Better algorithms for the multiplication and division functions

[AaronKutch]

This is a continuation of #18 and #27. I am tracking this in one issue, because the specific division algorithm I have in mind uses Karatsuba multiplication. I will post updates here as I experiment more. I will also fix the sprawling technical debt of the division function.

[AaronKutch]

As a reminder to myself, I need to check that my unrolling is actually making a difference in the multiplication function

[AaronKutch]

Based on problems I have encountered so far, the Karatsuba versions will have to wait for the better constructors and related PRs. However, I am preparing rewrites of the O(n^2) multiplication and division functions using lessons I learned from my slice rotation function. They should be dramatically faster and also reduce cognitive complexity problems.
I will not be PRing them until the reorganization PR is settled.

[Robbepop]

Maybe this can be some inspiration for you: paritytech/parity-common#126

[AaronKutch]

A new slice::fill method was added which will help me clean up certain parts