[avx][avx2][fma] Gate high-power-draw instructions behind additional cargo features

[Lokathor]

256-bit multiplies as well as all 512-bit ops draw more power and so clock down the core. Thus, we should potentially limit them behind a cargo feature so that you opt into them.

Also 512-bit multiplies are an additional level beyond that, so they can be separated again.

[Lokathor]

We'd want them to be off by default, so this'll be breaking, but not that breaking since people can just flip on the flag when they go to the next version if they still need it..

[Lokathor]

Upon thinking more, it's probably easier for the crate to have no features at all. A crate that wraps safe_arch could choose to make this exclusion, if desired.