1 // This file is part of Eigen, a lightweight C++ template library
2 // for linear algebra.
3 //
4 // Copyright (C) 2008 Gael Guennebaud <gael.guennebaud@inria.fr>
5 // Copyright (C) 2006-2008 Benoit Jacob <jacob.benoit.1@gmail.com>
6 //
7 // This Source Code Form is subject to the terms of the Mozilla
8 // Public License v. 2.0. If a copy of the MPL was not distributed
9 // with this file, You can obtain one at http://mozilla.org/MPL/2.0/.
10
11 // this hack is needed to make this file compiles with -pedantic (gcc)
12 #ifdef __GNUC__
13 #define throw(X)
14 #endif
15 // discard stack allocation as that too bypasses malloc
16 #define EIGEN_STACK_ALLOCATION_LIMIT 0
17 // any heap allocation will raise an assert
18 #define EIGEN_NO_MALLOC
19
20 #include "main.h"
21 #include <Eigen/Cholesky>
22 #include <Eigen/Eigenvalues>
23 #include <Eigen/LU>
24 #include <Eigen/QR>
25 #include <Eigen/SVD>
26
nomalloc(const MatrixType & m)27 template<typename MatrixType> void nomalloc(const MatrixType& m)
28 {
29 /* this test check no dynamic memory allocation are issued with fixed-size matrices
30 */
31 typedef typename MatrixType::Index Index;
32 typedef typename MatrixType::Scalar Scalar;
33 typedef Matrix<Scalar, MatrixType::RowsAtCompileTime, 1> VectorType;
34
35 Index rows = m.rows();
36 Index cols = m.cols();
37
38 MatrixType m1 = MatrixType::Random(rows, cols),
39 m2 = MatrixType::Random(rows, cols),
40 m3(rows, cols);
41
42 Scalar s1 = internal::random<Scalar>();
43
44 Index r = internal::random<Index>(0, rows-1),
45 c = internal::random<Index>(0, cols-1);
46
47 VERIFY_IS_APPROX((m1+m2)*s1, s1*m1+s1*m2);
48 VERIFY_IS_APPROX((m1+m2)(r,c), (m1(r,c))+(m2(r,c)));
49 VERIFY_IS_APPROX(m1.cwiseProduct(m1.block(0,0,rows,cols)), (m1.array()*m1.array()).matrix());
50 VERIFY_IS_APPROX((m1*m1.transpose())*m2, m1*(m1.transpose()*m2));
51
52 m2.col(0).noalias() = m1 * m1.col(0);
53 m2.col(0).noalias() -= m1.adjoint() * m1.col(0);
54 m2.col(0).noalias() -= m1 * m1.row(0).adjoint();
55 m2.col(0).noalias() -= m1.adjoint() * m1.row(0).adjoint();
56
57 m2.row(0).noalias() = m1.row(0) * m1;
58 m2.row(0).noalias() -= m1.row(0) * m1.adjoint();
59 m2.row(0).noalias() -= m1.col(0).adjoint() * m1;
60 m2.row(0).noalias() -= m1.col(0).adjoint() * m1.adjoint();
61 VERIFY_IS_APPROX(m2,m2);
62
63 m2.col(0).noalias() = m1.template triangularView<Upper>() * m1.col(0);
64 m2.col(0).noalias() -= m1.adjoint().template triangularView<Upper>() * m1.col(0);
65 m2.col(0).noalias() -= m1.template triangularView<Upper>() * m1.row(0).adjoint();
66 m2.col(0).noalias() -= m1.adjoint().template triangularView<Upper>() * m1.row(0).adjoint();
67
68 m2.row(0).noalias() = m1.row(0) * m1.template triangularView<Upper>();
69 m2.row(0).noalias() -= m1.row(0) * m1.adjoint().template triangularView<Upper>();
70 m2.row(0).noalias() -= m1.col(0).adjoint() * m1.template triangularView<Upper>();
71 m2.row(0).noalias() -= m1.col(0).adjoint() * m1.adjoint().template triangularView<Upper>();
72 VERIFY_IS_APPROX(m2,m2);
73
74 m2.col(0).noalias() = m1.template selfadjointView<Upper>() * m1.col(0);
75 m2.col(0).noalias() -= m1.adjoint().template selfadjointView<Upper>() * m1.col(0);
76 m2.col(0).noalias() -= m1.template selfadjointView<Upper>() * m1.row(0).adjoint();
77 m2.col(0).noalias() -= m1.adjoint().template selfadjointView<Upper>() * m1.row(0).adjoint();
78
79 m2.row(0).noalias() = m1.row(0) * m1.template selfadjointView<Upper>();
80 m2.row(0).noalias() -= m1.row(0) * m1.adjoint().template selfadjointView<Upper>();
81 m2.row(0).noalias() -= m1.col(0).adjoint() * m1.template selfadjointView<Upper>();
82 m2.row(0).noalias() -= m1.col(0).adjoint() * m1.adjoint().template selfadjointView<Upper>();
83 VERIFY_IS_APPROX(m2,m2);
84
85 m2.template selfadjointView<Lower>().rankUpdate(m1.col(0),-1);
86 m2.template selfadjointView<Lower>().rankUpdate(m1.row(0),-1);
87
88 // The following fancy matrix-matrix products are not safe yet regarding static allocation
89 // m1 += m1.template triangularView<Upper>() * m2.col(;
90 // m1.template selfadjointView<Lower>().rankUpdate(m2);
91 // m1 += m1.template triangularView<Upper>() * m2;
92 // m1 += m1.template selfadjointView<Lower>() * m2;
93 // VERIFY_IS_APPROX(m1,m1);
94 }
95
96 template<typename Scalar>
ctms_decompositions()97 void ctms_decompositions()
98 {
99 const int maxSize = 16;
100 const int size = 12;
101
102 typedef Eigen::Matrix<Scalar,
103 Eigen::Dynamic, Eigen::Dynamic,
104 0,
105 maxSize, maxSize> Matrix;
106
107 typedef Eigen::Matrix<Scalar,
108 Eigen::Dynamic, 1,
109 0,
110 maxSize, 1> Vector;
111
112 typedef Eigen::Matrix<std::complex<Scalar>,
113 Eigen::Dynamic, Eigen::Dynamic,
114 0,
115 maxSize, maxSize> ComplexMatrix;
116
117 const Matrix A(Matrix::Random(size, size)), B(Matrix::Random(size, size));
118 Matrix X(size,size);
119 const ComplexMatrix complexA(ComplexMatrix::Random(size, size));
120 const Matrix saA = A.adjoint() * A;
121 const Vector b(Vector::Random(size));
122 Vector x(size);
123
124 // Cholesky module
125 Eigen::LLT<Matrix> LLT; LLT.compute(A);
126 X = LLT.solve(B);
127 x = LLT.solve(b);
128 Eigen::LDLT<Matrix> LDLT; LDLT.compute(A);
129 X = LDLT.solve(B);
130 x = LDLT.solve(b);
131
132 // Eigenvalues module
133 Eigen::HessenbergDecomposition<ComplexMatrix> hessDecomp; hessDecomp.compute(complexA);
134 Eigen::ComplexSchur<ComplexMatrix> cSchur(size); cSchur.compute(complexA);
135 Eigen::ComplexEigenSolver<ComplexMatrix> cEigSolver; cEigSolver.compute(complexA);
136 Eigen::EigenSolver<Matrix> eigSolver; eigSolver.compute(A);
137 Eigen::SelfAdjointEigenSolver<Matrix> saEigSolver(size); saEigSolver.compute(saA);
138 Eigen::Tridiagonalization<Matrix> tridiag; tridiag.compute(saA);
139
140 // LU module
141 Eigen::PartialPivLU<Matrix> ppLU; ppLU.compute(A);
142 X = ppLU.solve(B);
143 x = ppLU.solve(b);
144 Eigen::FullPivLU<Matrix> fpLU; fpLU.compute(A);
145 X = fpLU.solve(B);
146 x = fpLU.solve(b);
147
148 // QR module
149 Eigen::HouseholderQR<Matrix> hQR; hQR.compute(A);
150 X = hQR.solve(B);
151 x = hQR.solve(b);
152 Eigen::ColPivHouseholderQR<Matrix> cpQR; cpQR.compute(A);
153 X = cpQR.solve(B);
154 x = cpQR.solve(b);
155 Eigen::FullPivHouseholderQR<Matrix> fpQR; fpQR.compute(A);
156 // FIXME X = fpQR.solve(B);
157 x = fpQR.solve(b);
158
159 // SVD module
160 Eigen::JacobiSVD<Matrix> jSVD; jSVD.compute(A, ComputeFullU | ComputeFullV);
161 }
162
test_nomalloc()163 void test_nomalloc()
164 {
165 // check that our operator new is indeed called:
166 VERIFY_RAISES_ASSERT(MatrixXd dummy(MatrixXd::Random(3,3)));
167 CALL_SUBTEST_1(nomalloc(Matrix<float, 1, 1>()) );
168 CALL_SUBTEST_2(nomalloc(Matrix4d()) );
169 CALL_SUBTEST_3(nomalloc(Matrix<float,32,32>()) );
170
171 // Check decomposition modules with dynamic matrices that have a known compile-time max size (ctms)
172 CALL_SUBTEST_4(ctms_decompositions<float>());
173
174 }
175