https://hgpu.org/?p=11232
A Framework for Productive, Efficient and Portable Parallel Computing