Detailed Information on Publication Record

During substructure formation in magnetized astrophysical plasma, dissipation of magnetic energy facilitated by magnetic reconnection affects the system dynamics by heating and accelerating the ejected plasmoids. Numerical simulations are a crucial tool for investigating such systems. In astrophysical simulations, the energy dissipation, reconnection rate, and substructure formation critically depend on the onset of reconnection of numerical or physical origin. In this paper, we hope to assess the reliability of the state-of-the-art numerical codes, pluto and koral by quantifying and discussing the impact of dimensionality, resolution, and code accuracy on magnetic energy dissipation, reconnection rate, and substructure formation. We quantitatively compare results obtained with relativistic and non-relativistic, resistive and non-resistive, as well as two- and three-dimensional set-ups performing the Orszag-Tang test problem. We find sufficient resolution in each model, for which numerical error is negligible and the resolution does not significantly affect the magnetic energy dissipation and reconnection rate. The non-relativistic simulations show that at sufficient resolution, magnetic and kinetic energies convert to internal energy and heat the plasma. In the relativistic system, energy components undergo mutual conversion during the simulation time, which leads to a substantial increase in magnetic energy at 20 per cent and 90 per cent of the total simulation time of 10 light-crossing times - the magnetic field is amplified by a factor of 5 due to relativistic shocks. We also show that the reconnection rate in all our simulations is higher than 0.1, indicating plasmoid-mediated regime. It is shown that in koral simulations more substructures are captured than in pluto simulations.