[lammps-users] Restart from .restart file

Axel and Ahmed,
Thank you so much for your responses. I missed that caveat of restarting jobs.
I will try your suggestions in the future.
Thanks!
~Kate

Date: Tue, 29 Jun 2010 08:47:28 -0400
From: Axel Kohlmeyer <[email protected]>
Subject: Re: [lammps-users] Restart from .restart file
To: "Kathleen E. Ryan" <[email protected]...>
Cc: LAMMPS <[email protected]>

Hello,

hi kate,

I have a question about restarting from a .restart file. I have a system that I

am

trying to equilibrate. I first quenched the system to 0K using the nvt

thermostat

which yielded a .restart file. It was my understanding that the .restart file
contains all the info about the atoms (positions, velocities, masses, etc) but

not

any of the input parameters such as the thermostat or temperature criterion.

So

I started a new job using the read_restart command with my .restart file, but
now changed the nvt thermostat to go to 300K instead of 0K. However, after

10

ps, the temperature is still at 0K. If, however, I use the restart2data.cpp tool
included in LAMMPS and convert my .restart to a .dat file, then restart the

job

with the nvt thermostat set to 300K, the temperature will increase. So my
question is, am I incorrect about the .restart file? Is it actually storing
information about the input parameters as well as about the atoms being

yes, the restart stores most of the input parameters and settings for
fixes and so on in order to facilitate clean restart on production
simulations. the details about what is written to a restart is given
for each fix, pair/bond/angle/dihedral/improper style and so on.

simulated? Or should I be able to restart a job using the .restart file as a
coordinate file while also changing the input parameters in my .in file?

just think of it as if you have a two part input. to change parameters
of a fix you have to unfix the fix id and then create a new fix.
to override potential parameters you have to reset them after
the restart and so on.

If the answer if yes to the latter, do you have any thoughts on why this might

not