![]() |
Octopus
|
Data Types | |
| type | hamiltonian_elec_t |
Functions/Subroutines | |
| subroutine, public | hamiltonian_elec_init (hm, namespace, space, gr, ions, ext_partners, st, theory_level, xc, mc, kpoints, need_exchange, xc_photons) |
| subroutine, public | hamiltonian_elec_end (hm) |
| subroutine | hamiltonian_elec_release_copy_owned (hm) |
| Release owned pointer targets created by hamiltonian_elec_copy. More... | |
| subroutine, public | hamiltonian_elec_copy (hm_out, hm_in) |
| Deep-copy a hamiltonian_elec_t snapshot. More... | |
| subroutine | hamiltonian_elec_copy_guard_supported (hm_in) |
| Hard-fail guard matrix for unsupported hamiltonian snapshots. More... | |
| logical function | hamiltonian_elec_hermitian (hm) |
| pure logical function | hamiltonian_elec_needs_mgga_term (hm, terms) |
| subroutine | hamiltonian_elec_span (hm, delta, emin, namespace) |
| pure logical function, public | hamiltonian_elec_inh_term (hm) |
| subroutine, public | hamiltonian_elec_set_inh (hm, st) |
| subroutine, public | hamiltonian_elec_remove_inh (hm) |
| subroutine, public | hamiltonian_elec_adjoint (hm) |
| subroutine, public | hamiltonian_elec_not_adjoint (hm) |
| subroutine | hamiltonian_elec_update (this, mesh, namespace, space, ext_partners, time) |
| (re-)build the Hamiltonian for the next application: More... | |
| subroutine, public | hamiltonian_elec_update_pot (this, mesh, accumulate) |
| Update the KS potential of the electronic Hamiltonian. More... | |
| subroutine, public | hamiltonian_elec_epot_generate (this, namespace, space, gr, ions, ext_partners, st, time) |
| real(real64) function, public | hamiltonian_elec_get_time (this) |
| pure logical function, public | hamiltonian_elec_apply_packed (this) |
| subroutine, public | zhamiltonian_elec_apply_atom (hm, namespace, space, latt, species, pos, ia, mesh, psi, vpsi) |
| subroutine, public | hamiltonian_elec_update_with_ext_pot (this, mesh, space, ext_partners, time, mu) |
| This is an extension of "hamiltonian_elec_update_pot" to be used by the CFM4 propagator. It updates the Hamiltonian by considering a weighted sum of the external potentials at times time(1) and time(2), weighted by alpha(1) and alpha(2). More... | |
| subroutine | hamiltonian_elec_build_projector_phases (this, nlpp, async) |
| Build the projector phases for a given nonlocal-pseudopotential set. More... | |
| logical function, public | hamiltonian_elec_needs_current (hm, states_are_real) |
| subroutine, public | zhamiltonian_elec_apply_all (hm, namespace, gr, st, hst) |
| logical function, public | hamiltonian_elec_has_kick (hm) |
| subroutine | hamiltonian_elec_set_mass (this, namespace, mass) |
| set the effective electron mass, checking whether it was previously redefined. More... | |
| subroutine, public | hamiltonian_elec_diagonal (hm, mesh, diag, ik) |
| subroutine | dhamiltonian_elec_apply (hm, namespace, mesh, psib, hpsib, terms, set_bc) |
| subroutine, public | dhamiltonian_elec_apply_batch (hm, namespace, mesh, psib, hpsib, terms, set_bc, set_phase, spsib) |
| subroutine, public | dhamiltonian_elec_apply_overlap_batch (hm, namespace, mesh, psib, spsib, set_bc, inverse, set_phase) |
| Apply the generalized ultrasoft overlap operator S (or its inverse) to a batch. More... | |
| subroutine, public | dhamiltonian_elec_apply_sinvh_batch (hm, namespace, mesh, psib, hspsib, terms, set_bc, set_phase) |
| Apply the operator A = S^{-1} H to a batch (for ultrasoft/PAW). More... | |
| subroutine | dhamiltonian_elec_external (this, mesh, psib, vpsib) |
| subroutine, public | dhamiltonian_elec_apply_single (hm, namespace, mesh, psi, hpsi, ist, ik, terms, set_bc, set_phase) |
| subroutine, public | dhamiltonian_elec_apply_sinvh_single (hm, namespace, mesh, psi, hspsi, ist, ik, terms, set_bc, set_phase) |
| Apply A = S^{-1} H to a single state (wrapper around the batch routine). More... | |
| subroutine, public | dhamiltonian_elec_apply_overlap_single (hm, namespace, mesh, psi, spsi, ist, ik, set_bc, set_phase, inverse) |
| Apply the overlap S (or its inverse) to a single state (wrapper around the batch routine). More... | |
| subroutine | dh_mgga_terms (hm, mesh, psib, hpsib, lapl_psib) |
| Computes the orbital-dependent part of the mGGA potential. More... | |
| subroutine, public | dvmask (mesh, hm, st) |
| subroutine | zhamiltonian_elec_apply (hm, namespace, mesh, psib, hpsib, terms, set_bc) |
| subroutine, public | zhamiltonian_elec_apply_batch (hm, namespace, mesh, psib, hpsib, terms, set_bc, set_phase, spsib) |
| subroutine, public | zhamiltonian_elec_apply_overlap_batch (hm, namespace, mesh, psib, spsib, set_bc, inverse, set_phase) |
| Apply the generalized ultrasoft overlap operator S (or its inverse) to a batch. More... | |
| subroutine, public | zhamiltonian_elec_apply_sinvh_batch (hm, namespace, mesh, psib, hspsib, terms, set_bc, set_phase) |
| Apply the operator A = S^{-1} H to a batch (for ultrasoft/PAW). More... | |
| subroutine, public | zhamiltonian_elec_external (this, mesh, psib, vpsib) |
| subroutine, public | zhamiltonian_elec_apply_single (hm, namespace, mesh, psi, hpsi, ist, ik, terms, set_bc, set_phase) |
| subroutine, public | zhamiltonian_elec_apply_sinvh_single (hm, namespace, mesh, psi, hspsi, ist, ik, terms, set_bc, set_phase) |
| Apply A = S^{-1} H to a single state (wrapper around the batch routine). More... | |
| subroutine, public | zhamiltonian_elec_apply_overlap_single (hm, namespace, mesh, psi, spsi, ist, ik, set_bc, set_phase, inverse) |
| Apply the overlap S (or its inverse) to a single state (wrapper around the batch routine). More... | |
| subroutine | zh_mgga_terms (hm, mesh, psib, hpsib, lapl_psib) |
| Computes the orbital-dependent part of the mGGA potential. More... | |
| subroutine, public | zvmask (mesh, hm, st) |
Variables | |
| integer, parameter, public | length = 1 |
| integer, parameter, public | velocity = 2 |
| subroutine, public hamiltonian_elec_oct_m::hamiltonian_elec_init | ( | type(hamiltonian_elec_t), intent(inout), target | hm, |
| type(namespace_t), intent(in) | namespace, | ||
| class(space_t), intent(in) | space, | ||
| type(grid_t), intent(inout), target | gr, | ||
| type(ions_t), intent(inout), target | ions, | ||
| type(partner_list_t), intent(inout) | ext_partners, | ||
| type(states_elec_t), intent(inout), target | st, | ||
| integer, intent(in) | theory_level, | ||
| type(xc_t), intent(in), target | xc, | ||
| type(multicomm_t), intent(in) | mc, | ||
| type(kpoints_t), intent(in), target | kpoints, | ||
| logical, intent(in), optional | need_exchange, | ||
| type(xc_photons_t), intent(in), optional, target | xc_photons | ||
| ) |
Definition at line 315 of file hamiltonian_elec.F90.
| subroutine, public hamiltonian_elec_oct_m::hamiltonian_elec_end | ( | type(hamiltonian_elec_t), intent(inout), target | hm | ) |
Definition at line 800 of file hamiltonian_elec.F90.
|
private |
Release owned pointer targets created by hamiltonian_elec_copy.
This avoids running the full hamiltonian finalizer on shallow-alias snapshots.
Definition at line 908 of file hamiltonian_elec.F90.
| subroutine, public hamiltonian_elec_oct_m::hamiltonian_elec_copy | ( | type(hamiltonian_elec_t), intent(out), target | hm_out, |
| type(hamiltonian_elec_t), intent(inout) | hm_in | ||
| ) |
Deep-copy a hamiltonian_elec_t snapshot.
Pointer semantics:
ions is cloned and owned by the copy.der, xc, xc_photons, kpoints) are aliased.zora is cloned when present. Definition at line 959 of file hamiltonian_elec.F90.
|
private |
Hard-fail guard matrix for unsupported hamiltonian snapshots.
Definition at line 1089 of file hamiltonian_elec.F90.
|
private |
Definition at line 1130 of file hamiltonian_elec.F90.
|
private |
Definition at line 1142 of file hamiltonian_elec.F90.
|
private |
Definition at line 1163 of file hamiltonian_elec.F90.
| pure logical function, public hamiltonian_elec_oct_m::hamiltonian_elec_inh_term | ( | type(hamiltonian_elec_t), intent(in) | hm | ) |
Definition at line 1185 of file hamiltonian_elec.F90.
| subroutine, public hamiltonian_elec_oct_m::hamiltonian_elec_set_inh | ( | type(hamiltonian_elec_t), intent(inout) | hm, |
| type(states_elec_t), intent(in) | st | ||
| ) |
Definition at line 1193 of file hamiltonian_elec.F90.
| subroutine, public hamiltonian_elec_oct_m::hamiltonian_elec_remove_inh | ( | type(hamiltonian_elec_t), intent(inout) | hm | ) |
Definition at line 1208 of file hamiltonian_elec.F90.
| subroutine, public hamiltonian_elec_oct_m::hamiltonian_elec_adjoint | ( | type(hamiltonian_elec_t), intent(inout) | hm | ) |
Definition at line 1222 of file hamiltonian_elec.F90.
| subroutine, public hamiltonian_elec_oct_m::hamiltonian_elec_not_adjoint | ( | type(hamiltonian_elec_t), intent(inout) | hm | ) |
Definition at line 1239 of file hamiltonian_elec.F90.
|
private |
(re-)build the Hamiltonian for the next application:
Definition at line 1257 of file hamiltonian_elec.F90.
| subroutine, public hamiltonian_elec_oct_m::hamiltonian_elec_update_pot | ( | type(hamiltonian_elec_t), intent(inout) | this, |
| class(mesh_t), intent(in) | mesh, | ||
| logical, intent(in), optional | accumulate | ||
| ) |
Update the KS potential of the electronic Hamiltonian.
Adds PCM contributions
Adds possible absorbing potential
Definition at line 1458 of file hamiltonian_elec.F90.
| subroutine, public hamiltonian_elec_oct_m::hamiltonian_elec_epot_generate | ( | type(hamiltonian_elec_t), intent(inout), target | this, |
| type(namespace_t), intent(in) | namespace, | ||
| class(electron_space_t), intent(in) | space, | ||
| type(grid_t), intent(in) | gr, | ||
| type(ions_t), intent(inout), target | ions, | ||
| type(partner_list_t), intent(in) | ext_partners, | ||
| type(states_elec_t), intent(inout) | st, | ||
| real(real64), intent(in), optional | time | ||
| ) |
Generates the real-space PCM potential due to nuclei which do not change during the SCF calculation.
Local field effects due to static electrostatic potentials (if they were). The laser and the kick are included in subroutine v_ks_hartree (module v_ks).
Definition at line 1585 of file hamiltonian_elec.F90.
| real(real64) function, public hamiltonian_elec_oct_m::hamiltonian_elec_get_time | ( | type(hamiltonian_elec_t), intent(inout) | this | ) |
Definition at line 1690 of file hamiltonian_elec.F90.
| pure logical function, public hamiltonian_elec_oct_m::hamiltonian_elec_apply_packed | ( | class(hamiltonian_elec_t), intent(in) | this | ) |
Definition at line 1698 of file hamiltonian_elec.F90.
| subroutine, public hamiltonian_elec_oct_m::zhamiltonian_elec_apply_atom | ( | type(hamiltonian_elec_t), intent(in) | hm, |
| type(namespace_t), intent(in) | namespace, | ||
| class(space_t), intent(in) | space, | ||
| type(lattice_vectors_t), intent(in) | latt, | ||
| class(species_t), intent(in) | species, | ||
| real(real64), dimension(1:space%dim), intent(in) | pos, | ||
| integer, intent(in) | ia, | ||
| class(mesh_t), intent(in) | mesh, | ||
| complex(real64), dimension(:,:), intent(in), contiguous | psi, | ||
| complex(real64), dimension(:,:), intent(out), contiguous | vpsi | ||
| ) |
| [in] | psi | (grnp_part, hmddim) |
| [out] | vpsi | (grnp, hmddim) |
Definition at line 1707 of file hamiltonian_elec.F90.
| subroutine, public hamiltonian_elec_oct_m::hamiltonian_elec_update_with_ext_pot | ( | type(hamiltonian_elec_t), intent(inout) | this, |
| class(mesh_t), intent(in) | mesh, | ||
| class(space_t), intent(in) | space, | ||
| type(partner_list_t), intent(in) | ext_partners, | ||
| real(real64), dimension(1:2), intent(in) | time, | ||
| real(real64), dimension(1:2), intent(in) | mu | ||
| ) |
This is an extension of "hamiltonian_elec_update_pot" to be used by the CFM4 propagator. It updates the Hamiltonian by considering a weighted sum of the external potentials at times time(1) and time(2), weighted by alpha(1) and alpha(2).
Definition at line 1745 of file hamiltonian_elec.F90.
|
private |
Build the projector phases for a given nonlocal-pseudopotential set.
Used for both the standard nonlocal potential (vnl) and, for ultrasoft/PAW, the generalized overlap operator (aug_overlap), which both project onto the same beta functions and therefore share the per-atom phase factors.
Definition at line 1918 of file hamiltonian_elec.F90.
| logical function, public hamiltonian_elec_oct_m::hamiltonian_elec_needs_current | ( | type(hamiltonian_elec_t), intent(in) | hm, |
| logical, intent(in) | states_are_real | ||
| ) |
Definition at line 1971 of file hamiltonian_elec.F90.
| subroutine, public hamiltonian_elec_oct_m::zhamiltonian_elec_apply_all | ( | type(hamiltonian_elec_t), intent(inout) | hm, |
| type(namespace_t), intent(in) | namespace, | ||
| type(grid_t), intent(in) | gr, | ||
| type(states_elec_t), intent(inout) | st, | ||
| type(states_elec_t), intent(inout) | hst | ||
| ) |
Definition at line 1989 of file hamiltonian_elec.F90.
| logical function, public hamiltonian_elec_oct_m::hamiltonian_elec_has_kick | ( | type(hamiltonian_elec_t), intent(in) | hm | ) |
Definition at line 2036 of file hamiltonian_elec.F90.
|
private |
set the effective electron mass, checking whether it was previously redefined.
Definition at line 2048 of file hamiltonian_elec.F90.
| subroutine, public hamiltonian_elec_oct_m::hamiltonian_elec_diagonal | ( | type(hamiltonian_elec_t), intent(in) | hm, |
| class(mesh_t), intent(in) | mesh, | ||
| real(real64), dimension(:,:), intent(out), contiguous | diag, | ||
| integer, intent(in) | ik | ||
| ) |
| [out] | diag | hpsi(grmeshnp, hmdnspin) |
Definition at line 2066 of file hamiltonian_elec.F90.
|
private |
| [in] | set_bc | If set to .false. the boundary conditions are assumed to be set previously. |
Definition at line 2186 of file hamiltonian_elec.F90.
| subroutine, public hamiltonian_elec_oct_m::dhamiltonian_elec_apply_batch | ( | type(hamiltonian_elec_t), intent(in) | hm, |
| type(namespace_t), intent(in) | namespace, | ||
| class(mesh_t), intent(in) | mesh, | ||
| type(wfs_elec_t), intent(inout), target | psib, | ||
| type(wfs_elec_t), intent(inout), target | hpsib, | ||
| integer, intent(in), optional | terms, | ||
| logical, intent(in), optional | set_bc, | ||
| logical, intent(in), optional | set_phase, | ||
| type(wfs_elec_t), intent(inout), optional | spsib | ||
| ) |
| [in] | set_bc | If set to .false. the boundary conditions are assumed to be set previously. |
| [in] | set_phase | override of whether to apply the phase (default: .not. psibhas_phase) |
| [in,out] | spsib | if present (generalized overlap only), also return S|psib>, reusing the V_NL projections <beta|psi>: the fused H+S apply costs one projection pass instead of two. On GPUs, S is applied separately. |
Definition at line 2215 of file hamiltonian_elec.F90.
| subroutine, public hamiltonian_elec_oct_m::dhamiltonian_elec_apply_overlap_batch | ( | type(hamiltonian_elec_t), intent(in) | hm, |
| type(namespace_t), intent(in) | namespace, | ||
| class(mesh_t), intent(in) | mesh, | ||
| type(wfs_elec_t), intent(inout), target | psib, | ||
| type(wfs_elec_t), intent(inout), optional, target | spsib, | ||
| logical, intent(in), optional | set_bc, | ||
| logical, intent(in), optional | inverse, | ||
| logical, intent(in), optional | set_phase | ||
| ) |
Apply the generalized ultrasoft overlap operator S (or its inverse) to a batch.
With spsib present the result is written there (spsib = S psib, the identity part copied first); WITHOUT spsib the overlap is applied IN PLACE (psib := S psib), which saves the identity copy and lets the caller skip a scratch batch when the input is not needed afterwards (e.g. S^{-1} on H|psi> inside SinvH). If the Hamiltonian does not require a generalized overlap, this reduces to a plain copy (or a no-op in place). With inverse=.true. the operator applied is S^{-1} instead of S.
| [in,out] | spsib | result S psib; when absent, in place |
| [in] | set_bc | If set to .false. the boundary conditions are assumed to be set previously. |
| [in] | inverse | If .true., apply S^{-1} instead of S. |
| [in] | set_phase | override of whether to apply the phase (default: .not. psibhas_phase) |
Definition at line 2479 of file hamiltonian_elec.F90.
| subroutine, public hamiltonian_elec_oct_m::dhamiltonian_elec_apply_sinvh_batch | ( | type(hamiltonian_elec_t), intent(in) | hm, |
| type(namespace_t), intent(in) | namespace, | ||
| class(mesh_t), intent(in) | mesh, | ||
| type(wfs_elec_t), intent(inout) | psib, | ||
| type(wfs_elec_t), intent(inout) | hspsib, | ||
| integer, intent(in), optional | terms, | ||
| logical, intent(in), optional | set_bc, | ||
| logical, intent(in), optional | set_phase | ||
| ) |
Apply the operator A = S^{-1} H to a batch (for ultrasoft/PAW).
This is the operator whose lowest eigenpairs are the generalized eigenpairs of (H, S). It is the operator that the Chebyshev filter must act with: filtering H alone targets the wrong subspace when S deviates from the identity. When the Hamiltonian does not require a generalized overlap, S^{-1} = 1 and this reduces to applying H.
| [in,out] | psib | input batch |
| [in,out] | hspsib | result = S^{-1} H psib |
Definition at line 2561 of file hamiltonian_elec.F90.
|
private |
Definition at line 2589 of file hamiltonian_elec.F90.
| subroutine, public hamiltonian_elec_oct_m::dhamiltonian_elec_apply_single | ( | type(hamiltonian_elec_t), intent(in) | hm, |
| type(namespace_t), intent(in) | namespace, | ||
| class(mesh_t), intent(in) | mesh, | ||
| real(real64), dimension(:,:), intent(inout), target, contiguous | psi, | ||
| real(real64), dimension(:,:), intent(inout), target, contiguous | hpsi, | ||
| integer, intent(in) | ist, | ||
| integer, intent(in) | ik, | ||
| integer, intent(in), optional | terms, | ||
| logical, intent(in), optional | set_bc, | ||
| logical, intent(in), optional | set_phase | ||
| ) |
| [in] | ist | the index of the state |
| [in] | ik | the index of the k-point |
| [in,out] | psi | (grmeshnp_part, hmddim) |
| [in,out] | hpsi | (grmeshnp, hmddim) |
Definition at line 2638 of file hamiltonian_elec.F90.
| subroutine, public hamiltonian_elec_oct_m::dhamiltonian_elec_apply_sinvh_single | ( | type(hamiltonian_elec_t), intent(in) | hm, |
| type(namespace_t), intent(in) | namespace, | ||
| class(mesh_t), intent(in) | mesh, | ||
| real(real64), dimension(:,:), intent(inout), target, contiguous | psi, | ||
| real(real64), dimension(:,:), intent(inout), target, contiguous | hspsi, | ||
| integer, intent(in) | ist, | ||
| integer, intent(in) | ik, | ||
| integer, intent(in), optional | terms, | ||
| logical, intent(in), optional | set_bc, | ||
| logical, intent(in), optional | set_phase | ||
| ) |
Apply A = S^{-1} H to a single state (wrapper around the batch routine).
| [in,out] | psi | (meshnp_part, hmddim) |
| [in,out] | hspsi | (meshnp, hmddim) |
Definition at line 2673 of file hamiltonian_elec.F90.
| subroutine, public hamiltonian_elec_oct_m::dhamiltonian_elec_apply_overlap_single | ( | type(hamiltonian_elec_t), intent(in) | hm, |
| type(namespace_t), intent(in) | namespace, | ||
| class(mesh_t), intent(in) | mesh, | ||
| real(real64), dimension(:,:), intent(inout), target, contiguous | psi, | ||
| real(real64), dimension(:,:), intent(inout), target, contiguous | spsi, | ||
| integer, intent(in) | ist, | ||
| integer, intent(in) | ik, | ||
| logical, intent(in), optional | set_bc, | ||
| logical, intent(in), optional | set_phase, | ||
| logical, intent(in), optional | inverse | ||
| ) |
Apply the overlap S (or its inverse) to a single state (wrapper around the batch routine).
| [in,out] | psi | (meshnp_part, hmddim) |
| [in,out] | spsi | (meshnp, hmddim) |
Definition at line 2707 of file hamiltonian_elec.F90.
|
private |
Computes the orbital-dependent part of the mGGA potential.
This reads as \( L\psi =-\nabla \cdot [\vtau \nabla \psi]\).
It is trivial to show that this operator is self-adjoint. However, this is never the case with finite differences. This is thus reimplemented as L\psi = -1/2(v_\tau \Delta \psi - \psi \Delta[v_\tau] + \Delta [v_\tau \psi]
The same expression was proposed in Richter et al., J. Chem. Phys. 159, 124117 (2023).
Definition at line 2751 of file hamiltonian_elec.F90.
| subroutine, public hamiltonian_elec_oct_m::dvmask | ( | class(mesh_t), intent(in) | mesh, |
| type(hamiltonian_elec_t), intent(in) | hm, | ||
| type(states_elec_t), intent(inout) | st | ||
| ) |
Definition at line 2801 of file hamiltonian_elec.F90.
|
private |
| [in] | set_bc | If set to .false. the boundary conditions are assumed to be set previously. |
Definition at line 2909 of file hamiltonian_elec.F90.
| subroutine, public hamiltonian_elec_oct_m::zhamiltonian_elec_apply_batch | ( | type(hamiltonian_elec_t), intent(in) | hm, |
| type(namespace_t), intent(in) | namespace, | ||
| class(mesh_t), intent(in) | mesh, | ||
| type(wfs_elec_t), intent(inout), target | psib, | ||
| type(wfs_elec_t), intent(inout), target | hpsib, | ||
| integer, intent(in), optional | terms, | ||
| logical, intent(in), optional | set_bc, | ||
| logical, intent(in), optional | set_phase, | ||
| type(wfs_elec_t), intent(inout), optional | spsib | ||
| ) |
| [in] | set_bc | If set to .false. the boundary conditions are assumed to be set previously. |
| [in] | set_phase | override of whether to apply the phase (default: .not. psibhas_phase) |
| [in,out] | spsib | if present (generalized overlap only), also return S|psib>, reusing the V_NL projections <beta|psi>: the fused H+S apply costs one projection pass instead of two. On GPUs, S is applied separately. |
Definition at line 2938 of file hamiltonian_elec.F90.
| subroutine, public hamiltonian_elec_oct_m::zhamiltonian_elec_apply_overlap_batch | ( | type(hamiltonian_elec_t), intent(in) | hm, |
| type(namespace_t), intent(in) | namespace, | ||
| class(mesh_t), intent(in) | mesh, | ||
| type(wfs_elec_t), intent(inout), target | psib, | ||
| type(wfs_elec_t), intent(inout), optional, target | spsib, | ||
| logical, intent(in), optional | set_bc, | ||
| logical, intent(in), optional | inverse, | ||
| logical, intent(in), optional | set_phase | ||
| ) |
Apply the generalized ultrasoft overlap operator S (or its inverse) to a batch.
With spsib present the result is written there (spsib = S psib, the identity part copied first); WITHOUT spsib the overlap is applied IN PLACE (psib := S psib), which saves the identity copy and lets the caller skip a scratch batch when the input is not needed afterwards (e.g. S^{-1} on H|psi> inside SinvH). If the Hamiltonian does not require a generalized overlap, this reduces to a plain copy (or a no-op in place). With inverse=.true. the operator applied is S^{-1} instead of S.
| [in,out] | spsib | result S psib; when absent, in place |
| [in] | set_bc | If set to .false. the boundary conditions are assumed to be set previously. |
| [in] | inverse | If .true., apply S^{-1} instead of S. |
| [in] | set_phase | override of whether to apply the phase (default: .not. psibhas_phase) |
Definition at line 3202 of file hamiltonian_elec.F90.
| subroutine, public hamiltonian_elec_oct_m::zhamiltonian_elec_apply_sinvh_batch | ( | type(hamiltonian_elec_t), intent(in) | hm, |
| type(namespace_t), intent(in) | namespace, | ||
| class(mesh_t), intent(in) | mesh, | ||
| type(wfs_elec_t), intent(inout) | psib, | ||
| type(wfs_elec_t), intent(inout) | hspsib, | ||
| integer, intent(in), optional | terms, | ||
| logical, intent(in), optional | set_bc, | ||
| logical, intent(in), optional | set_phase | ||
| ) |
Apply the operator A = S^{-1} H to a batch (for ultrasoft/PAW).
This is the operator whose lowest eigenpairs are the generalized eigenpairs of (H, S). It is the operator that the Chebyshev filter must act with: filtering H alone targets the wrong subspace when S deviates from the identity. When the Hamiltonian does not require a generalized overlap, S^{-1} = 1 and this reduces to applying H.
| [in,out] | psib | input batch |
| [in,out] | hspsib | result = S^{-1} H psib |
Definition at line 3284 of file hamiltonian_elec.F90.
| subroutine, public hamiltonian_elec_oct_m::zhamiltonian_elec_external | ( | type(hamiltonian_elec_t), intent(in) | this, |
| class(mesh_t), intent(in) | mesh, | ||
| type(wfs_elec_t), intent(in) | psib, | ||
| type(wfs_elec_t), intent(inout) | vpsib | ||
| ) |
Definition at line 3312 of file hamiltonian_elec.F90.
| subroutine, public hamiltonian_elec_oct_m::zhamiltonian_elec_apply_single | ( | type(hamiltonian_elec_t), intent(in) | hm, |
| type(namespace_t), intent(in) | namespace, | ||
| class(mesh_t), intent(in) | mesh, | ||
| complex(real64), dimension(:,:), intent(inout), target, contiguous | psi, | ||
| complex(real64), dimension(:,:), intent(inout), target, contiguous | hpsi, | ||
| integer, intent(in) | ist, | ||
| integer, intent(in) | ik, | ||
| integer, intent(in), optional | terms, | ||
| logical, intent(in), optional | set_bc, | ||
| logical, intent(in), optional | set_phase | ||
| ) |
| [in] | ist | the index of the state |
| [in] | ik | the index of the k-point |
| [in,out] | psi | (grmeshnp_part, hmddim) |
| [in,out] | hpsi | (grmeshnp, hmddim) |
Definition at line 3361 of file hamiltonian_elec.F90.
| subroutine, public hamiltonian_elec_oct_m::zhamiltonian_elec_apply_sinvh_single | ( | type(hamiltonian_elec_t), intent(in) | hm, |
| type(namespace_t), intent(in) | namespace, | ||
| class(mesh_t), intent(in) | mesh, | ||
| complex(real64), dimension(:,:), intent(inout), target, contiguous | psi, | ||
| complex(real64), dimension(:,:), intent(inout), target, contiguous | hspsi, | ||
| integer, intent(in) | ist, | ||
| integer, intent(in) | ik, | ||
| integer, intent(in), optional | terms, | ||
| logical, intent(in), optional | set_bc, | ||
| logical, intent(in), optional | set_phase | ||
| ) |
Apply A = S^{-1} H to a single state (wrapper around the batch routine).
| [in,out] | psi | (meshnp_part, hmddim) |
| [in,out] | hspsi | (meshnp, hmddim) |
Definition at line 3396 of file hamiltonian_elec.F90.
| subroutine, public hamiltonian_elec_oct_m::zhamiltonian_elec_apply_overlap_single | ( | type(hamiltonian_elec_t), intent(in) | hm, |
| type(namespace_t), intent(in) | namespace, | ||
| class(mesh_t), intent(in) | mesh, | ||
| complex(real64), dimension(:,:), intent(inout), target, contiguous | psi, | ||
| complex(real64), dimension(:,:), intent(inout), target, contiguous | spsi, | ||
| integer, intent(in) | ist, | ||
| integer, intent(in) | ik, | ||
| logical, intent(in), optional | set_bc, | ||
| logical, intent(in), optional | set_phase, | ||
| logical, intent(in), optional | inverse | ||
| ) |
Apply the overlap S (or its inverse) to a single state (wrapper around the batch routine).
| [in,out] | psi | (meshnp_part, hmddim) |
| [in,out] | spsi | (meshnp, hmddim) |
Definition at line 3430 of file hamiltonian_elec.F90.
|
private |
Computes the orbital-dependent part of the mGGA potential.
This reads as \( L\psi =-\nabla \cdot [\vtau \nabla \psi]\).
It is trivial to show that this operator is self-adjoint. However, this is never the case with finite differences. This is thus reimplemented as L\psi = -1/2(v_\tau \Delta \psi - \psi \Delta[v_\tau] + \Delta [v_\tau \psi]
The same expression was proposed in Richter et al., J. Chem. Phys. 159, 124117 (2023).
Definition at line 3474 of file hamiltonian_elec.F90.
| subroutine, public hamiltonian_elec_oct_m::zvmask | ( | class(mesh_t), intent(in) | mesh, |
| type(hamiltonian_elec_t), intent(in) | hm, | ||
| type(states_elec_t), intent(inout) | st | ||
| ) |
Definition at line 3524 of file hamiltonian_elec.F90.
| integer, parameter, public hamiltonian_elec_oct_m::length = 1 |
Definition at line 307 of file hamiltonian_elec.F90.
| integer, parameter, public hamiltonian_elec_oct_m::velocity = 2 |
Definition at line 307 of file hamiltonian_elec.F90.