is it an issue of number of significant digits?

I have two identical input files except one has more significant digits than the other in ‘region’ commands to compute surface energy of {111} surface in Al. They report different energies and one is correct. I used version ‘lammps-16Sep12’ on linux machine. Does anybody know what is wrong with the input files?

INPUT FILE I: