Not strictly an answer, but I just started learning low-level CUDA and this gives me the idea to implement it there as an exercise. Certainly lends itself to pa

Not strictly an answer, but I just started learning low-level CUDA and this gives me the idea to implement it there as an exercise. Certainly lends itself to parallelization.