David Widmann

Since passing around the integrator in OrdinaryDiffEq is not completely straightforward (at least it seems to me), I started playing around with something that's more centered around the use case...

I'll take a look at it and see if I can figure out the issue.

Hej! I had a look at your notebook, and you're right, there was an issue with ForwardDiff if the delays were parameters. Fortunately, the latest changes in DelayDiffEq had already...

Yes, that's clearly the biggest issue, and maybe one has to adapt or change some of the ideas. Maybe this would work: - define `DDEFunction` as outlined above - define...

> That could be done by allowing a choice of indices for which the `k` is then copied over as. It just needs an option and a handling in `savevalues!`....

I observed that (at least) one approximate check is required due to the fact that interpolating an ODE solution returns different values for no, scalar, and vector indices in some...

After removing `@muladd` in https://github.com/JuliaDiffEq/OrdinaryDiffEq.jl/blob/master/src/dense/interpolants.jl#L449, https://github.com/JuliaDiffEq/OrdinaryDiffEq.jl/blob/master/src/dense/interpolants.jl#L457, and https://github.com/JuliaDiffEq/OrdinaryDiffEq.jl/blob/master/src/dense/interpolants.jl#L466 I obtain the same result `-0.05731721456340372` in all three use cases. However, I don't understand why adding `@muladd` leads to inconsistent results...

> No it's not. That looks correct. You can even check it with: Ah, of course. Don't know why I didn't notice that `Tsit5` just exits early.

> Another thing we might want to look into is the extrapolation over discontinuities. When extrapolating over a discontinuity, the calculation becomes incorrect because of that discontinuity in the interval...

> So I am curious how Hairer and Gugliemi know that the value should be t = 16.8 since I can't seem to get an RK method to say it's...