Hi Marc

Did you check the CG-CG RDF at 300the step? How do they compare with the
reference RDF?

In my experience, I found that, in the case of bulk water, with the current
RE minimization method in VOTCA,
the CG potentials more or less converge to the values which reproduce
target RDF in less than 200 iterations.
However, due to statistical nature of short CG-MD simulations at every
update step, the values of CG potentials
fluctuate around the converged values. These fluctuations may or may not
satisfy the convergence check criteria, i.e., they may be smaller or larger
than the specified tolerance, and it depends on various setting parameters
such as length of CG-MD simulations, relaxation parameters, and number of
knot values.

Therefore, it is always good to check the CG particle RDFs periodically to
judge whether the RE minimization is
converged or not. I guess, in VOTCA, there is a way to use CG particle RDFs
as a measure to decide convergence. I will check how to do that and get
back to you.

We are also working on implementing alternative approach for RE
minimization, which is a deterministic approach. It is based on reweighting
strategy. You can read more about it in the article

Chaimovich, Aviel, and M. Scott Shell. “Coarse-Graining Errors and
Numerical Optimization Using a Relative Entropy Framework.” *The Journal of
Chemical Physics* 134, no. 9 (March 3, 2011): 094112. doi:10.1063/1.3557038

I hope it helps. Let us know if you have any further questions or
suggestions.

Thanks,
Sikandar






On Fri, May 2, 2014 at 4:31 PM, Marc Segovia <[email protected]> wrote:

> Hi
>
> Last days I run relative entropy tutorial, using votca 1.3, in spce water
> model
> After several days in my PC, finally the cg don't converge.
>
> step 300 done, needed 1083 secs
> Doing convergence check: default
> Iterations are not converged, going on
>
> My question is , this happens because a convergence of 0.001 is to small,
> or because numerical errors in my software or because the number of
> configuration to average is too low , 50 ? or something else.
>
> Thanking in advance by your help
>
> Best Regards
>
> Marc
>
> --
> You received this message because you are subscribed to the Google Groups
> "votca" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to [email protected].
> To post to this group, send email to [email protected].
> Visit this group at http://groups.google.com/group/votca.
> For more options, visit https://groups.google.com/d/optout.
>

-- 
You received this message because you are subscribed to the Google Groups 
"votca" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
To post to this group, send email to [email protected].
Visit this group at http://groups.google.com/group/votca.
For more options, visit https://groups.google.com/d/optout.

Reply via email to