KS energy is an abnormal value (NaN/Inf).

[Raymond Amador]

Good morning everyone,

I have been encountering somewhat of an odd problem on CP2K on both Piz Daint and LUMI, whereby the following error message is displayed:

 *******************************************************************************
 *   ___                                                                       *
 *  /   \                                                                      *
 * [ABORT]                                                                     *
 *  \___/                KS energy is an abnormal value (NaN/Inf).             *
 *    |                                                                        *
 *  O/|                                                                        *
 * /| |                                                                        *
 * / \                                                     qs_ks_methods.F:888 *
 *******************************************************************************

I first got the error message running some MD simulations, and then decided to just try a simple geometry optimization, and the problem still persisted. I have tried playing around with the convergence parameters and the basis set size, and, after around 15 OPTIMIZATION STEPS or so, the error prints.

Would anyone know how to solve this issue? Attached are the input and output files as well as the submission script on LUMI.

Thanks and best regards,

Ray

[fabia...@gmail.com]

Dear Raymond,

Have you checked how the structure evolves? I suspect it changes drastically. The energy change per iteration is ~10 Ha which is enormous (more than few 100 mHa is suspicious). From a glance at your input deck and the structure I can see a few issues.

1. the periodicity of the poisson solver and the cell don't match

2. there is very little vacuum between the molecule and the periodic image of the substrate

3. only the top surface of the substrate is passivated

You need to decide how you want to deal with periodicity and adapt the input and the structure accordingly.

Cheers,

Fabian