It seems that the system is exploding!
Check the atom ordering in your initial and final structures, they should be the same. For example, if you have a CO2 molecule and the order of atoms is “COO” in the reactant structure, it cannot be “OCO” for the product.
Example of bad input structures:
..
..
..
Ru 5.51 4.68 11.98
Ru 8.21 4.74 11.96
C 4.20 2.26 14.11
O 4.21 2.16 15.27
O 1.42 0.82 13.41..
..
..
Ru 5.51 4.71 11.95
Ru 8.20 4.72 11.95
O 2.17 1.22 13.75
C 3.31 1.94 13.89
O 3.82 2.30 14.93
The initial###.xyz and ISOMERS#### files’ numbers should match.
The ISOMERS file’s name and it’s content (ADD, BREAK, ANGLE and TORSION) should be in capital letters.
Make sure all the files are in correct directories. The job should be submitted from the upper directory of scratch directory.
(qsub scratch/go_gsm_dft.qsh)
The string does not contain a peak as the transition state and two nodes with similar energies form a flat line as maximum energy structures.
Increase the number of nodes to get better resolution in that area.
The jobs using VASP as calculator should be run in different directories since VASP generates output files with identical names.
The string reaches maximum iteration during the optimization.
MAX_OPT_ITERS should be increased to a higher value (default value is 100). Then, restart the job.