In place deinterleave

[Shnatsel]

Addresses #93

WIP. Works but needs tuning for parallelism to switch on/off at a given size. I don't think there's a way around measuring how each approach performs at runtime because there isn't a portable way to get cache information and there's a very sudden switch from single-threaded to multi-threaded being beneficial as we leave the cache and hit the memory wall.

There's also a very sudden transition from out-of-place being faster to in-place being faster as we leave the cache on Zen4, but Apple M4 is not affected by that and keeps out-of-place performance almost unchanged far past the advertised cache size. That is a rather surprising result.

[codecov-commenter]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 99.79%. Comparing base (6ca05ed) to head (d6bf80f).
⚠️ Report is 2 commits behind head on main.

Additional details and impacted files

@@           Coverage Diff           @@
##             main      #95   +/-   ##
=======================================
  Coverage   99.79%   99.79%           
=======================================
  Files           8        8           
  Lines        1441     1471   +30     
=======================================
+ Hits         1438     1468   +30     
  Misses          3        3           

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

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.