[lammps-users] Different results from diff versions and diff systems

Dear All,

I am minimizing a gold nanowire (boundary conditions: sss) in Linux system and Mac OSX.
LAMMPS version in Linux system is " 7 Jul 2009 " and in Mac OSX is " 5 Mar 2010 "
I am getting different results from the two systems.
I am attaching input file, data file, potential file herewith

Length before minimization Length after minimization
Linux system 304.39 305.87
Mac OSX 304.39 313.72

Could anyone help me


in.edensity.80x6x6 (1.81 KB)

data.80x6x6 (454 KB)

Au.eam.0.15.alloy (1.68 MB)

What is the initial/final energy of the 2 systems?
Assuming the initial energies are the same, do the 2 runs track each
other initially and then diverge?

Please use the most current LAMMPS.
Minimization is a tricky numerical business in atomistic systems
with respect to local minima, convergence criteria, etc. Lots
of changes have been made to the minimizers in LAMMPS in the
last year+, so comparing the two is not simple. The only thing
we can really help with, is whether the current version is
doing the right thing.