A version with the bond deleted (though the atom is still present and included in the exclusion list as if it were bonded) minimises and equilibrates regularly, leading me to believe that it is that short bond that causes instability.
Minimisation with pmemd (as opposed to pmemd.CUDA) proceeds normally, but equilibration gives a segmentation fault a few seconds in with shake off. If shake is on, it stops almost immediately, saying
"Coordinate resetting cannot be accomplished,
deviation is too large"
Performing the same simulation, with shake on and a much lower timestep (20 times as low, from 0.002 to 0.0001) makes it run to completion.
________________________________________
From: amber-bounces.ambermd.org [amber-bounces.ambermd.org] on behalf of Adrian Roitberg [roitberg.ufl.edu]
Sent: 27 March 2012 01:08
To: AMBER Mailing List
Subject: Re: [AMBER] Amber and Short Bonds
I would most certainly CHANGE the values of ntc and ntf to compute all
forces for the minimization. Put them back to normal after minimization.
Also, see if you can create a system WITHOUT solvent and ions and see if
it also dies. It is important to separate solvent issues from internal
ones when diagnosing.
Third, look at the manual for the xmin minimizer. It is much more
efficient than the regular minimizers.
Adrian
On 3/26/12 5:02 PM, Aronica, Pietro wrote:
> Yes.
> These are the minimisation input files I use:
> First minimisation: initial solvent + ions min
> &cntrl
> imin=1,
> maxcyc=1000,
> ncyc=500,
> ntpr=100,
> ntb=1,
> ntr=1,
> cut=8.0,
> /
>
> Second minimisation: full min
> &cntrl
> imin=1,
> maxcyc=2500,
> ncyc=1000,
> ntpr=100,
> ntb=1,
> ntr=0,
> cut=8.0,
> /
>
> Both give errors during the minimisation, saying .... RESTARTED DUE TO LINMIN FAILURE ...
> The first one reaches conclusion, but the second one has five such errors and crashes with ***** REPEATED LINMIN FAILURE *****
> There still is restart file, but proceding with heating and density equilibration, the simulation just stops.
> ________________________________________
> From: amber-bounces.ambermd.org [amber-bounces.ambermd.org] on behalf of Jason Swails [jason.swails.gmail.com]
> Sent: 27 March 2012 00:47
> To: AMBER Mailing List
> Subject: Re: [AMBER] Amber and Short Bonds
>
> But did you minimize first?
>
> --
> Jason M. Swails
> Quantum Theory Project,
> University of Florida
> Ph.D. Candidate
> 352-392-4032
>
> On Mar 26, 2012, at 7:18 PM, "Aronica, Pietro"<pietro.aronica07.imperial.ac.uk> wrote:
>
>>> Also, make sure you haven't drastically increased the bond force
>>> constants. The values used are right at the edge of stability. The fact
>>> that you have to go to shorter time steps suggests that you boosted the
>>> force constants for some reason...
>>
>> Hello,
>> no, the force constants are those of "normal" bonds, standard values of CT-CT and CT-HC. The only thing I've changed is the equilibrium bond length.
>> ________________________________________
>> From: amber-bounces.ambermd.org [amber-bounces.ambermd.org] on behalf of Thomas Cheatham [tec3.utah.edu]
>> Sent: 26 March 2012 21:21
>> To: jason.swails.gmail.com; AMBER Mailing List
>> Subject: Re: [AMBER] Amber and Short Bonds
>>
>>> The equilibrium bond distances should have little effect on the
>>> stability of your simulation, I would think. You don't have to worry
>>> about VDW and EEL clashes between bonded pairs (since those interactions
>>> are excluded from bond/angle partners).
>>
>> Also, make sure you haven't drastically increased the bond force
>> constants. The values used are right at the edge of stability. The fact
>> that you have to go to shorter time steps suggests that you boosted the
>> force constants for some reason...
>>
>>
>> _______________________________________________
>> AMBER mailing list
>> AMBER.ambermd.org
>> http://lists.ambermd.org/mailman/listinfo/amber
>>
>> _______________________________________________
>> AMBER mailing list
>> AMBER.ambermd.org
>> http://lists.ambermd.org/mailman/listinfo/amber
>
> _______________________________________________
> AMBER mailing list
> AMBER.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber
>
> _______________________________________________
> AMBER mailing list
> AMBER.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber
--
Dr. Adrian E. Roitberg
Professor
Quantum Theory Project, Department of Chemistry
University of Florida
roitberg.ufl.edu
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Tue Mar 27 2012 - 05:30:03 PDT
