Would wrap during MD make the result unreasonable?

[marvcks]

https://github.com/torchmd/torchmd/blob/78063fb38ec0cff366a7d42361947e62efe8112d/torchmd/wrapper.py#L4C1-L30C59 I am using TorchMD to run molecular dynamics simulations. I have a question regarding the setup. In my simulation, I am working with a small molecule surrounded by 500 water molecules. I am considering fixing the small molecule at the center of the simulation box (wrapidx is not None) and wrapping the water molecules accordingly. My concern is whether this approach might lead to unreasonable MD results. Or are there some paper use the same approach ? P.S. Due to my special FF, i must fix the small molecule at the center of the box.

[stefdoerr]

The only issue you might have are numerical issues from the extra operations but I think you should be fine if you go with torch.double precision (might be ok even with single)