Optimize Chaum Pedersen Proof Generation & Validation

As a user of electionguard I would like selection and contest proof generation and validation to be as fast as possible so that I can optimize my application for performance.

Currently the code in chaum_pedersen.cpp is a line-by-line port of the python code. While this improves readability, there are multiple allocations and deallocations happening inline which incur a performance cost. We should instead refactor this code to allocate all of the necessary memory on entry, execute the operations, and then deallocate the intermediate values only at the end.