Add some performance tips to the documentation

[Zentrik]

Adds a tutorial containing some julia specific performance optimisations and some resources for more information.

[Zentrik]

I left this as a draft just to skip the CI suite.

[maleadt]

Thanks, that will be a great addition to the docs!

[maleadt]

Another (more advanced) tip may be @cuda maxregs, even though that forces to spill it may push the register usage below a threshold, e.g., allowing the thread scheduler to double the amount of active threads.

[maleadt]

Great stuff, thanks! I'll merge this already so that it can be part of the docs of the upcoming release, but feel free to create follow-up PRs of course.

[maleadt]

Also, it might be good to attach [skip ci] or at least [skip tests] to commit messages, to avoid needless CI runs.