prg_normalize_mod Module Reference

The prg_normalize module. More...

Functions/Subroutines
subroutine, public	prg_normalize (h_bml)
	Normalize a Hamiltonian matrix prior to running the SP2 algorithm. More...
subroutine, public	prg_normalize_fermi (h_bml, h1, hN, mu)
	Normalize a Hamiltonian matrix prior to running the truncated SP2 algorithm. More...
subroutine, public	prg_normalize_implicit_fermi (h_bml, cnst, mu)
	Normalize a Hamiltonian matrix prior to running the implicit fermi dirac algorithm. More...
subroutine, public	prg_gershgorinreduction (gp)
	Determine gershgorin bounds across all parts, local and distributed. More...
subroutine, public	prg_normalize_cheb (h_bml, mu, emin, emax, alpha, scaledmu)
	Normalize a Hamiltonian matrix prior to running the Chebyshev algorithm. More...

Variables
integer, parameter	dp = kind(1.0d0)

Detailed Description

The prg_normalize module.

Function/Subroutine Documentation

prg_gershgorinreduction()

subroutine, public prg_normalize_mod::prg_gershgorinreduction

(

type(graph_partitioning_t), intent(inout)

gp

)

Determine gershgorin bounds across all parts, local and distributed.

Definition at line 97 of file prg_normalize_mod.F90.

prg_normalize()

subroutine, public prg_normalize_mod::prg_normalize

(

type(bml_matrix_t), intent(inout)

h_bml

)

Normalize a Hamiltonian matrix prior to running the SP2 algorithm.

X0 = (e_max * I - H) / (e_max - e_min)

where e_max and e_min are obtained sing the Gershgorin circle theorem.

Parameters

h_bml

Input/Output Hamiltonian matrix

Definition at line 32 of file prg_normalize_mod.F90.

prg_normalize_cheb()

subroutine, public prg_normalize_mod::prg_normalize_cheb	(	type(bml_matrix_t), intent(inout)	h_bml,
		real(dp), intent(in)	mu,
		real(dp), intent(in)	emin,
		real(dp), intent(in)	emax,
		real(dp), intent(inout)	alpha,
		real(dp), intent(inout)	scaledmu
	)

Normalize a Hamiltonian matrix prior to running the Chebyshev algorithm.

X0 = 2*(H - e_min*I) / (e_max - e_min) - I

where e_max and e_min are obtained sing the Gershgorin circle theorem.

Parameters

h_bml

Input/Output Hamiltonian matrix

Definition at line 126 of file prg_normalize_mod.F90.

prg_normalize_fermi()

subroutine, public prg_normalize_mod::prg_normalize_fermi	(	type(bml_matrix_t), intent(inout)	h_bml,
		real(dp), intent(in)	h1,
		real(dp), intent(in)	hN,
		real(dp), intent(in)	mu
	)

Normalize a Hamiltonian matrix prior to running the truncated SP2 algorithm.

X0 = ((hN-mu) * I - H) / (hN - h1) or X0 = (hN*I-H0-mu*I)/(hN-h1)

where h1 and hN are scaled Gershgorin bounds.

Parameters

H_bml	Hamiltonian matrix
h1	Scaled minimum Gershgorin bound.
hN	Scaled maximum Gershgorin bound.
mu	Chemical potential

Definition at line 59 of file prg_normalize_mod.F90.

prg_normalize_implicit_fermi()

subroutine, public prg_normalize_mod::prg_normalize_implicit_fermi	(	type(bml_matrix_t), intent(inout)	h_bml,
		real(dp), intent(in)	cnst,
		real(dp), intent(in)	mu
	)

Normalize a Hamiltonian matrix prior to running the implicit fermi dirac algorithm.

X0 = 0.5*II - cnst*(H0-mu0*II) or X0 = (0.5 + cnst * mu0)*II -cnst* H0

Parameters

H_bml	Hamiltonian matrix
cnst	Constant based on beta and steps
mu	Chemical potential

Definition at line 83 of file prg_normalize_mod.F90.

Variable Documentation

dp

integer, parameter prg_normalize_mod::dp = kind(1.0d0)

private

Definition at line 15 of file prg_normalize_mod.F90.