groups fairly general SCF methods, so that modules other than qs_scf can use them too split off from qs_scf to reduce dependencies More...

Data Types
interface	combine_ks_matrices

Functions/Subroutines
subroutine, public	scf_env_density_mixing (p_mix_new, mixing_store, rho_ao, para_env, iter_delta, iter_count, diis, invert)
	perform (if requested) a density mixing

subroutine, public	eigensolver (matrix_ks_fm, mo_set, ortho, work, cholesky_method, do_level_shift, level_shift, matrix_u_fm, use_jacobi)
	Diagonalise the Kohn-Sham matrix to get a new set of MO eigen- vectors and MO eigenvalues. ks will be modified.

subroutine, public	eigensolver_dbcsr (matrix_ks, matrix_ks_fm, mo_set, ortho_dbcsr, ksbuf1, ksbuf2)
	...

subroutine, public	eigensolver_symm (matrix_ks_fm, mo_set, ortho, work, do_level_shift, level_shift, matrix_u_fm, use_jacobi, jacobi_threshold, ortho_red, work_red, matrix_ks_fm_red, matrix_u_fm_red)
	...

subroutine, public	eigensolver_simple (matrix_ks, mo_set, work, do_level_shift, level_shift, use_jacobi, jacobi_threshold)
	...

subroutine, public	cp_sm_mix (m1, m2, p_mix, delta, para_env, m3)
	Perform a mixing of the given matrixes into the first matrix m1 = m2 + p_mix (m1-m2)

Detailed Description

groups fairly general SCF methods, so that modules other than qs_scf can use them too split off from qs_scf to reduce dependencies

History

Joost VandeVondele (03.2006)
combine_ks_matrices added (05.04.06,MK)
second ROKS scheme added (15.04.06,MK)
MO occupation management moved (29.08.2008,MK)
correct_mo_eigenvalues was moved from qs_mo_types; new subroutine shift_unocc_mos (03.2016, Sergey Chulkov)

Function/Subroutine Documentation

◆ scf_env_density_mixing()

subroutine, public qs_scf_methods::scf_env_density_mixing	(	type(dbcsr_p_type), dimension(:, :), pointer	p_mix_new,
		type(mixing_storage_type), pointer	mixing_store,
		type(dbcsr_p_type), dimension(:, :), pointer	rho_ao,
		type(mp_para_env_type), pointer	para_env,
		real(kind=dp), intent(inout)	iter_delta,
		integer, intent(in)	iter_count,
		logical, intent(in), optional	diis,
		logical, intent(in), optional	invert
	)

perform (if requested) a density mixing

Parameters

p_mix_new	New density matrices
mixing_store	...
rho_ao	Density environment
para_env	...
iter_delta	...
iter_count	...
diis	...
invert	Invert mixing

History: 02.2003 created [fawzi] 08.2014 adapted for kpoints [JGH]

Author: fawzi

Definition at line 92 of file qs_scf_methods.F.

Here is the call graph for this function:

Here is the caller graph for this function:

◆ eigensolver()

subroutine, public qs_scf_methods::eigensolver	(	type(cp_fm_type), intent(in)	matrix_ks_fm,
		type(mo_set_type), intent(in)	mo_set,
		type(cp_fm_type), intent(in)	ortho,
		type(cp_fm_type), intent(in)	work,
		integer, intent(inout)	cholesky_method,
		logical, intent(in)	do_level_shift,
		real(kind=dp), intent(in)	level_shift,
		type(cp_fm_type), intent(in), optional	matrix_u_fm,
		logical, intent(in)	use_jacobi
	)

Diagonalise the Kohn-Sham matrix to get a new set of MO eigen- vectors and MO eigenvalues. ks will be modified.

Parameters

matrix_ks_fm	...
mo_set	...
ortho	...
work	...
cholesky_method	...
do_level_shift	activate the level shifting technique
level_shift	amount of shift applied (in a.u.)
matrix_u_fm	matrix U : S (overlap matrix) = U^T * U
use_jacobi	...

Date: 01.05.2001

Author: Matthias Krack

Version: 1.0

Definition at line 162 of file qs_scf_methods.F.

Here is the call graph for this function:

Here is the caller graph for this function:

◆ eigensolver_dbcsr()

subroutine, public qs_scf_methods::eigensolver_dbcsr	(	type(dbcsr_type), intent(in)	matrix_ks,
		type(cp_fm_type), intent(inout)	matrix_ks_fm,
		type(mo_set_type), intent(in)	mo_set,
		type(dbcsr_type), intent(in)	ortho_dbcsr,
		type(dbcsr_type), intent(inout)	ksbuf1,
		type(dbcsr_type), intent(inout)	ksbuf2
	)

...

Parameters

matrix_ks	...
matrix_ks_fm	...
mo_set	...
ortho_dbcsr	...
ksbuf1	...
ksbuf2	...

Definition at line 264 of file qs_scf_methods.F.

Here is the call graph for this function:

Here is the caller graph for this function:

◆ eigensolver_symm()

subroutine, public qs_scf_methods::eigensolver_symm	(	type(cp_fm_type), intent(in)	matrix_ks_fm,
		type(mo_set_type), intent(in)	mo_set,
		type(cp_fm_type), intent(in)	ortho,
		type(cp_fm_type), intent(in)	work,
		logical, intent(in)	do_level_shift,
		real(kind=dp), intent(in)	level_shift,
		type(cp_fm_type), intent(in), optional	matrix_u_fm,
		logical, intent(in)	use_jacobi,
		real(kind=dp), intent(in)	jacobi_threshold,
		type(cp_fm_type), intent(inout), optional	ortho_red,
		type(cp_fm_type), intent(inout), optional	work_red,
		type(cp_fm_type), intent(inout), optional	matrix_ks_fm_red,
		type(cp_fm_type), intent(inout), optional	matrix_u_fm_red
	)

...

Parameters

matrix_ks_fm	...
mo_set	...
ortho	...
work	...
do_level_shift	activate the level shifting technique
level_shift	amount of shift applied (in a.u.)
matrix_u_fm	matrix U : S (overlap matrix) = U^T * U
use_jacobi	...
jacobi_threshold	...
ortho_red	...
work_red	...
matrix_ks_fm_red	...
matrix_u_fm_red	...

Definition at line 328 of file qs_scf_methods.F.

Here is the call graph for this function:

Here is the caller graph for this function:

◆ eigensolver_simple()

subroutine, public qs_scf_methods::eigensolver_simple	(	type(cp_fm_type), intent(in)	matrix_ks,
		type(mo_set_type), intent(in)	mo_set,
		type(cp_fm_type), intent(in)	work,
		logical, intent(in)	do_level_shift,
		real(kind=dp), intent(in)	level_shift,
		logical, intent(in)	use_jacobi,
		real(kind=dp), intent(in)	jacobi_threshold
	)

...

Parameters

matrix_ks	...
mo_set	...
work	...
do_level_shift	activate the level shifting technique
level_shift	amount of shift applied (in a.u.)
use_jacobi	...
jacobi_threshold	...

Definition at line 425 of file qs_scf_methods.F.

Here is the call graph for this function:

Here is the caller graph for this function:

◆ cp_sm_mix()

subroutine, public qs_scf_methods::cp_sm_mix	(	type(dbcsr_type), pointer	m1,
		type(dbcsr_type), pointer	m2,
		real(kind=dp), intent(in)	p_mix,
		real(kind=dp), intent(out)	delta,
		type(mp_para_env_type), pointer	para_env,
		type(dbcsr_type), optional, pointer	m3
	)

Perform a mixing of the given matrixes into the first matrix m1 = m2 + p_mix (m1-m2)

Parameters

m1	first (new) matrix, is modified
m2	the second (old) matrix
p_mix	how much m1 is conserved (0: none, 1: all)
delta	maximum norm of m1-m2
para_env	...
m3	...

History: 02.2003 rewamped [fawzi]

Author: fawzi

Note: if you what to store the result in m2 swap m1 and m2 an use (1-pmix) as pmix para_env should be removed (embedded in matrix)

Definition at line 499 of file qs_scf_methods.F.

Here is the call graph for this function:

Here is the caller graph for this function:

Data Types

Functions/Subroutines

Detailed Description

Function/Subroutine Documentation

◆ scf_env_density_mixing()

◆ eigensolver()

◆ eigensolver_dbcsr()

◆ eigensolver_symm()

◆ eigensolver_simple()

◆ cp_sm_mix()