From: Fangqiang Zhu (fzhu_at_ks.uiuc.edu)
Date: Thu Dec 04 2003 - 18:05:24 CST
Velocity rescaling is just one of a few temperature control algorithms
in NAMD. If it was shown to be problematic, you can always use other
alternatives (e.g., Langevin dynamics) to control temperature.
Zhu
Shantenu Jha wrote:
> > Regarding your Q1, first of all, how do you define the "global rotational
> > energy"?
>
> .. as rotational kinetic energy.
>
> FWIW, I'm concerned with the type of problem as mentioned in:
>
> @Article{ ,
> author = {Stephen C. Harvey and Robert K.-Z. Tan and Thomas E.
> Cheatham},
> title = {The Flying Ice Cube: Velocity Rescaling in Molecular
> Dynamics Leads to Violation of Energy Equipartition},
> journal = {Journal of Computational Chemistry},
> year = {1998},
> volume = {19},
> number = {7},
> pages = {726},
> }
>
> Shantenu
>
> > Shantenu Jha wrote:
> >
> > > Dear NAMD Junkies,
> > >
> > > I'm try to script a set of simulations so as to be able to remove the
> > > centre of mass and global rotational energy, say every 10E5 timesteps.
> > >
> > > The approach I'm taking is to a start a simulation, run for the 10E5
> > > steps, checkpoint (coor, vel and xsc) and then restart as a new
> > > simulation. I'm using the default value of COMmotion which takes care of
> > > the centre of mass motion at the beginning of each (sub)simulation.
> > >
> > > Q0: Is there a more elegant &/or efficient way to do this?
> > >
> > > Q1: How to remove global rotational energy?
> > >
> > > Q2: When I try to set a variable in my NAMD config file by reading in an
> > > environment variable, the environment variable is not being parsed (at
> > > all). This problem is unique to the Linux precompiled binary (2.5), as all
> > > works well using the precompiled Origin (MPI) binary. (The environment
> > > variable is being set by a driver perl script, but I've set it by hand and
> > > checked for correctness using 'echo $...')
> > >
> > > I've attached below the relevant part of the config file. Any ideas?
> > >
> > > Since I'm at it can I also ask, how NAMD goes about adjusting initial
> > > velocities to remove COM motion of the sytem on using 'COMmotion no'?
> > >
> > > Thanks.
> > > Shantenu
> > >
> > > ===
> > >
> > > set run_num $env(RUN_NUM)
> > > set first_timestep $env(FIRST_TIMESTEP)
> > > set run_step $env(RUN)
> > >
> > > restartfreq $run_step
> > >
> > > outputname evris24.$run_num
> > > extendedSystem evris24.0.$first_timestep.xsc
> > > bincoordinates evris24.0.$first_timestep.coor
> > > binvelocities evris24.0.$first_timestep.vel
> > >
> > > firsttimestep $first_timestep
> > > run $run_step
> >
> >
>
> --
> --Shantenu Jha, Research Fellow
> Centre for Computational Science
> University College London (UCL) http://www.realitygrid.org
> s.jha_at_ucl.ac.uk 020-7679-5300 PGP Key ID 0x78F823E6
This archive was generated by hypermail 2.1.6 : Wed Feb 29 2012 - 15:37:12 CST