bidiagonal_2x2_decomposition Struct Template Reference

Special SVD for the 2x2 case. More...

#include <svd.h>

List of all members.

Public Types
typedef FloatType	scalar_t
Public Member Functions
	bidiagonal_2x2_decomposition (scalar_t f, scalar_t g, scalar_t h, bool compute_singular_vectors)
Public Attributes
scalar_t	s_min
scalar_t	s_max
scalar_t	c_u
scalar_t	s_u
scalar_t	c_v
scalar_t	s_v

---

Detailed Description

template<typename FloatType>
struct scitbx::matrix::svd::bidiagonal_2x2_decomposition< FloatType >

Special SVD for the 2x2 case.

Reference: LAPACK DLASV2

---

Constructor & Destructor Documentation

bidiagonal_2x2_decomposition	(	scalar_t	f,
		scalar_t	g,
		scalar_t	h,
		bool	compute_singular_vectors
	)			[inline]

Singular values of [ f g ] [ 0 h ] If the argument compute_singular_vectors is false, only the singular values are computed and they are both non-negative (making that code more or less equivalent to DLAS2 in LAPACK). Otherwise, s_max and s_min have respectively the sign of f and g.

References bidiagonal_2x2_decomposition::c_u, bidiagonal_2x2_decomposition::c_v, and bidiagonal_2x2_decomposition::s_min.

---

Member Data Documentation

scalar_t s_min

singular values: |s_min| < |s_max| U^T A V = [ s_max 0 ] [ 0 s_min ]

Referenced by bidiagonal_2x2_decomposition::bidiagonal_2x2_decomposition().

scalar_t c_u

U = [ c_u -s_u ] [ s_u c_u ]

Referenced by bidiagonal_2x2_decomposition::bidiagonal_2x2_decomposition().

scalar_t c_v

V = [ c_v -s_v ] [ s_v c_v ]

Referenced by bidiagonal_2x2_decomposition::bidiagonal_2x2_decomposition().

---

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

• scitbx/matrix/svd.h