Implement point doubling for weierstrass curves

[capito27]

Quick, low hanging fruit for #10

Implement the exception free point doubling algorithm from the same source as the point addition algorithm.

It should be noted that on my system (Intel i7-7700k, Windows 10), criterion benchmarks were somewhat unconclusive, for P256, there was a 33% performance improvement with the doubling algorithm. For P384 however, there was a 30% performance degradation, whereas for p521, there was no difference.

As such, help with more in depth testing would be appreciated.

Also of note, adding a verification of equality when adding two points to instead call the point doubling function should they be equal (with both reference and inner field comparison) also produced unconclusive results, thus it was not added to the commit

PS : here is an archive containing the produced criterion graphs, should they be of use.
criterion_comparisons.zip

[armfazh]

Thanks for your input.