templateLapack.hpp File Reference

Declares and defines templatized wrappers for the Lapack library. More...

Declares and defines templatized wrappers for the Lapack library.

Definition in file templateLapack.hpp.

#include <complex>
#include <lapacke.h>

Go to the source code of this file.

Namespaces
namespace	mx
	The mxlib c++ namespace.

Functions
template<typename dataT>
dataT	mx::math::lamch (char CMACH)
	Determine machine parameters.
template<typename dataT>
MXLAPACK_INT	mx::math::potrf (char UPLO, MXLAPACK_INT N, dataT *A, MXLAPACK_INT LDA, MXLAPACK_INT &INFO)
	Compute the Cholesky factorization of a real symmetric positive definite matrix A.
template<typename dataT>
MXLAPACK_INT	mx::math::sytrd (char UPLO, MXLAPACK_INT N, dataT *A, MXLAPACK_INT LDA, dataT *D, dataT *E, dataT *TAU, dataT *WORK, MXLAPACK_INT LWORK, MXLAPACK_INT INFO)
	Reduce a real symmetric matrix to real symmetric tridiagonal form by an orthogonal similarity transformation.
template<typename dataT>
MXLAPACK_INT	mx::math::laed9 (dataT *D, dataT *Q, dataT *S, MXLAPACK_INT K, MXLAPACK_INT KSTART, MXLAPACK_INT KSTOP, MXLAPACK_INT N, MXLAPACK_INT LDQ, dataT RHO, dataT *DLAMDA, dataT *W, MXLAPACK_INT LDS)
	Solve selected roots of a diagonal-plus-rank-one secular equation and form its eigenvectors.
template<>
MXLAPACK_INT	mx::math::laed9< float > (float *D, float *Q, float *S, MXLAPACK_INT K, MXLAPACK_INT KSTART, MXLAPACK_INT KSTOP, MXLAPACK_INT N, MXLAPACK_INT LDQ, float RHO, float *DLAMDA, float *W, MXLAPACK_INT LDS)
	Float specialization of laed9, a wrapper for LAPACK SLAED9.
template<>
MXLAPACK_INT	mx::math::laed9< double > (double *D, double *Q, double *S, MXLAPACK_INT K, MXLAPACK_INT KSTART, MXLAPACK_INT KSTOP, MXLAPACK_INT N, MXLAPACK_INT LDQ, double RHO, double *DLAMDA, double *W, MXLAPACK_INT LDS)
	Double specialization of laed9, a wrapper for LAPACK DLAED9.
template<typename dataT>
MXLAPACK_INT	mx::math::syevr (char JOBZ, char RANGE, char UPLO, MXLAPACK_INT N, dataT *A, MXLAPACK_INT LDA, dataT VL, dataT VU, MXLAPACK_INT IL, MXLAPACK_INT IU, dataT ABSTOL, MXLAPACK_INT *M, dataT *W, dataT *Z, MXLAPACK_INT LDZ, MXLAPACK_INT *ISUPPZ, dataT *WORK, MXLAPACK_INT LWORK, MXLAPACK_INT *IWORK, MXLAPACK_INT LIWORK)
	Compute selected eigenvalues and, optionally, eigenvectors of a real symmetric matrix.
template<typename dataT>
MXLAPACK_INT	mx::math::gesvd (char JOBU, char JOBVT, MXLAPACK_INT M, MXLAPACK_INT N, dataT *A, MXLAPACK_INT LDA, dataT *S, dataT *U, MXLAPACK_INT LDU, dataT *VT, MXLAPACK_INT LDVT, dataT *WORK, MXLAPACK_INT LWORK)
	Compute the singular value decomposition (SVD) of a real matrix.
template<typename dataT>
MXLAPACK_INT	mx::math::gesdd (char JOBZ, MXLAPACK_INT M, MXLAPACK_INT N, dataT *A, MXLAPACK_INT LDA, dataT *S, dataT *U, MXLAPACK_INT LDU, dataT *VT, MXLAPACK_INT LDVT, dataT *WORK, MXLAPACK_INT LWORK, MXLAPACK_INT *IWORK)
	Compute the singular value decomposition (SVD) of a real matrix with GESDD.

Function Documentation

laed9< double >()

template<>

MXLAPACK_INT mx::math::laed9< double >	(	double *	D,
		double *	Q,
		double *	S,
		MXLAPACK_INT	K,
		MXLAPACK_INT	KSTART,
		MXLAPACK_INT	KSTOP,
		MXLAPACK_INT	N,
		MXLAPACK_INT	LDQ,
		double	RHO,
		double *	DLAMDA,
		double *	W,
		MXLAPACK_INT	LDS )

Double specialization of laed9, a wrapper for LAPACK DLAED9.

[in] leading dimension of S, at least max(1,K)

Parameters

[out]	D	N-element array of selected updated eigenvalues
[out]	Q	LDQ-by-N secular-equation workspace produced by LAPACK
[out]	S	LDS-by-K updated eigenvectors, stored column-wise
[in]	K	number of terms in the secular equation
[in]	KSTART	first updated eigenvalue, inclusive and one-based
[in]	KSTOP	last updated eigenvalue, inclusive and one-based
[in]	N	Q matrix dimension, at least K
[in]	LDQ	leading dimension of Q, at least max(1,N)
[in]	RHO	positive rank-one update weight
[in]	DLAMDA	K-element array of strictly increasing diagonal poles
[in,out]	W	K deflation-adjusted update components

Definition at line 246 of file templateLapack.cpp.

References laed9().

laed9< float >()

template<>

MXLAPACK_INT mx::math::laed9< float >	(	float *	D,
		float *	Q,
		float *	S,
		MXLAPACK_INT	K,
		MXLAPACK_INT	KSTART,
		MXLAPACK_INT	KSTOP,
		MXLAPACK_INT	N,
		MXLAPACK_INT	LDQ,
		float	RHO,
		float *	DLAMDA,
		float *	W,
		MXLAPACK_INT	LDS )

Float specialization of laed9, a wrapper for LAPACK SLAED9.

[in] leading dimension of S, at least max(1,K)

Parameters

[out]	D	N-element array of selected updated eigenvalues
[out]	Q	LDQ-by-N secular-equation workspace produced by LAPACK
[out]	S	LDS-by-K updated eigenvectors, stored column-wise
[in]	K	number of terms in the secular equation
[in]	KSTART	first updated eigenvalue, inclusive and one-based
[in]	KSTOP	last updated eigenvalue, inclusive and one-based
[in]	N	Q matrix dimension, at least K
[in]	LDQ	leading dimension of Q, at least max(1,N)
[in]	RHO	positive rank-one update weight
[in]	DLAMDA	K-element array of strictly increasing diagonal poles
[in,out]	W	K deflation-adjusted update components

Definition at line 225 of file templateLapack.cpp.

References laed9().

potrf()

template<typename dataT>

MXLAPACK_INT mx::math::potrf	(	char	UPLO,
		MXLAPACK_INT	N,
		dataT *	A,
		MXLAPACK_INT	LDA,
		MXLAPACK_INT &	INFO )

Compute the Cholesky factorization of a real symmetric positive definite matrix A.

The factorization has the form A = U**T * U, if UPLO = 'U', or A = L * L**T, if UPLO = 'L', where U is an upper triangular matrix and L is lower triangular.

Parameters

[in]	UPLO	'U' if upper triangle of A is stored, 'L' if lower triangle of A is stored.
[in]	N	The order of the matrix A, >= 0.
	A	[in.out] Symmetric matrix of dimension (LDA,N), stored as specified in UPLO. Note that the opposite half is not referenced.
[in]	LDA	The leading dimension of A.
[out]	INFO	0 on success, < 0 -INFO means the i-th argument had an illegal value, >0 the leading minor of order INFO is not positive definite, and the factorization could not be completed.