https://hgpu.org/?p=11480
Parallel dual tree traversal on multi-core and many-core architectures for astrophysical N-body simulations