pub fn dot_simd(a: &[f32], b: &[f32]) -> f32
Wrapper for dot product - f32, optimised for SIMD
a
b
Dot product