Energy in fit.out wrong

Alloy Theoretic Automated Toolkit (ATAT)
1

I have an issue with one value of the energies listed in fit.out which does not fit into the general energy landscape with values well below one eV/site. It is the structure 193 with apparently an energy of -20.600382 eV/site. I rechecked the energy as well the number of atoms of the structure in the 193 folder and compared it to e.g. the 188 structure. Their respective number of atoms for each atom type are the same, whereas their energies are only a factor of 2 apart. So I can’t explain how the -20.600382 value could possibly be extracted from the data.

energy of 193: -.37716378E+03
energy of 188: -.19234066E+03

I attach the relevant output files as well as the 193 and 188 folders.

Any thoughts? Thanks in advance!

2

There is nothing illogical about the total energy being a factor 2 apart and the formation energies being a factor ~300 apart. A reference energy is subtracted from the formation energies, so this can magnify energy differences.

Now, in this case, clearly the energy of str 193 calculated by VASP is physically implausible. But this is VASP problem, not maps one…