Re: [AMBER] SHAKE&Vlimit

From: David A Case <david.case.rutgers.edu>
Date: Mon, 21 Sep 2015 11:48:01 -0400

On Mon, Sep 21, 2015, 罗艳玲 wrote:

> Thanks, Jason! Maybe I didn't express it clearly.The work get shake
> erreos with a timestep of 1 fs. In order to see what's wrong, I have
> tried one production with a timestep of 0.5 fs and printed out the
> result every 1 step. But the similation ended up normally with a very
> large output file. And I couldn't open it. At that moment, I think there
> is nothing wrong with my system.

If the output file is too big (how big is it?) then use cpptraj to output
(say) only every 10th step, and look at that. Or, run a shorter simulation,
and look at that. If you are getting SHAKE errors with a 1 fs time step, it
is very likely that something is indeed wrong.

> So I redo the preoduction with the result printing out every 1000
> steps(timestep 0.5 fs), then there are many vlimit warnings. But
> the work continue. I wonder if the work succeeds, may I believe the
> result? And what can cause such vlimit warnings?

If you just get vlimit warnings during equilibration, and the go away later
on (and you can move to a 2 fs time step using SHAKE), the it's likely to be a
problem with just the initial structure, and you are probably OK. But if
vlimit errors persist, there is something wrong. The *usually* happens when
there is some non-standard species in the simulation, and reflects a
parameterization problem with that species. So, it is of interest to know
whether you have non-standard parts to your simulation.

...dac


_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Mon Sep 21 2015 - 09:00:03 PDT
Custom Search