perf: Improve var/cov/corr performance

[orlp]

This uses a new algorithm that should be more numerically stable, and is also 2x faster on my machine than the old implementation.

It's also written in such a way that it can be easily parallelized and integrated into the streaming engine now.

[codecov]

Codecov Report

Attention: Patch coverage is 91.66667% with 22 lines in your changes missing coverage. Please review.

Project coverage is 80.20%. Comparing base (27289b2) to head (ae91817).
Report is 8 commits behind head on main.

Files with missing lines	Patch %	Lines
crates/polars-compute/src/var_cov.rs	93.38%	16 Missing ⚠️
crates/polars-arrow/src/array/iterator.rs	0.00%	6 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main   #19381      +/-   ##
==========================================
+ Coverage   80.18%   80.20%   +0.01%     
==========================================
  Files        1523     1524       +1     
  Lines      209897   210203     +306     
  Branches     2434     2432       -2     
==========================================
+ Hits       168314   168597     +283     
- Misses      41028    41050      +22     
- Partials      555      556       +1     

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

[ritchie46]

micro-benchmark driven parallelism. :')

[alexander-beedie]

FYI: all things cov & corr related will be very popular here 😎👍