ASE running molecular dynamics errors

[bawanglong-2]

ASE running molecular dynamics errors

[ilyes319]

you clearly have not installed mace properly in your environment. Can you please install mace with pip and then do python and then import mace in the terminal.

[Chernjz]

Hi, I'd like to know, when doing molecular dynamic with LAMMPS(with/without kokkos) or ASE, which one is faster?

[Swiftily]

You can't use like this
python mace.py
Since there's already a mace package. Rename your python file.

[bawanglong-2]

Hello, I changed the name later, my .xyz file and .cif file ran molecular dynamics, found that it ignores periodicity, Unable to react in the box, should I add something to the executable?

[gabor1]

you need to map atoms back to the unit cell. Have you relaxed the initial configuration before you started the MD ?

[bawanglong-2]

Hello, may I ask how to remap atoms back to the unit cell? The structure I'm referring to is a relaxed structure.This is my files ,Iuse mace-mpa-0 to deal with [this.
mac111e.txt
666.txt

[gabor1]

It's just a visualisation thing. MACE fully respects the PBC you have set in ASE, so your simulation is fine from this perspective. If you visualise using OVITO for example, then there is a "wrap" option in the big menu that adds to the visualisation pipeline.

[bawanglong-2]

Thank you very much. Indeed, that's exactly the case.

[Swiftily]

Is MPA-0 a reation model? I'm not sure.

[gabor1]

I assume you are asking whether it is a "reactive" model. yes it is, all mace models are, there is no fixed topology.

[Swiftily]

I mean whether it can describe a ''chemistry reaction''?

[gabor1]

yes

[gabor1]

as a foundation model, it will give approximately the right behaviour, and will be smooth and continuous as atoms form and break bonds. for very high accuracy you will likely need fine-tuning.