Very large performance difference between Turing and Stan [Gaussian Process]

[dlakelan]

I remember having a huge slowdown with truncated normals…

Making Turing Fast with large numbers of parameters?

It’s a bit difficult to find, but PRs to the performance section of the website can be made against the https://github.com/TuringLang/Turing.jl/blob/master/docs/src/using-turing/performancetips.md file.

See if anything discussed there helps?