Minima Hopping with Vasp, error in MD initiation

ASE
1

Dear ASE Community,

I am attempting to run a minima hopping study for OH on Pt (I cannot attach files) using Vasp as a calculator. If anyone has examples of minima hopping with Vasp, I would appreciate them. Below are the problems I have.

The job performs an initial qn step fine, but when it enters its first md step, the positions of the adsorbates are set to NaN and the get_potential_energy calls fail. Accordingly, vasp fails with Orbital orthonormalization failed in the inversion of matrix. LAPACK: Routine ZPOTRF failed! I have attempted the job with variations of ISTART = 0, IBRION = 0, ISYM = 0, to no avail.

Any guidance or working examples of Minima Hopping with Vasp would be appreciated.

2

To anyone coming to this thread in the future, the source of the bug was identified to be duplicate FixAtoms constraints. This can be cause by writing additional FixAtoms to an existing Trajectory/Atoms object which has set FixAtoms.

I am not sure how this translates to the POSCAR adsorbates being written as NaN, but removing duplicate FixAtoms prevents that.