num::kernel::subspace::LinearOp Struct Reference

Abstract matrix-free linear operator: y = A*x. More...

#include <subspace.hpp>

Inheritance diagram for num::kernel::subspace::LinearOp:

Public Member Functions
virtual	~LinearOp ()=default
virtual void	apply (const Vector &x, Vector &y) const =0
	y = A*x (y must be pre-allocated to the correct size)
virtual idx	rows () const noexcept=0
virtual idx	cols () const noexcept=0

Detailed Description

Abstract matrix-free linear operator: y = A*x.

Subclass this or use DenseOp / SparseOp / CallableOp / make_op. Virtual dispatch overhead is O(1) per apply(); work inside is O(n) or O(nnz), so the overhead is negligible.

Definition at line 55 of file subspace.hpp.

Constructor & Destructor Documentation

~LinearOp()

virtual num::kernel::subspace::LinearOp::~LinearOp ( )

virtualdefault

Member Function Documentation

apply()

virtual void num::kernel::subspace::LinearOp::apply ( const Vector &  x, Vector &  y  ) const

pure virtual

y = A*x (y must be pre-allocated to the correct size)

Implemented in num::kernel::subspace::DenseOp, num::kernel::subspace::SparseOp, and num::kernel::subspace::CallableOp< F >.

Referenced by num::kernel::subspace::arnoldi_step().

cols()

virtual idx num::kernel::subspace::LinearOp::cols ( ) const

pure virtualnoexcept

Implemented in num::kernel::subspace::DenseOp, num::kernel::subspace::SparseOp, and num::kernel::subspace::CallableOp< F >.

rows()

virtual idx num::kernel::subspace::LinearOp::rows ( ) const

pure virtualnoexcept

Implemented in num::kernel::subspace::DenseOp, num::kernel::subspace::SparseOp, and num::kernel::subspace::CallableOp< F >.

---

The documentation for this struct was generated from the following file:

• /home/runner/work/numerics/numerics/include/kernel/subspace.hpp