QuantumATK Forum

QuantumATK => General Questions and Answers => Topic started by: MolFET93 on May 16, 2022, 09:43

Title: Computational time estimation for NEB?
Post by: MolFET93 on May 16, 2022, 09:43

Dear QuantumATK staff,

I launched few days ago a NEB calculation for 7 images of a geometry containing 29 atoms (32 CPU).
The simulation is still running, and I would like to estimate more or less how much time is still required by inspecting the log file. Is there information to understand it?

In attachment the .py script and log file.

Thank you,
Regards

Title: Re: Computational time estimation for NEB?
Post by: Anders Blom on June 2, 2022, 01:54

I think you need to go back to the setup of the initial NEB trajectory, it doesn't look correct, the atoms are moving all over the place. For a NEB calculation to be successful (and mean something) you basically want most atoms to remain at or close to the same position, and only some particular atoms to move. Might be that the indices of the atoms have become mixed up, but best to go back to the initial idea and set it up properly.