Solver comparison for Poisson-like equations on tokamak geometries

The solution of Poisson-like equations defined on a complex geometry is required for gyrokinetic simulations, which are important the modelling plasma turbulence in nuclear fusion devices such as ITER tokamak. In this paper, we compare three existing solvers finely tuned to solve problem, terms accuracy solution, and their computational efficiency. We also consider practical implementation aspects, including parallel efficiency code, potentially enabling an integration state-of-the-art first-principle simulation framework. first, Spline FEM solver, uses C1 polar splines construct finite elements method solves equation curvilinear coordinates. resulting linear system solved using conjugate gradient method. second, GMGPolar symmetric difference discretise differential equation. tailored geometric multigrid scheme, with combination zebra circle radial line smoothers, together implicit extrapolation scheme. third, Embedded Boundary volumes Cartesian coordinates embedded boundary solver shown be most accurate. use least memory. fastest cases. All capable solving realistic non-analytical geometry. additionally used attempt X-point