GS2: A fast, flexible, physicist's toolkit for
gyrokinetics GS2 logo

Find us on…

Bitbucket Download the Source

GS2

GS2 is a physics application, developed to study low-frequency turbulence in magnetized plasma. It is typically used to assess the microstability of plasmas produced in the laboratory and to calculate key properties of the turbulence which results from instabilities. It is also used to simulate turbulence in plasmas which occur in nature, such as in astrophysical and magnetospheric systems.

GS2 is:

  • open and accessible, including to new PhDs
  • an excellent tool for doing physics and developing new improved physics calculations
  • a great platform for trying out new algorithms
  • highly optimised and designed to scale to use as many cores as possible

Getting started: read the user manual.

Features

Fast microstability assessment

Linear microstability growth rates are calculated on a wavenumber-by-wavenumber basis with an implicit initial-value algorithm in the ballooning (or "flux-tube") limit. Linear and quasilinear properties of the fastest growing (or least damped) eigenmode at a given wavenumber may be calculated independently (and therefore reasonably quickly).

Fully gyrokinetic, nonlinear simulations

Nonlinear simulations of fully developed turbulence can be performed by users with access to a parallel computer. All plasma species are treated on an equal, gyrokinetic footing. Nonlinear simulations provide fluctuation spectra, anomalous (turbulent) heating rates, and species-by-species anomalous transport coefficients for particles, momentum and energy.

Good performance

GS2 is a parallel code which scales well to large numbers of processors. There are separate optimizations to work with for small (Beowulf-style) clusters and large supercomputers.

Flexible Simulation Geometry

Linear and nonlinear calculations may be carried out using a wide range of assumptions, including:

  • Local slab
  • Local cylinder
  • Local torus
  • Vacuum magnetic dipole
  • High aspect ratio torus (analytic axisymmetric equilibrium)
  • Numerically generated, local equilibrium (Miller-style)
  • Axisymmetric, numerically generated equilibria with arbitrary poloidal shaping from

Efficient computational grid

Turbulent structures in gyrokinetics are highly elongated along the magnetic field. GS2 uses field-line following (Clebsch) coordinates to resolve such structures with maximal efficiency, in a flux tube of modest extent. Pseudo-spectral algorithms are used in the spatial directions perpendicular to the field, and for the two velocity space coordinate grids (energy and pitch angle) for high accuracy on computable 5-D grids.

GS2 scaling

Get the code

The project is hosted on BitBucket. To download the source run:

git clone --recurse-submodules https://bitbucket.org/gyrokinetics/gs2.git

or download the latest stable version from here.

Dependencies

GS2 has no required external dependencies, but does have the following optional ones:

  • MPI
  • NetCDF (requires >= v4.2.0)
  • FFTW (support for v2 will be dropped in the near future)
  • LAPACK
  • NAG
  • python (required for code generation, tests and documentation, requires >= 3.6)

Building and running GS2

GS2 has been built on a variety of existing systems. To build on one of the existing supported systems, set GK_SYSTEM. For example, to build GS2 on the UK national supercomputer, Archer2, run:

make GK_SYSTEM=archer2

For further details on building GS2, including how to build it on a new system, see the documentation in the user manual

Citing GS2

If you use GS2 please remember to cite the following publications and the source code. For more details see CITATION.cff provided with GS2, or see https://doi.org/10.5281/zenodo.2551066

Documentation

This website has documentation built automatically using Ford. See below for details on individual source files, modules, procedures, and data types:

Developer Info

The GS2 team

Collection of developers from around the world. GS2 has been developed by dozens of developers over the years. For a complete list, please see https://doi.org/10.5281/zenodo.2551066