Big Changes
- Initial multi-arch support
- Iterator overhaul & optimization
- More SIMD iterators may be strided
- Switch dependency from ~coresimd~ to ~std::arch~ and ~std::simd~
- Significantly optimize certain strided iterators
- Significantly optimize zipped iterators
- Built-in performance inspector (--feature trace)

Features
- Vector population counts
- Vector destriding
- Vectorize downcasts on SSE4.1 and AVX2
- Saturating subtraction for u8s and u16s
- Significantly optimize reductive vector sums

Bugfixes & Minor Improvements
- Upgrade to Rust 2018
- Add many new tests for unsafe code
- Fix UB for certain iterator patterns