https://hgpu.org/?p=9676
Efficient computation of constrained parameterizations on parallel platforms