CONTROL

control section [Edit on GitHub]

Keywords

• BETA_CHUNK_SIZE

• BETA_ON_DEVICE

• CYCLIC_BLOCK_SIZE

• FFT_MODE

• GEN_EVP_SOLVER_NAME

• GVEC_CHUNK_SIZE

• MPI_GRID_DIMS

• NUM_BANDS_TO_PRINT

• ORTHO_RF

• OUTPUT

• PRINT_FORCES

• PRINT_NEIGHBORS

• PRINT_STRESS

• PROCESSING_UNIT

• REDUCE_GVEC

• RMT_MAX

• SAVE_RF

• SPGLIB_TOLERANCE

• STD_EVP_SOLVER_NAME

• USE_SECOND_VARIATION

• VERBOSITY

• VERIFICATION

Keyword descriptions

BETA_CHUNK_SIZE: integer = 256

Number of atoms in a chunk of beta-projectors. [Edit on GitHub]

BETA_ON_DEVICE: logical = F

Lone keyword: T

True if whole set of beta-projectors can be allocated on device for the entire run. [Edit on GitHub]

CYCLIC_BLOCK_SIZE: integer = -1

Block size for ScaLAPACK, ELPA, and DLA-Future. [Edit on GitHub]

FFT_MODE: enum = SERIAL

Valid values:

• SERIAL

• PARALLEL

Coarse grid FFT mode (serial or parallel). [Edit on GitHub]

GEN_EVP_SOLVER_NAME: enum = AUTO

Valid values:

• AUTO

• LAPACK

• SCALAPACK

• ELPA1

• ELPA2

• DLAF

• MAGMA

• MAGMA_GPU

• CUSOLVER

Generalized eigen-value solver to use. [Edit on GitHub]

GVEC_CHUNK_SIZE: integer = 500000

Split local G-vectors in chunks to reduce the GPU memory consumption of augmentation operator. [Edit on GitHub]

MPI_GRID_DIMS: integer[2] = 1 1

MPI grid is setting the parameters for blacs grid / band parallelisation, the rest is going to k-point parallelization. [Edit on GitHub]

NUM_BANDS_TO_PRINT: integer = 10

Number of eigen-values that are printed to the standard output. [Edit on GitHub]

ORTHO_RF: logical = F

Lone keyword: T

Orthogonalize LAPW radial functions. [Edit on GitHub]

OUTPUT: string

Type of the output stream (stdout:, file:name) [Edit on GitHub]

PRINT_FORCES: logical = F

Lone keyword: T

If true then the atomic forces are printed at the end of SCF run. [Edit on GitHub]

PRINT_NEIGHBORS: logical = F

Lone keyword: T

If true then the list of nearest neighbours for each atom is printed to the standard output. [Edit on GitHub]

PRINT_STRESS: logical = F

Lone keyword: T

If true then the stress tensor components are printed at the end of SCF run. [Edit on GitHub]

PROCESSING_UNIT: enum = AUTO

Valid values:

• AUTO

• CPU

• GPU

Main processing unit to use during the execution. [Edit on GitHub]

REDUCE_GVEC: logical = T

Lone keyword: T

Reduce G-vectors by inversion symmetry. [Edit on GitHub]

RMT_MAX: real = 2.20000000E+000

Maximum allowed muffin-tin radius in case of LAPW. [Edit on GitHub]

SAVE_RF: logical = F

Lone keyword: T

Save LAPW radial functions in text file for inspection. [Edit on GitHub]

SPGLIB_TOLERANCE: real = 1.00000000E-006

Tolerance of the spglib in finding crystal symmetries [Edit on GitHub]

STD_EVP_SOLVER_NAME: enum = AUTO

Valid values:

• AUTO

• LAPACK

• SCALAPACK

• ELPA1

• ELPA2

• DLAF

• MAGMA

• MAGMA_GPU

• CUSOLVER

Standard eigen-value solver to use. [Edit on GitHub]

USE_SECOND_VARIATION: logical = T

Lone keyword: T

True if second-variational diagonalization is used in LAPW method. [Edit on GitHub]

VERBOSITY: integer = 0

Level of verbosity. [Edit on GitHub]

VERIFICATION: integer = 0

Level of internal verification. [Edit on GitHub]