Optimise polynomial modulus

[teor2345]

In #9 we want to optimise polynomial multiplication.

We currently have a manual modulus implementation, but the existing library code for divide_with_q_and_r() might be faster.

Tasks:

• [ ] test that the manual implementation produces the same results as divide_with_q_and_r()
• [ ] benchmark both and see which is faster

[teor2345]

It looks like the manual implementation is 60% faster, so we don't need to change our code here. But I'll still add benchmarks for both versions, in case other code changes make the library code faster.

Manual implementation:

Benchmarking Cyclotomic multiplication: polynomial/Random input: Collecting 50 samples in estimated 5
Cyclotomic multiplication: polynomial/Random input
                        time:   [53.911 ms 54.057 ms 54.237 ms]
                        change: [-0.6539% -0.1373% +0.3285%] (p = 0.61 > 0.05)
                       No change in performance detected.
Found 4 outliers among 50 measurements (8.00%)
  1 (2.00%) high mild
  3 (6.00%) high severe

divide_with_q_and_r() call:

Benchmarking Cyclotomic multiplication: polynomial/Random input: Collecting 50 samples in estimated 8
Cyclotomic multiplication: polynomial/Random input
                        time:   [88.762 ms 88.957 ms 89.216 ms]
                        change: [+63.894% +64.562% +65.174%] (p = 0.00 < 0.05)
                        Performance has regressed.
Found 4 outliers among 50 measurements (8.00%)
  2 (4.00%) high mild
  2 (4.00%) high severe