proxsuite-nlp  0.11.0
A primal-dual augmented Lagrangian-type solver for nonlinear programming on manifolds.
 
Loading...
Searching...
No Matches
dense.hpp File Reference

Routines for Cholesky factorization. More...

Include dependency graph for dense.hpp:
This graph shows which files directly or indirectly include this file:

Go to the source code of this file.

Classes

struct  proxsuite::nlp::linalg::DenseLDLT< _Scalar >
 A fast, recursive divide-and-conquer LDLT algorithm. More...
 

Namespaces

namespace  proxsuite
 Main package namespace.
 
namespace  proxsuite::nlp::linalg
 Specific linear algebra routines.
 

Functions

template<typename Scalar>
void proxsuite::nlp::linalg::backend::update_sign_matrix (SignMatrix &sign, const Scalar &akk)
 
template<typename Derived>
bool proxsuite::nlp::linalg::backend::ldlt_in_place_unblocked (Eigen::MatrixBase< Derived > &a, SignMatrix &sign)
 
template<typename Derived>
bool proxsuite::nlp::linalg::backend::dense_ldlt_in_place (Eigen::MatrixBase< Derived > &a, SignMatrix &sign)
 
template<typename MatDerived, typename Rhs>
bool proxsuite::nlp::linalg::backend::dense_ldlt_solve_in_place (MatDerived &mat, Rhs &b)
 
template<typename Scalar>
void proxsuite::nlp::linalg::backend::dense_ldlt_reconstruct (typename math_types< Scalar >::ConstMatrixRef const &mat, typename math_types< Scalar >::MatrixRef res)
 

Detailed Description

Routines for Cholesky factorization.

Author
Sarah El-Kazdadi
Wilson Jallet

Definition in file dense.hpp.

Function Documentation

◆ update_sign_matrix()

template<typename Scalar>
void proxsuite::nlp::linalg::backend::update_sign_matrix ( SignMatrix & sign,
const Scalar & akk )

Definition at line 19 of file dense.hpp.

◆ ldlt_in_place_unblocked()

template<typename Derived>
bool proxsuite::nlp::linalg::backend::ldlt_in_place_unblocked ( Eigen::MatrixBase< Derived > & a,
SignMatrix & sign )
inline

At the end of the execution,

Parameters
acontains the lower-triangular matrix \(L\) in the LDLT decomposition. More precisely: a stores L -sans its diagonal which is all ones. The diagonal of
acontains the diagonal matrix \(D\).

Definition at line 45 of file dense.hpp.

◆ dense_ldlt_in_place()

template<typename Derived>
bool proxsuite::nlp::linalg::backend::dense_ldlt_in_place ( Eigen::MatrixBase< Derived > & a,
SignMatrix & sign )
inline

A recursive, in-place implementation of the LDLT decomposition. To be applied to dense blocks.

Definition at line 94 of file dense.hpp.

◆ dense_ldlt_solve_in_place()

template<typename MatDerived, typename Rhs>
bool proxsuite::nlp::linalg::backend::dense_ldlt_solve_in_place ( MatDerived & mat,
Rhs & b )
inline

Taking the decomposed LDLT matrix

Parameters
mat,solvethe original linear system.

Definition at line 129 of file dense.hpp.

◆ dense_ldlt_reconstruct()

template<typename Scalar>
void proxsuite::nlp::linalg::backend::dense_ldlt_reconstruct ( typename math_types< Scalar >::ConstMatrixRef const & mat,
typename math_types< Scalar >::MatrixRef res )
inline

Definition at line 150 of file dense.hpp.