https://hgpu.org/?p=2480
Parallel Processing of the Building-Cube Method on a GPU Platform