Feature: Optimizations by DistributedLab

[olegfomenko]

Feature: Optimizations by DistributedLab

Optimizing Keccak

• Replaced loops with pre-calculated slice indexes
• Using assembler xor for amd64 and arm64 platforms (taken from golang/x/crypto)

Optimizing FFT

• Taken parallel approach from gnark-crypto
• Adding BitReverse from gnark-crypto that should be faster for non-arm platforms

Optimizing util/parallel

• Use atomic uint instead of the chan of uints
• Adding thread pool based on chan that executes tasks with different execution time

Optimizing exponent

• Optimized exp function to user int power instead of big.Int

Checklist

• I wrote new tests for my new core changes.
• I have successfully ran tests, style checker and build against my new changes locally.
• I have informed the team of any breaking changes if there are any.