scrypt: parallel implementation

[smessmer]

Looking at the source code, it seems this crate always uses a sequential algorithm and only one core, even when the scrypt parameter p is larger than 1. For p>1, the derivation could be made faster by using a parallel algorithm. Are there plans to implement this?

[tarcieri]

Seems like a nice feature to add, potentially leveraging something like rayon as an optional dependency.

[newpavlov]

pbkdf2 has the parallel feature, which uses rayon for computing long output keys (though AFAIK such keys are usually not used in practice since they do not improve security). A PR which will add a similar feature to scrypt is welcomed!