 |
(git:6a2e663)
|
|
(git:6a2e663)
|
Methods using the PEXSI library to calculate the density matrix and related quantities using the Kohn-Sham and overlap matrices from the linear scaling quickstep SCF environment. More...
Functions/Subroutines | |
| subroutine, public | pexsi_init_read_input (pexsi_section, pexsi_env) |
| Read CP2K input section PEXSI and pass it to the PEXSI environment. More... | |
| subroutine, public | pexsi_init_scf (ks_env, pexsi_env, template_matrix) |
| Initializations needed before SCF. More... | |
| subroutine, public | pexsi_finalize_scf (pexsi_env, mu_spin) |
| Deallocations and post-processing after SCF. More... | |
| subroutine, public | density_matrix_pexsi (pexsi_env, matrix_p, matrix_w, kTS, matrix_ks, matrix_s, nelectron_exact, mu, iscf, ispin) |
| Calculate density matrix, energy-weighted density matrix and entropic energy contribution with the DFT driver of the PEXSI library. More... | |
| subroutine, public | pexsi_set_convergence_tolerance (pexsi_env, delta_scf, eps_scf, initialize, check_convergence) |
| Set PEXSI convergence tolerance (numElectronPEXSITolerance), adapted to the convergence error of the previous SCF step. The tolerance is calculated using an initial convergence threshold (tol_nel_initial) and a target convergence threshold (tol_nel_target): numElectronPEXSITolerance(delta_scf) = alpha*delta_scf+beta where alpha and beta are chosen such that numElectronPEXSITolerance(delta_scf_0) = tol_nel_initial numElectronPEXSITolerance(eps_scf) = tol_nel_target and delta_scf_0 is the initial SCF convergence error. More... | |
| subroutine, public | pexsi_to_qs (ls_scf_env, qs_env, kTS, matrix_w) |
| Pass energy weighted density matrix and entropic energy contribution to QS environment. More... | |
Methods using the PEXSI library to calculate the density matrix and related quantities using the Kohn-Sham and overlap matrices from the linear scaling quickstep SCF environment.
| subroutine, public pexsi_methods::pexsi_init_read_input | ( | type(section_vals_type), intent(in), pointer | pexsi_section, |
| type(lib_pexsi_env), intent(inout) | pexsi_env | ||
| ) |
Read CP2K input section PEXSI and pass it to the PEXSI environment.
| pexsi_section | ... |
| pexsi_env | ... |
Definition at line 80 of file pexsi_methods.F.
| subroutine, public pexsi_methods::pexsi_init_scf | ( | type(qs_ks_env_type), pointer | ks_env, |
| type(lib_pexsi_env), intent(inout) | pexsi_env, | ||
| type(dbcsr_type), intent(in) | template_matrix | ||
| ) |
Initializations needed before SCF.
| ks_env | ... |
| pexsi_env | ... |
| template_matrix | DBCSR matrix that defines the block structure and sparsity pattern of all matrices that are sent to PEXSI |
Definition at line 164 of file pexsi_methods.F.
| subroutine, public pexsi_methods::pexsi_finalize_scf | ( | type(lib_pexsi_env), intent(inout) | pexsi_env, |
| real(kind=dp), dimension(2), intent(in) | mu_spin | ||
| ) |
Deallocations and post-processing after SCF.
| pexsi_env | ... |
| mu_spin | ... |
Definition at line 238 of file pexsi_methods.F.
| subroutine, public pexsi_methods::density_matrix_pexsi | ( | type(lib_pexsi_env), intent(inout) | pexsi_env, |
| type(dbcsr_type), intent(inout) | matrix_p, | ||
| type(dbcsr_p_type), intent(inout) | matrix_w, | ||
| real(kind=dp), intent(out) | kTS, | ||
| type(dbcsr_type), intent(in), target | matrix_ks, | ||
| type(dbcsr_type), intent(in), target | matrix_s, | ||
| integer, intent(in) | nelectron_exact, | ||
| real(kind=dp), intent(out) | mu, | ||
| integer, intent(in) | iscf, | ||
| integer, intent(in) | ispin | ||
| ) |
Calculate density matrix, energy-weighted density matrix and entropic energy contribution with the DFT driver of the PEXSI library.
| [in,out] | pexsi_env | PEXSI environment |
| [in,out] | matrix_p | density matrix returned by PEXSI |
| [in,out] | matrix_w | energy-weighted density matrix returned by PEXSI |
| [out] | kTS | entropic energy contribution returned by PEXSI |
| [in] | matrix_ks | Kohn-Sham matrix from linear scaling QS environment |
| [in] | matrix_s | overlap matrix from linear scaling QS environment |
| [in] | nelectron_exact | exact number of electrons |
| [out] | mu | chemical potential calculated by PEXSI |
| [in] | iscf | SCF step |
| [in] | ispin | Number of spin |
Definition at line 301 of file pexsi_methods.F.
| subroutine, public pexsi_methods::pexsi_set_convergence_tolerance | ( | type(lib_pexsi_env), intent(inout) | pexsi_env, |
| real(kind=dp), intent(in) | delta_scf, | ||
| real(kind=dp), intent(in) | eps_scf, | ||
| logical, intent(in) | initialize, | ||
| logical, intent(out) | check_convergence | ||
| ) |
Set PEXSI convergence tolerance (numElectronPEXSITolerance), adapted to the convergence error of the previous SCF step. The tolerance is calculated using an initial convergence threshold (tol_nel_initial) and a target convergence threshold (tol_nel_target): numElectronPEXSITolerance(delta_scf) = alpha*delta_scf+beta where alpha and beta are chosen such that numElectronPEXSITolerance(delta_scf_0) = tol_nel_initial numElectronPEXSITolerance(eps_scf) = tol_nel_target and delta_scf_0 is the initial SCF convergence error.
| pexsi_env | ... |
| delta_scf | convergence error of previous SCF step |
| eps_scf | SCF convergence criterion |
| initialize | whether or not adaptive thresholing should be initialized with delta_scf as initial convergence error |
| check_convergence | is set to .FALSE. if convergence in number of electrons will not be achieved in next SCF step |
Definition at line 535 of file pexsi_methods.F.
| subroutine, public pexsi_methods::pexsi_to_qs | ( | type(ls_scf_env_type) | ls_scf_env, |
| type(qs_environment_type), intent(inout), pointer | qs_env, | ||
| real(kind=dp), dimension(:), intent(in), optional | kTS, | ||
| type(dbcsr_p_type), dimension(:), intent(in), optional | matrix_w | ||
| ) |
Pass energy weighted density matrix and entropic energy contribution to QS environment.
| ls_scf_env | ... |
| qs_env | ... |
| kTS | ... |
| matrix_w | ... |
Definition at line 581 of file pexsi_methods.F.
1.9.1