https://hgpu.org/?p=9523
Efficient Execution of AMR Computations on GPU Systems