Command line parameters. More...

#include <io.h>

Collaboration diagram for CommandLine:

Data Structures
struct	s_zipdata
	A B-spline expansion of a solution to "zip". See Bspline::zip . More...

Public Types
enum	HexEcsStg { StgNone = 0x00, StgRadial = 0x01, StgSolve = 0x02, StgExtract = 0x04 }
	Stages of the computations. More...

Public Member Functions
	CommandLine (int argc, char *argv[])

void	parse (int argc, char *argv[])
	Read options from command line. More...

Data Fields
std::ifstream	inputfile
	Alternative name for the input file. Default is "hex.inp". More...

bool	writegrid

struct CommandLine::s_zipdata	zipdata

bool	parallel
	Whether to use MPI. More...

std::string	preconditioner
	Preconditioner to use. More...

Real	droptol
	Drop tolerance for the ILU preconditioner. More...

int	itinerary
	Which parts of the computation to run. More...

bool	outofcore
	Whether to keep precomputed data only on disk and spare RAM. More...

bool	cont
	Whether to continue out-of-core computation from last computed solution (needs corresponding ooc files). More...

bool	wholematrix
	Whether to load full matrix from scratch disk at once when calculating dot product (and not by parts). More...

bool	cache_all_radint
	Whether to keep radial integrals in memory. More...

bool	cache_own_radint

Real	itertol
	Tolerance for terminating iterative solution. More...

Real	prec_itertol
	Tolerance for terminating block preconditioner. More...

bool	parallel_precondition
	Whether to use OpenMP parallelization to run preconditioner for several blocks simultaneously. More...

bool	gpu_large_data
	Keep large data in RAM instead of copying to the OpenCL compute device. More...

bool	lightweight_full
	Whether to avoid explicitly calculating big matrices and only apply them on the fly. More...

bool	lightweight_radial_cache

bool	shared_scratch
	Whether to compute only a subset of radial integrals in shared scratch architecture. More...

bool	reuse_dia_blocks
	Whether to use diagonal blocks as present in the scratch directory. (For debugging purposes only.) More...

bool	kpa_simple_rad
	Use simplified radial integral matrix for nested KPA iterations (experimental). More...

unsigned	ocl_platform
	Index of OpenCL platform to use. More...

unsigned	ocl_device
	Index of OpenCL device to use. More...

std::string	factorizer
	LU-factorizer. More...

int	groupsize
	Size of the local MPI communicator, used for distributed SuperLU. More...

int	parallel_factorization
	Allow parallel factorization. More...

int	parallel_extraction
	Allow parallel extraction. More...

int	ilu_max_iter
	Maximal number of ILU iterations for hybrid preconditioner. More...

int	max_sub_iter
	Maximal number of sub-preconditioner iterations. More...

bool	fail_on_sub_iter
	Stop with error message when sub-preconditioner fails to converge. More...

bool	carry_initial_guess
	Whether to use previous-energy solution as an initial guess. More...

bool	gpu_multiply
	Do the sparse matrix multiplication on the OpenCL device (memory intensive!). More...

bool	extract_extrapolate
	Whether to radially extrapolate the extracted T-matrix. More...

Real	extract_rho
	Extraction radius. More...

Real	extract_rho_begin
	Radial distance where to start radial averaging/extrapolation of the T-matrix. More...

int	extract_samples
	Extraction averaging/extrapolation sample count. More...

bool	refine_solution
	Load also existing solutions and check that they are within tolerance. More...

std::vector< std::string >	map_solution
	Map solution between different B-spline bases. More...

std::string	map_solution_target
	Target mapping basis. More...

Real	ssor
	Apply SSOR coupling. More...

bool	noluupdate
	Keep calculated LU also for next energy. More...

int	coupling_limit
	Maximal multipole considered by the coupled preconditioner. More...

bool	gpu_host_multiply
	Keep vectors in host memory when doing multiplication on GPU. More...

bool	mumps_outofcore
	MUMPS out of core. More...

int	mumps_verbose
	MUMPS diagnostic information. More...

Real	kpa_drop
	Whether to use drop tolerance for KPA preconditioner. More...

bool	write_intermediate_solutions
	Write intermediate solutions. More...

bool	fast_bessel
	Use faster Bessel function evaluation routine (not the Steed/Barnett) when calculating RHS. More...

int	hyb_additional_levels
	Additional levels to be solved by ILU preconditioner when using HYB preconditioner. More...

int	multigrid_depth
	Depth of the geometric multigrid. More...

int	multigrid_coarse_prec
	What preconditioner to use for solution of the coarse problem. More...

int	dom_x_panels
	Domain decomposition panels. More...

int	dom_y_panels

std::string	dom_preconditioner
	Domain preconditioner. More...

int	dom_sweeps
	Domain decomposition sweeps. More...

std::string	scratch
	Scratch directory for out-of-core data. More...

bool	analytic_eigenstates
	Use analytic eigenstates instead of those obtained by diagonalization. More...

bool	runtime_postprocess
	Calculate T-matrices after every iteration. More...

bool	sub_prec_verbose
	Verbosity of the sub-preconditioner. More...

Detailed Description

This class uses ParseCommandLine for parsing of the command line. The information on valid switches can be retrieved by executing

hex-ecs --help

Some more general comments on the switches are in the above mentioned class.

Member Enumeration Documentation

◆ HexEcsStg

enum CommandLine::HexEcsStg

Different stages of the computations, used to reflect user's choice from the command line. The corresponding mapping is:

Command line option	Program itinerary
(none)	StgRadial \| StgSolve \| StgExtract
--stg-integ	StgRadial
--stg-integ-solve	StgRadial \| StgSolve
--stg-extract	StgiExtract

Enumerator
StgNone	Do not start any intensive computations.
StgRadial	Compute only the radial integrals necessary for the construction of the equations.
StgSolve	Solve the equations.
StgExtract	Extract the T-matrices.

Constructor & Destructor Documentation

◆ CommandLine()

CommandLine::CommandLine	(	int	argc,
		char *	argv[]
	)

inline

Member Function Documentation

◆ parse()

void CommandLine::parse	(	int	argc,
		char *	argv[]
	)

Field Documentation

◆ analytic_eigenstates

bool CommandLine::analytic_eigenstates

◆ cache_all_radint

bool CommandLine::cache_all_radint

◆ cache_own_radint

bool CommandLine::cache_own_radint

◆ carry_initial_guess

bool CommandLine::carry_initial_guess

◆ cont

bool CommandLine::cont

◆ coupling_limit

int CommandLine::coupling_limit

◆ dom_preconditioner

std::string CommandLine::dom_preconditioner

◆ dom_sweeps

int CommandLine::dom_sweeps

◆ dom_x_panels

int CommandLine::dom_x_panels

◆ dom_y_panels

int CommandLine::dom_y_panels

◆ droptol

Real CommandLine::droptol

◆ extract_extrapolate

bool CommandLine::extract_extrapolate

◆ extract_rho

Real CommandLine::extract_rho

◆ extract_rho_begin

Real CommandLine::extract_rho_begin

◆ extract_samples

int CommandLine::extract_samples

◆ factorizer

std::string CommandLine::factorizer

◆ fail_on_sub_iter

bool CommandLine::fail_on_sub_iter

◆ fast_bessel

bool CommandLine::fast_bessel

◆ gpu_host_multiply

bool CommandLine::gpu_host_multiply

◆ gpu_large_data

bool CommandLine::gpu_large_data

◆ gpu_multiply

bool CommandLine::gpu_multiply

◆ groupsize

int CommandLine::groupsize

◆ hyb_additional_levels

int CommandLine::hyb_additional_levels

◆ ilu_max_iter

int CommandLine::ilu_max_iter

◆ inputfile

std::ifstream CommandLine::inputfile

◆ itertol

Real CommandLine::itertol

◆ itinerary

int CommandLine::itinerary

◆ kpa_drop

Real CommandLine::kpa_drop

◆ kpa_simple_rad

bool CommandLine::kpa_simple_rad

◆ lightweight_full

bool CommandLine::lightweight_full

◆ lightweight_radial_cache

bool CommandLine::lightweight_radial_cache

◆ map_solution

std::vector<std::string> CommandLine::map_solution

◆ map_solution_target

std::string CommandLine::map_solution_target

◆ max_sub_iter

int CommandLine::max_sub_iter

◆ multigrid_coarse_prec

int CommandLine::multigrid_coarse_prec

◆ multigrid_depth

int CommandLine::multigrid_depth

◆ mumps_outofcore

bool CommandLine::mumps_outofcore

◆ mumps_verbose

int CommandLine::mumps_verbose

◆ noluupdate

bool CommandLine::noluupdate

◆ ocl_device

unsigned CommandLine::ocl_device

◆ ocl_platform

unsigned CommandLine::ocl_platform

◆ outofcore

bool CommandLine::outofcore

◆ parallel

bool CommandLine::parallel

◆ parallel_extraction

int CommandLine::parallel_extraction

◆ parallel_factorization

int CommandLine::parallel_factorization

◆ parallel_precondition

bool CommandLine::parallel_precondition

◆ prec_itertol

Real CommandLine::prec_itertol

◆ preconditioner

std::string CommandLine::preconditioner

◆ refine_solution

bool CommandLine::refine_solution

◆ reuse_dia_blocks

bool CommandLine::reuse_dia_blocks

◆ runtime_postprocess

bool CommandLine::runtime_postprocess

◆ scratch

std::string CommandLine::scratch

◆ shared_scratch

bool CommandLine::shared_scratch

◆ ssor

Real CommandLine::ssor

◆ sub_prec_verbose

bool CommandLine::sub_prec_verbose

◆ wholematrix

bool CommandLine::wholematrix

◆ write_intermediate_solutions

bool CommandLine::write_intermediate_solutions

◆ writegrid

bool CommandLine::writegrid

◆ zipdata

struct CommandLine::s_zipdata CommandLine::zipdata

The documentation for this class was generated from the following files:

src/io.h
src/io.cpp

Data Structures

Public Types

Public Member Functions

Data Fields