Joel Clemmer

Adding an update, I believe the simplest way to patch this bug is to zero out the entire npartner array (replacing nall_neigh -> nmax on lines 354 and 409) in...

I'd be willing to work on this if there's still a need/interest. I've often wished there was an easier way to identify relevant examples. Condensing the above discussion, maybe a...

@sjplimp I updated the example and added links. I'll go ahead and write a python script to generate a list of all possible input scripts and start manually filtering out...

@akohlmey Thanks for the reminder, I'll go back and look at the docs.

I have an updated set of doc files in [this branch](https://github.com/jtclemm/lammps/tree/documentation-examples) . For all commands that were present in an example script (excluding some trivial commands like run, read_restart, etc)...

> First, because box size/shape change happens after force computation, then if a run is restarted, the newly computed forces will be different from those computed by lammps before the...

> In that scenario would it be "incorrect" to preserve the prior h_rate matrix? Yes, in that instance ghost atoms would initially assuming an incorrect old h_rate matrix instead of...

Thanks Dhairya for the contribution. It looks like a helpful addition to the granular package. Am I correct to assume there would be no reason to use the enhooke damping...

The normal model currently indicates whether the contact model is cohesive, `cohesive_flag`. Conflicts with cohesive contacts (e.g. in the Tsuji damping model) can then be defined using an `allow_cohesion` flag....

I think the best solution at the moment might be to add a user specified argument which controls the form of the prefactor (e.g. _hertz_ or _hooke_). We may then...