ViewVC Help
View File | Revision Log | Show Annotations | View Changeset | Root Listing
root/OpenMD/trunk/src/parallel/ForceMatrixDecomposition.cpp
Revision: 2061
Committed: Tue Mar 3 16:24:44 2015 UTC (10 years, 1 month ago) by gezelter
File size: 54574 byte(s)
Log Message:
Fixed a warning on PGI compilation

File Contents

# Content
1 /*
2 * Copyright (c) 2005 The University of Notre Dame. All Rights Reserved.
3 *
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
7 * that the following conditions are met:
8 *
9 * 1. Redistributions of source code must retain the above copyright
10 * notice, this list of conditions and the following disclaimer.
11 *
12 * 2. Redistributions in binary form must reproduce the above copyright
13 * notice, this list of conditions and the following disclaimer in the
14 * documentation and/or other materials provided with the
15 * distribution.
16 *
17 * This software is provided "AS IS," without a warranty of any
18 * kind. All express or implied conditions, representations and
19 * warranties, including any implied warranty of merchantability,
20 * fitness for a particular purpose or non-infringement, are hereby
21 * excluded. The University of Notre Dame and its licensors shall not
22 * be liable for any damages suffered by licensee as a result of
23 * using, modifying or distributing the software or its
24 * derivatives. In no event will the University of Notre Dame or its
25 * licensors be liable for any lost revenue, profit or data, or for
26 * direct, indirect, special, consequential, incidental or punitive
27 * damages, however caused and regardless of the theory of liability,
28 * arising out of the use of or inability to use software, even if the
29 * University of Notre Dame has been advised of the possibility of
30 * such damages.
31 *
32 * SUPPORT OPEN SCIENCE! If you use OpenMD or its source code in your
33 * research, please cite the appropriate papers when you publish your
34 * work. Good starting points are:
35 *
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, 234107 (2008).
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 // Row and colum scans must visit all surrounding cells
54 cellOffsets_.clear();
55 cellOffsets_.push_back( Vector3i(-1,-1,-1) );
56 cellOffsets_.push_back( Vector3i( 0,-1,-1) );
57 cellOffsets_.push_back( Vector3i( 1,-1,-1) );
58 cellOffsets_.push_back( Vector3i(-1, 0,-1) );
59 cellOffsets_.push_back( Vector3i( 0, 0,-1) );
60 cellOffsets_.push_back( Vector3i( 1, 0,-1) );
61 cellOffsets_.push_back( Vector3i(-1, 1,-1) );
62 cellOffsets_.push_back( Vector3i( 0, 1,-1) );
63 cellOffsets_.push_back( Vector3i( 1, 1,-1) );
64 cellOffsets_.push_back( Vector3i(-1,-1, 0) );
65 cellOffsets_.push_back( Vector3i( 0,-1, 0) );
66 cellOffsets_.push_back( Vector3i( 1,-1, 0) );
67 cellOffsets_.push_back( Vector3i(-1, 0, 0) );
68 cellOffsets_.push_back( Vector3i( 0, 0, 0) );
69 cellOffsets_.push_back( Vector3i( 1, 0, 0) );
70 cellOffsets_.push_back( Vector3i(-1, 1, 0) );
71 cellOffsets_.push_back( Vector3i( 0, 1, 0) );
72 cellOffsets_.push_back( Vector3i( 1, 1, 0) );
73 cellOffsets_.push_back( Vector3i(-1,-1, 1) );
74 cellOffsets_.push_back( Vector3i( 0,-1, 1) );
75 cellOffsets_.push_back( Vector3i( 1,-1, 1) );
76 cellOffsets_.push_back( Vector3i(-1, 0, 1) );
77 cellOffsets_.push_back( Vector3i( 0, 0, 1) );
78 cellOffsets_.push_back( Vector3i( 1, 0, 1) );
79 cellOffsets_.push_back( Vector3i(-1, 1, 1) );
80 cellOffsets_.push_back( Vector3i( 0, 1, 1) );
81 cellOffsets_.push_back( Vector3i( 1, 1, 1) );
82 }
83
84
85 /**
86 * distributeInitialData is essentially a copy of the older fortran
87 * SimulationSetup
88 */
89 void ForceMatrixDecomposition::distributeInitialData() {
90 snap_ = sman_->getCurrentSnapshot();
91 storageLayout_ = sman_->getStorageLayout();
92 ff_ = info_->getForceField();
93 nLocal_ = snap_->getNumberOfAtoms();
94
95 nGroups_ = info_->getNLocalCutoffGroups();
96 // gather the information for atomtype IDs (atids):
97 idents = info_->getIdentArray();
98 regions = info_->getRegions();
99 AtomLocalToGlobal = info_->getGlobalAtomIndices();
100 cgLocalToGlobal = info_->getGlobalGroupIndices();
101 vector<int> globalGroupMembership = info_->getGlobalGroupMembership();
102
103 massFactors = info_->getMassFactors();
104
105 PairList* excludes = info_->getExcludedInteractions();
106 PairList* oneTwo = info_->getOneTwoInteractions();
107 PairList* oneThree = info_->getOneThreeInteractions();
108 PairList* oneFour = info_->getOneFourInteractions();
109
110 if (needVelocities_)
111 snap_->cgData.setStorageLayout(DataStorage::dslPosition |
112 DataStorage::dslVelocity);
113 else
114 snap_->cgData.setStorageLayout(DataStorage::dslPosition);
115
116 #ifdef IS_MPI
117
118 MPI_Comm row = rowComm.getComm();
119 MPI_Comm col = colComm.getComm();
120
121 AtomPlanIntRow = new Plan<int>(row, nLocal_);
122 AtomPlanRealRow = new Plan<RealType>(row, nLocal_);
123 AtomPlanVectorRow = new Plan<Vector3d>(row, nLocal_);
124 AtomPlanMatrixRow = new Plan<Mat3x3d>(row, nLocal_);
125 AtomPlanPotRow = new Plan<potVec>(row, nLocal_);
126
127 AtomPlanIntColumn = new Plan<int>(col, nLocal_);
128 AtomPlanRealColumn = new Plan<RealType>(col, nLocal_);
129 AtomPlanVectorColumn = new Plan<Vector3d>(col, nLocal_);
130 AtomPlanMatrixColumn = new Plan<Mat3x3d>(col, nLocal_);
131 AtomPlanPotColumn = new Plan<potVec>(col, nLocal_);
132
133 cgPlanIntRow = new Plan<int>(row, nGroups_);
134 cgPlanVectorRow = new Plan<Vector3d>(row, nGroups_);
135 cgPlanIntColumn = new Plan<int>(col, nGroups_);
136 cgPlanVectorColumn = new Plan<Vector3d>(col, nGroups_);
137
138 nAtomsInRow_ = AtomPlanIntRow->getSize();
139 nAtomsInCol_ = AtomPlanIntColumn->getSize();
140 nGroupsInRow_ = cgPlanIntRow->getSize();
141 nGroupsInCol_ = cgPlanIntColumn->getSize();
142
143 // Modify the data storage objects with the correct layouts and sizes:
144 atomRowData.resize(nAtomsInRow_);
145 atomRowData.setStorageLayout(storageLayout_);
146 atomColData.resize(nAtomsInCol_);
147 atomColData.setStorageLayout(storageLayout_);
148 cgRowData.resize(nGroupsInRow_);
149 cgRowData.setStorageLayout(DataStorage::dslPosition);
150 cgColData.resize(nGroupsInCol_);
151 if (needVelocities_)
152 // we only need column velocities if we need them.
153 cgColData.setStorageLayout(DataStorage::dslPosition |
154 DataStorage::dslVelocity);
155 else
156 cgColData.setStorageLayout(DataStorage::dslPosition);
157
158 identsRow.resize(nAtomsInRow_);
159 identsCol.resize(nAtomsInCol_);
160
161 AtomPlanIntRow->gather(idents, identsRow);
162 AtomPlanIntColumn->gather(idents, identsCol);
163
164 regionsRow.resize(nAtomsInRow_);
165 regionsCol.resize(nAtomsInCol_);
166
167 AtomPlanIntRow->gather(regions, regionsRow);
168 AtomPlanIntColumn->gather(regions, regionsCol);
169
170 // allocate memory for the parallel objects
171 atypesRow.resize(nAtomsInRow_);
172 atypesCol.resize(nAtomsInCol_);
173
174 for (int i = 0; i < nAtomsInRow_; i++)
175 atypesRow[i] = ff_->getAtomType(identsRow[i]);
176 for (int i = 0; i < nAtomsInCol_; i++)
177 atypesCol[i] = ff_->getAtomType(identsCol[i]);
178
179 pot_row.resize(nAtomsInRow_);
180 pot_col.resize(nAtomsInCol_);
181
182 expot_row.resize(nAtomsInRow_);
183 expot_col.resize(nAtomsInCol_);
184
185 AtomRowToGlobal.resize(nAtomsInRow_);
186 AtomColToGlobal.resize(nAtomsInCol_);
187 AtomPlanIntRow->gather(AtomLocalToGlobal, AtomRowToGlobal);
188 AtomPlanIntColumn->gather(AtomLocalToGlobal, AtomColToGlobal);
189
190 cgRowToGlobal.resize(nGroupsInRow_);
191 cgColToGlobal.resize(nGroupsInCol_);
192 cgPlanIntRow->gather(cgLocalToGlobal, cgRowToGlobal);
193 cgPlanIntColumn->gather(cgLocalToGlobal, cgColToGlobal);
194
195 massFactorsRow.resize(nAtomsInRow_);
196 massFactorsCol.resize(nAtomsInCol_);
197 AtomPlanRealRow->gather(massFactors, massFactorsRow);
198 AtomPlanRealColumn->gather(massFactors, massFactorsCol);
199
200 groupListRow_.clear();
201 groupListRow_.resize(nGroupsInRow_);
202 for (int i = 0; i < nGroupsInRow_; i++) {
203 int gid = cgRowToGlobal[i];
204 for (int j = 0; j < nAtomsInRow_; j++) {
205 int aid = AtomRowToGlobal[j];
206 if (globalGroupMembership[aid] == gid)
207 groupListRow_[i].push_back(j);
208 }
209 }
210
211 groupListCol_.clear();
212 groupListCol_.resize(nGroupsInCol_);
213 for (int i = 0; i < nGroupsInCol_; i++) {
214 int gid = cgColToGlobal[i];
215 for (int j = 0; j < nAtomsInCol_; j++) {
216 int aid = AtomColToGlobal[j];
217 if (globalGroupMembership[aid] == gid)
218 groupListCol_[i].push_back(j);
219 }
220 }
221
222 excludesForAtom.clear();
223 excludesForAtom.resize(nAtomsInRow_);
224 toposForAtom.clear();
225 toposForAtom.resize(nAtomsInRow_);
226 topoDist.clear();
227 topoDist.resize(nAtomsInRow_);
228 for (int i = 0; i < nAtomsInRow_; i++) {
229 int iglob = AtomRowToGlobal[i];
230
231 for (int j = 0; j < nAtomsInCol_; j++) {
232 int jglob = AtomColToGlobal[j];
233
234 if (excludes->hasPair(iglob, jglob))
235 excludesForAtom[i].push_back(j);
236
237 if (oneTwo->hasPair(iglob, jglob)) {
238 toposForAtom[i].push_back(j);
239 topoDist[i].push_back(1);
240 } else {
241 if (oneThree->hasPair(iglob, jglob)) {
242 toposForAtom[i].push_back(j);
243 topoDist[i].push_back(2);
244 } else {
245 if (oneFour->hasPair(iglob, jglob)) {
246 toposForAtom[i].push_back(j);
247 topoDist[i].push_back(3);
248 }
249 }
250 }
251 }
252 }
253
254 #else
255 excludesForAtom.clear();
256 excludesForAtom.resize(nLocal_);
257 toposForAtom.clear();
258 toposForAtom.resize(nLocal_);
259 topoDist.clear();
260 topoDist.resize(nLocal_);
261
262 for (int i = 0; i < nLocal_; i++) {
263 int iglob = AtomLocalToGlobal[i];
264
265 for (int j = 0; j < nLocal_; j++) {
266 int jglob = AtomLocalToGlobal[j];
267
268 if (excludes->hasPair(iglob, jglob))
269 excludesForAtom[i].push_back(j);
270
271 if (oneTwo->hasPair(iglob, jglob)) {
272 toposForAtom[i].push_back(j);
273 topoDist[i].push_back(1);
274 } else {
275 if (oneThree->hasPair(iglob, jglob)) {
276 toposForAtom[i].push_back(j);
277 topoDist[i].push_back(2);
278 } else {
279 if (oneFour->hasPair(iglob, jglob)) {
280 toposForAtom[i].push_back(j);
281 topoDist[i].push_back(3);
282 }
283 }
284 }
285 }
286 }
287 #endif
288
289 // allocate memory for the parallel objects
290 atypesLocal.resize(nLocal_);
291
292 for (int i = 0; i < nLocal_; i++)
293 atypesLocal[i] = ff_->getAtomType(idents[i]);
294
295 groupList_.clear();
296 groupList_.resize(nGroups_);
297 for (int i = 0; i < nGroups_; i++) {
298 int gid = cgLocalToGlobal[i];
299 for (int j = 0; j < nLocal_; j++) {
300 int aid = AtomLocalToGlobal[j];
301 if (globalGroupMembership[aid] == gid) {
302 groupList_[i].push_back(j);
303 }
304 }
305 }
306 }
307
308 int ForceMatrixDecomposition::getTopologicalDistance(int atom1, int atom2) {
309 for (unsigned int j = 0; j < toposForAtom[atom1].size(); j++) {
310 if (toposForAtom[atom1][j] == atom2)
311 return topoDist[atom1][j];
312 }
313 return 0;
314 }
315
316 void ForceMatrixDecomposition::zeroWorkArrays() {
317 pairwisePot = 0.0;
318 embeddingPot = 0.0;
319 excludedPot = 0.0;
320 excludedSelfPot = 0.0;
321
322 #ifdef IS_MPI
323 if (storageLayout_ & DataStorage::dslForce) {
324 fill(atomRowData.force.begin(), atomRowData.force.end(), V3Zero);
325 fill(atomColData.force.begin(), atomColData.force.end(), V3Zero);
326 }
327
328 if (storageLayout_ & DataStorage::dslTorque) {
329 fill(atomRowData.torque.begin(), atomRowData.torque.end(), V3Zero);
330 fill(atomColData.torque.begin(), atomColData.torque.end(), V3Zero);
331 }
332
333 fill(pot_row.begin(), pot_row.end(),
334 Vector<RealType, N_INTERACTION_FAMILIES> (0.0));
335
336 fill(pot_col.begin(), pot_col.end(),
337 Vector<RealType, N_INTERACTION_FAMILIES> (0.0));
338
339 fill(expot_row.begin(), expot_row.end(),
340 Vector<RealType, N_INTERACTION_FAMILIES> (0.0));
341
342 fill(expot_col.begin(), expot_col.end(),
343 Vector<RealType, N_INTERACTION_FAMILIES> (0.0));
344
345 if (storageLayout_ & DataStorage::dslParticlePot) {
346 fill(atomRowData.particlePot.begin(), atomRowData.particlePot.end(),
347 0.0);
348 fill(atomColData.particlePot.begin(), atomColData.particlePot.end(),
349 0.0);
350 }
351
352 if (storageLayout_ & DataStorage::dslDensity) {
353 fill(atomRowData.density.begin(), atomRowData.density.end(), 0.0);
354 fill(atomColData.density.begin(), atomColData.density.end(), 0.0);
355 }
356
357 if (storageLayout_ & DataStorage::dslFunctional) {
358 fill(atomRowData.functional.begin(), atomRowData.functional.end(),
359 0.0);
360 fill(atomColData.functional.begin(), atomColData.functional.end(),
361 0.0);
362 }
363
364 if (storageLayout_ & DataStorage::dslFunctionalDerivative) {
365 fill(atomRowData.functionalDerivative.begin(),
366 atomRowData.functionalDerivative.end(), 0.0);
367 fill(atomColData.functionalDerivative.begin(),
368 atomColData.functionalDerivative.end(), 0.0);
369 }
370
371 if (storageLayout_ & DataStorage::dslSkippedCharge) {
372 fill(atomRowData.skippedCharge.begin(),
373 atomRowData.skippedCharge.end(), 0.0);
374 fill(atomColData.skippedCharge.begin(),
375 atomColData.skippedCharge.end(), 0.0);
376 }
377
378 if (storageLayout_ & DataStorage::dslFlucQForce) {
379 fill(atomRowData.flucQFrc.begin(),
380 atomRowData.flucQFrc.end(), 0.0);
381 fill(atomColData.flucQFrc.begin(),
382 atomColData.flucQFrc.end(), 0.0);
383 }
384
385 if (storageLayout_ & DataStorage::dslElectricField) {
386 fill(atomRowData.electricField.begin(),
387 atomRowData.electricField.end(), V3Zero);
388 fill(atomColData.electricField.begin(),
389 atomColData.electricField.end(), V3Zero);
390 }
391
392 if (storageLayout_ & DataStorage::dslSitePotential) {
393 fill(atomRowData.sitePotential.begin(),
394 atomRowData.sitePotential.end(), 0.0);
395 fill(atomColData.sitePotential.begin(),
396 atomColData.sitePotential.end(), 0.0);
397 }
398
399 #endif
400 // even in parallel, we need to zero out the local arrays:
401
402 if (storageLayout_ & DataStorage::dslParticlePot) {
403 fill(snap_->atomData.particlePot.begin(),
404 snap_->atomData.particlePot.end(), 0.0);
405 }
406
407 if (storageLayout_ & DataStorage::dslDensity) {
408 fill(snap_->atomData.density.begin(),
409 snap_->atomData.density.end(), 0.0);
410 }
411
412 if (storageLayout_ & DataStorage::dslFunctional) {
413 fill(snap_->atomData.functional.begin(),
414 snap_->atomData.functional.end(), 0.0);
415 }
416
417 if (storageLayout_ & DataStorage::dslFunctionalDerivative) {
418 fill(snap_->atomData.functionalDerivative.begin(),
419 snap_->atomData.functionalDerivative.end(), 0.0);
420 }
421
422 if (storageLayout_ & DataStorage::dslSkippedCharge) {
423 fill(snap_->atomData.skippedCharge.begin(),
424 snap_->atomData.skippedCharge.end(), 0.0);
425 }
426
427 if (storageLayout_ & DataStorage::dslElectricField) {
428 fill(snap_->atomData.electricField.begin(),
429 snap_->atomData.electricField.end(), V3Zero);
430 }
431 if (storageLayout_ & DataStorage::dslSitePotential) {
432 fill(snap_->atomData.sitePotential.begin(),
433 snap_->atomData.sitePotential.end(), 0.0);
434 }
435 }
436
437
438 void ForceMatrixDecomposition::distributeData() {
439
440 #ifdef IS_MPI
441
442 snap_ = sman_->getCurrentSnapshot();
443 storageLayout_ = sman_->getStorageLayout();
444
445 bool needsCG = true;
446 if(info_->getNCutoffGroups() != info_->getNAtoms())
447 needsCG = false;
448
449 // gather up the atomic positions
450 AtomPlanVectorRow->gather(snap_->atomData.position,
451 atomRowData.position);
452 AtomPlanVectorColumn->gather(snap_->atomData.position,
453 atomColData.position);
454
455 // gather up the cutoff group positions
456
457 if (needsCG) {
458 cgPlanVectorRow->gather(snap_->cgData.position,
459 cgRowData.position);
460
461 cgPlanVectorColumn->gather(snap_->cgData.position,
462 cgColData.position);
463 }
464
465
466 if (needVelocities_) {
467 // gather up the atomic velocities
468 AtomPlanVectorColumn->gather(snap_->atomData.velocity,
469 atomColData.velocity);
470
471 if (needsCG) {
472 cgPlanVectorColumn->gather(snap_->cgData.velocity,
473 cgColData.velocity);
474 }
475 }
476
477
478 // if needed, gather the atomic rotation matrices
479 if (storageLayout_ & DataStorage::dslAmat) {
480 AtomPlanMatrixRow->gather(snap_->atomData.aMat,
481 atomRowData.aMat);
482 AtomPlanMatrixColumn->gather(snap_->atomData.aMat,
483 atomColData.aMat);
484 }
485
486 // if needed, gather the atomic eletrostatic information
487 if (storageLayout_ & DataStorage::dslDipole) {
488 AtomPlanVectorRow->gather(snap_->atomData.dipole,
489 atomRowData.dipole);
490 AtomPlanVectorColumn->gather(snap_->atomData.dipole,
491 atomColData.dipole);
492 }
493
494 if (storageLayout_ & DataStorage::dslQuadrupole) {
495 AtomPlanMatrixRow->gather(snap_->atomData.quadrupole,
496 atomRowData.quadrupole);
497 AtomPlanMatrixColumn->gather(snap_->atomData.quadrupole,
498 atomColData.quadrupole);
499 }
500
501 // if needed, gather the atomic fluctuating charge values
502 if (storageLayout_ & DataStorage::dslFlucQPosition) {
503 AtomPlanRealRow->gather(snap_->atomData.flucQPos,
504 atomRowData.flucQPos);
505 AtomPlanRealColumn->gather(snap_->atomData.flucQPos,
506 atomColData.flucQPos);
507 }
508
509 #endif
510 }
511
512 /* collects information obtained during the pre-pair loop onto local
513 * data structures.
514 */
515 void ForceMatrixDecomposition::collectIntermediateData() {
516 snap_ = sman_->getCurrentSnapshot();
517 storageLayout_ = sman_->getStorageLayout();
518 #ifdef IS_MPI
519
520 if (storageLayout_ & DataStorage::dslDensity) {
521
522 AtomPlanRealRow->scatter(atomRowData.density,
523 snap_->atomData.density);
524
525 int n = snap_->atomData.density.size();
526 vector<RealType> rho_tmp(n, 0.0);
527 AtomPlanRealColumn->scatter(atomColData.density, rho_tmp);
528 for (int i = 0; i < n; i++)
529 snap_->atomData.density[i] += rho_tmp[i];
530 }
531
532 // this isn't necessary if we don't have polarizable atoms, but
533 // we'll leave it here for now.
534 if (storageLayout_ & DataStorage::dslElectricField) {
535
536 AtomPlanVectorRow->scatter(atomRowData.electricField,
537 snap_->atomData.electricField);
538
539 int n = snap_->atomData.electricField.size();
540 vector<Vector3d> field_tmp(n, V3Zero);
541 AtomPlanVectorColumn->scatter(atomColData.electricField,
542 field_tmp);
543 for (int i = 0; i < n; i++)
544 snap_->atomData.electricField[i] += field_tmp[i];
545 }
546 #endif
547 }
548
549 /*
550 * redistributes information obtained during the pre-pair loop out to
551 * row and column-indexed data structures
552 */
553 void ForceMatrixDecomposition::distributeIntermediateData() {
554 snap_ = sman_->getCurrentSnapshot();
555 storageLayout_ = sman_->getStorageLayout();
556 #ifdef IS_MPI
557 if (storageLayout_ & DataStorage::dslFunctional) {
558 AtomPlanRealRow->gather(snap_->atomData.functional,
559 atomRowData.functional);
560 AtomPlanRealColumn->gather(snap_->atomData.functional,
561 atomColData.functional);
562 }
563
564 if (storageLayout_ & DataStorage::dslFunctionalDerivative) {
565 AtomPlanRealRow->gather(snap_->atomData.functionalDerivative,
566 atomRowData.functionalDerivative);
567 AtomPlanRealColumn->gather(snap_->atomData.functionalDerivative,
568 atomColData.functionalDerivative);
569 }
570 #endif
571 }
572
573
574 void ForceMatrixDecomposition::collectData() {
575 snap_ = sman_->getCurrentSnapshot();
576 storageLayout_ = sman_->getStorageLayout();
577 #ifdef IS_MPI
578 int n = snap_->atomData.force.size();
579 vector<Vector3d> frc_tmp(n, V3Zero);
580
581 AtomPlanVectorRow->scatter(atomRowData.force, frc_tmp);
582 for (int i = 0; i < n; i++) {
583 snap_->atomData.force[i] += frc_tmp[i];
584 frc_tmp[i] = 0.0;
585 }
586
587 AtomPlanVectorColumn->scatter(atomColData.force, frc_tmp);
588 for (int i = 0; i < n; i++) {
589 snap_->atomData.force[i] += frc_tmp[i];
590 }
591
592 if (storageLayout_ & DataStorage::dslTorque) {
593
594 int nt = snap_->atomData.torque.size();
595 vector<Vector3d> trq_tmp(nt, V3Zero);
596
597 AtomPlanVectorRow->scatter(atomRowData.torque, trq_tmp);
598 for (int i = 0; i < nt; i++) {
599 snap_->atomData.torque[i] += trq_tmp[i];
600 trq_tmp[i] = 0.0;
601 }
602
603 AtomPlanVectorColumn->scatter(atomColData.torque, trq_tmp);
604 for (int i = 0; i < nt; i++)
605 snap_->atomData.torque[i] += trq_tmp[i];
606 }
607
608 if (storageLayout_ & DataStorage::dslSkippedCharge) {
609
610 int ns = snap_->atomData.skippedCharge.size();
611 vector<RealType> skch_tmp(ns, 0.0);
612
613 AtomPlanRealRow->scatter(atomRowData.skippedCharge, skch_tmp);
614 for (int i = 0; i < ns; i++) {
615 snap_->atomData.skippedCharge[i] += skch_tmp[i];
616 skch_tmp[i] = 0.0;
617 }
618
619 AtomPlanRealColumn->scatter(atomColData.skippedCharge, skch_tmp);
620 for (int i = 0; i < ns; i++)
621 snap_->atomData.skippedCharge[i] += skch_tmp[i];
622
623 }
624
625 if (storageLayout_ & DataStorage::dslFlucQForce) {
626
627 int nq = snap_->atomData.flucQFrc.size();
628 vector<RealType> fqfrc_tmp(nq, 0.0);
629
630 AtomPlanRealRow->scatter(atomRowData.flucQFrc, fqfrc_tmp);
631 for (int i = 0; i < nq; i++) {
632 snap_->atomData.flucQFrc[i] += fqfrc_tmp[i];
633 fqfrc_tmp[i] = 0.0;
634 }
635
636 AtomPlanRealColumn->scatter(atomColData.flucQFrc, fqfrc_tmp);
637 for (int i = 0; i < nq; i++)
638 snap_->atomData.flucQFrc[i] += fqfrc_tmp[i];
639
640 }
641
642 if (storageLayout_ & DataStorage::dslElectricField) {
643
644 int nef = snap_->atomData.electricField.size();
645 vector<Vector3d> efield_tmp(nef, V3Zero);
646
647 AtomPlanVectorRow->scatter(atomRowData.electricField, efield_tmp);
648 for (int i = 0; i < nef; i++) {
649 snap_->atomData.electricField[i] += efield_tmp[i];
650 efield_tmp[i] = 0.0;
651 }
652
653 AtomPlanVectorColumn->scatter(atomColData.electricField, efield_tmp);
654 for (int i = 0; i < nef; i++)
655 snap_->atomData.electricField[i] += efield_tmp[i];
656 }
657
658 if (storageLayout_ & DataStorage::dslSitePotential) {
659
660 int nsp = snap_->atomData.sitePotential.size();
661 vector<RealType> sp_tmp(nsp, 0.0);
662
663 AtomPlanRealRow->scatter(atomRowData.sitePotential, sp_tmp);
664 for (int i = 0; i < nsp; i++) {
665 snap_->atomData.sitePotential[i] += sp_tmp[i];
666 sp_tmp[i] = 0.0;
667 }
668
669 AtomPlanRealColumn->scatter(atomColData.sitePotential, sp_tmp);
670 for (int i = 0; i < nsp; i++)
671 snap_->atomData.sitePotential[i] += sp_tmp[i];
672 }
673
674 nLocal_ = snap_->getNumberOfAtoms();
675
676 vector<potVec> pot_temp(nLocal_,
677 Vector<RealType, N_INTERACTION_FAMILIES> (0.0));
678 vector<potVec> expot_temp(nLocal_,
679 Vector<RealType, N_INTERACTION_FAMILIES> (0.0));
680
681 // scatter/gather pot_row into the members of my column
682
683 AtomPlanPotRow->scatter(pot_row, pot_temp);
684 AtomPlanPotRow->scatter(expot_row, expot_temp);
685
686 for (int ii = 0; ii < pot_temp.size(); ii++ )
687 pairwisePot += pot_temp[ii];
688
689 for (int ii = 0; ii < expot_temp.size(); ii++ )
690 excludedPot += expot_temp[ii];
691
692 if (storageLayout_ & DataStorage::dslParticlePot) {
693 // This is the pairwise contribution to the particle pot. The
694 // embedding contribution is added in each of the low level
695 // non-bonded routines. In single processor, this is done in
696 // unpackInteractionData, not in collectData.
697 for (int ii = 0; ii < N_INTERACTION_FAMILIES; ii++) {
698 for (int i = 0; i < nLocal_; i++) {
699 // factor of two is because the total potential terms are divided
700 // by 2 in parallel due to row/ column scatter
701 snap_->atomData.particlePot[i] += 2.0 * pot_temp[i](ii);
702 }
703 }
704 }
705
706 fill(pot_temp.begin(), pot_temp.end(),
707 Vector<RealType, N_INTERACTION_FAMILIES> (0.0));
708 fill(expot_temp.begin(), expot_temp.end(),
709 Vector<RealType, N_INTERACTION_FAMILIES> (0.0));
710
711 AtomPlanPotColumn->scatter(pot_col, pot_temp);
712 AtomPlanPotColumn->scatter(expot_col, expot_temp);
713
714 for (int ii = 0; ii < pot_temp.size(); ii++ )
715 pairwisePot += pot_temp[ii];
716
717 for (int ii = 0; ii < expot_temp.size(); ii++ )
718 excludedPot += expot_temp[ii];
719
720 if (storageLayout_ & DataStorage::dslParticlePot) {
721 // This is the pairwise contribution to the particle pot. The
722 // embedding contribution is added in each of the low level
723 // non-bonded routines. In single processor, this is done in
724 // unpackInteractionData, not in collectData.
725 for (int ii = 0; ii < N_INTERACTION_FAMILIES; ii++) {
726 for (int i = 0; i < nLocal_; i++) {
727 // factor of two is because the total potential terms are divided
728 // by 2 in parallel due to row/ column scatter
729 snap_->atomData.particlePot[i] += 2.0 * pot_temp[i](ii);
730 }
731 }
732 }
733
734 if (storageLayout_ & DataStorage::dslParticlePot) {
735 int npp = snap_->atomData.particlePot.size();
736 vector<RealType> ppot_temp(npp, 0.0);
737
738 // This is the direct or embedding contribution to the particle
739 // pot.
740
741 AtomPlanRealRow->scatter(atomRowData.particlePot, ppot_temp);
742 for (int i = 0; i < npp; i++) {
743 snap_->atomData.particlePot[i] += ppot_temp[i];
744 }
745
746 fill(ppot_temp.begin(), ppot_temp.end(), 0.0);
747
748 AtomPlanRealColumn->scatter(atomColData.particlePot, ppot_temp);
749 for (int i = 0; i < npp; i++) {
750 snap_->atomData.particlePot[i] += ppot_temp[i];
751 }
752 }
753
754 for (int ii = 0; ii < N_INTERACTION_FAMILIES; ii++) {
755 RealType ploc1 = pairwisePot[ii];
756 RealType ploc2 = 0.0;
757 MPI_Allreduce(&ploc1, &ploc2, 1, MPI_REALTYPE, MPI_SUM, MPI_COMM_WORLD);
758 pairwisePot[ii] = ploc2;
759 }
760
761 for (int ii = 0; ii < N_INTERACTION_FAMILIES; ii++) {
762 RealType ploc1 = excludedPot[ii];
763 RealType ploc2 = 0.0;
764 MPI_Allreduce(&ploc1, &ploc2, 1, MPI_REALTYPE, MPI_SUM, MPI_COMM_WORLD);
765 excludedPot[ii] = ploc2;
766 }
767
768 // Here be dragons.
769 MPI_Comm col = colComm.getComm();
770
771 MPI_Allreduce(MPI_IN_PLACE,
772 &snap_->frameData.conductiveHeatFlux[0], 3,
773 MPI_REALTYPE, MPI_SUM, col);
774
775
776 #endif
777
778 }
779
780 /**
781 * Collects information obtained during the post-pair (and embedding
782 * functional) loops onto local data structures.
783 */
784 void ForceMatrixDecomposition::collectSelfData() {
785 snap_ = sman_->getCurrentSnapshot();
786 storageLayout_ = sman_->getStorageLayout();
787
788 #ifdef IS_MPI
789 for (int ii = 0; ii < N_INTERACTION_FAMILIES; ii++) {
790 RealType ploc1 = embeddingPot[ii];
791 RealType ploc2 = 0.0;
792 MPI_Allreduce(&ploc1, &ploc2, 1, MPI_REALTYPE, MPI_SUM, MPI_COMM_WORLD);
793 embeddingPot[ii] = ploc2;
794 }
795 for (int ii = 0; ii < N_INTERACTION_FAMILIES; ii++) {
796 RealType ploc1 = excludedSelfPot[ii];
797 RealType ploc2 = 0.0;
798 MPI_Allreduce(&ploc1, &ploc2, 1, MPI_REALTYPE, MPI_SUM, MPI_COMM_WORLD);
799 excludedSelfPot[ii] = ploc2;
800 }
801 #endif
802
803 }
804
805
806
807 int& ForceMatrixDecomposition::getNAtomsInRow() {
808 #ifdef IS_MPI
809 return nAtomsInRow_;
810 #else
811 return nLocal_;
812 #endif
813 }
814
815 /**
816 * returns the list of atoms belonging to this group.
817 */
818 vector<int>& ForceMatrixDecomposition::getAtomsInGroupRow(int cg1){
819 #ifdef IS_MPI
820 return groupListRow_[cg1];
821 #else
822 return groupList_[cg1];
823 #endif
824 }
825
826 vector<int>& ForceMatrixDecomposition::getAtomsInGroupColumn(int cg2){
827 #ifdef IS_MPI
828 return groupListCol_[cg2];
829 #else
830 return groupList_[cg2];
831 #endif
832 }
833
834 Vector3d ForceMatrixDecomposition::getIntergroupVector(int cg1, int cg2){
835 Vector3d d;
836
837 #ifdef IS_MPI
838 d = cgColData.position[cg2] - cgRowData.position[cg1];
839 #else
840 d = snap_->cgData.position[cg2] - snap_->cgData.position[cg1];
841 #endif
842
843 if (usePeriodicBoundaryConditions_) {
844 snap_->wrapVector(d);
845 }
846 return d;
847 }
848
849 Vector3d& ForceMatrixDecomposition::getGroupVelocityColumn(int cg2){
850 #ifdef IS_MPI
851 return cgColData.velocity[cg2];
852 #else
853 return snap_->cgData.velocity[cg2];
854 #endif
855 }
856
857 Vector3d& ForceMatrixDecomposition::getAtomVelocityColumn(int atom2){
858 #ifdef IS_MPI
859 return atomColData.velocity[atom2];
860 #else
861 return snap_->atomData.velocity[atom2];
862 #endif
863 }
864
865
866 Vector3d ForceMatrixDecomposition::getAtomToGroupVectorRow(int atom1, int cg1){
867
868 Vector3d d;
869
870 #ifdef IS_MPI
871 d = cgRowData.position[cg1] - atomRowData.position[atom1];
872 #else
873 d = snap_->cgData.position[cg1] - snap_->atomData.position[atom1];
874 #endif
875 if (usePeriodicBoundaryConditions_) {
876 snap_->wrapVector(d);
877 }
878 return d;
879 }
880
881 Vector3d ForceMatrixDecomposition::getAtomToGroupVectorColumn(int atom2, int cg2){
882 Vector3d d;
883
884 #ifdef IS_MPI
885 d = cgColData.position[cg2] - atomColData.position[atom2];
886 #else
887 d = snap_->cgData.position[cg2] - snap_->atomData.position[atom2];
888 #endif
889 if (usePeriodicBoundaryConditions_) {
890 snap_->wrapVector(d);
891 }
892 return d;
893 }
894
895 RealType& ForceMatrixDecomposition::getMassFactorRow(int atom1) {
896 #ifdef IS_MPI
897 return massFactorsRow[atom1];
898 #else
899 return massFactors[atom1];
900 #endif
901 }
902
903 RealType& ForceMatrixDecomposition::getMassFactorColumn(int atom2) {
904 #ifdef IS_MPI
905 return massFactorsCol[atom2];
906 #else
907 return massFactors[atom2];
908 #endif
909
910 }
911
912 Vector3d ForceMatrixDecomposition::getInteratomicVector(int atom1, int atom2){
913 Vector3d d;
914
915 #ifdef IS_MPI
916 d = atomColData.position[atom2] - atomRowData.position[atom1];
917 #else
918 d = snap_->atomData.position[atom2] - snap_->atomData.position[atom1];
919 #endif
920 if (usePeriodicBoundaryConditions_) {
921 snap_->wrapVector(d);
922 }
923 return d;
924 }
925
926 vector<int>& ForceMatrixDecomposition::getExcludesForAtom(int atom1) {
927 return excludesForAtom[atom1];
928 }
929
930 /**
931 * We need to exclude some overcounted interactions that result from
932 * the parallel decomposition.
933 */
934 bool ForceMatrixDecomposition::skipAtomPair(int atom1, int atom2, int cg1, int cg2) {
935 int unique_id_1, unique_id_2;
936
937 #ifdef IS_MPI
938 // in MPI, we have to look up the unique IDs for each atom
939 unique_id_1 = AtomRowToGlobal[atom1];
940 unique_id_2 = AtomColToGlobal[atom2];
941 // group1 = cgRowToGlobal[cg1];
942 // group2 = cgColToGlobal[cg2];
943 #else
944 unique_id_1 = AtomLocalToGlobal[atom1];
945 unique_id_2 = AtomLocalToGlobal[atom2];
946 int group1 = cgLocalToGlobal[cg1];
947 int group2 = cgLocalToGlobal[cg2];
948 #endif
949
950 if (unique_id_1 == unique_id_2) return true;
951
952 #ifdef IS_MPI
953 // this prevents us from doing the pair on multiple processors
954 if (unique_id_1 < unique_id_2) {
955 if ((unique_id_1 + unique_id_2) % 2 == 0) return true;
956 } else {
957 if ((unique_id_1 + unique_id_2) % 2 == 1) return true;
958 }
959 #endif
960
961 #ifndef IS_MPI
962 if (group1 == group2) {
963 if (unique_id_1 < unique_id_2) return true;
964 }
965 #endif
966
967 return false;
968 }
969
970 /**
971 * We need to handle the interactions for atoms who are involved in
972 * the same rigid body as well as some short range interactions
973 * (bonds, bends, torsions) differently from other interactions.
974 * We'll still visit the pairwise routines, but with a flag that
975 * tells those routines to exclude the pair from direct long range
976 * interactions. Some indirect interactions (notably reaction
977 * field) must still be handled for these pairs.
978 */
979 bool ForceMatrixDecomposition::excludeAtomPair(int atom1, int atom2) {
980
981 // excludesForAtom was constructed to use row/column indices in the MPI
982 // version, and to use local IDs in the non-MPI version:
983
984 for (vector<int>::iterator i = excludesForAtom[atom1].begin();
985 i != excludesForAtom[atom1].end(); ++i) {
986 if ( (*i) == atom2 ) return true;
987 }
988
989 return false;
990 }
991
992
993 void ForceMatrixDecomposition::addForceToAtomRow(int atom1, Vector3d fg){
994 #ifdef IS_MPI
995 atomRowData.force[atom1] += fg;
996 #else
997 snap_->atomData.force[atom1] += fg;
998 #endif
999 }
1000
1001 void ForceMatrixDecomposition::addForceToAtomColumn(int atom2, Vector3d fg){
1002 #ifdef IS_MPI
1003 atomColData.force[atom2] += fg;
1004 #else
1005 snap_->atomData.force[atom2] += fg;
1006 #endif
1007 }
1008
1009 // filling interaction blocks with pointers
1010 void ForceMatrixDecomposition::fillInteractionData(InteractionData &idat,
1011 int atom1, int atom2,
1012 bool newAtom1) {
1013
1014 idat.excluded = excludeAtomPair(atom1, atom2);
1015
1016 if (newAtom1) {
1017
1018 #ifdef IS_MPI
1019 idat.atid1 = identsRow[atom1];
1020 idat.atid2 = identsCol[atom2];
1021
1022 if (regionsRow[atom1] >= 0 && regionsCol[atom2] >= 0) {
1023 idat.sameRegion = (regionsRow[atom1] == regionsCol[atom2]);
1024 } else {
1025 idat.sameRegion = false;
1026 }
1027
1028 if (storageLayout_ & DataStorage::dslAmat) {
1029 idat.A1 = &(atomRowData.aMat[atom1]);
1030 idat.A2 = &(atomColData.aMat[atom2]);
1031 }
1032
1033 if (storageLayout_ & DataStorage::dslTorque) {
1034 idat.t1 = &(atomRowData.torque[atom1]);
1035 idat.t2 = &(atomColData.torque[atom2]);
1036 }
1037
1038 if (storageLayout_ & DataStorage::dslDipole) {
1039 idat.dipole1 = &(atomRowData.dipole[atom1]);
1040 idat.dipole2 = &(atomColData.dipole[atom2]);
1041 }
1042
1043 if (storageLayout_ & DataStorage::dslQuadrupole) {
1044 idat.quadrupole1 = &(atomRowData.quadrupole[atom1]);
1045 idat.quadrupole2 = &(atomColData.quadrupole[atom2]);
1046 }
1047
1048 if (storageLayout_ & DataStorage::dslDensity) {
1049 idat.rho1 = &(atomRowData.density[atom1]);
1050 idat.rho2 = &(atomColData.density[atom2]);
1051 }
1052
1053 if (storageLayout_ & DataStorage::dslFunctional) {
1054 idat.frho1 = &(atomRowData.functional[atom1]);
1055 idat.frho2 = &(atomColData.functional[atom2]);
1056 }
1057
1058 if (storageLayout_ & DataStorage::dslFunctionalDerivative) {
1059 idat.dfrho1 = &(atomRowData.functionalDerivative[atom1]);
1060 idat.dfrho2 = &(atomColData.functionalDerivative[atom2]);
1061 }
1062
1063 if (storageLayout_ & DataStorage::dslParticlePot) {
1064 idat.particlePot1 = &(atomRowData.particlePot[atom1]);
1065 idat.particlePot2 = &(atomColData.particlePot[atom2]);
1066 }
1067
1068 if (storageLayout_ & DataStorage::dslSkippedCharge) {
1069 idat.skippedCharge1 = &(atomRowData.skippedCharge[atom1]);
1070 idat.skippedCharge2 = &(atomColData.skippedCharge[atom2]);
1071 }
1072
1073 if (storageLayout_ & DataStorage::dslFlucQPosition) {
1074 idat.flucQ1 = &(atomRowData.flucQPos[atom1]);
1075 idat.flucQ2 = &(atomColData.flucQPos[atom2]);
1076 }
1077
1078 #else
1079
1080 idat.atid1 = idents[atom1];
1081 idat.atid2 = idents[atom2];
1082
1083 if (regions[atom1] >= 0 && regions[atom2] >= 0) {
1084 idat.sameRegion = (regions[atom1] == regions[atom2]);
1085 } else {
1086 idat.sameRegion = false;
1087 }
1088
1089 if (storageLayout_ & DataStorage::dslAmat) {
1090 idat.A1 = &(snap_->atomData.aMat[atom1]);
1091 idat.A2 = &(snap_->atomData.aMat[atom2]);
1092 }
1093
1094 if (storageLayout_ & DataStorage::dslTorque) {
1095 idat.t1 = &(snap_->atomData.torque[atom1]);
1096 idat.t2 = &(snap_->atomData.torque[atom2]);
1097 }
1098
1099 if (storageLayout_ & DataStorage::dslDipole) {
1100 idat.dipole1 = &(snap_->atomData.dipole[atom1]);
1101 idat.dipole2 = &(snap_->atomData.dipole[atom2]);
1102 }
1103
1104 if (storageLayout_ & DataStorage::dslQuadrupole) {
1105 idat.quadrupole1 = &(snap_->atomData.quadrupole[atom1]);
1106 idat.quadrupole2 = &(snap_->atomData.quadrupole[atom2]);
1107 }
1108
1109 if (storageLayout_ & DataStorage::dslDensity) {
1110 idat.rho1 = &(snap_->atomData.density[atom1]);
1111 idat.rho2 = &(snap_->atomData.density[atom2]);
1112 }
1113
1114 if (storageLayout_ & DataStorage::dslFunctional) {
1115 idat.frho1 = &(snap_->atomData.functional[atom1]);
1116 idat.frho2 = &(snap_->atomData.functional[atom2]);
1117 }
1118
1119 if (storageLayout_ & DataStorage::dslFunctionalDerivative) {
1120 idat.dfrho1 = &(snap_->atomData.functionalDerivative[atom1]);
1121 idat.dfrho2 = &(snap_->atomData.functionalDerivative[atom2]);
1122 }
1123
1124 if (storageLayout_ & DataStorage::dslParticlePot) {
1125 idat.particlePot1 = &(snap_->atomData.particlePot[atom1]);
1126 idat.particlePot2 = &(snap_->atomData.particlePot[atom2]);
1127 }
1128
1129 if (storageLayout_ & DataStorage::dslSkippedCharge) {
1130 idat.skippedCharge1 = &(snap_->atomData.skippedCharge[atom1]);
1131 idat.skippedCharge2 = &(snap_->atomData.skippedCharge[atom2]);
1132 }
1133
1134 if (storageLayout_ & DataStorage::dslFlucQPosition) {
1135 idat.flucQ1 = &(snap_->atomData.flucQPos[atom1]);
1136 idat.flucQ2 = &(snap_->atomData.flucQPos[atom2]);
1137 }
1138 #endif
1139
1140 } else {
1141 // atom1 is not new, so don't bother updating properties of that atom:
1142 #ifdef IS_MPI
1143 idat.atid2 = identsCol[atom2];
1144
1145 if (regionsRow[atom1] >= 0 && regionsCol[atom2] >= 0) {
1146 idat.sameRegion = (regionsRow[atom1] == regionsCol[atom2]);
1147 } else {
1148 idat.sameRegion = false;
1149 }
1150
1151 if (storageLayout_ & DataStorage::dslAmat) {
1152 idat.A2 = &(atomColData.aMat[atom2]);
1153 }
1154
1155 if (storageLayout_ & DataStorage::dslTorque) {
1156 idat.t2 = &(atomColData.torque[atom2]);
1157 }
1158
1159 if (storageLayout_ & DataStorage::dslDipole) {
1160 idat.dipole2 = &(atomColData.dipole[atom2]);
1161 }
1162
1163 if (storageLayout_ & DataStorage::dslQuadrupole) {
1164 idat.quadrupole2 = &(atomColData.quadrupole[atom2]);
1165 }
1166
1167 if (storageLayout_ & DataStorage::dslDensity) {
1168 idat.rho2 = &(atomColData.density[atom2]);
1169 }
1170
1171 if (storageLayout_ & DataStorage::dslFunctional) {
1172 idat.frho2 = &(atomColData.functional[atom2]);
1173 }
1174
1175 if (storageLayout_ & DataStorage::dslFunctionalDerivative) {
1176 idat.dfrho2 = &(atomColData.functionalDerivative[atom2]);
1177 }
1178
1179 if (storageLayout_ & DataStorage::dslParticlePot) {
1180 idat.particlePot2 = &(atomColData.particlePot[atom2]);
1181 }
1182
1183 if (storageLayout_ & DataStorage::dslSkippedCharge) {
1184 idat.skippedCharge2 = &(atomColData.skippedCharge[atom2]);
1185 }
1186
1187 if (storageLayout_ & DataStorage::dslFlucQPosition) {
1188 idat.flucQ2 = &(atomColData.flucQPos[atom2]);
1189 }
1190
1191 #else
1192 idat.atid2 = idents[atom2];
1193
1194 if (regions[atom1] >= 0 && regions[atom2] >= 0) {
1195 idat.sameRegion = (regions[atom1] == regions[atom2]);
1196 } else {
1197 idat.sameRegion = false;
1198 }
1199
1200 if (storageLayout_ & DataStorage::dslAmat) {
1201 idat.A2 = &(snap_->atomData.aMat[atom2]);
1202 }
1203
1204 if (storageLayout_ & DataStorage::dslTorque) {
1205 idat.t2 = &(snap_->atomData.torque[atom2]);
1206 }
1207
1208 if (storageLayout_ & DataStorage::dslDipole) {
1209 idat.dipole2 = &(snap_->atomData.dipole[atom2]);
1210 }
1211
1212 if (storageLayout_ & DataStorage::dslQuadrupole) {
1213 idat.quadrupole2 = &(snap_->atomData.quadrupole[atom2]);
1214 }
1215
1216 if (storageLayout_ & DataStorage::dslDensity) {
1217 idat.rho2 = &(snap_->atomData.density[atom2]);
1218 }
1219
1220 if (storageLayout_ & DataStorage::dslFunctional) {
1221 idat.frho2 = &(snap_->atomData.functional[atom2]);
1222 }
1223
1224 if (storageLayout_ & DataStorage::dslFunctionalDerivative) {
1225 idat.dfrho2 = &(snap_->atomData.functionalDerivative[atom2]);
1226 }
1227
1228 if (storageLayout_ & DataStorage::dslParticlePot) {
1229 idat.particlePot2 = &(snap_->atomData.particlePot[atom2]);
1230 }
1231
1232 if (storageLayout_ & DataStorage::dslSkippedCharge) {
1233 idat.skippedCharge2 = &(snap_->atomData.skippedCharge[atom2]);
1234 }
1235
1236 if (storageLayout_ & DataStorage::dslFlucQPosition) {
1237 idat.flucQ2 = &(snap_->atomData.flucQPos[atom2]);
1238 }
1239
1240 #endif
1241 }
1242 }
1243
1244 void ForceMatrixDecomposition::unpackInteractionData(InteractionData &idat,
1245 int atom1, int atom2) {
1246 #ifdef IS_MPI
1247 pot_row[atom1] += RealType(0.5) * *(idat.pot);
1248 pot_col[atom2] += RealType(0.5) * *(idat.pot);
1249 expot_row[atom1] += RealType(0.5) * *(idat.excludedPot);
1250 expot_col[atom2] += RealType(0.5) * *(idat.excludedPot);
1251
1252 atomRowData.force[atom1] += *(idat.f1);
1253 atomColData.force[atom2] -= *(idat.f1);
1254
1255 if (storageLayout_ & DataStorage::dslFlucQForce) {
1256 atomRowData.flucQFrc[atom1] -= *(idat.dVdFQ1);
1257 atomColData.flucQFrc[atom2] -= *(idat.dVdFQ2);
1258 }
1259
1260 if (storageLayout_ & DataStorage::dslElectricField) {
1261 atomRowData.electricField[atom1] += *(idat.eField1);
1262 atomColData.electricField[atom2] += *(idat.eField2);
1263 }
1264
1265 if (storageLayout_ & DataStorage::dslSitePotential) {
1266 atomRowData.sitePotential[atom1] += *(idat.sPot1);
1267 atomColData.sitePotential[atom2] += *(idat.sPot2);
1268 }
1269
1270 #else
1271 pairwisePot += *(idat.pot);
1272 excludedPot += *(idat.excludedPot);
1273
1274 snap_->atomData.force[atom1] += *(idat.f1);
1275 snap_->atomData.force[atom2] -= *(idat.f1);
1276
1277 if (idat.doParticlePot) {
1278 // This is the pairwise contribution to the particle pot. The
1279 // embedding contribution is added in each of the low level
1280 // non-bonded routines. In parallel, this calculation is done
1281 // in collectData, not in unpackInteractionData.
1282 snap_->atomData.particlePot[atom1] += *(idat.vpair) * *(idat.sw);
1283 snap_->atomData.particlePot[atom2] += *(idat.vpair) * *(idat.sw);
1284 }
1285
1286 if (storageLayout_ & DataStorage::dslFlucQForce) {
1287 snap_->atomData.flucQFrc[atom1] -= *(idat.dVdFQ1);
1288 snap_->atomData.flucQFrc[atom2] -= *(idat.dVdFQ2);
1289 }
1290
1291 if (storageLayout_ & DataStorage::dslElectricField) {
1292 snap_->atomData.electricField[atom1] += *(idat.eField1);
1293 snap_->atomData.electricField[atom2] += *(idat.eField2);
1294 }
1295
1296 if (storageLayout_ & DataStorage::dslSitePotential) {
1297 snap_->atomData.sitePotential[atom1] += *(idat.sPot1);
1298 snap_->atomData.sitePotential[atom2] += *(idat.sPot2);
1299 }
1300
1301 #endif
1302
1303 }
1304
1305 /*
1306 * buildNeighborList
1307 *
1308 * Constructs the Verlet neighbor list for a force-matrix
1309 * decomposition. In this case, each processor is responsible for
1310 * row-site interactions with column-sites.
1311 *
1312 * neighborList is returned as a packed array of neighboring
1313 * column-ordered CutoffGroups. The starting position in
1314 * neighborList for each row-ordered CutoffGroup is given by the
1315 * returned vector point.
1316 */
1317 void ForceMatrixDecomposition::buildNeighborList(vector<int>& neighborList,
1318 vector<int>& point) {
1319 neighborList.clear();
1320 point.clear();
1321 int len = 0;
1322
1323 bool doAllPairs = false;
1324
1325 Snapshot* snap_ = sman_->getCurrentSnapshot();
1326 Mat3x3d box;
1327 Mat3x3d invBox;
1328
1329 Vector3d rs, scaled, dr;
1330 Vector3i whichCell;
1331 int cellIndex;
1332
1333 #ifdef IS_MPI
1334 cellListRow_.clear();
1335 cellListCol_.clear();
1336 point.resize(nGroupsInRow_+1);
1337 #else
1338 cellList_.clear();
1339 point.resize(nGroups_+1);
1340 #endif
1341
1342 if (!usePeriodicBoundaryConditions_) {
1343 box = snap_->getBoundingBox();
1344 invBox = snap_->getInvBoundingBox();
1345 } else {
1346 box = snap_->getHmat();
1347 invBox = snap_->getInvHmat();
1348 }
1349
1350 Vector3d A = box.getColumn(0);
1351 Vector3d B = box.getColumn(1);
1352 Vector3d C = box.getColumn(2);
1353
1354 // Required for triclinic cells
1355 Vector3d AxB = cross(A, B);
1356 Vector3d BxC = cross(B, C);
1357 Vector3d CxA = cross(C, A);
1358
1359 // unit vectors perpendicular to the faces of the triclinic cell:
1360 AxB.normalize();
1361 BxC.normalize();
1362 CxA.normalize();
1363
1364 // A set of perpendicular lengths in triclinic cells:
1365 RealType Wa = abs(dot(A, BxC));
1366 RealType Wb = abs(dot(B, CxA));
1367 RealType Wc = abs(dot(C, AxB));
1368
1369 nCells_.x() = int( Wa / rList_ );
1370 nCells_.y() = int( Wb / rList_ );
1371 nCells_.z() = int( Wc / rList_ );
1372
1373 // handle small boxes where the cell offsets can end up repeating cells
1374 if (nCells_.x() < 3) doAllPairs = true;
1375 if (nCells_.y() < 3) doAllPairs = true;
1376 if (nCells_.z() < 3) doAllPairs = true;
1377
1378 int nCtot = nCells_.x() * nCells_.y() * nCells_.z();
1379
1380 #ifdef IS_MPI
1381 cellListRow_.resize(nCtot);
1382 cellListCol_.resize(nCtot);
1383 #else
1384 cellList_.resize(nCtot);
1385 #endif
1386
1387 if (!doAllPairs) {
1388
1389 #ifdef IS_MPI
1390
1391 for (int i = 0; i < nGroupsInRow_; i++) {
1392 rs = cgRowData.position[i];
1393
1394 // scaled positions relative to the box vectors
1395 scaled = invBox * rs;
1396
1397 // wrap the vector back into the unit box by subtracting integer box
1398 // numbers
1399 for (int j = 0; j < 3; j++) {
1400 scaled[j] -= roundMe(scaled[j]);
1401 scaled[j] += 0.5;
1402 // Handle the special case when an object is exactly on the
1403 // boundary (a scaled coordinate of 1.0 is the same as
1404 // scaled coordinate of 0.0)
1405 if (scaled[j] >= 1.0) scaled[j] -= 1.0;
1406 }
1407
1408 // find xyz-indices of cell that cutoffGroup is in.
1409 whichCell.x() = nCells_.x() * scaled.x();
1410 whichCell.y() = nCells_.y() * scaled.y();
1411 whichCell.z() = nCells_.z() * scaled.z();
1412
1413 // find single index of this cell:
1414 cellIndex = Vlinear(whichCell, nCells_);
1415
1416 // add this cutoff group to the list of groups in this cell;
1417 cellListRow_[cellIndex].push_back(i);
1418 }
1419 for (int i = 0; i < nGroupsInCol_; i++) {
1420 rs = cgColData.position[i];
1421
1422 // scaled positions relative to the box vectors
1423 scaled = invBox * rs;
1424
1425 // wrap the vector back into the unit box by subtracting integer box
1426 // numbers
1427 for (int j = 0; j < 3; j++) {
1428 scaled[j] -= roundMe(scaled[j]);
1429 scaled[j] += 0.5;
1430 // Handle the special case when an object is exactly on the
1431 // boundary (a scaled coordinate of 1.0 is the same as
1432 // scaled coordinate of 0.0)
1433 if (scaled[j] >= 1.0) scaled[j] -= 1.0;
1434 }
1435
1436 // find xyz-indices of cell that cutoffGroup is in.
1437 whichCell.x() = nCells_.x() * scaled.x();
1438 whichCell.y() = nCells_.y() * scaled.y();
1439 whichCell.z() = nCells_.z() * scaled.z();
1440
1441 // find single index of this cell:
1442 cellIndex = Vlinear(whichCell, nCells_);
1443
1444 // add this cutoff group to the list of groups in this cell;
1445 cellListCol_[cellIndex].push_back(i);
1446 }
1447
1448 #else
1449 for (int i = 0; i < nGroups_; i++) {
1450 rs = snap_->cgData.position[i];
1451
1452 // scaled positions relative to the box vectors
1453 scaled = invBox * rs;
1454
1455 // wrap the vector back into the unit box by subtracting integer box
1456 // numbers
1457 for (int j = 0; j < 3; j++) {
1458 scaled[j] -= roundMe(scaled[j]);
1459 scaled[j] += 0.5;
1460 // Handle the special case when an object is exactly on the
1461 // boundary (a scaled coordinate of 1.0 is the same as
1462 // scaled coordinate of 0.0)
1463 if (scaled[j] >= 1.0) scaled[j] -= 1.0;
1464 }
1465
1466 // find xyz-indices of cell that cutoffGroup is in.
1467 whichCell.x() = int(nCells_.x() * scaled.x());
1468 whichCell.y() = int(nCells_.y() * scaled.y());
1469 whichCell.z() = int(nCells_.z() * scaled.z());
1470
1471 // find single index of this cell:
1472 cellIndex = Vlinear(whichCell, nCells_);
1473
1474 // add this cutoff group to the list of groups in this cell;
1475 cellList_[cellIndex].push_back(i);
1476 }
1477
1478 #endif
1479
1480 #ifdef IS_MPI
1481 for (int j1 = 0; j1 < nGroupsInRow_; j1++) {
1482 rs = cgRowData.position[j1];
1483 #else
1484
1485 for (int j1 = 0; j1 < nGroups_; j1++) {
1486 rs = snap_->cgData.position[j1];
1487 #endif
1488 point[j1] = len;
1489
1490 // scaled positions relative to the box vectors
1491 scaled = invBox * rs;
1492
1493 // wrap the vector back into the unit box by subtracting integer box
1494 // numbers
1495 for (int j = 0; j < 3; j++) {
1496 scaled[j] -= roundMe(scaled[j]);
1497 scaled[j] += 0.5;
1498 // Handle the special case when an object is exactly on the
1499 // boundary (a scaled coordinate of 1.0 is the same as
1500 // scaled coordinate of 0.0)
1501 if (scaled[j] >= 1.0) scaled[j] -= 1.0;
1502 }
1503
1504 // find xyz-indices of cell that cutoffGroup is in.
1505 whichCell.x() = nCells_.x() * scaled.x();
1506 whichCell.y() = nCells_.y() * scaled.y();
1507 whichCell.z() = nCells_.z() * scaled.z();
1508
1509 // find single index of this cell:
1510 int m1 = Vlinear(whichCell, nCells_);
1511
1512 for (vector<Vector3i>::iterator os = cellOffsets_.begin();
1513 os != cellOffsets_.end(); ++os) {
1514
1515 Vector3i m2v = whichCell + (*os);
1516
1517 if (m2v.x() >= nCells_.x()) {
1518 m2v.x() = 0;
1519 } else if (m2v.x() < 0) {
1520 m2v.x() = nCells_.x() - 1;
1521 }
1522
1523 if (m2v.y() >= nCells_.y()) {
1524 m2v.y() = 0;
1525 } else if (m2v.y() < 0) {
1526 m2v.y() = nCells_.y() - 1;
1527 }
1528
1529 if (m2v.z() >= nCells_.z()) {
1530 m2v.z() = 0;
1531 } else if (m2v.z() < 0) {
1532 m2v.z() = nCells_.z() - 1;
1533 }
1534 int m2 = Vlinear (m2v, nCells_);
1535 #ifdef IS_MPI
1536 for (vector<int>::iterator j2 = cellListCol_[m2].begin();
1537 j2 != cellListCol_[m2].end(); ++j2) {
1538
1539 // In parallel, we need to visit *all* pairs of row
1540 // & column indicies and will divide labor in the
1541 // force evaluation later.
1542 dr = cgColData.position[(*j2)] - rs;
1543 if (usePeriodicBoundaryConditions_) {
1544 snap_->wrapVector(dr);
1545 }
1546 if (dr.lengthSquare() < rListSq_) {
1547 neighborList.push_back( (*j2) );
1548 ++len;
1549 }
1550 }
1551 #else
1552 for (vector<int>::iterator j2 = cellList_[m2].begin();
1553 j2 != cellList_[m2].end(); ++j2) {
1554
1555 // Always do this if we're in different cells or if
1556 // we're in the same cell and the global index of
1557 // the j2 cutoff group is greater than or equal to
1558 // the j1 cutoff group. Note that Rappaport's code
1559 // has a "less than" conditional here, but that
1560 // deals with atom-by-atom computation. OpenMD
1561 // allows atoms within a single cutoff group to
1562 // interact with each other.
1563
1564 if ( (*j2) >= j1 ) {
1565
1566 dr = snap_->cgData.position[(*j2)] - rs;
1567 if (usePeriodicBoundaryConditions_) {
1568 snap_->wrapVector(dr);
1569 }
1570 if ( dr.lengthSquare() < rListSq_) {
1571 neighborList.push_back( (*j2) );
1572 ++len;
1573 }
1574 }
1575 }
1576 #endif
1577 }
1578 }
1579 } else {
1580 // branch to do all cutoff group pairs
1581 #ifdef IS_MPI
1582 for (int j1 = 0; j1 < nGroupsInRow_; j1++) {
1583 point[j1] = len;
1584 rs = cgRowData.position[j1];
1585 for (int j2 = 0; j2 < nGroupsInCol_; j2++) {
1586 dr = cgColData.position[j2] - rs;
1587 if (usePeriodicBoundaryConditions_) {
1588 snap_->wrapVector(dr);
1589 }
1590 if (dr.lengthSquare() < rListSq_) {
1591 neighborList.push_back( j2 );
1592 ++len;
1593 }
1594 }
1595 }
1596 #else
1597 // include all groups here.
1598 for (int j1 = 0; j1 < nGroups_; j1++) {
1599 point[j1] = len;
1600 rs = snap_->cgData.position[j1];
1601 // include self group interactions j2 == j1
1602 for (int j2 = j1; j2 < nGroups_; j2++) {
1603 dr = snap_->cgData.position[j2] - rs;
1604 if (usePeriodicBoundaryConditions_) {
1605 snap_->wrapVector(dr);
1606 }
1607 if (dr.lengthSquare() < rListSq_) {
1608 neighborList.push_back( j2 );
1609 ++len;
1610 }
1611 }
1612 }
1613 #endif
1614 }
1615
1616 #ifdef IS_MPI
1617 point[nGroupsInRow_] = len;
1618 #else
1619 point[nGroups_] = len;
1620 #endif
1621
1622 // save the local cutoff group positions for the check that is
1623 // done on each loop:
1624 saved_CG_positions_.clear();
1625 saved_CG_positions_.reserve(nGroups_);
1626 for (int i = 0; i < nGroups_; i++)
1627 saved_CG_positions_.push_back(snap_->cgData.position[i]);
1628 }
1629 } //end namespace OpenMD