nonlinear

The source code of lyapunov was one of the very first functions of DynamicalSystems.jl. It does not use the set_state! function for parallel integrators. That is a mistake (and in fact made us miss a bug).

One should re-write the rescale! function of the code to simply generate the new reference state and the just call set_state!.

We now have uniform look for 2D and 3D: http://makie.juliaplots.org/stable/makielayout/axis3.html

Would be nice to update examples that use 3D into the new feel. Will probably also simplify source code.

TSP claims to be type agnostic.
There should be tests for this.
In particular Float32 and Complex numbers will be of Interest.

Want to back this issue? **[Post a bounty on it!](https://www.bountysource.com/issues/64400386-add-tests-for-different-number-types-float32-complex?utm_campaign=plugin&utm_content=tracker%2F89129719&utm_medium=issues&utm_source=githu

I was really young programming-wise when I wrote parallel_integrator of in-place continuous systems. I wrote it as having a matrix state, where each column is each state of the dynamical system. I'm not sure why I went with this route instead of the route of having it a vector of states. I'm certain that I have benchmarked both cases extensively and chose the one with most performance.

But th

We have established the validity of our unified embedding methods because we even wrote a paper utilizing them in detail and comparing them with each other. There isn't any doubt that they work. However, the way we have written our tests is brittle, because (due to chaos) running the test on different machines may make them fail. For example while working on PR #93 I've seen things like:

M