This is a follow up post for the questions in this post. Since there are some questions related to simulations tips in the original post so I think it makes sense to create a new one here in the FDTD section.
If you are interested in this type of simulation, I would recommend you to look at this page in the KB for OLED simulation methodology. I will also encourage you to start with a 2D setup like this in the first place, if possible. A 2D simulation should be a lot faster than that of in 3D.
As far as suggestions specific to you file in the original post goes, I think the most suspicious problem is the thickness of the Ni layer. Is it supposed to be 0.0001 nm? For a super thin layer like this, it may make sense to consider using a 2D rectangle object with the sampled 2D material model. It might also be related to the engine error you observed, although i am not 100% sure.
To speed up the simulation, you could try to start with a smaller simulation x and y span to include a smaller number of nanowire, say 5x5. This will speed up the initial simulations a lot and you can certainly increase the spans to include more wires down the road.
One more note, you can also try to switch the PML type to the new SC-PML. It should have a better performance in general.