Mixing class can mixing different steps of data to solver the iteration problem. For equation x = f(x), we can use iterative process to solve it: x_{n+1} = f(x_n), n = 0, 1, 2, ... To acceralte the convergence, we can use mixing method, we need information including x_in, x_out=f(x_in) in each iterative step. More...

#include <mixing.h>

Inheritance diagram for Base_Mixing::Mixing:

Collaboration diagram for Base_Mixing::Mixing:

Public Member Functions
	Mixing ()

virtual	~Mixing ()

virtual void	init_mixing_data (Mixing_Data &mdata, const int &length, const size_t &type_size) const
	init mixing data

virtual void	push_data (Mixing_Data &mdata, const double data_in, const double data_out, std::function< void(double )> screen, std::function< void(double , const double , const double )> mix, const bool &need_calcoef)=0

virtual void	push_data (Mixing_Data &mdata, const std::complex< double > data_in, const std::complex< double > data_out, std::function< void(std::complex< double > )> screen, std::function< void(std::complex< double > , const std::complex< double > , const std::complex< double > )> mix, const bool &need_calcoef)=0

virtual void	push_data (Mixing_Data &mdata, const double data_in, const double data_out, std::function< void(double *)> screen, const bool &need_calcoef)

virtual void	push_data (Mixing_Data &mdata, const std::complex< double > data_in, const std::complex< double > data_out, std::function< void(std::complex< double > *)> screen, const bool &need_calcoef)

virtual void	cal_coef (const Mixing_Data &mdata, std::function< double(double , double )> inner_product)=0
	calculate coeficients for mixing

virtual void	cal_coef (const Mixing_Data &mdata, std::function< double(std::complex< double > , std::complex< double > )> inner_product)=0

void	mix_data (const Mixing_Data &mdata, double *data_mix)
	calculate the mixing data

void	mix_data (const Mixing_Data &mdata, std::complex< double > *data_mix)

virtual void	reset ()=0
	reset mixing

Public Attributes
double	mixing_beta = 0.7

std::vector< double >	coef

int	data_ndim = 1

Detailed Description

Mixing class can mixing different steps of data to solver the iteration problem. For equation x = f(x), we can use iterative process to solve it: x_{n+1} = f(x_n), n = 0, 1, 2, ... To acceralte the convergence, we can use mixing method, we need information including x_in, x_out=f(x_in) in each iterative step.

Constructor & Destructor Documentation

◆ Mixing()

Base_Mixing::Mixing::Mixing ( )

inline

◆ ~Mixing()

virtual Base_Mixing::Mixing::~Mixing ( )

inlinevirtual

Member Function Documentation

◆ cal_coef() [1/2]

virtual void Base_Mixing::Mixing::cal_coef	(	const Mixing_Data &	mdata,
		std::function< double(double , double )>	inner_product
	)

pure virtual

calculate coeficients for mixing

Parameters

mdata	Mixing_Data
inner_product	pointer to the inner dot function

Implemented in Base_Mixing::Broyden_Mixing, Base_Mixing::Plain_Mixing, and Base_Mixing::Pulay_Mixing.

Here is the caller graph for this function:

◆ cal_coef() [2/2]

virtual void Base_Mixing::Mixing::cal_coef	(	const Mixing_Data &	mdata,
		std::function< double(std::complex< double > , std::complex< double > )>	inner_product
	)

pure virtual

Implemented in Base_Mixing::Broyden_Mixing, Base_Mixing::Plain_Mixing, and Base_Mixing::Pulay_Mixing.

◆ init_mixing_data()

virtual void Base_Mixing::Mixing::init_mixing_data	(	Mixing_Data &	mdata,
		const int &	length,
		const size_t &	type_size
	)		const

inlinevirtual

init mixing data

Parameters

mdata	mixing data
length	the length of each vector
type_size	size of type

Here is the call graph for this function:

Here is the caller graph for this function:

◆ mix_data() [1/2]

void Base_Mixing::Mixing::mix_data	(	const Mixing_Data &	mdata,
		double *	data_mix
	)

calculate the mixing data

Parameters

mdata	Mixing_Data
data_mix	output data

Here is the caller graph for this function:

◆ mix_data() [2/2]

void Base_Mixing::Mixing::mix_data	(	const Mixing_Data &	mdata,
		std::complex< double > *	data_mix
	)

◆ push_data() [1/4]

void Base_Mixing::Mixing::push_data	(	Mixing_Data &	mdata,
		const double *	data_in,
		const double *	data_out,
		std::function< void(double *)>	screen,
		const bool &	need_calcoef
	)

virtual

Parameters

mdata	store information of this iterative step
data_in	x_in
data_out	x_out = f(x_in)
screen	pointer to the screen function for Ker-Ker mixing
need_calcoef	whether need to calculate the coef

Reimplemented in Base_Mixing::Broyden_Mixing, and Base_Mixing::Plain_Mixing.

Here is the call graph for this function:

◆ push_data() [2/4]

virtual void Base_Mixing::Mixing::push_data	(	Mixing_Data &	mdata,
		const double *	data_in,
		const double *	data_out,
		std::function< void(double *)>	screen,
		std::function< void(double , const double , const double *)>	mix,
		const bool &	need_calcoef
	)

pure virtual

Parameters

mdata	store information of this iterative step
data_in	x_in
data_out	x_out = f(x_in)
screen	pointer to the screen function for Ker-Ker mixing
mix	(double* out, const double* in, const double* residual) calculate 'out' with 'in' and residual
need_calcoef	whether need to calculate the coef

Implemented in Base_Mixing::Broyden_Mixing, Base_Mixing::Plain_Mixing, Base_Mixing::Pulay_Mixing, Base_Mixing::Broyden_Mixing, and Base_Mixing::Plain_Mixing.

Here is the caller graph for this function:

◆ push_data() [3/4]

void Base_Mixing::Mixing::push_data	(	Mixing_Data &	mdata,
		const std::complex< double > *	data_in,
		const std::complex< double > *	data_out,
		std::function< void(std::complex< double > *)>	screen,
		const bool &	need_calcoef
	)

virtual

Reimplemented in Base_Mixing::Broyden_Mixing, and Base_Mixing::Plain_Mixing.

Here is the call graph for this function:

◆ push_data() [4/4]

virtual void Base_Mixing::Mixing::push_data	(	Mixing_Data &	mdata,
		const std::complex< double > *	data_in,
		const std::complex< double > *	data_out,
		std::function< void(std::complex< double > *)>	screen,
		std::function< void(std::complex< double > , const std::complex< double > , const std::complex< double > *)>	mix,
		const bool &	need_calcoef
	)

pure virtual

Implemented in Base_Mixing::Broyden_Mixing, Base_Mixing::Plain_Mixing, Base_Mixing::Pulay_Mixing, Base_Mixing::Broyden_Mixing, and Base_Mixing::Plain_Mixing.

◆ reset()

virtual void Base_Mixing::Mixing::reset ( )

pure virtual

reset mixing

Implemented in Base_Mixing::Broyden_Mixing, Base_Mixing::Plain_Mixing, and Base_Mixing::Pulay_Mixing.

Here is the caller graph for this function:

Member Data Documentation

◆ coef

std::vector<double> Base_Mixing::Mixing::coef

◆ data_ndim

int Base_Mixing::Mixing::data_ndim = 1

◆ mixing_beta

double Base_Mixing::Mixing::mixing_beta = 0.7

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

/home/runner/work/abacus-develop/abacus-develop/source/source_base/module_mixing/mixing.h
/home/runner/work/abacus-develop/abacus-develop/source/source_base/module_mixing/mixing.cpp

Public Member Functions

Public Attributes

Detailed Description

Constructor & Destructor Documentation

◆ Mixing()

◆ ~Mixing()

Member Function Documentation

◆ cal_coef() [1/2]

◆ cal_coef() [2/2]

◆ init_mixing_data()

◆ mix_data() [1/2]

◆ mix_data() [2/2]

◆ push_data() [1/4]

◆ push_data() [2/4]

◆ push_data() [3/4]

◆ push_data() [4/4]

◆ reset()

Member Data Documentation

◆ coef

◆ data_ndim

◆ mixing_beta