Performance: parallelising kimchi prover more

[volhovm]

A bunch of performance improvements to the kimchi prover aiming to utilise more parallelism where available.

As of now, the save-up seems to be as follows:

TODOs:

• Check how much faster it is on mina side. Adjust the benchmark if necessary (more columns? lookups?)
• Make sure there are no regressions for the verifier side.
• Check if parallel evaluations can be faster -- do we need to clone the vector and then interpolate?
• Check that perm_aggreg optimisation preserves correctness
• Last step: run the threading profiler once again to make sure I didn't miss anything

[codecov]

Codecov Report

Attention: Patch coverage is 97.11934% with 7 lines in your changes missing coverage. Please review.

Project coverage is 76.93%. Comparing base (fe10025) to head (f8ddd85).
Report is 21 commits behind head on master.

Files with missing lines	Patch %	Lines
kimchi/src/prover.rs	84.78%	7 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #2969      +/-   ##
==========================================
+ Coverage   76.85%   76.93%   +0.08%     
==========================================
  Files         258      261       +3     
  Lines       61202    61569     +367     
==========================================
+ Hits        47036    47370     +334     
- Misses      14166    14199      +33     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.