README updates

[Shnatsel]

No description provided.

[codecov-commenter]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 99.79%. Comparing base (8fe9395) to head (736c212).
⚠️ Report is 8 commits behind head on main.

Additional details and impacted files

@@           Coverage Diff           @@
##             main      #70   +/-   ##
=======================================
  Coverage   99.78%   99.79%           
=======================================
  Files          13       13           
  Lines        2846     2892   +46     
=======================================
+ Hits         2840     2886   +46     
  Misses          6        6           

☔ 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.

[smu160]

We may want to keep this as it would allow us to allow for non-power-of-2 input sizes.

The fourier crate seems to have a good reference implementation: https://github.com/calebzulawski/fourier/blob/master/fourier-algorithms/src/bluesteins.rs

[Shnatsel]

The problem is that non-power-of-2 FFTs have much worse performance, up to 10x, even in RustFFT that has multiple different algorithms to deal with them, including a dedicated path for prime numbers.

zero-padding ought to be much faster despite the extra copy, but if that's not always applicable then yeah I guess we'll need code for arbitrary-sized FFTs too just like RustFFT.