[gmx-developers] Energy not conserved in GROMACS replicate

[Berk Hess]

Hi,

The total energy will never be conserved perfectly. How good it is 
conserved depends on the time step.
But it seems, although it's not clear from your mail, that you are 
talking about your own program. This is a mailing list for problems with 
Gromacs relatead problems, not for help with your own code. All I can 
add is that most bugs in an MD code will cause to total energy to increase.

Cheers,

Berk

On 05/10/2016 05:57 AM, #CHEN WENJIE# wrote:
> Hi GROMACS team,
>
> I am an undergraduate student from Nanyang Technological University. I 
> am working on my MD physics engine project that replicates some of the 
> GROMACS features. The programming language is Java. The project is to 
> simulate some small molecules in atomic level. The topology file is 
> generated by GROMACS with pdb2gmx program. The force field I choose is 
> AMBER03, and  force field parameters are also from GROMACS. I use 
> Velocity Verlet integrator.
> The problem I encounter is that when I run the simulation, the total 
> energy doesn't conserve. It increases a little bit every frame and 
> eventually becomes huge and the molecule deforms.
>
> Attached is my GRO file and TOP file, and some log output files for 
> the simulation of water molecule. One only involves bond stretching 
> energy. One only involves angle rotation energy. One involves both 
> bond and angle energy. For each frame, I log the kinetic energy, 
> potential energy, total energy and position of each atom. Although 
> when Ep increases, Ek will decrease, and vice versa, the total energy 
> increases a little bit each frame.
>
> I really want to know if there is any part I miss out or didn't do it 
> correctly. I really appreciate it if you can give me some advice. 
> Thank you in advice.
>
> Best regards,
> Wenjie
>
>
>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://maillist.sys.kth.se/pipermail/gromacs.org_gmx-developers/attachments/20160510/f661fa7b/attachment.html>