KZG Polynomial Commitment

[mratsim]

Eth2 phase 1 is currently exploring Kate Polynomial Commitment at:

• https://dankradfeist.de/ethereum/2020/06/16/kate-polynomial-commitments.html
• https://github.com/protolambda/go-kate

We could use constantine as an alternate implementation to get an idea of the overhead involved using Assembly primitives.

[mratsim]

From @protolambda

Generally I followed the python code, and implemented the same test cases to run through them step-by-step to compare computations, input sizes, etc. to make sure I didn't get anything different. It's primarily just a lot of data transforms, and although the python code is readable, it didn't translate straightforward to idiomatic Go, so that took some time. The crypto itself is not too bad after that.
Theory:

• https://dankradfeist.de/ethereum/2020/06/16/kate-polynomial-commitments.html
• FK20: https://github.com/khovratovich/Kate/blob/master/Kate_amortized.pdf
  You can find the python code by Dankrad here:
• https://github.com/ethereum/research/tree/master/polynomial_reconstruction
• https://github.com/ethereum/research/tree/master/kzg_data_availability
  And Go code (with tests and benchmarks): https://github.com/protolambda/go-kate

[mratsim]

Closed by #239

Follow-up in #252