 + #else
926 +    return snap_->atomData.velocity[atom2];
927 + #endif
928 +  }
929 +
930 +
931 +  Vector3d ForceMatrixDecomposition::getAtomToGroupVectorRow(int atom1, int cg1){
932 +
933 +    Vector3d d;
934 +    
935 + #ifdef IS_MPI
936 +    d = cgRowData.position[cg1] - atomRowData.position[atom1];
937 + #else
938 +    d = snap_->cgData.position[cg1] - snap_->atomData.position[atom1];
939 + #endif
940 +
941 +    snap_->wrapVector(d);
942 +    return d;    
943 +  }
944 +  
945 +  Vector3d ForceMatrixDecomposition::getAtomToGroupVectorColumn(int atom2, int cg2){
946 +    Vector3d d;
947 +    
948 + #ifdef IS_MPI
949 +    d = cgColData.position[cg2] - atomColData.position[atom2];
950 + #else
951 +    d = snap_->cgData.position[cg2] - snap_->atomData.position[atom2];
952 + #endif
953 +    
954 +    snap_->wrapVector(d);
955 +    return d;    
956 +  }
957 +
958 +  RealType ForceMatrixDecomposition::getMassFactorRow(int atom1) {
959 + #ifdef IS_MPI
960 +    return massFactorsRow[atom1];
961 + #else
962 +    return massFactors[atom1];
963 + #endif
964 +  }
965 +
966 +  RealType ForceMatrixDecomposition::getMassFactorColumn(int atom2) {
967 + #ifdef IS_MPI
968 +    return massFactorsCol[atom2];
969 + #else
970 +    return massFactors[atom2];
971 + #endif
972 +
973 +  }
974 +    
975 +  Vector3d ForceMatrixDecomposition::getInteratomicVector(int atom1, int atom2){
976 +    Vector3d d;
977 +    
978 + #ifdef IS_MPI
979 +    d = atomColData.position[atom2] - atomRowData.position[atom1];
980 + #else
981 +    d = snap_->atomData.position[atom2] - snap_->atomData.position[atom1];
982 + #endif
983 +
984 +    snap_->wrapVector(d);
985 +    return d;    
986 +  }
987 +
988 +  vector<int> ForceMatrixDecomposition::getExcludesForAtom(int atom1) {
989 +    return excludesForAtom[atom1];
990 +  }
991 +
992 +  /**
993 +   * We need to exclude some overcounted interactions that result from
994 +   * the parallel decomposition.
995 +   */
996 +  bool ForceMatrixDecomposition::skipAtomPair(int atom1, int atom2) {
997 +    int unique_id_1, unique_id_2;
998 +        
999 + #ifdef IS_MPI
1000 +    // in MPI, we have to look up the unique IDs for each atom
1001 +    unique_id_1 = AtomRowToGlobal[atom1];
1002 +    unique_id_2 = AtomColToGlobal[atom2];
1003 + #else
1004 +    unique_id_1 = AtomLocalToGlobal[atom1];
1005 +    unique_id_2 = AtomLocalToGlobal[atom2];
1006 + #endif  
1007 +
1008 +    if (unique_id_1 == unique_id_2) return true;
1009 +
1010 + #ifdef IS_MPI
1011 +    // this prevents us from doing the pair on multiple processors
1012 +    if (unique_id_1 < unique_id_2) {
1013 +      if ((unique_id_1 + unique_id_2) % 2 == 0) return true;
1014 +    } else {
1015 +      if ((unique_id_1 + unique_id_2) % 2 == 1) return true;
1016 +    }
1017 + #endif
1018 +    
1019 +    return false;
1020 +  }
1021 +
1022 +  /**
1023 +   * We need to handle the interactions for atoms who are involved in
1024 +   * the same rigid body as well as some short range interactions
1025 +   * (bonds, bends, torsions) differently from other interactions.
1026 +   * We'll still visit the pairwise routines, but with a flag that
1027 +   * tells those routines to exclude the pair from direct long range
1028 +   * interactions.  Some indirect interactions (notably reaction
1029 +   * field) must still be handled for these pairs.
1030 +   */
1031 +  bool ForceMatrixDecomposition::excludeAtomPair(int atom1, int atom2) {
1032 +
1033 +    // excludesForAtom was constructed to use row/column indices in the MPI
1034 +    // version, and to use local IDs in the non-MPI version:
1035 +    
1036 +    for (vector<int>::iterator i = excludesForAtom[atom1].begin();
1037 +         i != excludesForAtom[atom1].end(); ++i) {
1038 +      if ( (*i) == atom2 ) return true;
1039 +    }
1040 +
1041 +    return false;
1042 +  }
1043 +
1044 +
1045 +  void ForceMatrixDecomposition::addForceToAtomRow(int atom1, Vector3d fg){
1046 + #ifdef IS_MPI
1047 +    atomRowData.force[atom1] += fg;
1048 + #else
1049 +    snap_->atomData.force[atom1] += fg;
1050 + #endif
1051 +  }
1052 +
1053 +  void ForceMatrixDecomposition::addForceToAtomColumn(int atom2, Vector3d fg){
1054 + #ifdef IS_MPI
1055 +    atomColData.force[atom2] += fg;
1056 + #else
1057 +    snap_->atomData.force[atom2] += fg;
1058 + #endif
1059 +  }
1060 +
1061 +    // filling interaction blocks with pointers
1062 +  void ForceMatrixDecomposition::fillInteractionData(InteractionData &idat,
1063 +                                                     int atom1, int atom2) {
1064 +
1065 +    idat.excluded = excludeAtomPair(atom1, atom2);
1066 +  
1067 + #ifdef IS_MPI
1068 +    idat.atypes = make_pair( atypesRow[atom1], atypesCol[atom2]);
1069 +    //idat.atypes = make_pair( ff_->getAtomType(identsRow[atom1]),
1070 +    //                         ff_->getAtomType(identsCol[atom2]) );
1071 +    
1072 +    if (storageLayout_ & DataStorage::dslAmat) {
1073 +      idat.A1 = &(atomRowData.aMat[atom1]);
1074 +      idat.A2 = &(atomColData.aMat[atom2]);
1075 +    }
1076 +    
1077 +    if (storageLayout_ & DataStorage::dslElectroFrame) {
1078 +      idat.eFrame1 = &(atomRowData.electroFrame[atom1]);
1079 +      idat.eFrame2 = &(atomColData.electroFrame[atom2]);
1080 +    }
1081 +
1082 +    if (storageLayout_ & DataStorage::dslTorque) {
1083 +      idat.t1 = &(atomRowData.torque[atom1]);
1084 +      idat.t2 = &(atomColData.torque[atom2]);
1085 +    }
1086 +
1087 +    if (storageLayout_ & DataStorage::dslDensity) {
1088 +      idat.rho1 = &(atomRowData.density[atom1]);
1089 +      idat.rho2 = &(atomColData.density[atom2]);
1090 +    }
1091 +
1092 +    if (storageLayout_ & DataStorage::dslFunctional) {
1093 +      idat.frho1 = &(atomRowData.functional[atom1]);
1094 +      idat.frho2 = &(atomColData.functional[atom2]);
1095 +    }
1096 +
1097 +    if (storageLayout_ & DataStorage::dslFunctionalDerivative) {
1098 +      idat.dfrho1 = &(atomRowData.functionalDerivative[atom1]);
1099 +      idat.dfrho2 = &(atomColData.functionalDerivative[atom2]);
1100 +    }
1101 +
1102 +    if (storageLayout_ & DataStorage::dslParticlePot) {
1103 +      idat.particlePot1 = &(atomRowData.particlePot[atom1]);
1104 +      idat.particlePot2 = &(atomColData.particlePot[atom2]);
1105 +    }
1106 +
1107 +    if (storageLayout_ & DataStorage::dslSkippedCharge) {              
1108 +      idat.skippedCharge1 = &(atomRowData.skippedCharge[atom1]);
1109 +      idat.skippedCharge2 = &(atomColData.skippedCharge[atom2]);
1110 +    }
1111 +
1112 +    if (storageLayout_ & DataStorage::dslFlucQPosition) {              
1113 +      idat.flucQ1 = &(atomRowData.flucQPos[atom1]);
1114 +      idat.flucQ2 = &(atomColData.flucQPos[atom2]);
1115 +    }
1116 +
1117 + #else
1118 +    
1119 +
1120 +    // cerr << "atoms = " << atom1 << " " << atom2 << "\n";
1121 +    // cerr << "pos1 = " << snap_->atomData.position[atom1] << "\n";
1122 +    // cerr << "pos2 = " << snap_->atomData.position[atom2] << "\n";
1123 +
1124 +    idat.atypes = make_pair( atypesLocal[atom1], atypesLocal[atom2]);
1125 +    //idat.atypes = make_pair( ff_->getAtomType(idents[atom1]),
1126 +    //                         ff_->getAtomType(idents[atom2]) );
1127 +
1128 +    if (storageLayout_ & DataStorage::dslAmat) {
1129 +      idat.A1 = &(snap_->atomData.aMat[atom1]);
1130 +      idat.A2 = &(snap_->atomData.aMat[atom2]);
1131 +    }
1132 +
1133 +    if (storageLayout_ & DataStorage::dslElectroFrame) {
1134 +      idat.eFrame1 = &(snap_->atomData.electroFrame[atom1]);
1135 +      idat.eFrame2 = &(snap_->atomData.electroFrame[atom2]);
1136 +    }
1137 +
1138 +    if (storageLayout_ & DataStorage::dslTorque) {
1139 +      idat.t1 = &(snap_->atomData.torque[atom1]);
1140 +      idat.t2 = &(snap_->atomData.torque[atom2]);
1141 +    }
1142 +
1143 +    if (storageLayout_ & DataStorage::dslDensity) {    
1144 +      idat.rho1 = &(snap_->atomData.density[atom1]);
1145 +      idat.rho2 = &(snap_->atomData.density[atom2]);
1146 +    }
1147 +
1148 +    if (storageLayout_ & DataStorage::dslFunctional) {
1149 +      idat.frho1 = &(snap_->atomData.functional[atom1]);
1150 +      idat.frho2 = &(snap_->atomData.functional[atom2]);
1151 +    }
1152 +
1153 +    if (storageLayout_ & DataStorage::dslFunctionalDerivative) {
1154 +      idat.dfrho1 = &(snap_->atomData.functionalDerivative[atom1]);
1155 +      idat.dfrho2 = &(snap_->atomData.functionalDerivative[atom2]);
1156 +    }
1157 +
1158 +    if (storageLayout_ & DataStorage::dslParticlePot) {
1159 +      idat.particlePot1 = &(snap_->atomData.particlePot[atom1]);
1160 +      idat.particlePot2 = &(snap_->atomData.particlePot[atom2]);
1161 +    }
1162 +
1163 +    if (storageLayout_ & DataStorage::dslSkippedCharge) {
1164 +      idat.skippedCharge1 = &(snap_->atomData.skippedCharge[atom1]);
1165 +      idat.skippedCharge2 = &(snap_->atomData.skippedCharge[atom2]);
1166 +    }
1167 +
1168 +    if (storageLayout_ & DataStorage::dslFlucQPosition) {              
1169 +      idat.flucQ1 = &(snap_->atomData.flucQPos[atom1]);
1170 +      idat.flucQ2 = &(snap_->atomData.flucQPos[atom2]);
1171 +    }
1172 +
1173 + #endif
1174 +  }
1175 +
1176 +  
1177 +  void ForceMatrixDecomposition::unpackInteractionData(InteractionData &idat, int atom1, int atom2) {    
1178 + #ifdef IS_MPI
1179 +    pot_row[atom1] += RealType(0.5) *  *(idat.pot);
1180 +    pot_col[atom2] += RealType(0.5) *  *(idat.pot);
1181 +
1182 +    atomRowData.force[atom1] += *(idat.f1);
1183 +    atomColData.force[atom2] -= *(idat.f1);
1184 +
1185 +    if (storageLayout_ & DataStorage::dslFlucQForce) {              
1186 +      atomRowData.flucQFrc[atom1] += *(idat.dVdFQ1);
1187 +      atomColData.flucQFrc[atom2] += *(idat.dVdFQ2);
1188 +    }
1189 +
1190 +    if (storageLayout_ & DataStorage::dslElectricField) {              
1191 +      atomRowData.electricField[atom1] += *(idat.eField1);
1192 +      atomColData.electricField[atom2] += *(idat.eField2);
1193 +    }
1194 +
1195 + #else
1196 +    pairwisePot += *(idat.pot);
1197 +
1198 +    snap_->atomData.force[atom1] += *(idat.f1);
1199 +    snap_->atomData.force[atom2] -= *(idat.f1);
1200 +
1201 +    if (idat.doParticlePot) {
1202 +      // This is the pairwise contribution to the particle pot.  The
1203 +      // embedding contribution is added in each of the low level
1204 +      // non-bonded routines.  In parallel, this calculation is done
1205 +      // in collectData, not in unpackInteractionData.
1206 +      snap_->atomData.particlePot[atom1] += *(idat.vpair) * *(idat.sw);
1207 +      snap_->atomData.particlePot[atom2] += *(idat.vpair) * *(idat.sw);
1208 +    }
1209 +    
1210 +    if (storageLayout_ & DataStorage::dslFlucQForce) {              
1211 +      snap_->atomData.flucQFrc[atom1] += *(idat.dVdFQ1);
1212 +      snap_->atomData.flucQFrc[atom2] -= *(idat.dVdFQ2);
1213 +    }
1214 +
1215 +    if (storageLayout_ & DataStorage::dslElectricField) {              
1216 +      snap_->atomData.electricField[atom1] += *(idat.eField1);
1217 +      snap_->atomData.electricField[atom2] += *(idat.eField2);
1218 +    }
1219 +
1220 + #endif
1221 +    
1222 +  }
1223 +
1224 +  /*
1225 +   * buildNeighborList
1226 +   *
1227 +   * first element of pair is row-indexed CutoffGroup
1228 +   * second element of pair is column-indexed CutoffGroup
1229 +   */
1230 +  vector<pair<int, int> > ForceMatrixDecomposition::buildNeighborList() {
1231 +      
1232 +    vector<pair<int, int> > neighborList;
1233 +    groupCutoffs cuts;
1234 +    bool doAllPairs = false;
1235 +
1236 + #ifdef IS_MPI
1237 +    cellListRow_.clear();
1238 +    cellListCol_.clear();
1239 + #else
1240 +    cellList_.clear();
1241 + #endif
1242 +
1243 +    RealType rList_ = (largestRcut_ + skinThickness_);
1244 +    RealType rl2 = rList_ * rList_;
1245 +    Snapshot* snap_ = sman_->getCurrentSnapshot();
1246 +    Mat3x3d Hmat = snap_->getHmat();
1247 +    Vector3d Hx = Hmat.getColumn(0);
1248 +    Vector3d Hy = Hmat.getColumn(1);
1249 +    Vector3d Hz = Hmat.getColumn(2);
1250 +
1251 +    nCells_.x() = (int) ( Hx.length() )/ rList_;
1252 +    nCells_.y() = (int) ( Hy.length() )/ rList_;
1253 +    nCells_.z() = (int) ( Hz.length() )/ rList_;
1254 +
1255 +    // handle small boxes where the cell offsets can end up repeating cells
1256 +    
1257 +    if (nCells_.x() < 3) doAllPairs = true;
1258 +    if (nCells_.y() < 3) doAllPairs = true;
1259 +    if (nCells_.z() < 3) doAllPairs = true;
1260 +
1261 +    Mat3x3d invHmat = snap_->getInvHmat();
1262 +    Vector3d rs, scaled, dr;
1263 +    Vector3i whichCell;
1264 +    int cellIndex;
1265 +    int nCtot = nCells_.x() * nCells_.y() * nCells_.z();
1266 +
1267 + #ifdef IS_MPI
1268 +    cellListRow_.resize(nCtot);
1269 +    cellListCol_.resize(nCtot);
1270 + #else
1271 +    cellList_.resize(nCtot);
1272 + #endif
1273 +
1274 +    if (!doAllPairs) {
1275 + #ifdef IS_MPI
1276 +
1277 +      for (int i = 0; i < nGroupsInRow_; i++) {
1278 +        rs = cgRowData.position[i];
1279 +        
1280 +        // scaled positions relative to the box vectors
1281 +        scaled = invHmat * rs;
1282 +        
1283 +        // wrap the vector back into the unit box by subtracting integer box
1284 +        // numbers
1285 +        for (int j = 0; j < 3; j++) {
1286 +          scaled[j] -= roundMe(scaled[j]);
1287 +          scaled[j] += 0.5;
1288 +        }
1289 +        
1290 +        // find xyz-indices of cell that cutoffGroup is in.
1291 +        whichCell.x() = nCells_.x() * scaled.x();
1292 +        whichCell.y() = nCells_.y() * scaled.y();
1293 +        whichCell.z() = nCells_.z() * scaled.z();
1294 +        
1295 +        // find single index of this cell:
1296 +        cellIndex = Vlinear(whichCell, nCells_);
1297 +        
1298 +        // add this cutoff group to the list of groups in this cell;
1299 +        cellListRow_[cellIndex].push_back(i);
1300 +      }
1301 +      for (int i = 0; i < nGroupsInCol_; i++) {
1302 +        rs = cgColData.position[i];
1303 +        
1304 +        // scaled positions relative to the box vectors
1305 +        scaled = invHmat * rs;
1306 +        
1307 +        // wrap the vector back into the unit box by subtracting integer box
1308 +        // numbers
1309 +        for (int j = 0; j < 3; j++) {
1310 +          scaled[j] -= roundMe(scaled[j]);
1311 +          scaled[j] += 0.5;
1312 +        }
1313 +        
1314 +        // find xyz-indices of cell that cutoffGroup is in.
1315 +        whichCell.x() = nCells_.x() * scaled.x();
1316 +        whichCell.y() = nCells_.y() * scaled.y();
1317 +        whichCell.z() = nCells_.z() * scaled.z();
1318 +        
1319 +        // find single index of this cell:
1320 +        cellIndex = Vlinear(whichCell, nCells_);
1321 +        
1322 +        // add this cutoff group to the list of groups in this cell;
1323 +        cellListCol_[cellIndex].push_back(i);
1324 +      }
1325 +    
1326 + #else
1327 +      for (int i = 0; i < nGroups_; i++) {
1328 +        rs = snap_->cgData.position[i];
1329 +        
1330 +        // scaled positions relative to the box vectors
1331 +        scaled = invHmat * rs;
1332 +        
1333 +        // wrap the vector back into the unit box by subtracting integer box
1334 +        // numbers
1335 +        for (int j = 0; j < 3; j++) {
1336 +          scaled[j] -= roundMe(scaled[j]);
1337 +          scaled[j] += 0.5;
1338 +        }
1339 +        
1340 +        // find xyz-indices of cell that cutoffGroup is in.
1341 +        whichCell.x() = nCells_.x() * scaled.x();
1342 +        whichCell.y() = nCells_.y() * scaled.y();
1343 +        whichCell.z() = nCells_.z() * scaled.z();
1344 +        
1345 +        // find single index of this cell:
1346 +        cellIndex = Vlinear(whichCell, nCells_);
1347 +        
1348 +        // add this cutoff group to the list of groups in this cell;
1349 +        cellList_[cellIndex].push_back(i);
1350 +      }
1351 +
1352 + #endif
1353 +
1354 +      for (int m1z = 0; m1z < nCells_.z(); m1z++) {
1355 +        for (int m1y = 0; m1y < nCells_.y(); m1y++) {
1356 +          for (int m1x = 0; m1x < nCells_.x(); m1x++) {
1357 +            Vector3i m1v(m1x, m1y, m1z);
1358 +            int m1 = Vlinear(m1v, nCells_);
1359 +            
1360 +            for (vector<Vector3i>::iterator os = cellOffsets_.begin();
1361 +                 os != cellOffsets_.end(); ++os) {
1362 +              
1363 +              Vector3i m2v = m1v + (*os);
1364 +            
1365 +
1366 +              if (m2v.x() >= nCells_.x()) {
1367 +                m2v.x() = 0;          
1368 +              } else if (m2v.x() < 0) {
1369 +                m2v.x() = nCells_.x() - 1;
1370 +              }
1371 +              
1372 +              if (m2v.y() >= nCells_.y()) {
1373 +                m2v.y() = 0;          
1374 +              } else if (m2v.y() < 0) {
1375 +                m2v.y() = nCells_.y() - 1;
1376 +              }
1377 +              
1378 +              if (m2v.z() >= nCells_.z()) {
1379 +                m2v.z() = 0;          
1380 +              } else if (m2v.z() < 0) {
1381 +                m2v.z() = nCells_.z() - 1;
1382 +              }
1383 +
1384 +              int m2 = Vlinear (m2v, nCells_);
1385 +              
1386 + #ifdef IS_MPI
1387 +              for (vector<int>::iterator j1 = cellListRow_[m1].begin();
1388 +                   j1 != cellListRow_[m1].end(); ++j1) {
1389 +                for (vector<int>::iterator j2 = cellListCol_[m2].begin();
1390 +                     j2 != cellListCol_[m2].end(); ++j2) {
1391 +                  
1392 +                  // In parallel, we need to visit *all* pairs of row
1393 +                  // & column indicies and will divide labor in the
1394 +                  // force evaluation later.
1395 +                  dr = cgColData.position[(*j2)] - cgRowData.position[(*j1)];
1396 +                  snap_->wrapVector(dr);
1397 +                  cuts = getGroupCutoffs( (*j1), (*j2) );
1398 +                  if (dr.lengthSquare() < cuts.third) {
1399 +                    neighborList.push_back(make_pair((*j1), (*j2)));
1400 +                  }                  
1401 +                }
1402 +              }
1403 + #else
1404 +              for (vector<int>::iterator j1 = cellList_[m1].begin();
1405 +                   j1 != cellList_[m1].end(); ++j1) {
1406 +                for (vector<int>::iterator j2 = cellList_[m2].begin();
1407 +                     j2 != cellList_[m2].end(); ++j2) {
1408 +    
1409 +                  // Always do this if we're in different cells or if
1410 +                  // we're in the same cell and the global index of
1411 +                  // the j2 cutoff group is greater than or equal to
1412 +                  // the j1 cutoff group.  Note that Rappaport's code
1413 +                  // has a "less than" conditional here, but that
1414 +                  // deals with atom-by-atom computation.  OpenMD
1415 +                  // allows atoms within a single cutoff group to
1416 +                  // interact with each other.
1417 +
1418 +
1419 +
1420 +                  if (m2 != m1 || (*j2) >= (*j1) ) {
1421 +
1422 +                    dr = snap_->cgData.position[(*j2)] - snap_->cgData.position[(*j1)];
1423 +                    snap_->wrapVector(dr);
1424 +                    cuts = getGroupCutoffs( (*j1), (*j2) );
1425 +                    if (dr.lengthSquare() < cuts.third) {
1426 +                      neighborList.push_back(make_pair((*j1), (*j2)));
1427 +                    }
1428 +                  }
1429 +                }
1430 +              }
1431 + #endif
1432 +            }
1433 +          }
1434 +        }
1435 +      }
1436 +    } else {
1437 +      // branch to do all cutoff group pairs
1438 + #ifdef IS_MPI
1439 +      for (int j1 = 0; j1 < nGroupsInRow_; j1++) {
1440 +        for (int j2 = 0; j2 < nGroupsInCol_; j2++) {    
1441 +          dr = cgColData.position[j2] - cgRowData.position[j1];
1442 +          snap_->wrapVector(dr);
1443 +          cuts = getGroupCutoffs( j1, j2 );
1444 +          if (dr.lengthSquare() < cuts.third) {
1445 +            neighborList.push_back(make_pair(j1, j2));
1446 +          }
1447 +        }
1448 +      }      
1449 + #else
1450 +      // include all groups here.
1451 +      for (int j1 = 0; j1 < nGroups_; j1++) {
1452 +        // include self group interactions j2 == j1
1453 +        for (int j2 = j1; j2 < nGroups_; j2++) {
1454 +          dr = snap_->cgData.position[j2] - snap_->cgData.position[j1];
1455 +          snap_->wrapVector(dr);
1456 +          cuts = getGroupCutoffs( j1, j2 );
1457 +          if (dr.lengthSquare() < cuts.third) {
1458 +            neighborList.push_back(make_pair(j1, j2));
1459 +          }
1460 +        }    
1461 +      }
1462 + #endif
1463 +    }
1464 +      
1465 +    // save the local cutoff group positions for the check that is
1466 +    // done on each loop:
1467 +    saved_CG_positions_.clear();
1468 +    for (int i = 0; i < nGroups_; i++)
1469 +      saved_CG_positions_.push_back(snap_->cgData.position[i]);
1470 +    
1471 +    return neighborList;
1472 +  }
1473 + } //end namespace OpenMD

Diff Legend

Removed lines
+ Added lines
< Changed lines
> Changed lines