devtools/eigen
1
0
Fork 0
mirror of https://gitlab.com/libeigen/eigen.git synced 2026-04-10 11:34:33 +08:00
Code Issues Packages Projects Releases Wiki Activity

Clang-format tests, examples, libraries, benchmarks, etc.

Browse Source
This commit is contained in:
Antonio Sánchez
2023-12-05 21:22:55 +00:00
committed by Rasmus Munk Larsen
parent 3252ecc7a4
commit 46e9cdb7fe
876 changed files with 33453 additions and 37795 deletions
Download Patch File Download Diff File Expand all files Collapse all files

119
test/sparseqr.cpp
View File

@@ -9,50 +9,48 @@
#include "sparse.h"
#include <Eigen/SparseQR>
template<typename MatrixType,typename DenseMat>
int generate_sparse_rectangular_problem(MatrixType& A, DenseMat& dA, int maxRows = 300, int maxCols = 150)
{
template <typename MatrixType, typename DenseMat>
int generate_sparse_rectangular_problem(MatrixType& A, DenseMat& dA, int maxRows = 300, int maxCols = 150) {
eigen_assert(maxRows >= maxCols);
typedef typename MatrixType::Scalar Scalar;
int rows = internal::random<int>(1,maxRows);
int cols = internal::random<int>(1,maxCols);
double density = (std::max)(8./(rows*cols), 0.01);
A.resize(rows,cols);
dA.resize(rows,cols);
initSparse<Scalar>(density, dA, A,ForceNonZeroDiag);
int rows = internal::random<int>(1, maxRows);
int cols = internal::random<int>(1, maxCols);
double density = (std::max)(8. / (rows * cols), 0.01);
A.resize(rows, cols);
dA.resize(rows, cols);
initSparse<Scalar>(density, dA, A, ForceNonZeroDiag);
A.makeCompressed();
int nop = internal::random<int>(0, internal::random<double>(0,1) > 0.5 ? cols/2 : 0);
for(int k=0; k<nop; ++k)
{
int j0 = internal::random<int>(0,cols-1);
int j1 = internal::random<int>(0,cols-1);
int nop = internal::random<int>(0, internal::random<double>(0, 1) > 0.5 ? cols / 2 : 0);
for (int k = 0; k < nop; ++k) {
int j0 = internal::random<int>(0, cols - 1);
int j1 = internal::random<int>(0, cols - 1);
Scalar s = internal::random<Scalar>();
A.col(j0) = s * A.col(j1);
A.col(j0) = s * A.col(j1);
dA.col(j0) = s * dA.col(j1);
}
// if(rows<cols) {
// A.conservativeResize(cols,cols);
// dA.conservativeResize(cols,cols);
// dA.bottomRows(cols-rows).setZero();
// }
// if(rows<cols) {
// A.conservativeResize(cols,cols);
// dA.conservativeResize(cols,cols);
// dA.bottomRows(cols-rows).setZero();
// }
return rows;
}
template<typename Scalar> void test_sparseqr_scalar()
{
template <typename Scalar>
void test_sparseqr_scalar() {
typedef typename NumTraits<Scalar>::Real RealScalar;
typedef SparseMatrix<Scalar,ColMajor> MatrixType;
typedef Matrix<Scalar,Dynamic,Dynamic> DenseMat;
typedef Matrix<Scalar,Dynamic,1> DenseVector;
typedef SparseMatrix<Scalar, ColMajor> MatrixType;
typedef Matrix<Scalar, Dynamic, Dynamic> DenseMat;
typedef Matrix<Scalar, Dynamic, 1> DenseVector;
MatrixType A;
DenseMat dA;
DenseVector refX,x,b;
SparseQR<MatrixType, COLAMDOrdering<int> > solver;
generate_sparse_rectangular_problem(A,dA);
DenseVector refX, x, b;
SparseQR<MatrixType, COLAMDOrdering<int> > solver;
generate_sparse_rectangular_problem(A, dA);
b = dA * DenseVector::Random(A.cols());
solver.compute(A);
@@ -65,29 +63,25 @@ template<typename Scalar> void test_sparseqr_scalar()
VERIFY_IS_EQUAL(solver.matrixR().cols(), A.cols());
// Q and R can be multiplied
DenseMat recoveredA = solver.matrixQ()
* DenseMat(solver.matrixR().template triangularView<Upper>())
* solver.colsPermutation().transpose();
DenseMat recoveredA = solver.matrixQ() * DenseMat(solver.matrixR().template triangularView<Upper>()) *
solver.colsPermutation().transpose();
VERIFY_IS_EQUAL(recoveredA.rows(), A.rows());
VERIFY_IS_EQUAL(recoveredA.cols(), A.cols());
// and in the full rank case the original matrix is recovered
if (solver.rank() == A.cols())
{
VERIFY_IS_APPROX(A, recoveredA);
if (solver.rank() == A.cols()) {
VERIFY_IS_APPROX(A, recoveredA);
}
if(internal::random<float>(0,1)>0.5f)
if (internal::random<float>(0, 1) > 0.5f)
solver.factorize(A); // this checks that calling analyzePattern is not needed if the pattern do not change.
if (solver.info() != Success)
{
if (solver.info() != Success) {
std::cerr << "sparse QR factorization failed\n";
exit(0);
return;
}
x = solver.solve(b);
if (solver.info() != Success)
{
if (solver.info() != Success) {
std::cerr << "sparse QR factorization failed\n";
exit(0);
return;
@@ -96,15 +90,14 @@ template<typename Scalar> void test_sparseqr_scalar()
// Compare with a dense QR solver
ColPivHouseholderQR<DenseMat> dqr(dA);
refX = dqr.solve(b);
bool rank_deficient = A.cols()>A.rows() || dqr.rank()<A.cols();
if(rank_deficient)
{
bool rank_deficient = A.cols() > A.rows() || dqr.rank() < A.cols();
if (rank_deficient) {
// rank deficient problem -> we might have to increase the threshold
// to get a correct solution.
RealScalar th = RealScalar(20)*dA.colwise().norm().maxCoeff()*(A.rows()+A.cols()) * NumTraits<RealScalar>::epsilon();
for(Index k=0; (k<16) && !test_isApprox(A*x,b); ++k)
{
RealScalar th =
RealScalar(20) * dA.colwise().norm().maxCoeff() * (A.rows() + A.cols()) * NumTraits<RealScalar>::epsilon();
for (Index k = 0; (k < 16) && !test_isApprox(A * x, b); ++k) {
th *= RealScalar(10);
solver.setPivotThreshold(th);
solver.compute(A);
@@ -113,37 +106,35 @@ template<typename Scalar> void test_sparseqr_scalar()
}
VERIFY_IS_APPROX(A * x, b);
// For rank deficient problem, the estimated rank might
// be slightly off, so let's only raise a warning in such cases.
if(rank_deficient) ++g_test_level;
if (rank_deficient) ++g_test_level;
VERIFY_IS_EQUAL(solver.rank(), dqr.rank());
if(rank_deficient) --g_test_level;
if (rank_deficient) --g_test_level;
if(solver.rank()==A.cols()) // full rank
if (solver.rank() == A.cols()) // full rank
VERIFY_IS_APPROX(x, refX);
// else
// VERIFY((dA * refX - b).norm() * 2 > (A * x - b).norm() );
// else
// VERIFY((dA * refX - b).norm() * 2 > (A * x - b).norm() );
// Compute explicitly the matrix Q
MatrixType Q, QtQ, idM;
Q = solver.matrixQ();
//Check ||Q' * Q - I ||
// Check ||Q' * Q - I ||
QtQ = Q * Q.adjoint();
idM.resize(Q.rows(), Q.rows()); idM.setIdentity();
idM.resize(Q.rows(), Q.rows());
idM.setIdentity();
VERIFY(idM.isApprox(QtQ));
// Q to dense
DenseMat dQ;
dQ = solver.matrixQ();
VERIFY_IS_APPROX(Q, dQ);
}
EIGEN_DECLARE_TEST(sparseqr)
{
for(int i=0; i<g_repeat; ++i)
{
EIGEN_DECLARE_TEST(sparseqr) {
for (int i = 0; i < g_repeat; ++i) {
CALL_SUBTEST_1(test_sparseqr_scalar<double>());
CALL_SUBTEST_2(test_sparseqr_scalar<std::complex<double> >());
}
}