Re: [ESPResSo-users] continue an interrupted simulation which used peptideB

[Stefan Kesselheim]

Hi,

On Feb 28, 2013, at 2:53 PM, pdam bright <address@hidden> wrote:

> Sorry, from the last pdb file!
> 
> Thanks Stephan,
> 
> My main question is how I can feed the last configuration before interruption 
> (the last pdb file), to the codes of peptideB? As you know there is two main 
> tcl files peptidebuilder.tcl and mypeptide.tcl. The second file needs the 
> sequence of my peptide not the position of atoms in the chain!

I have never used peptideB, but I believe the following would work: 
You create a new peptide with exactly the same command(s) as in the original 
simulation. Then you just move all the particles to the positions stored in the 
pdb, e.g. 
part 0 pos $oldpos_x $oldpos_y $oldpos_z

But: You also need the velocities in order to continue the run. Usually they 
are not in the pdb, right? Then I'm afraid it won't work.

Cheers
Stefan
> 
> Thanks again,
> Pee  
> 
> 
> On Thu, Feb 28, 2013 at 5:12 PM, Stefan Kesselheim <address@hidden> wrote:
> Hi Pee,
> 
> On Feb 28, 2013, at 2:35 PM, pdam bright <address@hidden> wrote:
> 
> > Dear all,
> >
> > I am using peptideB to simulate a peptide. But my computer has been shut 
> > down and I have to run it from the beginning. Is there any way to continue 
> > the simulation from where it has been interrupted using the last VMD file?
> 
> In principle this might be possible, but surely is not intended that way. A 
> good method for continue simulations is using the blockfile format to 
> regularly store the necessary particle data (and other parameters) and write 
> the simulation script in two variants: One for the initial setup and one that 
> allows to continue the simulation from the data stored in a block file.
> We do not support "checkpointing" in the sense that you can have an 
> automatically created file that contains everything necessary to continue 
> running. One reason is that this would include the necessity to store the 
> full state of the TCL interpreter.
> 
> If it is just one afternoon of computation, it is probably easier to re-run 
> the simulation :-). The particle positions can however be extracted from a 
> vtf file, if necessary, but that is manual work.
> 
> Cheers and good luck,
> Stefan
>