total energy conservation

LAMMPS LAMMPS Mailing List Mirror
#1

Hi to all,

Thanks for the fast reply.

Please find the attachment of 1 united atom nonane with trappe model.

flooding all mailing list subscribers with a 15MB e-mail attachment is
not the way to get your problem resolved. you will just massively
increase global entropy and probably also upset a lot of people.

folder consists of log.lammps ran without fix_modify energy yes &
log.lammps_with_fix_modify

Please look at "TotEng" as a function of time step.

without fix_modify energy yes : energy keep fluctuating around mean position

with fix_modify energy yes : energy keeps drifting

nobody will look at these files. as i was mentioning before: step 1 is
that you *yourself* validate your input data and parameters by running
(after suitable equilibration) in NVE ensemble and confirm that your
settings conserve energy.

once you can confirm that, you should construct a small(!) test case
that can easily and quickly(!) reproduce where you believe LAMMPS is
wrong.

just dumping your simulation on hundreds of people is not going to do
that for you. that is not how mailing lists work.

sorry,
    axel.

#2

Besides, u must first confirm your problem in every possible way, viz changing the timestep(.5 ps in your case), playing with the Tdamp values(Notice the large fluctuation in your temperature) etc.