Enumerations

Enumerations
enum class	Spectra::CompInfo {   CompInfo::Successful ,   CompInfo::NotComputed ,   CompInfo::NotConverging ,   CompInfo::NumericalIssue }
enum class	Spectra::GEigsMode {   GEigsMode::Cholesky ,   GEigsMode::RegularInverse ,   GEigsMode::ShiftInvert ,   GEigsMode::Buckling ,   GEigsMode::Cayley }
enum class	Spectra::SortRule {   SortRule::LargestMagn ,   SortRule::LargestReal ,   SortRule::LargestImag ,   SortRule::LargestAlge ,   SortRule::SmallestMagn ,   SortRule::SmallestReal ,   SortRule::SmallestImag ,   SortRule::SmallestAlge ,   SortRule::BothEnds }

Detailed Description

Enumeration types for the selection rule of eigenvalues.

Enumeration Type Documentation

CompInfo

enum class Spectra::CompInfo

strong

The enumeration to report the status of computation.

Enumerator
Successful	Computation was successful.
NotComputed	Used in eigen solvers, indicating that computation has not been conducted. Users should call the compute() member function of solvers.
NotConverging	Used in eigen solvers, indicating that some eigenvalues did not converge. The compute() function returns the number of converged eigenvalues.
NumericalIssue	Used in various matrix factorization classes, for example in Cholesky decomposition it indicates that the matrix is not positive definite.

Definition at line 17 of file CompInfo.h.

GEigsMode

enum class Spectra::GEigsMode

strong

The enumeration to specify the mode of generalized eigenvalue solver.

Enumerator
Cholesky	Using Cholesky decomposition to solve generalized eigenvalues.
RegularInverse	Regular inverse mode for generalized eigenvalue solver.
ShiftInvert	Shift-and-invert mode for generalized eigenvalue solver.
Buckling	Buckling mode for generalized eigenvalue solver.
Cayley	Cayley transformation mode for generalized eigenvalue solver.

Definition at line 17 of file GEigsMode.h.

SortRule

enum class Spectra::SortRule

strong

The enumeration of selection rules of desired eigenvalues.

Enumerator
LargestMagn	Select eigenvalues with largest magnitude. Magnitude means the absolute value for real numbers and norm for complex numbers. Applies to both symmetric and general eigen solvers.
LargestReal	Select eigenvalues with largest real part. Only for general eigen solvers.
LargestImag	Select eigenvalues with largest imaginary part (in magnitude). Only for general eigen solvers.
LargestAlge	Select eigenvalues with largest algebraic value, considering any negative sign. Only for symmetric eigen solvers.
SmallestMagn	Select eigenvalues with smallest magnitude. Applies to both symmetric and general eigen solvers.
SmallestReal	Select eigenvalues with smallest real part. Only for general eigen solvers.
SmallestImag	Select eigenvalues with smallest imaginary part (in magnitude). Only for general eigen solvers.
SmallestAlge	Select eigenvalues with smallest algebraic value. Only for symmetric eigen solvers.
BothEnds	Select eigenvalues half from each end of the spectrum. When nev is odd, compute more from the high end. Only for symmetric eigen solvers.

Definition at line 33 of file SelectionRule.h.