tridiagonal_eigensolver_cl()

template<bool need_eigenvectors = true>

void stan::math::internal::tridiagonal_eigensolver_cl ( const matrix_cl< double > &  diagonal_cl, const matrix_cl< double > &  subdiagonal_cl, matrix_cl< double > &  eigenvalues_cl, matrix_cl< double > &  eigenvectors_cl, const double  split_threshold = 1e-15  )

inline

Calculates eigenvalues and eigenvectors of a tridiagonal matrix T using MRRR algorithm.

If a subdiagonal element is close to zero compared to neighbors on diagonal, the problem can be split into smaller ones.

Parameters

	diagonal_cl	Diagonal of of T.
	subdiagonal_cl	Subdiagonal of T.
[out]	eigenvalues_cl	Eigenvlues.
[out]	eigenvectors_cl	Eigenvectors.
	split_threshold	Threshold for splitting the problem

Definition at line 452 of file mrrr.hpp.