From: Brian Radak (bradak_at_anl.gov)
Date: Mon Aug 28 2017 - 09:12:40 CDT
This all depends on what you want to simulate. Changing the thermostat
temperature (usually langevinTemp) only changes that parameter within
the integrator and thus only affects future dynamics.
If there is a mismatch between your integrator temperature and the
notional temperature at which a configuration was sampled (i.e. the mean
kinetic energy per degree of freedom from the velocity file) there is no
problem /per se/. However, if you are trying to determine an equilibrium
property then this may result in deviations from the target distribution.
Note that velocity relaxation may be rather fast (i.e. the kinetic
energy may stabilize quickly after the thermostat temperature changes),
but this is not necessarily the same as the system "re-equilibrating,"
since it gives no indication as to whether or not the structure is good
or not. For example, consider running a simulation of frozen water at
100 K and then raise the thermostat to 300 K - the temperature will rise
just fine, but I doubt the lattice will melt on that same timescale.
On 08/26/2017 02:20 AM, t.shivam_at_iitg.ernet.in wrote:
> Dear namd users
> If I equilibrate my system at a certain temperature and after that i wish
> to run a simulation at a temperature different than that of my
> equilibrated system, do i need to equilibrate my system to this new
> temperature or i can directly run a production run to that new
-- Brian Radak Postdoctoral Appointee Leadership Computing Facility Argonne National Laboratory 9700 South Cass Avenue, Bldg. 240 Argonne, IL 60439-4854 (630) 252-8643 brian.radak_at_anl.gov
This archive was generated by hypermail 2.1.6 : Sun Dec 31 2017 - 23:21:35 CST