1 | // This file is part of Eigen, a lightweight C++ template library
|
---|
2 | // for linear algebra.
|
---|
3 | //
|
---|
4 | // Copyright (C) 2008-2011 Gael Guennebaud <gael.guennebaud@inria.fr>
|
---|
5 | //
|
---|
6 | // This Source Code Form is subject to the terms of the Mozilla
|
---|
7 | // Public License v. 2.0. If a copy of the MPL was not distributed
|
---|
8 | // with this file, You can obtain one at http://mozilla.org/MPL/2.0/.
|
---|
9 |
|
---|
10 | #ifndef EIGEN_TESTSPARSE_H
|
---|
11 | #define EIGEN_TESTSPARSE_H
|
---|
12 |
|
---|
13 | #define EIGEN_YES_I_KNOW_SPARSE_MODULE_IS_NOT_STABLE_YET
|
---|
14 |
|
---|
15 | #include "main.h"
|
---|
16 |
|
---|
17 | #if EIGEN_GNUC_AT_LEAST(4,0) && !defined __ICC && !defined(__clang__)
|
---|
18 |
|
---|
19 | #ifdef min
|
---|
20 | #undef min
|
---|
21 | #endif
|
---|
22 |
|
---|
23 | #ifdef max
|
---|
24 | #undef max
|
---|
25 | #endif
|
---|
26 |
|
---|
27 | #include <tr1/unordered_map>
|
---|
28 | #define EIGEN_UNORDERED_MAP_SUPPORT
|
---|
29 | namespace std {
|
---|
30 | using std::tr1::unordered_map;
|
---|
31 | }
|
---|
32 | #endif
|
---|
33 |
|
---|
34 | #ifdef EIGEN_GOOGLEHASH_SUPPORT
|
---|
35 | #include <google/sparse_hash_map>
|
---|
36 | #endif
|
---|
37 |
|
---|
38 | #include <Eigen/Cholesky>
|
---|
39 | #include <Eigen/LU>
|
---|
40 | #include <Eigen/Sparse>
|
---|
41 |
|
---|
42 | enum {
|
---|
43 | ForceNonZeroDiag = 1,
|
---|
44 | MakeLowerTriangular = 2,
|
---|
45 | MakeUpperTriangular = 4,
|
---|
46 | ForceRealDiag = 8
|
---|
47 | };
|
---|
48 |
|
---|
49 | /* Initializes both a sparse and dense matrix with same random values,
|
---|
50 | * and a ratio of \a density non zero entries.
|
---|
51 | * \param flags is a union of ForceNonZeroDiag, MakeLowerTriangular and MakeUpperTriangular
|
---|
52 | * allowing to control the shape of the matrix.
|
---|
53 | * \param zeroCoords and nonzeroCoords allows to get the coordinate lists of the non zero,
|
---|
54 | * and zero coefficients respectively.
|
---|
55 | */
|
---|
56 | template<typename Scalar,int Opt1,int Opt2,typename Index> void
|
---|
57 | initSparse(double density,
|
---|
58 | Matrix<Scalar,Dynamic,Dynamic,Opt1>& refMat,
|
---|
59 | SparseMatrix<Scalar,Opt2,Index>& sparseMat,
|
---|
60 | int flags = 0,
|
---|
61 | std::vector<Matrix<Index,2,1> >* zeroCoords = 0,
|
---|
62 | std::vector<Matrix<Index,2,1> >* nonzeroCoords = 0)
|
---|
63 | {
|
---|
64 | enum { IsRowMajor = SparseMatrix<Scalar,Opt2,Index>::IsRowMajor };
|
---|
65 | sparseMat.setZero();
|
---|
66 | //sparseMat.reserve(int(refMat.rows()*refMat.cols()*density));
|
---|
67 | sparseMat.reserve(VectorXi::Constant(IsRowMajor ? refMat.rows() : refMat.cols(), int((1.5*density)*(IsRowMajor?refMat.cols():refMat.rows()))));
|
---|
68 |
|
---|
69 | for(Index j=0; j<sparseMat.outerSize(); j++)
|
---|
70 | {
|
---|
71 | //sparseMat.startVec(j);
|
---|
72 | for(Index i=0; i<sparseMat.innerSize(); i++)
|
---|
73 | {
|
---|
74 | int ai(i), aj(j);
|
---|
75 | if(IsRowMajor)
|
---|
76 | std::swap(ai,aj);
|
---|
77 | Scalar v = (internal::random<double>(0,1) < density) ? internal::random<Scalar>() : Scalar(0);
|
---|
78 | if ((flags&ForceNonZeroDiag) && (i==j))
|
---|
79 | {
|
---|
80 | v = internal::random<Scalar>()*Scalar(3.);
|
---|
81 | v = v*v + Scalar(5.);
|
---|
82 | }
|
---|
83 | if ((flags & MakeLowerTriangular) && aj>ai)
|
---|
84 | v = Scalar(0);
|
---|
85 | else if ((flags & MakeUpperTriangular) && aj<ai)
|
---|
86 | v = Scalar(0);
|
---|
87 |
|
---|
88 | if ((flags&ForceRealDiag) && (i==j))
|
---|
89 | v = numext::real(v);
|
---|
90 |
|
---|
91 | if (v!=Scalar(0))
|
---|
92 | {
|
---|
93 | //sparseMat.insertBackByOuterInner(j,i) = v;
|
---|
94 | sparseMat.insertByOuterInner(j,i) = v;
|
---|
95 | if (nonzeroCoords)
|
---|
96 | nonzeroCoords->push_back(Matrix<Index,2,1> (ai,aj));
|
---|
97 | }
|
---|
98 | else if (zeroCoords)
|
---|
99 | {
|
---|
100 | zeroCoords->push_back(Matrix<Index,2,1> (ai,aj));
|
---|
101 | }
|
---|
102 | refMat(ai,aj) = v;
|
---|
103 | }
|
---|
104 | }
|
---|
105 | //sparseMat.finalize();
|
---|
106 | }
|
---|
107 |
|
---|
108 | template<typename Scalar,int Opt1,int Opt2,typename Index> void
|
---|
109 | initSparse(double density,
|
---|
110 | Matrix<Scalar,Dynamic,Dynamic, Opt1>& refMat,
|
---|
111 | DynamicSparseMatrix<Scalar, Opt2, Index>& sparseMat,
|
---|
112 | int flags = 0,
|
---|
113 | std::vector<Matrix<Index,2,1> >* zeroCoords = 0,
|
---|
114 | std::vector<Matrix<Index,2,1> >* nonzeroCoords = 0)
|
---|
115 | {
|
---|
116 | enum { IsRowMajor = DynamicSparseMatrix<Scalar,Opt2,Index>::IsRowMajor };
|
---|
117 | sparseMat.setZero();
|
---|
118 | sparseMat.reserve(int(refMat.rows()*refMat.cols()*density));
|
---|
119 | for(int j=0; j<sparseMat.outerSize(); j++)
|
---|
120 | {
|
---|
121 | sparseMat.startVec(j); // not needed for DynamicSparseMatrix
|
---|
122 | for(int i=0; i<sparseMat.innerSize(); i++)
|
---|
123 | {
|
---|
124 | int ai(i), aj(j);
|
---|
125 | if(IsRowMajor)
|
---|
126 | std::swap(ai,aj);
|
---|
127 | Scalar v = (internal::random<double>(0,1) < density) ? internal::random<Scalar>() : Scalar(0);
|
---|
128 | if ((flags&ForceNonZeroDiag) && (i==j))
|
---|
129 | {
|
---|
130 | v = internal::random<Scalar>()*Scalar(3.);
|
---|
131 | v = v*v + Scalar(5.);
|
---|
132 | }
|
---|
133 | if ((flags & MakeLowerTriangular) && aj>ai)
|
---|
134 | v = Scalar(0);
|
---|
135 | else if ((flags & MakeUpperTriangular) && aj<ai)
|
---|
136 | v = Scalar(0);
|
---|
137 |
|
---|
138 | if ((flags&ForceRealDiag) && (i==j))
|
---|
139 | v = numext::real(v);
|
---|
140 |
|
---|
141 | if (v!=Scalar(0))
|
---|
142 | {
|
---|
143 | sparseMat.insertBackByOuterInner(j,i) = v;
|
---|
144 | if (nonzeroCoords)
|
---|
145 | nonzeroCoords->push_back(Matrix<Index,2,1> (ai,aj));
|
---|
146 | }
|
---|
147 | else if (zeroCoords)
|
---|
148 | {
|
---|
149 | zeroCoords->push_back(Matrix<Index,2,1> (ai,aj));
|
---|
150 | }
|
---|
151 | refMat(ai,aj) = v;
|
---|
152 | }
|
---|
153 | }
|
---|
154 | sparseMat.finalize();
|
---|
155 | }
|
---|
156 |
|
---|
157 | template<typename Scalar,int Options,typename Index> void
|
---|
158 | initSparse(double density,
|
---|
159 | Matrix<Scalar,Dynamic,1>& refVec,
|
---|
160 | SparseVector<Scalar,Options,Index>& sparseVec,
|
---|
161 | std::vector<int>* zeroCoords = 0,
|
---|
162 | std::vector<int>* nonzeroCoords = 0)
|
---|
163 | {
|
---|
164 | sparseVec.reserve(int(refVec.size()*density));
|
---|
165 | sparseVec.setZero();
|
---|
166 | for(Index i=0; i<refVec.size(); i++)
|
---|
167 | {
|
---|
168 | Scalar v = (internal::random<double>(0,1) < density) ? internal::random<Scalar>() : Scalar(0);
|
---|
169 | if (v!=Scalar(0))
|
---|
170 | {
|
---|
171 | sparseVec.insertBack(i) = v;
|
---|
172 | if (nonzeroCoords)
|
---|
173 | nonzeroCoords->push_back(i);
|
---|
174 | }
|
---|
175 | else if (zeroCoords)
|
---|
176 | zeroCoords->push_back(i);
|
---|
177 | refVec[i] = v;
|
---|
178 | }
|
---|
179 | }
|
---|
180 |
|
---|
181 | template<typename Scalar,int Options,typename Index> void
|
---|
182 | initSparse(double density,
|
---|
183 | Matrix<Scalar,1,Dynamic>& refVec,
|
---|
184 | SparseVector<Scalar,Options,Index>& sparseVec,
|
---|
185 | std::vector<int>* zeroCoords = 0,
|
---|
186 | std::vector<int>* nonzeroCoords = 0)
|
---|
187 | {
|
---|
188 | sparseVec.reserve(int(refVec.size()*density));
|
---|
189 | sparseVec.setZero();
|
---|
190 | for(int i=0; i<refVec.size(); i++)
|
---|
191 | {
|
---|
192 | Scalar v = (internal::random<double>(0,1) < density) ? internal::random<Scalar>() : Scalar(0);
|
---|
193 | if (v!=Scalar(0))
|
---|
194 | {
|
---|
195 | sparseVec.insertBack(i) = v;
|
---|
196 | if (nonzeroCoords)
|
---|
197 | nonzeroCoords->push_back(i);
|
---|
198 | }
|
---|
199 | else if (zeroCoords)
|
---|
200 | zeroCoords->push_back(i);
|
---|
201 | refVec[i] = v;
|
---|
202 | }
|
---|
203 | }
|
---|
204 |
|
---|
205 |
|
---|
206 | #include <unsupported/Eigen/SparseExtra>
|
---|
207 | #endif // EIGEN_TESTSPARSE_H
|
---|