vectorize() + #[inline(always)] is error-prone and hard to teach

Nobody has claimed this yet.

After seeing an HPC expert who literally wrote a book on HPC and SIMD struggle with it (and entirely failing to get it to work when combined with criterion benchmarks), I'm convinced that we need to provide a more robust and convenient way to handle this.

Ideally I want to have some sort of annotation that you just put on every function generic over `Simd` and do not think about inlining at all, ever. The compiler just does the right thing automatically.

I see multiversion's `#[multiversion(targets=simd)]` ([PR adding it](https://github.com/calebzulawski/multiversion/pull/52)) and [thermite](https://github.com/raygon-renderer/thermite)'s `#[thermite::dispatch(V)]` as prior art that does this well.

AFAIK (I haven't re-checked yet) `multiversion` has an indirect dispatch through a cached atomic pointer on every function call, but doesn't require functions to be generic over `Simd`. I haven't looked into how `thermite` works yet.

Start by reviewing the linked multiversion PR and the thermite dispatch approach, then compare their dispatch and inlining behavior with the current vectorize() and #[inline(always)] workflow. The issue does not name files or tests, and the desired annotation and completion criteria still need to be specified.

Written by the indexing model from the issue text.