1 |
|
#ifdef IS_MPI |
2 |
|
#include <iostream> |
3 |
< |
#include <cstdlib> |
4 |
< |
#include <cstring> |
5 |
< |
#include <cmath> |
3 |
> |
#include <stdlib.h> |
4 |
> |
#include <string.h> |
5 |
> |
#include <math.h> |
6 |
|
#include <mpi.h> |
7 |
|
|
8 |
|
#include "mpiSimulation.hpp" |
10 |
|
#include "fortranWrappers.hpp" |
11 |
|
#include "randomSPRNG.hpp" |
12 |
|
|
13 |
– |
#define BASE_SEED 123456789 |
14 |
– |
|
13 |
|
mpiSimulation* mpiSim; |
14 |
|
|
15 |
|
mpiSimulation::mpiSimulation(SimInfo* the_entryPlug) |
16 |
|
{ |
17 |
|
entryPlug = the_entryPlug; |
18 |
< |
mpiPlug = new mpiSimData; |
18 |
> |
parallelData = new mpiSimData; |
19 |
|
|
20 |
< |
MPI_Comm_size(MPI_COMM_WORLD, &(mpiPlug->numberProcessors) ); |
21 |
< |
mpiPlug->myNode = worldRank; |
20 |
> |
MPI_Comm_size(MPI_COMM_WORLD, &(parallelData->nProcessors) ); |
21 |
> |
parallelData->myNode = worldRank; |
22 |
|
|
23 |
|
MolToProcMap = new int[entryPlug->n_mol]; |
24 |
|
MolComponentType = new int[entryPlug->n_mol]; |
25 |
|
AtomToProcMap = new int[entryPlug->n_atoms]; |
26 |
+ |
GroupToProcMap = new int[entryPlug->ngroup]; |
27 |
|
|
28 |
|
mpiSim = this; |
29 |
|
wrapMeSimParallel( this ); |
35 |
|
delete[] MolToProcMap; |
36 |
|
delete[] MolComponentType; |
37 |
|
delete[] AtomToProcMap; |
38 |
+ |
delete[] GroupToProcMap; |
39 |
|
|
40 |
< |
delete mpiPlug; |
40 |
> |
delete parallelData; |
41 |
|
// perhaps we should let fortran know the party is over. |
42 |
|
|
43 |
|
} |
44 |
|
|
45 |
< |
int* mpiSimulation::divideLabor( void ){ |
45 |
> |
void mpiSimulation::divideLabor( ){ |
46 |
|
|
47 |
– |
int* globalIndex; |
48 |
– |
|
47 |
|
int nComponents; |
48 |
|
MoleculeStamp** compStamps; |
49 |
|
randomSPRNG *myRandom; |
50 |
|
int* componentsNmol; |
51 |
|
int* AtomsPerProc; |
52 |
+ |
int* GroupsPerProc; |
53 |
|
|
54 |
|
double numerator; |
55 |
|
double denominator; |
56 |
|
double precast; |
57 |
|
double x, y, a; |
58 |
|
int old_atoms, add_atoms, new_atoms; |
59 |
+ |
int old_groups, add_groups, new_groups; |
60 |
|
|
61 |
|
int nTarget; |
62 |
< |
int molIndex, atomIndex, compIndex, compStart; |
62 |
> |
int molIndex, atomIndex, groupIndex; |
63 |
|
int done; |
64 |
< |
int nLocal, molLocal; |
65 |
< |
int i, j, loops, which_proc, nmol_local, natoms_local; |
66 |
< |
int nmol_global, natoms_global; |
67 |
< |
int local_index, index; |
68 |
< |
int smallDiff, bigDiff; |
64 |
> |
int i, j, loops, which_proc; |
65 |
> |
int nmol_global, nmol_local; |
66 |
> |
int ngroups_global, ngroups_local; |
67 |
> |
int natoms_global, natoms_local; |
68 |
> |
int ncutoff_groups, nAtomsInGroups; |
69 |
> |
int local_index; |
70 |
|
int baseSeed = entryPlug->getSeed(); |
71 |
+ |
CutoffGroupStamp* cg; |
72 |
|
|
71 |
– |
int testSum; |
72 |
– |
|
73 |
|
nComponents = entryPlug->nComponents; |
74 |
|
compStamps = entryPlug->compStamps; |
75 |
|
componentsNmol = entryPlug->componentsNmol; |
76 |
< |
AtomsPerProc = new int[mpiPlug->numberProcessors]; |
76 |
> |
AtomsPerProc = new int[parallelData->nProcessors]; |
77 |
> |
GroupsPerProc = new int[parallelData->nProcessors]; |
78 |
|
|
79 |
< |
mpiPlug->nAtomsGlobal = entryPlug->n_atoms; |
80 |
< |
mpiPlug->nBondsGlobal = entryPlug->n_bonds; |
81 |
< |
mpiPlug->nBendsGlobal = entryPlug->n_bends; |
82 |
< |
mpiPlug->nTorsionsGlobal = entryPlug->n_torsions; |
83 |
< |
mpiPlug->nSRIGlobal = entryPlug->n_SRI; |
84 |
< |
mpiPlug->nMolGlobal = entryPlug->n_mol; |
79 |
> |
parallelData->nAtomsGlobal = entryPlug->n_atoms; |
80 |
> |
parallelData->nBondsGlobal = entryPlug->n_bonds; |
81 |
> |
parallelData->nBendsGlobal = entryPlug->n_bends; |
82 |
> |
parallelData->nTorsionsGlobal = entryPlug->n_torsions; |
83 |
> |
parallelData->nSRIGlobal = entryPlug->n_SRI; |
84 |
> |
parallelData->nGroupsGlobal = entryPlug->ngroup; |
85 |
> |
parallelData->nMolGlobal = entryPlug->n_mol; |
86 |
|
|
85 |
– |
|
87 |
|
myRandom = new randomSPRNG( baseSeed ); |
88 |
|
|
89 |
< |
a = 3.0 * (double)mpiPlug->nMolGlobal / (double)mpiPlug->nAtomsGlobal; |
89 |
> |
a = 3.0 * (double)parallelData->nMolGlobal / (double)parallelData->nAtomsGlobal; |
90 |
|
|
91 |
|
// Initialize things that we'll send out later: |
92 |
< |
for (i = 0; i < mpiPlug->numberProcessors; i++ ) { |
92 |
> |
for (i = 0; i < parallelData->nProcessors; i++ ) { |
93 |
|
AtomsPerProc[i] = 0; |
94 |
+ |
GroupsPerProc[i] = 0; |
95 |
|
} |
96 |
< |
for (i = 0; i < mpiPlug->nMolGlobal; i++ ) { |
96 |
> |
for (i = 0; i < parallelData->nMolGlobal; i++ ) { |
97 |
|
// default to an error condition: |
98 |
|
MolToProcMap[i] = -1; |
99 |
|
MolComponentType[i] = -1; |
100 |
|
} |
101 |
< |
for (i = 0; i < mpiPlug->nAtomsGlobal; i++ ) { |
101 |
> |
for (i = 0; i < parallelData->nAtomsGlobal; i++ ) { |
102 |
|
// default to an error condition: |
103 |
|
AtomToProcMap[i] = -1; |
104 |
|
} |
105 |
+ |
for (i = 0; i < parallelData->nGroupsGlobal; i++ ) { |
106 |
+ |
// default to an error condition: |
107 |
+ |
GroupToProcMap[i] = -1; |
108 |
+ |
} |
109 |
|
|
110 |
< |
if (mpiPlug->myNode == 0) { |
110 |
> |
if (parallelData->myNode == 0) { |
111 |
|
numerator = (double) entryPlug->n_atoms; |
112 |
< |
denominator = (double) mpiPlug->numberProcessors; |
112 |
> |
denominator = (double) parallelData->nProcessors; |
113 |
|
precast = numerator / denominator; |
114 |
|
nTarget = (int)( precast + 0.5 ); |
115 |
|
|
123 |
|
} |
124 |
|
|
125 |
|
atomIndex = 0; |
126 |
+ |
groupIndex = 0; |
127 |
|
|
128 |
|
for (i = 0; i < molIndex; i++ ) { |
129 |
|
|
135 |
|
|
136 |
|
// Pick a processor at random |
137 |
|
|
138 |
< |
which_proc = (int) (myRandom->getRandom() * mpiPlug->numberProcessors); |
138 |
> |
which_proc = (int) (myRandom->getRandom() * parallelData->nProcessors); |
139 |
|
|
140 |
|
// How many atoms does this processor have? |
141 |
|
|
143 |
|
add_atoms = compStamps[MolComponentType[i]]->getNAtoms(); |
144 |
|
new_atoms = old_atoms + add_atoms; |
145 |
|
|
146 |
+ |
old_groups = GroupsPerProc[which_proc]; |
147 |
+ |
ncutoff_groups = compStamps[MolComponentType[i]]->getNCutoffGroups(); |
148 |
+ |
nAtomsInGroups = 0; |
149 |
+ |
for (j = 0; j < ncutoff_groups; j++) { |
150 |
+ |
cg = compStamps[MolComponentType[i]]->getCutoffGroup(j); |
151 |
+ |
nAtomsInGroups += cg->getNMembers(); |
152 |
+ |
} |
153 |
+ |
add_groups = add_atoms - nAtomsInGroups + ncutoff_groups; |
154 |
+ |
new_groups = old_groups + add_groups; |
155 |
+ |
|
156 |
|
// If we've been through this loop too many times, we need |
157 |
|
// to just give up and assign the molecule to this processor |
158 |
|
// and be done with it. |
172 |
|
AtomToProcMap[atomIndex] = which_proc; |
173 |
|
atomIndex++; |
174 |
|
} |
175 |
+ |
GroupsPerProc[which_proc] += add_groups; |
176 |
+ |
for (j=0; j < add_groups; j++) { |
177 |
+ |
GroupToProcMap[groupIndex] = which_proc; |
178 |
+ |
groupIndex++; |
179 |
+ |
} |
180 |
|
done = 1; |
181 |
|
continue; |
182 |
|
} |
191 |
|
AtomToProcMap[atomIndex] = which_proc; |
192 |
|
atomIndex++; |
193 |
|
} |
194 |
+ |
GroupsPerProc[which_proc] += add_groups; |
195 |
+ |
for (j=0; j < add_groups; j++) { |
196 |
+ |
GroupToProcMap[groupIndex] = which_proc; |
197 |
+ |
groupIndex++; |
198 |
+ |
} |
199 |
|
done = 1; |
200 |
|
continue; |
201 |
|
} |
219 |
|
AtomToProcMap[atomIndex] = which_proc; |
220 |
|
atomIndex++; |
221 |
|
} |
222 |
+ |
GroupsPerProc[which_proc] += add_groups; |
223 |
+ |
for (j=0; j < add_groups; j++) { |
224 |
+ |
GroupToProcMap[groupIndex] = which_proc; |
225 |
+ |
groupIndex++; |
226 |
+ |
} |
227 |
|
done = 1; |
228 |
|
continue; |
229 |
|
} else { |
233 |
|
} |
234 |
|
} |
235 |
|
|
236 |
+ |
|
237 |
|
// Spray out this nonsense to all other processors: |
238 |
|
|
239 |
< |
MPI_Bcast(MolToProcMap, mpiPlug->nMolGlobal, |
239 |
> |
//std::cerr << "node 0 mol2proc = \n"; |
240 |
> |
//for (i = 0; i < parallelData->nMolGlobal; i++) |
241 |
> |
// std::cerr << i << "\t" << MolToProcMap[i] << "\n"; |
242 |
> |
|
243 |
> |
MPI_Bcast(MolToProcMap, parallelData->nMolGlobal, |
244 |
|
MPI_INT, 0, MPI_COMM_WORLD); |
245 |
|
|
246 |
< |
MPI_Bcast(AtomToProcMap, mpiPlug->nAtomsGlobal, |
246 |
> |
MPI_Bcast(AtomToProcMap, parallelData->nAtomsGlobal, |
247 |
|
MPI_INT, 0, MPI_COMM_WORLD); |
248 |
|
|
249 |
< |
MPI_Bcast(MolComponentType, mpiPlug->nMolGlobal, |
249 |
> |
MPI_Bcast(GroupToProcMap, parallelData->nGroupsGlobal, |
250 |
|
MPI_INT, 0, MPI_COMM_WORLD); |
251 |
|
|
252 |
< |
MPI_Bcast(AtomsPerProc, mpiPlug->numberProcessors, |
252 |
> |
MPI_Bcast(MolComponentType, parallelData->nMolGlobal, |
253 |
> |
MPI_INT, 0, MPI_COMM_WORLD); |
254 |
> |
|
255 |
> |
MPI_Bcast(AtomsPerProc, parallelData->nProcessors, |
256 |
|
MPI_INT, 0, MPI_COMM_WORLD); |
257 |
+ |
|
258 |
+ |
MPI_Bcast(GroupsPerProc, parallelData->nProcessors, |
259 |
+ |
MPI_INT, 0, MPI_COMM_WORLD); |
260 |
|
} else { |
261 |
|
|
262 |
|
// Listen to your marching orders from processor 0: |
263 |
|
|
264 |
< |
MPI_Bcast(MolToProcMap, mpiPlug->nMolGlobal, |
264 |
> |
MPI_Bcast(MolToProcMap, parallelData->nMolGlobal, |
265 |
|
MPI_INT, 0, MPI_COMM_WORLD); |
266 |
|
|
267 |
< |
MPI_Bcast(AtomToProcMap, mpiPlug->nAtomsGlobal, |
267 |
> |
MPI_Bcast(AtomToProcMap, parallelData->nAtomsGlobal, |
268 |
|
MPI_INT, 0, MPI_COMM_WORLD); |
269 |
|
|
270 |
< |
MPI_Bcast(MolComponentType, mpiPlug->nMolGlobal, |
270 |
> |
MPI_Bcast(GroupToProcMap, parallelData->nGroupsGlobal, |
271 |
|
MPI_INT, 0, MPI_COMM_WORLD); |
272 |
+ |
|
273 |
+ |
MPI_Bcast(MolComponentType, parallelData->nMolGlobal, |
274 |
+ |
MPI_INT, 0, MPI_COMM_WORLD); |
275 |
|
|
276 |
< |
MPI_Bcast(AtomsPerProc, mpiPlug->numberProcessors, |
276 |
> |
MPI_Bcast(AtomsPerProc, parallelData->nProcessors, |
277 |
|
MPI_INT, 0, MPI_COMM_WORLD); |
278 |
|
|
279 |
+ |
MPI_Bcast(GroupsPerProc, parallelData->nProcessors, |
280 |
+ |
MPI_INT, 0, MPI_COMM_WORLD); |
281 |
|
|
282 |
+ |
|
283 |
|
} |
284 |
|
|
236 |
– |
|
285 |
|
// Let's all check for sanity: |
286 |
|
|
287 |
|
nmol_local = 0; |
288 |
< |
for (i = 0 ; i < mpiPlug->nMolGlobal; i++ ) { |
289 |
< |
if (MolToProcMap[i] == mpiPlug->myNode) { |
288 |
> |
for (i = 0 ; i < parallelData->nMolGlobal; i++ ) { |
289 |
> |
if (MolToProcMap[i] == parallelData->myNode) { |
290 |
|
nmol_local++; |
291 |
|
} |
292 |
|
} |
293 |
|
|
294 |
|
natoms_local = 0; |
295 |
< |
for (i = 0; i < mpiPlug->nAtomsGlobal; i++) { |
296 |
< |
if (AtomToProcMap[i] == mpiPlug->myNode) { |
295 |
> |
for (i = 0; i < parallelData->nAtomsGlobal; i++) { |
296 |
> |
if (AtomToProcMap[i] == parallelData->myNode) { |
297 |
|
natoms_local++; |
298 |
|
} |
299 |
|
} |
300 |
|
|
301 |
+ |
ngroups_local = 0; |
302 |
+ |
for (i = 0; i < parallelData->nGroupsGlobal; i++) { |
303 |
+ |
if (GroupToProcMap[i] == parallelData->myNode) { |
304 |
+ |
ngroups_local++; |
305 |
+ |
} |
306 |
+ |
} |
307 |
+ |
|
308 |
|
MPI_Allreduce(&nmol_local,&nmol_global,1,MPI_INT,MPI_SUM, |
309 |
|
MPI_COMM_WORLD); |
310 |
+ |
|
311 |
|
MPI_Allreduce(&natoms_local,&natoms_global,1,MPI_INT, |
312 |
|
MPI_SUM, MPI_COMM_WORLD); |
313 |
+ |
|
314 |
+ |
MPI_Allreduce(&ngroups_local,&ngroups_global,1,MPI_INT, |
315 |
+ |
MPI_SUM, MPI_COMM_WORLD); |
316 |
|
|
317 |
|
if( nmol_global != entryPlug->n_mol ){ |
318 |
|
sprintf( painCave.errMsg, |
332 |
|
simError(); |
333 |
|
} |
334 |
|
|
335 |
+ |
if( ngroups_global != entryPlug->ngroup ){ |
336 |
+ |
sprintf( painCave.errMsg, |
337 |
+ |
"The sum of all ngroups_local, %d, did not equal the " |
338 |
+ |
"total number of cutoffGroups, %d.\n", |
339 |
+ |
ngroups_global, entryPlug->ngroup ); |
340 |
+ |
painCave.isFatal = 1; |
341 |
+ |
simError(); |
342 |
+ |
} |
343 |
+ |
|
344 |
|
sprintf( checkPointMsg, |
345 |
|
"Successfully divided the molecules among the processors.\n" ); |
346 |
|
MPIcheckPoint(); |
347 |
|
|
348 |
< |
mpiPlug->myNMol = nmol_local; |
349 |
< |
mpiPlug->myNlocal = natoms_local; |
348 |
> |
parallelData->nMolLocal = nmol_local; |
349 |
> |
parallelData->nAtomsLocal = natoms_local; |
350 |
> |
parallelData->nGroupsLocal = ngroups_local; |
351 |
|
|
352 |
< |
globalIndex = new int[mpiPlug->myNlocal]; |
352 |
> |
globalAtomIndex.resize(parallelData->nAtomsLocal); |
353 |
> |
globalToLocalAtom.resize(parallelData->nAtomsGlobal); |
354 |
|
local_index = 0; |
355 |
< |
for (i = 0; i < mpiPlug->nAtomsGlobal; i++) { |
356 |
< |
if (AtomToProcMap[i] == mpiPlug->myNode) { |
357 |
< |
globalIndex[local_index] = i; |
355 |
> |
for (i = 0; i < parallelData->nAtomsGlobal; i++) { |
356 |
> |
if (AtomToProcMap[i] == parallelData->myNode) { |
357 |
> |
globalAtomIndex[local_index] = i; |
358 |
> |
|
359 |
> |
globalToLocalAtom[i] = local_index; |
360 |
|
local_index++; |
361 |
+ |
|
362 |
|
} |
363 |
+ |
else |
364 |
+ |
globalToLocalAtom[i] = -1; |
365 |
|
} |
366 |
+ |
|
367 |
+ |
globalGroupIndex.resize(parallelData->nGroupsLocal); |
368 |
+ |
globalToLocalGroup.resize(parallelData->nGroupsGlobal); |
369 |
+ |
local_index = 0; |
370 |
+ |
for (i = 0; i < parallelData->nGroupsGlobal; i++) { |
371 |
+ |
if (GroupToProcMap[i] == parallelData->myNode) { |
372 |
+ |
globalGroupIndex[local_index] = i; |
373 |
+ |
|
374 |
+ |
globalToLocalGroup[i] = local_index; |
375 |
+ |
local_index++; |
376 |
+ |
|
377 |
+ |
} |
378 |
+ |
else |
379 |
+ |
globalToLocalGroup[i] = -1; |
380 |
+ |
} |
381 |
+ |
|
382 |
+ |
globalMolIndex.resize(parallelData->nMolLocal); |
383 |
+ |
globalToLocalMol.resize(parallelData->nMolGlobal); |
384 |
+ |
local_index = 0; |
385 |
+ |
for (i = 0; i < parallelData->nMolGlobal; i++) { |
386 |
+ |
if (MolToProcMap[i] == parallelData->myNode) { |
387 |
+ |
globalMolIndex[local_index] = i; |
388 |
+ |
globalToLocalMol[i] = local_index; |
389 |
+ |
local_index++; |
390 |
+ |
} |
391 |
+ |
else |
392 |
+ |
globalToLocalMol[i] = -1; |
393 |
+ |
} |
394 |
|
|
292 |
– |
return globalIndex; |
395 |
|
} |
396 |
|
|
397 |
|
|
398 |
|
void mpiSimulation::mpiRefresh( void ){ |
399 |
|
|
400 |
|
int isError, i; |
401 |
< |
int *globalIndex = new int[mpiPlug->myNlocal]; |
401 |
> |
int *localToGlobalAtomIndex = new int[parallelData->nAtomsLocal]; |
402 |
> |
int *localToGlobalGroupIndex = new int[parallelData->nGroupsLocal]; |
403 |
|
|
404 |
< |
// Fortran indexing needs to be increased by 1 in order to get the 2 languages to |
405 |
< |
// not barf |
404 |
> |
// Fortran indexing needs to be increased by 1 in order to get the 2 |
405 |
> |
// languages to not barf |
406 |
|
|
407 |
< |
for(i=0; i<mpiPlug->myNlocal; i++) globalIndex[i] = entryPlug->atoms[i]->getGlobalIndex()+1; |
407 |
> |
for(i = 0; i < parallelData->nAtomsLocal; i++) |
408 |
> |
localToGlobalAtomIndex[i] = globalAtomIndex[i] + 1; |
409 |
|
|
410 |
+ |
for(i = 0; i < parallelData->nGroupsLocal; i++) |
411 |
+ |
localToGlobalGroupIndex[i] = globalGroupIndex[i] + 1; |
412 |
|
|
413 |
|
isError = 0; |
414 |
< |
setFsimParallel( mpiPlug, &(entryPlug->n_atoms), globalIndex, &isError ); |
414 |
> |
|
415 |
> |
setFsimParallel( parallelData, |
416 |
> |
&(parallelData->nAtomsLocal), localToGlobalAtomIndex, |
417 |
> |
&(parallelData->nGroupsLocal), localToGlobalGroupIndex, |
418 |
> |
&isError ); |
419 |
> |
|
420 |
|
if( isError ){ |
421 |
|
|
422 |
|
sprintf( painCave.errMsg, |
425 |
|
simError(); |
426 |
|
} |
427 |
|
|
428 |
< |
delete[] globalIndex; |
428 |
> |
delete[] localToGlobalGroupIndex; |
429 |
> |
delete[] localToGlobalAtomIndex; |
430 |
|
|
431 |
+ |
|
432 |
|
sprintf( checkPointMsg, |
433 |
|
" mpiRefresh successful.\n" ); |
434 |
|
MPIcheckPoint(); |