[gmx-users] Re:(Mark) The X-size of the box times triclinic skew factor is smaller than ...

[wuxiao]

Dear Mark,
  Thank you very much for the quick reply. According to your suggesstion, I had a look at the structure or trajectory using ngmx. Only to find that there are some big hole in the center, which seems to indicate the system is very inhomogeneous. And the system volume  decreased slowly.  when I using cutoff instead of PME for columbic force, and change the cutoff from 1.4 to 1.0, the mdrun finished 300 ps successfully. Another 3ns is started to try whether it can succeed too. more help from these descriptions?
Best regards,
xiaowu
 

wuxiao wrote:
> Dear Mark,
> Thank you for your reply. The system I simulated was one Polymer 
> chain solvated by 200 water that had totally about 3000 atoms. I had 
> defined the system density as 1000 using editconf.

That just does a scaling. It might not fix a gross problem, so do look
at the box dimensions before and after.

> This density was very 
> the experimental one. After enough minimization (convergent to 0.1 
> kj/mol) and 3 ns of NVT MD, the NPT MD was performed as describe in the 
> original question. It happened to terminate with the fatal error after 
> about 300 ps. It should be noted that I ran the mpimdrun using the 
> following command:
> mpirun -np 4 mpimdrun -s 03md.tpr -c poly_md.gro -v
> I have a four-core CPU. Can you give more help according to these 
> descriptions?

Your preparation seems good enough not to be the source of the problem.

Have a look at the trajectory. What does the structure look like? What
is happening to the box dimensions and/or density (look in the .log file
and/or use g_energy)?.

I believe that the error message you're getting is consistent with a
drastic decrease in the box size. Since EM and NVT don't break and NPT
does, the most likely hypothesis is that it started far too large.

Mark

> Best regards,
> xiaowu
> 
> wuxiao wrote:
> > >Dear gmx users,
> > >While doing a NPT molecular dynamics(MD), A fatal error was given:
> > >The X-size of the box times triclinic skew factor is smaller than the
> > >number of DD cells times the smallest allowed cell size.
> > >However, either NVT or NVE MD can work well. Can you give any help to
> > >deal with this problem?
> 
> >It sounds like you might have started with a density that is too low for
> >your box size.
> 
> >A more complete description of your system and what you've done with it
> >might help you get useful advice.
> 
> Mark
> 
> 
> ------------------------------------------------------------------------



把MSN装进手机，更多聊天乐趣等你挖掘！ 立刻下载！

_________________________________________________________________
Live Search视频搜索，快速检索视频的利器！
http://www.live.com/?scope=video
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://maillist.sys.kth.se/pipermail/gromacs.org_gmx-users/attachments/20090427/9f9781b5/attachment.html>