we are actually using. Whether the additional complexity this adds is worth it depends on the rest of the project, but here we can see that we still have rather nice performance. Unsurprisingly, this means that we are still using packed vector instructions, but we'll also find that we fulfilled
4x Code Performance with SIMD