1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
|
// Copyright (C) 2021 EDF
// All Rights Reserved
// This code is published under the GNU Lesser General Public License (GNU LGPL)
#ifdef USE_MPI
#include <functional>
#include <memory>
#include <boost/mpi.hpp>
#ifdef _OPENMP
#include <omp.h>
#include "StOpt/core/utils/OpenmpException.h"
#endif
#include <Eigen/Dense>
#include "geners/BinaryFileArchive.hh"
#include "geners/Record.hh"
#include "geners/vectorIO.hh"
#include "StOpt/core/utils/types.h"
#include "StOpt/core/parallelism/ParallelComputeGridSplitting.h"
#include "StOpt/core/grids/FullRegularIntGridIterator.h"
#include "StOpt/core/grids/RegularSpaceIntGrid.h"
#include "StOpt/core/utils/eigenGeners.h"
#include "StOpt/regression/BaseRegressionGeners.h"
#include "StOpt/dp/OptimizerSwitchBase.h"
#include "StOpt/dp/TransitionStepRegressionSwitchDist.h"
using namespace StOpt;
using namespace Eigen;
using namespace std;
TransitionStepRegressionSwitchDist::TransitionStepRegressionSwitchDist(const vector< shared_ptr<RegularSpaceIntGrid> > &p_pGridCurrent,
const vector< shared_ptr<RegularSpaceIntGrid> > &p_pGridPrevious,
const shared_ptr<OptimizerSwitchBase > &p_pOptimize,
const boost::mpi::communicator &p_world): m_pGridCurrent(p_pGridCurrent),
m_pGridPrevious(p_pGridPrevious), m_pOptimize(p_pOptimize), m_paral(p_pGridCurrent.size()), m_gridCurrentProc(p_pGridCurrent.size()), m_gridExtendPreviousStep(p_pGridCurrent.size()), m_world(p_world)
{
vector< Array< bool, Dynamic, 1> > dimToSplit = m_pOptimize->getDimensionToSplit();
for (size_t iReg = 0; iReg < p_pGridCurrent.size(); ++iReg)
{
// initial and previous dimensions
ArrayXi initialDimension = p_pGridCurrent[iReg]->getDimensions();
ArrayXi initialDimensionPrev = p_pGridPrevious[iReg]->getDimensions();
// organize the hypercube splitting for parallel
ArrayXi splittingRatio = paraOptimalSplitting(initialDimension, dimToSplit[iReg], m_world);
ArrayXi splittingRatioPrev = paraOptimalSplitting(initialDimensionPrev, dimToSplit[iReg], m_world);
// cone value
auto fMesh = [iReg, p_pGridCurrent, p_pGridPrevious, p_pOptimize](const SubMeshIntCoord & p_intMesh)->SubMeshIntCoord
{
vector< array<int, 2 > > intMeshUsed(p_intMesh.size());
for (int i = 0; i < p_intMesh.size(); ++i)
{
// from grid starting at 0 to real grid position
intMeshUsed[i][0] = p_intMesh(i)[0] + p_pGridCurrent[iReg]->getLowValueDim(i);
intMeshUsed[i][1] = p_intMesh(i)[1] + p_pGridCurrent[iReg]->getLowValueDim(i) - 1; // last is outside the grid
}
vector< array<int, 2 > > retCone = p_pOptimize->getCone(iReg, intMeshUsed);
// cap to max, min
SubMeshIntCoord ret(retCone.size());
for (int i = 0; i < p_intMesh.size(); ++i)
{
ret(i)[0] = max(retCone[i][0], p_pGridPrevious[iReg]->getLowValueDim(i)) ;
ret(i)[1] = min(retCone[i][1], p_pGridPrevious[iReg]->getMaxValueDim(i));
}
for (int i = 0; i < p_intMesh.size(); ++i)
{
ret(i)[1] += 1; // border should be outside
// from real position to position starting at 0
ret(i)[0] -= p_pGridPrevious[iReg]->getLowValueDim(i);
ret(i)[1] -= p_pGridPrevious[iReg]->getLowValueDim(i);
}
return ret;
};
// ParallelComputeGridsSplitting objects
m_paral[iReg] = make_shared<ParallelComputeGridSplitting>(initialDimension, initialDimensionPrev,
function < SubMeshIntCoord(const SubMeshIntCoord &) >(fMesh),
splittingRatio, splittingRatioPrev, m_world);
// get back grid treated by current processor
SubMeshIntCoord gridLocal = m_paral[iReg]->getCurrentCalculationGrid();
// Construct local sub grid
m_gridCurrentProc[iReg] = m_pGridCurrent[iReg]->getSubGrid(gridLocal);
// only if the grid is not empty
if (m_gridCurrentProc[iReg]->getNbPoints() > 0)
{
// get back grid extended on previous step
SubMeshIntCoord gridLocalExtended = m_paral[iReg]->getExtendedGridProcOldGrid();
m_gridExtendPreviousStep [iReg] = m_pGridPrevious[iReg]->getSubGrid(gridLocalExtended);
}
}
}
vector< shared_ptr< ArrayXXd >> TransitionStepRegressionSwitchDist::oneStep(const vector< shared_ptr< ArrayXXd > > &p_phiIn,
const shared_ptr< BaseRegression> &p_condExp) const
{
// number of regimes at current time
int nbRegimes = m_pOptimize->getNbRegime();
vector< shared_ptr< ArrayXXd > > phiOut(nbRegimes);
// only if the processor is working
vector < shared_ptr< ArrayXXd > > phiInExtended(p_phiIn.size());
// Organize the data splitting : spread the incoming values on an extended grid
for (int iReg = 0; iReg < nbRegimes ; ++iReg)
{
// utilitary
ArrayXXd emptyArray;
if (p_phiIn[iReg])
{
phiInExtended[iReg] = make_shared< ArrayXXd >(m_paral[iReg]->runOneStep(*p_phiIn[iReg])) ;
}
else
phiInExtended[iReg] = make_shared< ArrayXXd >(m_paral[iReg]->runOneStep(emptyArray)) ;
if (m_gridCurrentProc[iReg]->getNbPoints() > 0)
{
// allocate for solution
phiOut[iReg] = make_shared< ArrayXXd >(p_condExp->getNbSimul(), m_gridCurrentProc[iReg]->getNbPoints());
}
}
for (int iReg = 0; iReg < nbRegimes ; ++iReg)
{
if (m_gridCurrentProc[iReg]->getNbPoints() > 0)
{
// number of thread
#ifdef _OPENMP
int nbThreads = omp_get_max_threads();
#else
int nbThreads = 1;
#endif
// create iterator on current grid treated for processor
int iThread = 0 ;
#ifdef _OPENMP
OpenmpException excep; // deal with exception in openmp
#pragma omp parallel for private(iThread)
#endif
for (iThread = 0; iThread < nbThreads; ++iThread)
{
#ifdef _OPENMP
excep.run([&]
{
#endif
FullRegularIntGridIterator iterGridPoint = m_gridCurrentProc[iReg]->getGridIterator();
// account fo threads
iterGridPoint.jumpToAndInc(0, 1, iThread);
// iterates on points of the grid
while (iterGridPoint.isValid())
{
ArrayXi pointCoord = iterGridPoint.getIntCoordinate();
// optimize the current point and the set of regimes
ArrayXd solution = m_pOptimize->stepOptimize(m_gridExtendPreviousStep, iReg, pointCoord, p_condExp, phiInExtended);
// copie solution
(*phiOut[iReg]).col(iterGridPoint.getCount()) = solution;
iterGridPoint.nextInc(nbThreads);
}
#ifdef _OPENMP
});
#endif
}
#ifdef _OPENMP
excep.rethrow();
#endif
}
}
return phiOut;
}
void TransitionStepRegressionSwitchDist::dumpContinuationValues(shared_ptr<gs::BinaryFileArchive> p_ar, const string &p_name, const int &p_iStep,
const vector< shared_ptr< ArrayXXd > > &p_phiInPrev,
const shared_ptr<BaseRegression> &p_condExp) const
{
string stepString = boost::lexical_cast<string>(p_iStep) ;
vector< Array< bool, Dynamic, 1> > dimToSplit = m_pOptimize->getDimensionToSplit();
if (m_world.rank() == 0)
// store regressor
*p_ar << gs::Record(dynamic_cast<const BaseRegression &>(*p_condExp), "regressor", stepString.c_str()) ;
for (size_t iReg = 0; iReg < m_paral.size(); ++iReg)
{
ArrayXi initialDimensionPrev = m_pGridPrevious[iReg]->getDimensions();
ArrayXi initialDimension = m_pGridCurrent[iReg]->getDimensions();
// utilitary
SubMeshIntCoord gridOnProc0Prev(initialDimensionPrev.size());
for (int id = 0; id < initialDimensionPrev.size(); ++id)
{
gridOnProc0Prev(id)[0] = 0 ;
gridOnProc0Prev(id)[1] = initialDimensionPrev(id) ;
}
ArrayXi splittingRatioPrev = paraOptimalSplitting(initialDimensionPrev, dimToSplit[iReg], m_world);
ParallelComputeGridSplitting paralObjectPrev(initialDimensionPrev, splittingRatioPrev, m_world);
ArrayXXd reconstructedArray ;
if (m_world.rank() < m_paral[iReg]->getNbProcessorUsedPrev())
reconstructedArray = paralObjectPrev.reconstruct(*p_phiInPrev[iReg], gridOnProc0Prev);
if (m_world.rank() == 0)
{
ArrayXXd transposeCont = reconstructedArray.transpose();
ArrayXXd basisValues = p_condExp->getCoordBasisFunctionMultiple(transposeCont).transpose();
*p_ar << gs::Record(basisValues, (p_name + "basisValues").c_str(), stepString.c_str()) ;
}
}
if (m_world.rank() == 0)
p_ar->flush() ; // necessary for python mapping
m_world.barrier() ; // onlyt to prevent the reading in simualtion before the end of writting
}
#endif
|
