From: Axel Kohlmeyer (akohlmey_at_gmail.com)
Date: Tue Aug 12 2014 - 14:26:05 CDT
On Tue, Aug 12, 2014 at 3:11 PM, Viswanath Pasumarthi
<v.pasumarthi_at_iitg.ernet.in> wrote:
[...]
> As Alex previously pointed out, the variations reported for O-H bond
> length are corresponding to the minimization stage. More importantly, I
> have observed that the deviations for the same O-H bond length during MD
> step are in the order of 10^-4, implying rigidBonds constraint working.
> Please correct me if I'm wrong, for some reason since NAMD doesn't support
> constrained minimization, these distortions are possible during
> minimization and for this reason, it is necessary to specify bond and
> angle force constants in the parameter file despite having rigidBonds
> constraint active.
SHAKE and RATTLE are derived for doing MD with velocity Verlet.
Deriving a constraint algorithm for minimization is a highly
complicated procedure and not worth the effort.
for most MD codes, especially something that is tuned for doing MD
efficiently, it is irrelevant that the minimization is not "exact" as
its purpose is solely to remove excess potential energy from initial
geometries. thus people simply rely on having reasonably large force
constants so the structure doesn't deform too much.
if you absolutely do want to do a minimization with constraints, you
have to do a simulated annealing procedure.
i think at this point it is safe to advise you to go back to studying
the NAMD manual closely and also revisiting whatever text books you
have at hand about MD with specific focus on the subjects touched on
in this discussion before doing anything else.
it is also worth pointing out that with a more complete, detailed and
accurate description of the problem you are facing, you could have
saved everybody (including you) a bit of time and grief.
axel.
-- Dr. Axel Kohlmeyer akohlmey_at_gmail.com http://goo.gl/1wk0 College of Science & Technology, Temple University, Philadelphia PA, USA International Centre for Theoretical Physics, Trieste. Italy.
This archive was generated by hypermail 2.1.6 : Wed Dec 31 2014 - 23:22:43 CST