Faster Rust with SIMD

This repository contains the code for the Faster Rust with SIMD blog post. It contains an incomplete bitset implementation using an abstract block for the bits. The block is implemented in three different ways - using a simple usize, using u64x4 from the standard library (nightly-only), and using u64x4 from wide.

Running the benchmarks

Make sure to run the benchmarks with all features enabled:

cargo bench --all-features

As the benchmarks use the experimental SIMD APIs in the standard library, they require nightly Rust.