Voronoi neighbor lists on subsets of particles?

[drayer587]

I am working with neighbor de-correlation functions for the first time. I would like to look at neighbors of only a subset of the particles in the system (i.e., ignore some particles as being neighbors).

In trying to do so, building the Voronoi neighbor list throws a segmentation fault. Is this something that's out of scope? I am trying the neighbor list build with all atoms now, but this really isn't the desired calculation, e.g., I want to ignore hydrogen atoms, not to mention that this would likely be slower as I'm doing a lot of calculation I don't think I want in principle. Is it necessary to build those neighbors even if I don't want them? Is there something else I'm missing?

A test case could be built out here using the all-atom polystyrene that is one of the canonical systems for testing; as above, it would be nice to see the difference between all neighbor de-correlation versus ignoring hydrogen.

[pkawak]

I don't think that AMDAT is able to build a Voronoi list without considering all atoms in the box. I believe that's because @dssimmons hadn't considered a case where that would be useful but I may be wrong. The case you bring up sounds like a good argument to have that be an option and a good test system to explore its effect.

As an aside, in the mean time, you can probably use OVITO or some other trajectory editing tool to filter out your Hydrogen atoms and save the trajectory without the Hydrogens for AMDAT analysis.