Application of Mesh-refinement for Particle-in-Cell Simulations of Relativistic Magnetic Reconnection

Relativistic magnetic reconnection is a fundamental process where the topology of the magnetic fields rapidly rearrange (they break and reconnect) converting energy stored in the strong magnetic fields to non-thermal particle energy. This process is often invoked to explain particle energization in many high-energy astrophysical systems. Particle-in-cell (PIC) methods are commonly used for simulating reconnection from first principles. While significant studies have been performed to investigate reconnection physics, especially in 2D, adoption of advanced algorithms especially for efficiently modeling such systems have been limited. We leverage the GPU-accelerated WarpX code to perform PIC simulations of reconnection. In this talk, we will present the mesh-refinement strategy implemented in WarpX, and applied to magnetic reconnection. We demonstrate the method on 2D systems and compare the accuracy of the implementation by comparing the current sheet evolution, energy conversion and particle acceleration with the high-resolution uniform grid simulation. We also compare the total node hours saved. The methods presented can also be applied to 3D where the performance gains with mesh-refinement are expected to be higher due to few particles required.