Thanks, Ray, for all your helps. I tried Nickel EAM potential, gold Morse
and all didn't produce correct results using LAMMPS fix NPT...
My suggestion is still to make sure that these potentials reproduces
reasonable elastic properties at finite temperatures.
I've really tried hard to exclude reasons that I can think of. That's why I
looking at LAMMPS algorithm and code itself...I'd like to pose a question:
Is LAMMPS fix NPT capable of produce fluctuation that is consistent with
mechanics rules, irregardless of human tunable parameters such as Pdamp?
You can post this question as a new thread on to the mailing list.