Is it possible to parallelize `lmbtr.create` when working on one `ase.Atoms` object? about dscribe

[lauri-codes]

Hi @ARostekMU,

The parallellization is only effective when working on multiple systems. We do not parallelize the computations for a single system, as this is much trickier to implement. Our documentation should be updated to point out this fact better.

[ARostekMU]

Thank you for your answer.

If I were to split my ase.Atoms into multiple ase.Atoms and pass them to create in a list, would that give incorrect results?

[lauri-codes]

Depends how you do the splitting. The systems that you feed in still need to contain all of the neighbouring atoms within the cutoff radius from the LMBTR centers.

Such splitting might help if you have problems with memory, but I don't think it will make the computation faster as compared to feeding the system as a whole.