Re: How to get around with "atoms moving too fast"

From: yi wang (yiwang_at_ks.uiuc.edu)
Date: Fri Oct 19 2007 - 22:34:32 CDT

Hi Yinglong,
      Would it help if you first minimize your new configurations
after you generated them using your functions? The purpose of
increasing velocity limit would be, I guess, to let NAMD run and get
rid of those "strange" configurations, e.g., stretched bonds. This
kind of problem is often well solved by a few thousand steps of
minimization. Another option might be to add some constraints in your
function to prevent it from generating too twisted/strange structures.

Yi

On Oct 19, 2007, at 11:45 AM, Yinglong Miao wrote:

> Hi, NAMD developers and users,
>
> I am introducing a new method into NAMD to accelerate MD simulations.
> For the protein system in vacuum I have been testing, it works out
> well
> for the direct MD simulations by following minimization of the initial
> PDB structure, thermalization and finally a MD product run.
>
> But I have been getting error messages, like "FATAL ERROR: Bad global
> exclusion count!" and "atoms moving too fast", when implementing my
> method. The method for accelerating MD involves a set of simulation
> cycles, in which a few number of MD time steps are run first, then
> with
> the output structure from MD run we apply some mathematical functions
> to atomic coordinates to get a new configuration and put it back to
> the
> MD run as input to start another cycle. Depending on the parameters
> and
> functions I use to generate the new configuration, the simulation can
> be done for as many cycles as I want, but can also crash with the
> above
> error messages after several cycles when the configuration
> generated is
> too strange with long stretched bonds, too close atoms, etc.
>
> So I am wondering whether I can make the atom velocity limit larger
> in NAMD to get around the error or run an additional MD run with a
> smaller timestep after getting the new configuration. But it seems
> I can set only one timestep in NAMD configure file. Is there a way
> to set different timesteps for multiple MD runs?
> And are there any other NAMD techniques I can make use of to tackle
> my problem? Your comments and suggestions are greatly appreciated.
>
> Thanks,
> Long
>
> --
> Yinglong Miao
> Ph.D. Candidate
> Center for Cell and Virus Theory
> Chemistry Department, Indiana University
> 800 E Kirkwood Ave Room C203A, Bloomington, IN 47405
> 1-812-856-0981(office); http://ylmiao.dict.cn/mypage/
>

This archive was generated by hypermail 2.1.6 : Wed Feb 29 2012 - 15:45:24 CST