Hash digest performance (Rust vs. Python)

Looks like the sha2 crate which the sha256 crate is a wrapper around only uses SIMD when both the sha and sse4.1 target features are supported by your cpu. If you are using linux what does cat /proc/cpuinfo show? You might also want to try ring::digest::digest(&ring::digest::SHA256, data) from the ring crate. It has several optimized implementations written in assembly for different cpu feature levels.

1 Like