AMBER Archive (2002)

Subject: Re: SHAKE troubles

From: Yong Duan (yduan_at_chem.udel.edu)
Date: Wed May 29 2002 - 07:33:14 CDT

SHAKE failure is typically linked to instability of the system.
If this happens at the beginning of the simulation, it may imply that
atoms are too close to one another which can be resulted from initial
model building. If this happens in the middle of a simulation, it may
imply coding/algorithmic problem. In the later case, we'd appreciate if
you can provide detailed information to help us to track the problem.

yong

On Wed, 29 May 2002, Arvid Soederhaell wrote:

> Dear all,
> I have some problems with SHAKE that I do not understand. I get the error
> message
>
> COORDINATE RESETTING CANNOT BE ACCOMPLISHED,
> DEVIATION IS TOO LARGE
> NITER, NIT, LL, I AND J ARE : 0 0 1963 3195 3196
>
> no matter what i do???!!! The error message comes from the shake.f
> subprogram. The message implyes that i take a too long timestep, but even
> a timestep of 0.5 fs is too long! (Normally one should be able to use a 2
> fs timestep.) The simulation is stable if i turn off SHAKE and use a
> 0.2 fs timestep. The temperature in the simulation never exceeds 350 K.
>
> Any suggestions to what could be the problem?
>
> (The system is a DPPC lipid membrane with TIP3 water, and I use amber6)
>
> Arvid
>