https://hgpu.org/?p=5976
Parallelization of the distinct lattice spring model