Modules

Provides random forcing for Alfven wave problem Originally by Hammett, Dorland and Quataert, Dec. 5, 2001

Provides wrappers to some simple array operations. We use this indirection in order to be able to consistently treat performance concerns. Currently this is limited to adding OpenMP, although in the future it might allow different approaches as appropriate.

Helper functions for running benchmarks

Reads in the geometry using a CHEASE output file. The CHEASE output file is read using the helper module read_chease.

Routines for implementing the model collision operator defined by Barnes, Abel et al. 2009 or on arxiv. The collision operator causes physically motivated smoothing of structure in velocity space which is necessary to prevent buildup of structure at fine scales in velocity space, while conserving energy and momentum.

This module provides 10,000 fixed random numbers used for initialising tests, so as always to give the same result when starting from noise, regardless of compiler/system. These should not be used for production runs.

Convert from complex variable a(d1,d2,d3, ...) to a real variable ar(2,d1,d2,d3,...) and back.
This is necessary for saving complex variables in NetCDF format

A module to calculate the properties of the antenna used to drive turbulence

This module contains functions for writing quantities that have the dimensions of the fields, including volume averaged quantities, etc.

Module which contains functions for calculating and writing out the fluxes of heat and momentum etc.

A module which writes out the electrostatic drive of zonal flows. The code is based on a Python3 post-processing tool from Stephen Biggs-Fox which can be found at [gitlab.com/stephen-biggs-fox/zonal-transfer-functions]

A module which contains methods for checking whether nonlinear runs have reached a saturated steady state

A module for printing summary information to stdout during the simulation

A module which writes out quantities which writes out diagnostic quantities which assess whether the velocity space resolution is sufficient.

The principal function of this module is to evolve the distribution function, that is, to advance the discrete gyrokinetic equation. This involves calculating the source and dealing with the complexities of the parallel boundary conditions. In addition it contains a routine for implementing perpendicular velocity shear and calculating the right-hand side of the field equation, as well as a host of other functions. Initializes a limited selection of arrays, for example g, gnew, vperp2, typically those which are needed by other modules that don't need dist_fn to be fully initialized (e.g. nonlinear_terms) This initialization level depends on grid sizes.

This module is a submodule of geometry which handles reading from the ascii EQDSK format output by EFIT, but also now by other codes. This file contains psi on an R,Z grid, as well as other quantities such as q (safety factor), I (a.k.a. f) and p (pressure) on a psi grid.

A stub module representing the eigenvalue solver

Provides some wrappers for managing FFTW plans.

FIXME : Add documentation

Provides storage for the EM potentials and some related data

This module is based on the fields_local module (which is in term a re-implementation of the fields implicit functionality). fields_gf_local implements the implicit field calculation using a gf_lo based data decomposition to improve performance on large process counts. It utilises a smaller set of processes to reduce communication costs (at the expense of significant load imbalance in the fields calculations). It assumes gf_lo velocity space integration and stores field data in gf_lo format (i.e. ik,it are decomposed over processes, everthing else is local).

This module implements the same implicit algorithm as fields_implicit. However, it is an almost complete rewrite of the implementation which can give large gains in the speed of calculating the response matrix, as well as significant gains to the advance steps

FIXME : Add documentation

A module which contains a series of high level tests used in the linear and nonlinear test cases, as well as a driver function which runs GS2. Check that the relative error of the growth rates as a function of ky with respect to the given rslt is less than err. rslt should be an array of length naky which contains the growth rates as a function of ky. Returns .true. for pass and .false. for fail

Allows a general weight function (as specified by an external function) for integration. Returns the Gaussian quadrature weights and abscissae corresponding to the set of orthogonal polynomials associated with this weight. See: Gnauchi, in: "Orthogonal Polynomials: Theory and Practice", Nevai (Ed.). Kluwer, 1990. Wheeler, Rocky Mountain Journal of Mathematics, Vol 4, p. 287 (1974)

Deals with most types of equilibria to generate GS2's expected geometrical parameters Input: rhoc :: minor radius of surface of interest (see irho below) use_large_aspect :: default false. Choose true to get "s-alpha" equilibria (see local_eq) Equilibrium options : Choose 1 of: local_eq : Analytic parameterization of local equilibrium ppl_eq = .true. :: use PPL style NetCDF equilibrium (Menard) transp_eq = .true. :: use TRXPL style NetCDF equilibrium (Menard) gen_eq = .true. :: use GA style NetCDF equilibrium (TOQ) chs_eq = .true. :: Use CHEASE generated equilibrium efit_eq = .true. :: use EFIT equilibrium dfit_eq = .true. :: use dipole equilibrium -------------------> [note: if any of the above except local_eq are true set eqfile = input file name] irho :: choose flux surface label case (1) :: sqrt(toroidal flux)/sqrt(toroidal flux of LCFS) case (2) :: diameter/(diameter of LCFS). recommended case (3) :: poloidal flux/(poloidal flux of LCFS) case (4) :: rho_mid? Only used for the vacuum ring dipole, dfit_eq = T

Explicit interfaces for the functions in git_hash.fpp

Given B{theta_i} attempt to produce an optimal set of theta values with the same size as the input {theta_i}. The output {theta} set should include all extrema (global and local) and as many other grid points as required to meet the requested number.

A module for calculating and writing gs2 outputs. It can write these outputs both to a netcdf file .cdf or to ascii text files. It is controlled via the namelist diagnostics_config. This module is intended to replace the old gs2_diagnostics module with a simpler and more structured interface.

This module is analogous to the init() function in Linux-based operating systems: it initialises gs2 to a certain init_level. At a given init level, certain modules are initialised and certain are not.

A module for writing geometry to netcdf.

This module provides the external interface to gs2. It contains functions to initialize gs2, functions to run gs2, functions to finalize gs2 and functions to override/tweak gs2 parameters.

This module sits on top of gs2_main and provides a toolkit for measuring and optimising performance

FIXME : Add documentation

FIXME : Add documentation

hdf5 wrapper subroutines written by Tomo Tatsuno once ver. 1.8.0 becomes available, we can remove this module and use those High-level APIs.

FIXME : Add documentation

This module contains the subroutines which set the initial value of the fields and the distribution function.

FIXME : Add documentation

Set the perpendicular box size and resolution for linear or nonlinear runs.

Set up values of kx and ky for linear runs that use a single k_perp mode.

A small module just designed to provide interfaces to commonly used lapack routines so that we can forget about the precision of the arguments at the calling site and the compiler has some information about the lapack routine arguments.

FIXME : Add documentation

Analytic local equilibria via generalised forms of the Miller model

A module to hold different routines which can be used to invert a matrix

Provide some small utilty functions for trying to monitor memory use.

Mersenne-Twister pseudorandom number generator (PRNG), specifically MT19937

FIXME : Add documentation

A module for handling the configuration of the optimisation module via the namelist optimisation_config.

A module which defines the override types. These types are used within the init object (which itself is contained within the gs2_program_state object) to override values of the specified parameters (i.e. modify their values from what is specified in the input file). The appropriate "prepare_..." function from gs2_main must always be called before setting overrides.

Data module holding variables associated with the parameter scan functionality. Held separately from parameter_scan module to avoid complicating dependencies.

FIXME : Add documentation!

A module to read in datafiles from CHEASE.

A small helper module for storing and providing RK schemes

This module is intended to be used for runtime tests which interrogate what is functional/what compile time options were enabled/disabled, and also any enviroment variables which affect what happens at runtime.

FIXME : Add documentation

FIXME : Add documentation

FIXME : Add documentation

Contains functions for creating a standard header for files and output

FIXME : Add documentation

FIXME : Add documentation

FIXME : Add documentation

FIXME : Add documentation

Create a random (version 4) Universally Unique Identifier (UUID)

Small functions for making semantically-correct code that doesn't raise compiler warnings (or at least, fewer). The main downside of this is requiring a function call, which can destroy vectorisation of loops.