Read_data for atom_style hybrid ellipsoid line

[springboy]

Hello, I am attempting to build a 3D simulation containing ellipsoids and lines. I am attempting to implement my workflow of reading in objects of those kinds but I am running into a segmentation fault error with my read_data step. Attached are the simulation file (sim_buggy.in), the input file giving an error at the read_data step (demo_ellipsoid_line_hybrid.in), the commands I use to execute the simulation (RunSim.sh), and the resulting error message (error.log). I’m using the LAMMPS version made on Aug 2 2023.

When I input a file that is formatted similarly to the first input file but has only ellipsoids and no lines, the simulation seems to move through the read_data command just fine. (This similarly formatted input file, demo_ellipsoid_only.in, is also attached.) So, I’m inclined to think the issue is my Lines section. However, based on my read of the documentation for read_data, it appears all that section requires is the atom-ID of the segment then the x1 y1 x2 y2 of the segment endpoints. Am I missing something for how to define the Line objects? I’m very sorry if this was enumerated somewhere in the documentation and I just missed it!

Disclaimer: sim_buggy.in still crashes even with demo_ellipsoid_only.in; I’m trying to build the minimal simulation for my purposes step by step and will add in a complete set of pair coeffs after I get the data input step working.

Thank you for whatever help and time you can provide!

springboy

demo_ellipsoid_line_hybrid.in (221 Bytes)
demo_ellipsoid_only.in (160 Bytes)
error.log (2.9 KB)
RunSim.sh (149 Bytes)
sim_buggy.in (976 Bytes)

[akohlmey]

@springboy

I am attempting to build a 3D simulation containing ellipsoids and lines.

How would a line particle “know” how to interact with an ellipsoid?

According to their documentation, all they are able to interact with are other line particles or point particles with a regular Lennard-Jones potential.

@springboy

So, I’m inclined to think the issue is my Lines section. However, based on my read of the documentation for read_data, it appears all that section requires is the atom-ID of the segment then the x1 y1 x2 y2 of the segment endpoints. Am I missing something for how to define the Line objects? I’m very sorry if this was enumerated somewhere in the documentation and I just missed it!

Before trying to build a hybrid system combining multiple atom styles, I suggest you first set up a system with only line particles. There is an example in the examples tree for guidance.

Then (and only then) after you have successfully run an input with ellipsoids and also an input with lines, you can consider setting up a system with one ellipsoid an one line. The rules for how you have to write the Atoms section for hybrid atom styles are explained at great length and with high detail in the documentation for the read_data command. If you feel that something is missing, read it again. It is all there. Just don’t expect a tutorial telling you exactly how to solve your specific problem.

But then again, before you invest more time, you need to solve the fundamental problem of how to compute all interactions correctly.