Pull-coord1-geometry = direction vs pull-coord1-geometry = direction-periodic

What issues would you need to watch out for when using direction-periodic as opposed to direction? I want to apply a pulling simulation on a rather large polymer and if possible I want to avoid setting too large a box size. I saw from this post that direction-periodic can be used to pull for lengths greater than half the box length, but saw some warnings on the manual regarding how the box can’t be dynamic and that the force is not added to the virial. Does this mean that if the box size is kept constant throughout the simulation, that direction-periodic can be used instead of direction without any effect on the results of the simulation? Or is there some issue that I’m not understanding and still need to watch out for?

Direction-periodic should only be used for cases where you want to pull over distances of more than half the box length. Such cases are very uncommon. Pulling a large polymer could be a valid use case. With an NVT simulation things should be fine. But you probably want to pull to a distance of slightly less than the full box size to avoid interactions between periodic images.

Lets say I run a simulation where I pull a polymer up to a certain length then hold it there with an umbrella potential for the duration of the simulation. Up to now I’d been using direction, so each box always had to be at least twice the length of the extension at which the polymer is held, even if the polymer is contracting and not exploring the lengths beyond the initial extension. If I were to use direction-periodic instead so as to use a smaller box, would there be any difference in the simulation results I should watch out for?

I think that should work fine. The only thing I am not completely sure of whether the initial distance is handled correctly, but I think it is.