Re: [AMBER] iwrap on very long simulation

From: Jason Swails <>
Date: Mon, 08 Sep 2014 10:02:46 -0400

On Mon, 2014-09-08 at 19:22 +0530, wrote:
> > On Mon, 2014-09-08 at 18:53 +0530, wrote:
> >> Thank you Jason.
> >> When I was not using iwrap=1, after ~ 60 ns or so I was getting
> coordinates overflow warning. Now I will try to use ioutfm=1 and ntxo=2
> for my new simulations. Hoping it won't give any problem. If we get
> such
> >> problem again can we change it back to iwrap=1?
> > Yes. But do you have a specific reason you don't want to use iwrap=1
> from the beginning? I always use iwrap=1 (several other MD programs
> effectively require setting iwrap=1...)
> There is no specific reason. Somewhere, I read the diffusivity and other
> property can not be determined if we use iwrap=1.

MSD and diffusion cannot be computed with wrapped trajectories (since
particles can appear to *jump* the entire length of a box). However,
you really need to run NVE to compute diffusion properties (thermostats
ruin temporal measurements).

That said, cpptraj has an "unwrap" command that can be used to reverse
the effect of iwrap=1 if you need to later.


Jason M. Swails
Rutgers University
Postdoctoral Researcher
AMBER mailing list
Received on Mon Sep 08 2014 - 07:00:08 PDT
Custom Search