stella v1.0

The $\texttt{stella}$ code has been extensively cleaned and reorganised by H. Thienpondt and G. Acton. Numerous files have been moved, renamed, split, and streamlined. Additional comments and headers have been introduced throughout the code to enhance readability and clarity. Each folder now includes a dedicated README file providing an overview of its contents. Furthermore, new automated tests have been added to strengthen reliability.

Restructuring of the code in July 2024

The cleanup of the $\texttt{stella}$ code began in July 2024. To preserve the pre-cleanup version, it was released as stella release v0.6. Subsequently, the electromagnetic extension was implemented and released as stella release v0.7, which introduced the following restructurings and improvements:

• The code has been organised into subfolders:

  • AUTOMATIC_TESTS
  • COMPILATION
  • DOCUMENTATION
  • EXTERNALS
  • POST_PROCESSING
  • STELLA_CODE
    • calculations
    • diagnostics
    • dissipation
    • fields
    • geometry
    • grids
    • gyrokinetic_terms
    • neoclassical
    • parameters
    • radial_variation
    • utils
    • stella.f90

• Most importantly, the main $\texttt{stella}$ code has been centralised in the "STELLA_CODE" folder and further organised into subfolders. This structure makes it easier for users to search for keywords, to locate key sections, and to focus on the core scripts.

• The compilation process using make has been fully revamped and moved to the "COMPILATION" folder, preventing compiled object files from cluttering the main code.

• An automatic testing infrastructure has been implemented to ensure code reliability.

This initial cleaned version of $\texttt{stella}$ has been saved and released as stella release v0.8.

Restructuring of the code in September 2025

The cleanup of the $\texttt{stella}$ code was resumed in the summer of 2025. Key updates include:

• The input variables have been fully reorganised, with most variables and namelists renamed.

• The main stella code within "STELLA_CODE" has been split up further into the following subfolders:

  • arrays
  • calculations
  • diagnostics
  • dissipation
  • field_equations
  • geometry
  • grids
  • gyrokinetic_equation
  • neoclassical
  • parallelisation
  • parameters
  • radial_variation
  • read_namelists_from_input_file
  • utils
  • init_stella.f90
  • stella.f90

• Many modules, routine and variable names have been updated to improve clarity and readability.

• Many more comments have been added throughout the code to help others follow what the code is doing.

• Additional abort statements have been added throughout the code to prevent incorrect usage.

• Added an Acknowledgments section to the main README.md file to keep track of contributions to the $\texttt{stella}$ code. This list is incomplete and it is the responsibility of the various authors to add their contributions.

• Various bugs have been fixed.

New input variables

Many of the namelists and variable names have been changed to make them more intuitive. An up-to-date default $\texttt{stella}$ input file can be found at:

STELLA_CODE/read_namelists_from_input_file/default_input_file.in

To convert old $\texttt{stella}$ input files to the new format, run the following command in the folder containing the input file:

python3 $STELLA/AUTOMATIC_TESTS/convert_input_files/convert_inputFile.py

Restructuring of the $\texttt{stella}$ scripts

The main $\texttt{stella}$ script has been cleaned up and divided into two separate files::

• stella.f90
• init_stella.f90

Added a new directory called “read_namelists_from_input_file” which reads in the input file. All namelists are read here now.

The script that evolves the distribution function has been split up and renamed:

• The time_advance.f90 script has been replaced by:

  • gyrokinetic_equation/gyrokinetic_equation_initialisation.f90
  • gyrokinetic_equation/gyrokinetic_equation_explicit.f90
  • gyrokinetic_equation/gyrokinetic_equation_implicit.f90
  • gyrokinetic_equation/gk_magnetic_drift.f90
  • gyrokinetic_equation/gk_drive.f90
  • gyrokinetic_equation/gk_ExB_nonlinearity.f90
  • gyrokinetic_equation/gk_parallel_nonlinearity.f90

• Several routines from time_advance.f90 have been separated into:

  • calculations/calculations_checksim.f90
  • calculations/calculations_add_explicit_terms.f90
  • calculations/calculations_timestep.f90

The following scripts have been renamed:

• ffs_solve.f90 → gyrokinetic_equation/gk_ffs_solve.f90
• flow_shear.f90 → gyrokinetic_equation/gk_flow_shear.f90
• parallel_streaming.f90 → gyrokinetic_equation/gk_parallel_streaming.f90
• implicit_solve.f90 → gyrokinetic_equation/gk_implicit_terms.f90

The gyro_averages.f90 script has been split up:

• calculations_gyro_averages.f90
• arrays_gyro_averages.f90

The fields scripts have been renamed to field_equations, and the following files have been moved out of "fields":

• fields/dist_fn.f90 → arrays/initialise_arrays.f90
• fields/init_g.f90 → arrays/initialise_distribution_function.f90

The vpamu_grids.f90 script has been renamed and split into:

• calculations/calculations_velocity_integrals.f90
• grids/grids_velocity.f90

Other moves and name changes:

• grids/stella_time.f90 → grids/grids_time.f90
• grids/arrays_dist_fn.f90 → arrays/arrays_distribution_function.f90
• grids/arrays_fields.f90 → arrays/arrays_fields.f90
• grids/common_types.f90 → parallelisation/common_types.f90
• grids/stella_layouts.f90 → parallelisation/parallelisation_layouts.f90
• utils/redistribute.f90 → parallelisation/redistribute.f90
• calculations/dist_redistribute.f90 → parallelisation/initialise_redistribute.f90

New files:

• parallelisation/timers.f90

The scripts related to the radially global version of $\texttt{stella}$ have been isolated:

• radial_variation/gk_radial_variation.f90
• radial_variation/gk_sources.fpp
• radial_variation/multibox.f90

In addition, within the other scripts, code specific to the radially global version has been moved into subroutines, ensuring a clearer separation from the main $\texttt{stella}$ code. Specifically,

• dist_fn.f90/init_gxyz() → initialise_distribution_function/add_corrections_to_g_for_radial_variation()

Important name changes

• The geometry variables have been redefined:

  • bmag(ia,iz) = $B / B_{ref}$
  • gradx_dot_gradx(ia,iz) = $|\nabla x|^2$
  • grady_dot_grady(ia,iz) = $|\nabla y|^2$
  • gradx_dot_grady(ia,iz) = $\nabla x \cdot \nabla y$
  • B_times_gradB_dot_gradx(ia,iz) = $B \times \nabla B \cdot \nabla x \ (a*B_{ref}/B^3)$
  • B_times_gradB_dot_grady(ia,iz) = $B \times \nabla B \cdot \nabla y \ (a*B_{ref}/B^3)$
  • B_times_kappa_dot_gradx(ia,iz) = $B \times \kappa \cdot \nabla x \ (a*B_{ref}/B^2)$
  • B_times_kappa_dot_grady(ia,iz) = $B \times \kappa \cdot \nabla y \ (a*B_{ref}/B^2)$
  • b_dot_gradz(ia,iz) = $b \cdot \nabla z$
  • b_dot_gradz_avg(iz) = $\sum_\alpha b \cdot \nabla z \mathcal{J} d \alpha / \sum_\alpha \mathcal{J} d \alpha$

• Therefore, the following name changes have been implemented:

  • gds22 → gradx_dot_gradx * shat * shat
  • gds2 → grady_dot_grady
  • gds21 → gradx_dot_grady * shat
  • gbdrift0 → B_times_gradB_dot_gradx * 2. * shat
  • gbdrift → B_times_gradB_dot_grady * 2.
  • cvdrift0 → B_times_kappa_dot_gradx * 2. * shat
  • cvdrift → B_times_kappa_dot_grady * 2.
  • gradpar → b_dot_gradz_avg
  • dgradpardrho → d_bdotgradz_drho
  • gradpar_c → b_dot_gradz_centredinz (in gk_parallel_streaming.f90)
  • d_gradydotgrady_drho → d_gradydotgrady_drho
  • d_gradxdotgradx_drho → d_gradxdotgradx_drho
  • d_gradxdotgrady_drho → d_gradxdotgrady_drho

Automatic tests

To ensure that the stella code is functioning correctly, a suite of numerical tests has been implemented and is automatically run on every push to GitHub. These tests cover a wide range of routines and options within stella. The automatic testing infrastructure was implemented by H. Thienpondt, who also designed most of the tests, while the full-flux-surface tests were added by G. Acton and the electromagnetic tests were contributed by M. Hardman. The numerical tests are organised into seven categories. Tests 1–5 use the electrostatic flux-tube version of stella, while tests 6 and 7 target the full-flux-surface and electromagnetic versions, respectively.

• Test 1: Confirms that the stella executable exists and that stella runs correctly by
  checking that the executable produces output files.
• Test 2: Verifies that the magnetic geometries are implemented correctly. This includes
  geometries based on Miller parameters, VMEC equilibria, and slab geometry.
• Test 3: Checks each term of the gyrokinetic equation independently. It confirms that
  the electrostatic potential remains constant when no terms (nor collisions or dissipation)
  are included; validates the initialsation options for the distribution function, and verifies
  the correct evolution of the potential for each term. The (kx , ky ) grid options (“box” and
  “range”) are also tested.
• Test 4: Tests the parallel boundary conditions: (1) standard twist-and-shift, (2) stellarator-
  symmetric, (3) periodic, and (4) zero boundary conditions. Additionally, multiple input flags
  are tested simultaneously. In the future, each input parameter would be tested individually.
• Test 5: Validates the diagnostics, including growth rates, fluxes, density and temperature,
  distribution function, and electrostatic potential.
• Test 6: Checks the electrostatic full-flux-surface version of stella. Geometric quantities
  are verified first...

stella before big 2025 clean-up (v0.8)

This master branch has been in use since September 2024 until September 2025. Stella is undergoing a big clean-up so this version is released as a back-up.

Electromagnetic stella

The electromagnetic version of stella.

stella before big 2024 clean-up

This release serves as a back-up, since many big changes will be implemented.

v0.5.1

This is the version of stella associated with the upcoming publication dealing with the radially global implementation.

v0.5

Numerous changes, fixes and additions:

Full-flux-surface capability
New physical Fokker-Planck collision operator
Periodic boundary conditions for simulations with low magnetic shear
Added option to calculate drift terms implicitly (as an extra implicit step)
Better support for shared memory in the response matrix framework
Consistent handling of the k_perp = 0 mode for simulations with radial variation
Numerous performance enhancements, done by optimizing the use of CPU cache
Added the option to build with CMake, rather than make
Parallel nonlinearity now functions properly when radial profile variation is included
git commit number is now printed out in the stella splash screen