RE: "the error of movement of atoms"

From: soroush ziaei (soroush.ziaei7337_at_gmail.com)
Date: Thu Mar 14 2019 - 02:59:31 CDT

thanks a lot, dr.radak.
your suggestion was very useful.
by the way, does this error is related to incorrect epsilon amount or another parameter of Lennard-jones in the force field parameters?
actually, all of these errors are related to water and I have chosen force constant and length of initial bond correctly.
another idea might solve this issue without increasing margin or another parameter...

From: Brian Radak
Sent: Saturday, March 2, 2019 1:48 AM
To: namd-l; soroush ziaei
Subject: Re: namd-l: "the error of movement of atoms"

I would recommend against lots of minimization. For a well prepared structure 200 - 1000 cycles should be more than enough (I've never done more than 200 in all of my years using NAMD). A large number of steps can cause trouble with periodic boundaries (something with the patch updates?). You can also try increasing margin (5 usually works for me) during the initial constant pressure equilibration and then set it back to the default (overall performance might be bad otherwise).

On Fri, Mar 1, 2019 at 4:05 PM soroush ziaei <soroush.ziaei7337_at_gmail.com> wrote:
hello everybody,
At first, I have to say the number of atoms of my simulation is 140000.
And after running the first stage of simulation that means minimization at 10000 steps in order to check the error before running a full simulation,
but at the next stage ( temp & pressure coupling ) an error appears in the log file that says some of the atoms with related numbers are unstable and moving too fast, why?
I think it's because of the short time of minimization, am I right? if not, please guide me.
 
Regards,
soroush

This archive was generated by hypermail 2.1.6 : Sat Dec 14 2019 - 23:20:33 CST