The NEB calculation is performed by the following three steps:
where in the three calculations users have to keep the same computational parameters such as unit cell, cutoff energy, basis functions, pseudopotentials, and electronic temperatures to avoid numerical inconsistency. After the calculations 1 and 2, files *.dat# are generated. By using the atomic coordinates in the files *.dat#, one can easily construct an input file for the calculation 3. Once you have an input file for the calculation 3, the execution of the NEB calculation is same as for the conventional OpenMX calculation such as
% mpirun -np 32 openmx input.dat -nt 4