# Inquiry about obtaining the neighbor list in compute coord/atom

I have been using the find_computes_neighlist function, and based on your comments, I was able to retrieve the neighbor list after executing the run command. However, it appears that I was unable to obtain the neighbor list within the compute coord/atom commands. It is more likely that the neighbor list is associated with the pair style. Actually, I am wondering if the find_computes_neighlist and get_neighlist functions can be used to obtain the index of the neighbor list in the compute coord/atom command, as I require these values to update the force in my simulations.

Thank you for your time and assistance.

This is a pretty strong statement. From looking at the source code of how the neighbor list access is implemented, that is impossible. Putting that aside, why does it make a difference whether you get the neighbor list for the pair style or the compute?

…and I am wondering, how you come to this speculation.

But since it is you that is claiming that the code is not working according to its documentation, the burden of proof is with you.

That is an unusual claim as well. You need to explain how this is supposed to work as well.

It seems to be a misunderstanding on my part. I might have thought that find_compute_neighlist is used to find the index of a neighbor list of a compute. Upon carefully reviewing the documentation, I realized that it is actually used to find the index of a neighbor list requested by a compute. Is there any possible means to obtain the computed atomic coordinates from compute coord/atom?

The information you are looking for are the indices of the N closest atoms. That is not computed by compute coord/atom only the number N. You would have to implement this as a different compute style in LAMMPS in C++ or use the neighbor list you have requested to compute this in python. That is straightforward to do in serial but tricky business in parallel.

Okay, I understand. So, is it relatively easy to retrieve the indices of the N closest atoms using the existing “compute coord/atom” function?

No.