MatrixBatchTridiag< ExecSpace > Class Template Reference

A structure for solving a set of independant tridiagonal systems using a direct method. More...

Inheritance diagram for MatrixBatchTridiag< ExecSpace >:

Public Member Functions
	MatrixBatchTridiag (const int batch_size, const int mat_size, DKokkosView2D const aa, DKokkosView2D const bb, DKokkosView2D const cc)
	Creates an instance of the MatrixBatchTridiag class.
bool	check_stability () const
	Check if the matrices are in the stability area of the solver.
void	setup_solver () final
	Perform a pre-process operation on the solver.
void	solve (BatchedRHS const b) const final
	Solve the batched linear problem Ax=b.
Public Member Functions inherited from MatrixBatch< ExecSpace >
virtual	~MatrixBatch ()=default
	Destruct.
std::size_t	size () const
	Get the size of the square matrix corresponding to a single batch in one of its dimensions.
std::size_t	batch_size () const
	Get the batch size of the linear problem.

Additional Inherited Members
Public Types inherited from MatrixBatch< ExecSpace >
using	BatchedRHS = Kokkos::View< double **, Kokkos::LayoutRight, ExecSpace >
	The type of a Kokkos::View storing batched right-hand sides. Second dimenion is batch dimension.
Protected Member Functions inherited from MatrixBatch< ExecSpace >
	MatrixBatch (const std::size_t batch_size, const std::size_t mat_size)
	The constructor for MatrixBatch class.

Detailed Description

template<class ExecSpace>
class MatrixBatchTridiag< ExecSpace >

A structure for solving a set of independant tridiagonal systems using a direct method.

The parallelism operates on the whole collection by dispatching to threads. Each problem is treated sequentially, by the tridiagonal matrix algorithm (TDMA). This solver is stable for tridiagonal matrices which satisfy one of the following conditions:

• Diagonally Dominant.
• Symmetric positive-definite. Diagonally Dominant property is fully checked. Only symmetry property is checked, positivity-definiteness is not.

  Template Parameters

  ExecSpace

  The execution space related to Kokkos.

Constructor & Destructor Documentation

MatrixBatchTridiag()

template<class ExecSpace >

MatrixBatchTridiag< ExecSpace >::MatrixBatchTridiag ( const int  batch_size, const int  mat_size, DKokkosView2D const  aa, DKokkosView2D const  bb, DKokkosView2D const  cc  )

inlineexplicit

Creates an instance of the MatrixBatchTridiag class.

First dimension is the batch, second one refers to matrix entries indexed by line. The entries aa,bb,cc are 2D Kokkos views and have the same dimensions. LayoutRight: means that the "last" dimension is the contiguous one. aa(batch_idx,0) and cc(batch_idx,mat_size-1) are not used for any values of batch_idx.

Parameters

[in]	batch_size	The size of the set of linear problems.
[in]	mat_size	The common size of each individual matrix .
[in]	aa	2d Kokkos View which stores subdiagonal components for all matrices.
[in]	bb	2d Kokkos View which stores diagonal components for all matrices.
[in]	cc	2d Kokkos View which stores upper diagonal components for all matrices.

Member Function Documentation

check_stability()

template<class ExecSpace >

bool MatrixBatchTridiag< ExecSpace >::check_stability ( ) const

inline

Check if the matrices are in the stability area of the solver.

It checks if each matrix of the batch has one of the following structures:

• Diagonally Dominant.
• Symmetric. If assertion fails, there is at least one of the matrices which does not verify these conditions. Positivity-definiteness is not checked, it is needed to fully achieve TDMA algorithm requirement: symmetric positive definite.

Returns

A boolean which indicates if the stability condition may be verified.

setup_solver()

template<class ExecSpace >

void MatrixBatchTridiag< ExecSpace >::setup_solver ( )

inlinefinalvirtual

Perform a pre-process operation on the solver.

Must be called after filling the matrix.

It calls check_stability function to verify if the matrices data is in range of validity of the solver.

The stopping criterion is a reduction factor ||Ax-b||/||b||<tol with max_iter maximum iterations.

Implements MatrixBatch< ExecSpace >.

solve()

template<class ExecSpace >

void MatrixBatchTridiag< ExecSpace >::solve ( BatchedRHS const  b ) const

inlinefinalvirtual

Solve the batched linear problem Ax=b.

Parameters

[in,out]

b

A 2D Kokkos::View storing the batched right-hand sides of the problem and receiving the corresponding solutions.

Implements MatrixBatch< ExecSpace >.

---

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

• /home/runner/work/gyselalibxx/gyselalibxx/code_branch/vendor/sll/include/sll/matrix_batch_tridiag.hpp