Brian Silverman | 72890c2 | 2015-09-19 14:37:37 -0400 | [diff] [blame^] | 1 | // This file is part of Eigen, a lightweight C++ template library |
| 2 | // for linear algebra. |
| 3 | // |
| 4 | // Copyright (C) 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 | #include "sparse.h" |
| 11 | |
| 12 | template<int OtherStorage, typename SparseMatrixType> void sparse_permutations(const SparseMatrixType& ref) |
| 13 | { |
| 14 | typedef typename SparseMatrixType::Index Index; |
| 15 | |
| 16 | const Index rows = ref.rows(); |
| 17 | const Index cols = ref.cols(); |
| 18 | typedef typename SparseMatrixType::Scalar Scalar; |
| 19 | typedef typename SparseMatrixType::Index Index; |
| 20 | typedef SparseMatrix<Scalar, OtherStorage, Index> OtherSparseMatrixType; |
| 21 | typedef Matrix<Scalar,Dynamic,Dynamic> DenseMatrix; |
| 22 | typedef Matrix<Index,Dynamic,1> VectorI; |
| 23 | |
| 24 | double density = (std::max)(8./(rows*cols), 0.01); |
| 25 | |
| 26 | SparseMatrixType mat(rows, cols), up(rows,cols), lo(rows,cols); |
| 27 | OtherSparseMatrixType res; |
| 28 | DenseMatrix mat_d = DenseMatrix::Zero(rows, cols), up_sym_d, lo_sym_d, res_d; |
| 29 | |
| 30 | initSparse<Scalar>(density, mat_d, mat, 0); |
| 31 | |
| 32 | up = mat.template triangularView<Upper>(); |
| 33 | lo = mat.template triangularView<Lower>(); |
| 34 | |
| 35 | up_sym_d = mat_d.template selfadjointView<Upper>(); |
| 36 | lo_sym_d = mat_d.template selfadjointView<Lower>(); |
| 37 | |
| 38 | VERIFY_IS_APPROX(mat, mat_d); |
| 39 | VERIFY_IS_APPROX(up, DenseMatrix(mat_d.template triangularView<Upper>())); |
| 40 | VERIFY_IS_APPROX(lo, DenseMatrix(mat_d.template triangularView<Lower>())); |
| 41 | |
| 42 | PermutationMatrix<Dynamic> p, p_null; |
| 43 | VectorI pi; |
| 44 | randomPermutationVector(pi, cols); |
| 45 | p.indices() = pi; |
| 46 | |
| 47 | res = mat*p; |
| 48 | res_d = mat_d*p; |
| 49 | VERIFY(res.isApprox(res_d) && "mat*p"); |
| 50 | |
| 51 | res = p*mat; |
| 52 | res_d = p*mat_d; |
| 53 | VERIFY(res.isApprox(res_d) && "p*mat"); |
| 54 | |
| 55 | res = mat*p.inverse(); |
| 56 | res_d = mat*p.inverse(); |
| 57 | VERIFY(res.isApprox(res_d) && "mat*inv(p)"); |
| 58 | |
| 59 | res = p.inverse()*mat; |
| 60 | res_d = p.inverse()*mat_d; |
| 61 | VERIFY(res.isApprox(res_d) && "inv(p)*mat"); |
| 62 | |
| 63 | res = mat.twistedBy(p); |
| 64 | res_d = (p * mat_d) * p.inverse(); |
| 65 | VERIFY(res.isApprox(res_d) && "p*mat*inv(p)"); |
| 66 | |
| 67 | |
| 68 | res = mat.template selfadjointView<Upper>().twistedBy(p_null); |
| 69 | res_d = up_sym_d; |
| 70 | VERIFY(res.isApprox(res_d) && "full selfadjoint upper to full"); |
| 71 | |
| 72 | res = mat.template selfadjointView<Lower>().twistedBy(p_null); |
| 73 | res_d = lo_sym_d; |
| 74 | VERIFY(res.isApprox(res_d) && "full selfadjoint lower to full"); |
| 75 | |
| 76 | |
| 77 | res = up.template selfadjointView<Upper>().twistedBy(p_null); |
| 78 | res_d = up_sym_d; |
| 79 | VERIFY(res.isApprox(res_d) && "upper selfadjoint to full"); |
| 80 | |
| 81 | res = lo.template selfadjointView<Lower>().twistedBy(p_null); |
| 82 | res_d = lo_sym_d; |
| 83 | VERIFY(res.isApprox(res_d) && "lower selfadjoint full"); |
| 84 | |
| 85 | |
| 86 | res = mat.template selfadjointView<Upper>(); |
| 87 | res_d = up_sym_d; |
| 88 | VERIFY(res.isApprox(res_d) && "full selfadjoint upper to full"); |
| 89 | |
| 90 | res = mat.template selfadjointView<Lower>(); |
| 91 | res_d = lo_sym_d; |
| 92 | VERIFY(res.isApprox(res_d) && "full selfadjoint lower to full"); |
| 93 | |
| 94 | res = up.template selfadjointView<Upper>(); |
| 95 | res_d = up_sym_d; |
| 96 | VERIFY(res.isApprox(res_d) && "upper selfadjoint to full"); |
| 97 | |
| 98 | res = lo.template selfadjointView<Lower>(); |
| 99 | res_d = lo_sym_d; |
| 100 | VERIFY(res.isApprox(res_d) && "lower selfadjoint full"); |
| 101 | |
| 102 | |
| 103 | res.template selfadjointView<Upper>() = mat.template selfadjointView<Upper>(); |
| 104 | res_d = up_sym_d.template triangularView<Upper>(); |
| 105 | VERIFY(res.isApprox(res_d) && "full selfadjoint upper to upper"); |
| 106 | |
| 107 | res.template selfadjointView<Lower>() = mat.template selfadjointView<Upper>(); |
| 108 | res_d = up_sym_d.template triangularView<Lower>(); |
| 109 | VERIFY(res.isApprox(res_d) && "full selfadjoint upper to lower"); |
| 110 | |
| 111 | res.template selfadjointView<Upper>() = mat.template selfadjointView<Lower>(); |
| 112 | res_d = lo_sym_d.template triangularView<Upper>(); |
| 113 | VERIFY(res.isApprox(res_d) && "full selfadjoint lower to upper"); |
| 114 | |
| 115 | res.template selfadjointView<Lower>() = mat.template selfadjointView<Lower>(); |
| 116 | res_d = lo_sym_d.template triangularView<Lower>(); |
| 117 | VERIFY(res.isApprox(res_d) && "full selfadjoint lower to lower"); |
| 118 | |
| 119 | |
| 120 | |
| 121 | res.template selfadjointView<Upper>() = mat.template selfadjointView<Upper>().twistedBy(p); |
| 122 | res_d = ((p * up_sym_d) * p.inverse()).eval().template triangularView<Upper>(); |
| 123 | VERIFY(res.isApprox(res_d) && "full selfadjoint upper twisted to upper"); |
| 124 | |
| 125 | res.template selfadjointView<Upper>() = mat.template selfadjointView<Lower>().twistedBy(p); |
| 126 | res_d = ((p * lo_sym_d) * p.inverse()).eval().template triangularView<Upper>(); |
| 127 | VERIFY(res.isApprox(res_d) && "full selfadjoint lower twisted to upper"); |
| 128 | |
| 129 | res.template selfadjointView<Lower>() = mat.template selfadjointView<Lower>().twistedBy(p); |
| 130 | res_d = ((p * lo_sym_d) * p.inverse()).eval().template triangularView<Lower>(); |
| 131 | VERIFY(res.isApprox(res_d) && "full selfadjoint lower twisted to lower"); |
| 132 | |
| 133 | res.template selfadjointView<Lower>() = mat.template selfadjointView<Upper>().twistedBy(p); |
| 134 | res_d = ((p * up_sym_d) * p.inverse()).eval().template triangularView<Lower>(); |
| 135 | VERIFY(res.isApprox(res_d) && "full selfadjoint upper twisted to lower"); |
| 136 | |
| 137 | |
| 138 | res.template selfadjointView<Upper>() = up.template selfadjointView<Upper>().twistedBy(p); |
| 139 | res_d = ((p * up_sym_d) * p.inverse()).eval().template triangularView<Upper>(); |
| 140 | VERIFY(res.isApprox(res_d) && "upper selfadjoint twisted to upper"); |
| 141 | |
| 142 | res.template selfadjointView<Upper>() = lo.template selfadjointView<Lower>().twistedBy(p); |
| 143 | res_d = ((p * lo_sym_d) * p.inverse()).eval().template triangularView<Upper>(); |
| 144 | VERIFY(res.isApprox(res_d) && "lower selfadjoint twisted to upper"); |
| 145 | |
| 146 | res.template selfadjointView<Lower>() = lo.template selfadjointView<Lower>().twistedBy(p); |
| 147 | res_d = ((p * lo_sym_d) * p.inverse()).eval().template triangularView<Lower>(); |
| 148 | VERIFY(res.isApprox(res_d) && "lower selfadjoint twisted to lower"); |
| 149 | |
| 150 | res.template selfadjointView<Lower>() = up.template selfadjointView<Upper>().twistedBy(p); |
| 151 | res_d = ((p * up_sym_d) * p.inverse()).eval().template triangularView<Lower>(); |
| 152 | VERIFY(res.isApprox(res_d) && "upper selfadjoint twisted to lower"); |
| 153 | |
| 154 | |
| 155 | res = mat.template selfadjointView<Upper>().twistedBy(p); |
| 156 | res_d = (p * up_sym_d) * p.inverse(); |
| 157 | VERIFY(res.isApprox(res_d) && "full selfadjoint upper twisted to full"); |
| 158 | |
| 159 | res = mat.template selfadjointView<Lower>().twistedBy(p); |
| 160 | res_d = (p * lo_sym_d) * p.inverse(); |
| 161 | VERIFY(res.isApprox(res_d) && "full selfadjoint lower twisted to full"); |
| 162 | |
| 163 | res = up.template selfadjointView<Upper>().twistedBy(p); |
| 164 | res_d = (p * up_sym_d) * p.inverse(); |
| 165 | VERIFY(res.isApprox(res_d) && "upper selfadjoint twisted to full"); |
| 166 | |
| 167 | res = lo.template selfadjointView<Lower>().twistedBy(p); |
| 168 | res_d = (p * lo_sym_d) * p.inverse(); |
| 169 | VERIFY(res.isApprox(res_d) && "lower selfadjoint twisted to full"); |
| 170 | } |
| 171 | |
| 172 | template<typename Scalar> void sparse_permutations_all(int size) |
| 173 | { |
| 174 | CALL_SUBTEST(( sparse_permutations<ColMajor>(SparseMatrix<Scalar, ColMajor>(size,size)) )); |
| 175 | CALL_SUBTEST(( sparse_permutations<ColMajor>(SparseMatrix<Scalar, RowMajor>(size,size)) )); |
| 176 | CALL_SUBTEST(( sparse_permutations<RowMajor>(SparseMatrix<Scalar, ColMajor>(size,size)) )); |
| 177 | CALL_SUBTEST(( sparse_permutations<RowMajor>(SparseMatrix<Scalar, RowMajor>(size,size)) )); |
| 178 | } |
| 179 | |
| 180 | void test_sparse_permutations() |
| 181 | { |
| 182 | for(int i = 0; i < g_repeat; i++) { |
| 183 | int s = Eigen::internal::random<int>(1,50); |
| 184 | CALL_SUBTEST_1(( sparse_permutations_all<double>(s) )); |
| 185 | CALL_SUBTEST_2(( sparse_permutations_all<std::complex<double> >(s) )); |
| 186 | } |
| 187 | } |