Implement more efficient decomposition in G2 for GLV

[jules]

    Something that we might want to follow up on is that for G2 there exists an even more efficient decomposition relying on the GLS endomorphism (https://eprint.iacr.org/2008/194)

Originally posted by @Pratyush in https://github.com/AleoHQ/snarkVM/pull/1084#discussion_r987115395

[howardwu]

I think we may have implemented this in PR https://github.com/AleoHQ/snarkVM/pull/1084.

Could @julesdesmit or @Pratyush confirm this? (feel free to close this issue if so)

[howardwu]

On a related note, do we want to open a Github issue regarding the MSM optimization? @Pratyush

[Pratyush]

We haven't implemented the more efficient GLS optimization in G2. However it's also less impactful for us since we don't do many G2 muls.