MPI-SCATCI 2.0
An MPI version of SCATCI
Loading...
Searching...
No Matches
consts_mpi_ci Module Reference

MPI SCATCI Constants module. More...

Variables

integer, parameter mxint = 1000
 Maximal length of integer arrays in input namelists.
integer, parameter nidx = 2
 Number of long integers used to store up to 8 (packed) integral indices.
integer, parameter max_neig = 400
 Default limit on the number of eigenvalues printed.
real(wp), parameter default_archer_memory = 2.5
 A default memory value (in GiB), we use the default given for per proc in archer.
integer, parameter name_len_max = 120
 The maximum length of a name.
integer, parameter symtype_cinfv = 0
 This describes C_inf_v symmetries.
integer, parameter symtype_dinfh = 1
 This describes D_inf_h symmetries.
integer, parameter symtype_d2h = 2
 This describes D_2_h symmetries.
integer, parameter no_ci_target = 0
 No Ci target contraction (target only run)
integer, parameter do_ci_target_contract = 1
 Perform CI target contraction (target+scattering run)
integer, parameter normal_csf = 0
 Use configuration state functions as is.
integer, parameter prototype_csf = 1
 The configuration state functions are prototypes and therefore require expansion.
integer, parameter generate_target_wf = 0
 Generate target wavefunction parameter.
integer, parameter matrix_evaluate_diff = 0
 Matrix is evaluated as a difference from the first element.
integer, parameter matrix_evaluate_full = 1
 Matrix is evaluated as full.
integer, parameter no_pure_target_integral_diff = 2
 Same as diff but dont evalute integrals belonging to the target state.
integer, parameter no_pure_target_integral = 3
 Same as full but dont evalute integrals belonging to the target state.
integer, parameter no_diagonalization = 0
 No diagonalization.
integer, parameter diagonalize = 1
 Diagonalize.
integer, parameter diagonalize_no_restart = 2
 Diagonalize but with no restart.
integer, parameter read_from_file = 3
 Do not diagonalize, just read from file.
integer, parameter normal_phase = -1
 No phase correction.
integer, parameter neglect_continuum = 0
 Phase correct by negelcting continuum orbitals.
integer, parameter scatci_decide = -10
 SCATCI chooses the diagonalizer.
integer, parameter dense_diag = 1
 Force a dense diagonalizer (all eigenvalues)
integer, parameter davidson_diag = 0
 Force a Davidson diagonalizer.
integer, parameter iterative_diag = -1
 Force an iterative (like ARPACK) diagonalizer.
integer, parameter save_continuum_coeffs = 1
 Omit L2 section of the eigenvectors (only keep continuum channels)
integer, parameter pass_to_cdenprop = 2
 Eigenpairs will be saved in memory and passed to CDENPROP and outer interface.
integer, parameter save_l2_coeffs = 4
 Keep only L2 section of the eigenvectors (omit continuum channels)
integer, parameter save_all_coeffs = SAVE_CONTINUUM_COEFFS + SAVE_L2_COEFFS
 Do not discard any coefficients.
integer, parameter mat_dense = 0
 Matrix is dense.
integer, parameter mat_sparse = 1
 Matrix is sparse.
real(wp), parameter default_matelem_threshold = 1e-15
 Matrix element storage threshold.
integer, parameter matrix_index_c = 0
 Matrix uses C indexing.
integer, parameter matrix_index_fortran = 1
 Matrix uses FORTRAN indexing.
integer, parameter matrix_order_ij = 1
 Sort via IJ ordering.
integer, parameter matrix_order_ji = 2
 Sort via JI ordering.
integer, parameter default_expand_size = 100000
 Default expansion size for the cache.
integer, parameter main_hamiltonian = 0
 Hamiltonain is a scattering one.
integer, parameter target_hamiltonian = 1
 Hamiltonain is target one.
real(wp), parameter default_integral_threshold = 0
 Default threshold to store integrals.
integer, parameter ecp_type_null = 0
integer, parameter ecp_type_molpro = 1
integer, parameter slepc_max_eigenvalues = 100

Detailed Description

MPI SCATCI Constants module.

Authors
A Al-Refaie
Date
2017

This module provides constants used by MPI SCATCI.

Note
30/01/2017 - Ahmed Al-Refaie: Initial revision.
16/01/2019 - Jakub Benda: Unifom coding style and expanded documentation.

Variable Documentation

◆ davidson_diag

integer, parameter consts_mpi_ci::davidson_diag = 0

Force a Davidson diagonalizer.

Definition at line 108 of file consts_mpi_ci.f90.

◆ default_archer_memory

real(wp), parameter consts_mpi_ci::default_archer_memory = 2.5

A default memory value (in GiB), we use the default given for per proc in archer.

Definition at line 49 of file consts_mpi_ci.f90.

◆ default_expand_size

integer, parameter consts_mpi_ci::default_expand_size = 100000

Default expansion size for the cache.

Definition at line 138 of file consts_mpi_ci.f90.

◆ default_integral_threshold

real(wp), parameter consts_mpi_ci::default_integral_threshold = 0

Default threshold to store integrals.

Definition at line 147 of file consts_mpi_ci.f90.

◆ default_matelem_threshold

real(wp), parameter consts_mpi_ci::default_matelem_threshold = 1e-15

Matrix element storage threshold.

Definition at line 127 of file consts_mpi_ci.f90.

◆ dense_diag

integer, parameter consts_mpi_ci::dense_diag = 1

Force a dense diagonalizer (all eigenvalues)

Definition at line 106 of file consts_mpi_ci.f90.

◆ diagonalize

integer, parameter consts_mpi_ci::diagonalize = 1

Diagonalize.

Definition at line 90 of file consts_mpi_ci.f90.

◆ diagonalize_no_restart

integer, parameter consts_mpi_ci::diagonalize_no_restart = 2

Diagonalize but with no restart.

Definition at line 92 of file consts_mpi_ci.f90.

◆ do_ci_target_contract

integer, parameter consts_mpi_ci::do_ci_target_contract = 1

Perform CI target contraction (target+scattering run)

Definition at line 66 of file consts_mpi_ci.f90.

◆ ecp_type_molpro

integer, parameter consts_mpi_ci::ecp_type_molpro = 1

Definition at line 151 of file consts_mpi_ci.f90.

◆ ecp_type_null

integer, parameter consts_mpi_ci::ecp_type_null = 0

Definition at line 150 of file consts_mpi_ci.f90.

◆ generate_target_wf

integer, parameter consts_mpi_ci::generate_target_wf = 0

Generate target wavefunction parameter.

Definition at line 74 of file consts_mpi_ci.f90.

◆ iterative_diag

integer, parameter consts_mpi_ci::iterative_diag = -1

Force an iterative (like ARPACK) diagonalizer.

Definition at line 110 of file consts_mpi_ci.f90.

◆ main_hamiltonian

integer, parameter consts_mpi_ci::main_hamiltonian = 0

Hamiltonain is a scattering one.

Definition at line 142 of file consts_mpi_ci.f90.

◆ mat_dense

integer, parameter consts_mpi_ci::mat_dense = 0

Matrix is dense.

Definition at line 123 of file consts_mpi_ci.f90.

◆ mat_sparse

integer, parameter consts_mpi_ci::mat_sparse = 1

Matrix is sparse.

Definition at line 125 of file consts_mpi_ci.f90.

◆ matrix_evaluate_diff

integer, parameter consts_mpi_ci::matrix_evaluate_diff = 0

Matrix is evaluated as a difference from the first element.

Definition at line 78 of file consts_mpi_ci.f90.

◆ matrix_evaluate_full

integer, parameter consts_mpi_ci::matrix_evaluate_full = 1

Matrix is evaluated as full.

Definition at line 80 of file consts_mpi_ci.f90.

◆ matrix_index_c

integer, parameter consts_mpi_ci::matrix_index_c = 0

Matrix uses C indexing.

Definition at line 129 of file consts_mpi_ci.f90.

◆ matrix_index_fortran

integer, parameter consts_mpi_ci::matrix_index_fortran = 1

Matrix uses FORTRAN indexing.

Definition at line 131 of file consts_mpi_ci.f90.

◆ matrix_order_ij

integer, parameter consts_mpi_ci::matrix_order_ij = 1

Sort via IJ ordering.

Definition at line 134 of file consts_mpi_ci.f90.

◆ matrix_order_ji

integer, parameter consts_mpi_ci::matrix_order_ji = 2

Sort via JI ordering.

Definition at line 136 of file consts_mpi_ci.f90.

◆ max_neig

integer, parameter consts_mpi_ci::max_neig = 400

Default limit on the number of eigenvalues printed.

Definition at line 46 of file consts_mpi_ci.f90.

◆ mxint

integer, parameter consts_mpi_ci::mxint = 1000

Maximal length of integer arrays in input namelists.

Definition at line 40 of file consts_mpi_ci.f90.

◆ name_len_max

integer, parameter consts_mpi_ci::name_len_max = 120

The maximum length of a name.

Definition at line 52 of file consts_mpi_ci.f90.

◆ neglect_continuum

integer, parameter consts_mpi_ci::neglect_continuum = 0

Phase correct by negelcting continuum orbitals.

Definition at line 100 of file consts_mpi_ci.f90.

◆ nidx

integer, parameter consts_mpi_ci::nidx = 2

Number of long integers used to store up to 8 (packed) integral indices.

Definition at line 43 of file consts_mpi_ci.f90.

◆ no_ci_target

integer, parameter consts_mpi_ci::no_ci_target = 0

No Ci target contraction (target only run)

Definition at line 64 of file consts_mpi_ci.f90.

◆ no_diagonalization

integer, parameter consts_mpi_ci::no_diagonalization = 0

No diagonalization.

Definition at line 88 of file consts_mpi_ci.f90.

◆ no_pure_target_integral

integer, parameter consts_mpi_ci::no_pure_target_integral = 3

Same as full but dont evalute integrals belonging to the target state.

Definition at line 84 of file consts_mpi_ci.f90.

◆ no_pure_target_integral_diff

integer, parameter consts_mpi_ci::no_pure_target_integral_diff = 2

Same as diff but dont evalute integrals belonging to the target state.

Definition at line 82 of file consts_mpi_ci.f90.

◆ normal_csf

integer, parameter consts_mpi_ci::normal_csf = 0

Use configuration state functions as is.

Definition at line 70 of file consts_mpi_ci.f90.

◆ normal_phase

integer, parameter consts_mpi_ci::normal_phase = -1

No phase correction.

Definition at line 98 of file consts_mpi_ci.f90.

◆ pass_to_cdenprop

integer, parameter consts_mpi_ci::pass_to_cdenprop = 2

Eigenpairs will be saved in memory and passed to CDENPROP and outer interface.

Definition at line 115 of file consts_mpi_ci.f90.

◆ prototype_csf

integer, parameter consts_mpi_ci::prototype_csf = 1

The configuration state functions are prototypes and therefore require expansion.

Definition at line 72 of file consts_mpi_ci.f90.

◆ read_from_file

integer, parameter consts_mpi_ci::read_from_file = 3

Do not diagonalize, just read from file.

Definition at line 94 of file consts_mpi_ci.f90.

◆ save_all_coeffs

integer, parameter consts_mpi_ci::save_all_coeffs = SAVE_CONTINUUM_COEFFS + SAVE_L2_COEFFS

Do not discard any coefficients.

Definition at line 119 of file consts_mpi_ci.f90.

◆ save_continuum_coeffs

integer, parameter consts_mpi_ci::save_continuum_coeffs = 1

Omit L2 section of the eigenvectors (only keep continuum channels)

Definition at line 113 of file consts_mpi_ci.f90.

◆ save_l2_coeffs

integer, parameter consts_mpi_ci::save_l2_coeffs = 4

Keep only L2 section of the eigenvectors (omit continuum channels)

Definition at line 117 of file consts_mpi_ci.f90.

◆ scatci_decide

integer, parameter consts_mpi_ci::scatci_decide = -10

SCATCI chooses the diagonalizer.

Definition at line 104 of file consts_mpi_ci.f90.

◆ slepc_max_eigenvalues

integer, parameter consts_mpi_ci::slepc_max_eigenvalues = 100

Definition at line 153 of file consts_mpi_ci.f90.

◆ symtype_cinfv

integer, parameter consts_mpi_ci::symtype_cinfv = 0

This describes C_inf_v symmetries.

Definition at line 56 of file consts_mpi_ci.f90.

◆ symtype_d2h

integer, parameter consts_mpi_ci::symtype_d2h = 2

This describes D_2_h symmetries.

Definition at line 60 of file consts_mpi_ci.f90.

◆ symtype_dinfh

integer, parameter consts_mpi_ci::symtype_dinfh = 1

This describes D_inf_h symmetries.

Definition at line 58 of file consts_mpi_ci.f90.

◆ target_hamiltonian

integer, parameter consts_mpi_ci::target_hamiltonian = 1

Hamiltonain is target one.

Definition at line 144 of file consts_mpi_ci.f90.