ldlt-allocator.hpp File Reference

Utility function to allocate an LDLT solver for the Newton iterations. More...

#include "proxsuite-nlp/linalg/block-ldlt.hpp"
#include "proxsuite-nlp/linalg/bunchkaufman.hpp"
#include <boost/variant.hpp>
#include <array>

Include dependency graph for ldlt-allocator.hpp:

This graph shows which files directly or indirectly include this file:

Go to the source code of this file.

Classes
struct	proxsuite::nlp::ComputeSignatureVisitor

Namespaces
namespace	proxsuite
	Main package namespace.

Typedefs
template<typename Scalar, class MatrixType = typename math_types<Scalar>::MatrixXs>
using	proxsuite::nlp::LDLTVariant
using	proxsuite::nlp::isize

Enumerations
enum class	proxsuite::nlp::LDLTChoice {   DENSE , BUNCHKAUFMAN , BLOCKSPARSE , EIGEN ,   PROXSUITE }

Functions
linalg::SymbolicBlockMatrix	proxsuite::nlp::create_default_block_structure (const std::vector< isize > &dims_primal, const std::vector< isize > &dims_dual)
isize	proxsuite::nlp::get_total_dim_helper (const std::vector< isize > &nprims, const std::vector< isize > &nduals)
template<typename Scalar>
LDLTVariant< Scalar >	proxsuite::nlp::allocate_ldlt_from_sizes (const std::vector< isize > &nprims, const std::vector< isize > &nduals, LDLTChoice choice)
std::array< int, 3 >	proxsuite::nlp::computeInertiaTuple (const Eigen::Ref< Eigen::VectorXi const > &signature)

Detailed Description

Utility function to allocate an LDLT solver for the Newton iterations.

Copyright

Copyright (C) 2022 LAAS-CNRS, INRIA

Definition in file ldlt-allocator.hpp.

Typedef Documentation

LDLTVariant

template<typename Scalar, class MatrixType = typename math_types<Scalar>::MatrixXs>

using proxsuite::nlp::LDLTVariant

Initial value:

 
    boost::variant<linalg::DenseLDLT<Scalar>, linalg::BlockLDLT<Scalar>,
                   Eigen::LDLT<MatrixType>, BunchKaufman<MatrixType>
 
 
 
 
                   >

Definition at line 37 of file ldlt-allocator.hpp.

isize

using proxsuite::nlp::linalg::isize

Definition at line 18 of file block-kind.hpp.

Enumeration Type Documentation

LDLTChoice

enum class proxsuite::nlp::LDLTChoice

strong

Enumerator
DENSE	Use our dense LDLT.
BUNCHKAUFMAN	Use Bunch-Kaufman factorization.
BLOCKSPARSE	Use blocked LDLT.
EIGEN	Use Eigen's implementation.
PROXSUITE	Use Proxsuite's LDLT.

Definition at line 22 of file ldlt-allocator.hpp.

Function Documentation

create_default_block_structure()

linalg::SymbolicBlockMatrix proxsuite::nlp::create_default_block_structure ( const std::vector< isize > & dims_primal, const std::vector< isize > & dims_dual )

inline

Definition at line 47 of file ldlt-allocator.hpp.

get_total_dim_helper()

isize proxsuite::nlp::get_total_dim_helper ( const std::vector< isize > & nprims, const std::vector< isize > & nduals )

inline

Definition at line 95 of file ldlt-allocator.hpp.

allocate_ldlt_from_sizes()

template<typename Scalar>

LDLTVariant< Scalar > proxsuite::nlp::allocate_ldlt_from_sizes	(	const std::vector< isize > &	nprims,
		const std::vector< isize > &	nduals,
		LDLTChoice	choice )

Definition at line 102 of file ldlt-allocator.hpp.

computeInertiaTuple()

std::array< int, 3 > proxsuite::nlp::computeInertiaTuple ( const Eigen::Ref< Eigen::VectorXi const > & signature )

inline

Definition at line 197 of file ldlt-allocator.hpp.