Optimisations (inlining, …)

[Tamschi]

Is your feature request related to a problem? Please describe.

This crate is currently barely optimised.
As it contains extremely low-level building-blocks, it seems sensible to optimise it very thoroughly.

Describe the solution you'd like

Before beginning with the optimisation, a way to benchmark the crate (and to monitor regressions in this regard) needs to be added.

The optimisations should ideally not affect the crate's API.

Describe alternatives you've considered

None.

Additional context

It's likely that most or all of the actual optimisations can be copied directly from Rust's standard library, as its Arc (and Rc) are very similar.