- Configure FRI with a list of arities that's more appropriate for a 2^14 instance. The previous config resulted in a huge final polynomial.
- Log the blinding factors, and other logging tweaks.
As in plonky1. The semantics of virtual targets in plonky1 were rather weird, but I think it's somewhat better here, since we already separate `generate_copy` and `assert_equal` methods. Users now make more of an explicit choice -- they can use a `VirtualTarget` for the witness generation only using `generate_copy`, or they can involve it in copy constraints.
* Use built-in `reverse_bits`; remove duplicate `reverse_index_bits`.
* Reduce precomputation time/space complexity from quadratic to linear.
* Several working cache-friendly FFTs.
* Fix to allow FFT of constant polynomial.
* Simplify FFT strategy choice.
* Add PrimeField and CHARACTERISTIC properties to Fields.
* Add faster method for inverse of 2^m.
* Pre-compute some of the roots; tidy up loop iteration.
* Precomputation for both FFT variants.
* Refactor precomputation; add optional parameters; rename some things.
* Unrolled version with zero tail.
* Iterative version of Unrolled precomputation.
* Test zero tail algo.
* Restore default degree.
* Address comments from @dlubarov and @wborgeaud.
- `self.gates` -> `self.gate_instances`
- Some tests were using a single binary FRI reduction, which doesn't provide enough succinctness for our blinding scheme to work. This caused `blinding_counts` to continue until it overflowed.
