People have had success rolling their own explicit discretizations: Costs/benefits of solving large ODE systems as systems of difference equations/in discrete time
It’s things like this that make me interested in direct method stuff. I do like the exact gradients of the approximation thing, though it’s hard to know exactly how much trouble those gradients cause.
Edit: Last sentence is meant to say that I haven’t tried and I don’t know how to diagnose this, not that it isn’t important or can’t be diagnosed somehow.