CheMPS2::ConjugateGradient Class Reference

#include <ConjugateGradient.h>

Public Member Functions
	ConjugateGradient (const int veclength_in, const double RTOL_in, const double DIAG_CUTOFF_in, const bool print_in)
	Constructor. More...
virtual	~ConjugateGradient ()
	Destructor.
char	step (double **pointers)
	The iterator to converge the ground state vector. More...
int	get_num_matvec () const
	Get the number of matrix vector multiplications which have been performed. More...

Detailed Description

Conjugate gradient class.

Author

Sebastian Wouters sebas.nosp@m.tian.nosp@m.woute.nosp@m.rs@g.nosp@m.mail..nosp@m.com

Date

January 27, 2016

The ConjugateGradient class implements the conjugate gradient algorithm to solve the symmetric linear problem
.
With precon = 1 / sqrt( diag( operator ) ), the problem is turned into

.

Definition at line 37 of file ConjugateGradient.h.

Constructor & Destructor Documentation

CheMPS2::ConjugateGradient::ConjugateGradient	(	const int	veclength_in,
		const double	RTOL_in,
		const double	DIAG_CUTOFF_in,
		const bool	print_in
	)

Constructor.

Parameters

veclength_in	Linear dimension of the symmetric matrix
RTOL_in	The tolerance for the two-norm of the residual
DIAG_CUTOFF_in	The cutoff to truncate the diagonal elements of operator
print_in	Whether or not to print

Definition at line 29 of file ConjugateGradient.cpp.

Member Function Documentation

int CheMPS2::ConjugateGradient::get_num_matvec

(

)

const

Get the number of matrix vector multiplications which have been performed.

Returns

The number of matrix vector multiplications which have been performed

Definition at line 61 of file ConjugateGradient.cpp.

Here is the caller graph for this function:

char CheMPS2::ConjugateGradient::step

(

double **

pointers

)

The iterator to converge the ground state vector.

Parameters

pointers

Array of double* of length 3 to return pointers to vectors to the caller

Returns

Instruction character. 'A' means copy the initial guess to pointers[0], the diagonal of the symmetric matrix to pointers[1], and the right-hand side of the problem to pointers[2]. 'B' means calculate pointers[1] = symmetric matrix times pointers[0]. 'C' means that the converged solution can be copied back from pointers[0], and the residual norm from pointers[1][0]. 'D' means that an error has occurred.

Definition at line 63 of file ConjugateGradient.cpp.

Here is the caller graph for this function:

---

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

• ConjugateGradient.h
• ConjugateGradient.cpp