Re: Charm++ Fatal error ! Periodic cell has become too small for original patch grid!

From: Roman Petrenko (rpetrenko_at_gmail.com)
Date: Fri Nov 07 2008 - 13:57:33 CST

I think the "solvate" package which is used (i believe) by all
beginners in namd should be replaced by something more sophisticated,
because as it was mentioned before there are big cavities between a
protein and water after the solvation process.

The solution i use is to solvate a system in twice larger water box
than it should be. Then i run very short namd simulation without any
boundary conditions, which allows waters to quickly fill in those gaps
in protein. Only after that the box is trimmed to the size just to
cover the protein and PBC is used in a new simulation run.

Of course, it is not the smartest solution, but having a nice
solvation function would eliminate all this hustle and would allow one
to do PBC right away.

On Fri, Nov 7, 2008 at 1:45 PM, Peter Freddolino <petefred_at_ks.uiuc.edu> wrote:
>
> Dave Rogers wrote:
>>
>> But it would be nice if someone threw in a function call to
>> reinitialize the patch grid instead of breaking when this happens...
>>
>> ~ David Rogers
>>
>
> That might be helpful in some cases, but frequently this error occurs
> because someone has a system that is badly undersolvated, or is doing a
> constant pressure simulation when they should be doing constant volume. I
> believe the latter case is what is going on here; from Vlad's description it
> sounds like there's only a single molecule in the simulation box, in which
> case the box will collapse if a barostat is used. That isn't to say that
> automatically reinitializing the patch grid wouldn't be a valid option
> (although the reason that it's very difficult is you have to essentially go
> through many of the startup phases all over again, at which point you're
> essentially starting a new run anyway), but the presence of this error
> frequently means there's something wrong with the input system, so silencing
> it in all cases would not be a good idea.
> Peter
>
>

-- 
Roman Petrenko.
Physics Department
University of Cincinnati

This archive was generated by hypermail 2.1.6 : Wed Feb 29 2012 - 15:50:04 CST