prg_xlbokernel_mod Module Reference

Pre-conditioned O(N) calculation of the kernel for XL-BOMD. More...

Functions/Subroutines
subroutine	invert (A, AI, N)
subroutine, public	prg_kernel_multirank_latte (KRes, KK0_bml, Res, FelTol, L, LMAX, NUMRANK, HO_bml, mu, beta, RXYZ, Box, Hubbard_U, Element_Pointer, Nr_atoms, HDIM, Max_Nr_Neigh, Coulomb_acc, nebcoul, totnebcoul, Hinxlist, S_bml, Z_bml, Nocc, Inv_bml, H1_bml, DO_bml, D1_bml, m_rec, threshold, Nr_elem, mdstep, update)
	Compute low rank approximation of (K0*J)^(-1)*K0*(q[n]-n)(for LATTE) More...
subroutine	prg_update_preconditioner (K0, v, fv, Nr_atoms, threshold)
subroutine, public	prg_kernel_multirank (KRes, KK0_bml, Res, FelTol, L, LMAX, HO_bml, mu, beta, RX, RY, RZ, LBox, Hubbard_U, Element_Type, Nr_atoms, HDIM, Max_Nr_Neigh, Coulomb_acc, TIMERATIO, nnRx, nnRy, nnRz, nrnnlist, nnType, H_INDEX_START, H_INDEX_END, S_bml, Z_bml, Nocc, Znuc, Inv_bml, H1_bml, X_bml, Y_bml, DO_bml, D1_bml, m_rec, threshold)
subroutine, public	prg_full_kernel_latte (KK, DO_bml, mu0, RXYZ, Box, Hubbard_U, Element_Pointer, Nr_atoms, HDIM, Max_Nr_Neigh, Coulomb_acc, Hinxlist, S_bml, Z_bml, Inv_bml, D1_bml, H1_bml, HO_bml, Nocc, m_rec, threshold, beta, Nr_elem, nebcoul, totnebcoul)
	Compute full inverse Jacobian of q[n]-n (for LATTE) More...
subroutine, public	prg_full_kernel (KK, DO_bml, mu0, RX, RY, RZ, LBox, Hubbard_U, Element_Type, Nr_atoms, HDIM, Max_Nr_Neigh, Coulomb_acc, TIMERATIO, nnRx, nnRy, nnRz, nrnnlist, nnType, H_INDEX_START, H_INDEX_END, S_bml, Z_bml, Inv_bml, D1_bml, H1_bml, HO_bml, Y_bml, X_bml, Nocc, Znuc, m_rec, threshold, beta, diagonal)
	Compute full inverse Jacobian of q[n]-n (for development code) More...
subroutine, public	prg_kernel_matrix_multirank (KRes, KK0_bml, Res, FelTol, L, LMAX, HO_bml, mu, beta, RX, RY, RZ, LBox, Hubbard_U, Element_Type, Nr_atoms, HDIM, Max_Nr_Neigh, Coulomb_acc, TIMERATIO, nnRx, nnRy, nnRz, nrnnlist, nnType, H_INDEX_START, H_INDEX_END, S_bml, Z_bml, Nocc, Znuc, Inv_bml, H1_bml, X_bml, Y_bml, DO_bml, D1_bml, m_rec, threshold)

Variables
integer, parameter	dp = kind(1.0d0)

Detailed Description

Pre-conditioned O(N) calculation of the kernel for XL-BOMD.

Here are subroutines implementing Niklasson's scheme for low-rank, Krylov subspace approximation of the kernel.

Function/Subroutine Documentation

invert()

subroutine prg_xlbokernel_mod::invert ( real(prec), dimension(n,n), intent(in)  A, real(prec), dimension(n,n), intent(out)  AI, integer, intent(in)  N  )

private

Definition at line 30 of file prg_xlbokernel_mod.F90.

prg_full_kernel()

subroutine, public prg_xlbokernel_mod::prg_full_kernel	(	real(prec), dimension(nr_atoms,nr_atoms), intent(out)	KK,
		type(bml_matrix_t), intent(inout)	DO_bml,
		real(prec), intent(inout)	mu0,
		real(prec), dimension(nr_atoms), intent(in)	RX,
		real(prec), dimension(nr_atoms), intent(in)	RY,
		real(prec), dimension(nr_atoms), intent(in)	RZ,
		real(prec), dimension(3), intent(in)	LBox,
		real(prec), dimension(nr_atoms), intent(in)	Hubbard_U,
		character(10), dimension(nr_atoms), intent(in)	Element_Type,
		integer, intent(in)	Nr_atoms,
		integer, intent(in)	HDIM,
		integer, intent(in)	Max_Nr_Neigh,
		real(prec), intent(in)	Coulomb_acc,
		real(prec), intent(in)	TIMERATIO,
		real(prec), dimension(nr_atoms,max_nr_neigh), intent(in)	nnRx,
		real(prec), dimension(nr_atoms,max_nr_neigh), intent(in)	nnRy,
		real(prec), dimension(nr_atoms,max_nr_neigh), intent(in)	nnRz,
		integer, dimension(nr_atoms), intent(in)	nrnnlist,
		integer, dimension(nr_atoms,max_nr_neigh), intent(in)	nnType,
		integer, dimension(nr_atoms), intent(in)	H_INDEX_START,
		integer, dimension(nr_atoms), intent(in)	H_INDEX_END,
		type(bml_matrix_t), intent(in)	S_bml,
		type(bml_matrix_t), intent(in)	Z_bml,
		type(bml_matrix_t), dimension(m_rec), intent(in)	Inv_bml,
		type(bml_matrix_t), intent(inout)	D1_bml,
		type(bml_matrix_t), intent(inout)	H1_bml,
		type(bml_matrix_t), intent(in)	HO_bml,
		type(bml_matrix_t), intent(inout)	Y_bml,
		type(bml_matrix_t), intent(inout)	X_bml,
		integer, intent(in)	Nocc,
		real(prec), dimension(nr_atoms), intent(in)	Znuc,
		integer, intent(in)	m_rec,
		real(prec), intent(in)	threshold,
		real(prec), intent(in)	beta,
		real(prec), dimension(hdim), intent(inout)	diagonal
	)

Compute full inverse Jacobian of q[n]-n (for development code)

Parameters

KK	The inverse Jacobian.
DO_bml	Orthogonalized density matrix.
mu0	The chemical potiential.
RX,RY,RZ	Nuclear coordinates.
Lbox	Box dimensions.
Hubbard_U	Hubbard U list.
Element_Type	List to keep track of elements.
Nr_atoms	The number of atoms.
HDIM	Hamiltonian matrix dimension.
Max_Nr_Neigh	Max neighbours for Ewald.
Coulomb_acc	Coulomb accuracy
TIMERATIO	Parameter for Ewald
nnRx,nnRy,nnRz	Neighbour lists.
nrnnlist	Number of neighbours list.
nnType	Refers to original order of atoms.
H_INDEX_START,H_INDEX_END	Lists to keep track of atomic positions in the Hamiltonian.
S_bml	The S matrix.
Z_bml,The	Z matrix.
Inv_bml,Inverses	generated by prg_implicit_fermi_save_inverse.
HO_bml,Orthogonalized	Hamiltonian matrix.
D1_bml,H1_bml,Y_bml,X_bml	Auxillary matrices.
Nocc	Occupation.
Znuc	List of nuclear charges.
m_rec	Number of recursion steps.
threshold	Threshold value for matrix truncation.
beta	Scaled inverse temperature.
diagonal	Auxillary vector.

Definition at line 680 of file prg_xlbokernel_mod.F90.

prg_full_kernel_latte()

subroutine, public prg_xlbokernel_mod::prg_full_kernel_latte	(	real(prec), dimension(nr_atoms,nr_atoms), intent(out)	KK,
		type(bml_matrix_t), intent(inout)	DO_bml,
		real(prec), intent(inout)	mu0,
		real(prec), dimension(3,nr_atoms), intent(in)	RXYZ,
		real(prec), dimension(3,3), intent(in)	Box,
		real(prec), dimension(nr_elem), intent(in)	Hubbard_U,
		integer, dimension(nr_atoms), intent(in)	Element_Pointer,
		integer, intent(in)	Nr_atoms,
		integer, intent(in)	HDIM,
		integer, intent(in)	Max_Nr_Neigh,
		real(prec), intent(in)	Coulomb_acc,
		integer, dimension(hdim), intent(in)	Hinxlist,
		type(bml_matrix_t), intent(in)	S_bml,
		type(bml_matrix_t), intent(in)	Z_bml,
		type(bml_matrix_t), dimension(m_rec), intent(in)	Inv_bml,
		type(bml_matrix_t), intent(inout)	D1_bml,
		type(bml_matrix_t), intent(inout)	H1_bml,
		type(bml_matrix_t), intent(in)	HO_bml,
		real(prec), intent(in)	Nocc,
		integer, intent(in)	m_rec,
		real(prec), intent(in)	threshold,
		real(prec), intent(in)	beta,
		integer, intent(in)	Nr_elem,
		integer, dimension(4,max_nr_neigh,nr_atoms), intent(in)	nebcoul,
		integer, dimension(nr_atoms), intent(in)	totnebcoul
	)

Compute full inverse Jacobian of q[n]-n (for LATTE)

Parameters

KK	The inverse Jacobian.
DO_bml	Orthogonalized density matrix.
mu0	The chemical potiential.
RXYZ	Nuclear coordinates.
Box	Box dimensions.
Hubbard_U	Hubbard U list.
Element_Pointer	List to keep track of elements.
Nr_atoms	The number of atoms.
HDIM	Hamiltonian matrix dimension.
Max_Nr_Neigh	Max neighbours for Ewald.
Coulomb_acc	Coulomb accuracy
Hinxlist	List to keep track of atomic positions in the Hamiltonian.
S_bml	The S matrix.
Z_bml,The	Z matrix.
Inv_bml,Inverses	generated by prg_implicit_fermi_save_inverse.
HO_bml,Orthogonalized	Hamiltonian matrix.
D1_bml,H1_bml,Y_bml,X_bml	Auxillary matrices.
Nocc	Occupation.
m_rec	Number of recursion steps.
threshold	Threshold value for matrix truncation.
beta	Scaled inverse temperature.
Nr_elem	Number of elements in Hubbard list.

Definition at line 511 of file prg_xlbokernel_mod.F90.

prg_kernel_matrix_multirank()

subroutine, public prg_xlbokernel_mod::prg_kernel_matrix_multirank	(	real(dp), dimension(nr_atoms), intent(out)	KRes,
		type(bml_matrix_t), intent(inout)	KK0_bml,
		real(dp), dimension(nr_atoms), intent(in)	Res,
		real(dp), intent(in)	FelTol,
		integer, intent(out)	L,
		integer, intent(in)	LMAX,
		type(bml_matrix_t), intent(in)	HO_bml,
		real(dp), intent(in)	mu,
		real(dp), intent(in)	beta,
		real(dp), dimension(nr_atoms), intent(in)	RX,
		real(dp), dimension(nr_atoms), intent(in)	RY,
		real(dp), dimension(nr_atoms), intent(in)	RZ,
		real(dp), dimension(3), intent(in)	LBox,
		real(dp), dimension(nr_atoms), intent(in)	Hubbard_U,
		character(10), dimension(nr_atoms), intent(in)	Element_Type,
		integer, intent(in)	Nr_atoms,
		integer, intent(in)	HDIM,
		integer, intent(in)	Max_Nr_Neigh,
		real(dp), intent(in)	Coulomb_acc,
		real(dp), intent(in)	TIMERATIO,
		real(dp), dimension(nr_atoms,max_nr_neigh), intent(in)	nnRx,
		real(dp), dimension(nr_atoms,max_nr_neigh), intent(in)	nnRy,
		real(dp), dimension(nr_atoms,max_nr_neigh), intent(in)	nnRz,
		integer, dimension(nr_atoms), intent(in)	nrnnlist,
		integer, dimension(nr_atoms,max_nr_neigh), intent(in)	nnType,
		integer, dimension(nr_atoms), intent(in)	H_INDEX_START,
		integer, dimension(nr_atoms), intent(in)	H_INDEX_END,
		type(bml_matrix_t), intent(in)	S_bml,
		type(bml_matrix_t), intent(in)	Z_bml,
		real(dp), intent(in)	Nocc,
		real(dp), dimension(nr_atoms), intent(in)	Znuc,
		type(bml_matrix_t), dimension(m_rec), intent(in)	Inv_bml,
		type(bml_matrix_t), intent(inout)	H1_bml,
		type(bml_matrix_t), intent(inout)	X_bml,
		type(bml_matrix_t), intent(inout)	Y_bml,
		type(bml_matrix_t), intent(inout)	DO_bml,
		type(bml_matrix_t), intent(inout)	D1_bml,
		integer, intent(in)	m_rec,
		real(dp), intent(in)	threshold
	)

Definition at line 786 of file prg_xlbokernel_mod.F90.

prg_kernel_multirank()

subroutine, public prg_xlbokernel_mod::prg_kernel_multirank	(	real(dp), dimension(nr_atoms), intent(inout)	KRes,
		type(bml_matrix_t), intent(in)	KK0_bml,
		real(dp), dimension(nr_atoms), intent(in)	Res,
		real(dp), intent(in)	FelTol,
		integer, intent(out)	L,
		integer, intent(in)	LMAX,
		type(bml_matrix_t), intent(in)	HO_bml,
		real(dp), intent(in)	mu,
		real(dp), intent(in)	beta,
		real(dp), dimension(nr_atoms), intent(in)	RX,
		real(dp), dimension(nr_atoms), intent(in)	RY,
		real(dp), dimension(nr_atoms), intent(in)	RZ,
		real(dp), dimension(3), intent(in)	LBox,
		real(dp), dimension(nr_atoms), intent(in)	Hubbard_U,
		character(10), dimension(nr_atoms), intent(in)	Element_Type,
		integer, intent(in)	Nr_atoms,
		integer, intent(in)	HDIM,
		integer, intent(in)	Max_Nr_Neigh,
		real(dp), intent(in)	Coulomb_acc,
		real(dp), intent(in)	TIMERATIO,
		real(dp), dimension(nr_atoms,max_nr_neigh), intent(in)	nnRx,
		real(dp), dimension(nr_atoms,max_nr_neigh), intent(in)	nnRy,
		real(dp), dimension(nr_atoms,max_nr_neigh), intent(in)	nnRz,
		integer, dimension(nr_atoms), intent(in)	nrnnlist,
		integer, dimension(nr_atoms,max_nr_neigh), intent(in)	nnType,
		integer, dimension(nr_atoms), intent(in)	H_INDEX_START,
		integer, dimension(nr_atoms), intent(in)	H_INDEX_END,
		type(bml_matrix_t), intent(in)	S_bml,
		type(bml_matrix_t), intent(in)	Z_bml,
		integer, intent(in)	Nocc,
		real(dp), dimension(nr_atoms), intent(in)	Znuc,
		type(bml_matrix_t), dimension(m_rec), intent(in)	Inv_bml,
		type(bml_matrix_t), intent(inout)	H1_bml,
		type(bml_matrix_t), intent(inout)	X_bml,
		type(bml_matrix_t), intent(inout)	Y_bml,
		type(bml_matrix_t), intent(inout)	DO_bml,
		type(bml_matrix_t), intent(inout)	D1_bml,
		integer, intent(in)	m_rec,
		real(dp), intent(in)	threshold
	)

Definition at line 334 of file prg_xlbokernel_mod.F90.

prg_kernel_multirank_latte()

subroutine, public prg_xlbokernel_mod::prg_kernel_multirank_latte	(	real(dp), dimension(nr_atoms), intent(inout)	KRes,
		type(bml_matrix_t), intent(inout)	KK0_bml,
		real(dp), dimension(nr_atoms), intent(in)	Res,
		real(dp), intent(in)	FelTol,
		integer, intent(out)	L,
		integer, intent(in)	LMAX,
		integer, intent(in)	NUMRANK,
		type(bml_matrix_t), intent(inout)	HO_bml,
		real(dp), intent(in)	mu,
		real(dp), intent(in)	beta,
		real(dp), dimension(3,nr_atoms), intent(in)	RXYZ,
		real(dp), dimension(3,3), intent(in)	Box,
		real(dp), dimension(nr_elem), intent(in)	Hubbard_U,
		integer, dimension(nr_atoms), intent(in)	Element_Pointer,
		integer, intent(in)	Nr_atoms,
		integer, intent(in)	HDIM,
		integer, intent(in)	Max_Nr_Neigh,
		real(dp), intent(in)	Coulomb_acc,
		integer, dimension(4,max_nr_neigh,nr_atoms), intent(in)	nebcoul,
		integer, dimension(nr_atoms), intent(in)	totnebcoul,
		integer, dimension(hdim), intent(in)	Hinxlist,
		type(bml_matrix_t), intent(inout)	S_bml,
		type(bml_matrix_t), intent(inout)	Z_bml,
		real(dp), intent(in)	Nocc,
		type(bml_matrix_t), dimension(m_rec), intent(inout)	Inv_bml,
		type(bml_matrix_t), intent(inout)	H1_bml,
		type(bml_matrix_t), intent(inout)	DO_bml,
		type(bml_matrix_t), intent(inout)	D1_bml,
		integer, intent(in)	m_rec,
		real(dp), intent(in)	threshold,
		integer, intent(in)	Nr_elem,
		integer, intent(in)	mdstep,
		integer, intent(in)	update
	)

Compute low rank approximation of (K0*J)^(-1)*K0*(q[n]-n)(for LATTE)

Parameters

KRes	The low rank approximation
KK0_bml	The pre-conditioner K0.
Res	The residual q[n]-n
FelTol	Relative error tolerance for approximation
L	Number of vectors used.
LMAX	Maximum nr of vectors to use.
NUMRANK	Nr of vectors to use.
HO_bml,Orthogonalized	Hamiltonian matrix.
mu	The chemical potiential.
beta	Scaled inverse temperature.
RXYZ	Nuclear coordinates.
Box	Box dimensions.
Hubbard_U	Hubbard U list.
Element_Pointer	List to keep track of elements.
Nr_atoms	The number of atoms.
HDIM	Hamiltonian matrix dimension.
Max_Nr_Neigh	Max neighbours for Ewald.
Coulomb_acc	Coulomb accuracy.
nebcoul	Neighbour lists.
totnebcoul	Number of neighbours list.
Hinxlist	List to keep track of atomic positions in the Hamiltonian.
S_bml	The S matrix.
Z_bml,The	Z matrix.
Nocc	Occupation.
Inv_bml,Inverses	generated by prg_implicit_fermi_save_inverse.
DO_bml,D1_bml,H1_bml,Y_bml,X_bml	Auxillary matrices.
m_rec	Number of recursion steps.
threshold	Threshold value for matrix truncation.
Nr_elem	Number of elements in Hubbard list.

Definition at line 84 of file prg_xlbokernel_mod.F90.

prg_update_preconditioner()

subroutine prg_xlbokernel_mod::prg_update_preconditioner ( type(bml_matrix_t), intent(inout)  K0, real(dp), dimension(nr_atoms), intent(in)  v, real(dp), dimension(nr_atoms), intent(in)  fv, integer, intent(in)  Nr_atoms, real(dp), intent(in)  threshold  )

private

Definition at line 265 of file prg_xlbokernel_mod.F90.

Variable Documentation

dp

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

private

Definition at line 20 of file prg_xlbokernel_mod.F90.