Pressure-Correction and Lammps: General Question

[frank Zack]

Hi all,

Finally I've managed to use votca in combination with lammps and the results are quite impressive. The only problem I could not solve is the pressure correction. My naive approach is to equilibrate my parametrized coarse grained polymer and then apply my target pressure for some steps. The final configuration is then used again as input in ibi. This way I was able to get quite close to the All-Atom-forcefield density. However, I guess the pressure-correction-algorithm does a better job.

Thats why i'm thinking about implementing the simple pressure correction algorithm in conjunction with lammps as simulator.
From the documentation of votca it is not 100% clear how it works though.
What's the actual Input, the pressure correction needs? As far as i get it, its just the average pressure of the latest md-run, right?
SO all I need is calculate the average pressure during my md-runs and the rdf,right?
Which script-files do i have to edit (is it only postupd_pressure.sh) ?

Thanks for any advice,

regards,
frank.

[Christoph Junghans]

2015-10-18 5:05 GMT-06:00 frank Zack <daybre...@gmail.com>:
> Hi all,
>
> Finally I've managed to use votca in combination with lammps and the results
> are quite impressive.

Any chances to the code you want to contribute?

> The only problem I could not solve is the pressure
> correction. My naive approach is to equilibrate my parametrized coarse
> grained polymer and then apply my target pressure for some steps. The final
> configuration is then used again as input in ibi. This way I was able to get
> quite close to the All-Atom-forcefield density. However, I guess the
> pressure-correction-algorithm does a better job.

The standard pressure correction never changes the density, but just
adds an additional potential correction to change the pressure.

>
> Thats why i'm thinking about implementing the simple pressure correction
> algorithm in conjunction with lammps as simulator.
> From the documentation of votca it is not 100% clear how it works though.
> What's the actual Input, the pressure correction needs? As far as i get it,
> its just the average pressure of the latest md-run, right?
> SO all I need is calculate the average pressure during my md-runs and the
> rdf,right?
> Which script-files do i have to edit (is it only postupd_pressure.sh) ?

You will need to write a script like calc_pressure_gromacs.sh for LAMMPS.
This calc_pressure_lammps.sh takes one argument - a filename, which
contains one line
"Pressure=XXX"
after it got executed.

See Section 9.1 of the manual to understand how to add a script in user space.

Christoph

> Thanks for any advice,
>
> regards,
> frank.
>

> --
> 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 votca+un...@googlegroups.com.
> To post to this group, send email to vo...@googlegroups.com.
> Visit this group at http://groups.google.com/group/votca.
> For more options, visit https://groups.google.com/d/optout.



--
Christoph Junghans
Web: http://www.compphys.de

[frank Zack]

Hi Christoph,

thx again for your support this great tool. Im glad I didnt give up.
Meanwhile I've tested my new cg-potential in a tensile test of 40 chains of PP each having 2400 repeats units. The result is surprisignly close to the experiment (simulation is still running) and i can use a timestep of 30 fs.

I surley can share my little code-modifications. But its very sloppy and I just built a  little "workaround". However, I've appended the modified files and if you replace the original scripts, votca will only work with lammps and no pressure correction is implemented yet. So do a backup of those files.
csg_stat is replaced by a simple bash script. The rdfs and distributions are calculated within lammps as defined in the lammps-input-script. The lammps-script calls another bash-script which converts the tables from gromacs-units to units-real (see lammps doc:distance=Angstrom, Energy=kcal/mole). This file has to be added into the <filelist> -section in your settings.xml. The options for the histrograms need to be adjusted in the lammps-input script. The values defined in settings.xml are ignored.

Keep in mind: Im not a professional, this is only a little hack. Its not a full implementation and im not a about to learn perl to do so. Its working for me and maybe it will work for somebody else too..

regards,
frank.

[Christoph Junghans]

Thanks for sharing you scripts!

I incorporated parts of it in VOTCA:
<https://github.com/votca/csg/commit/fb9f982e27cc6ddc3285d46dbbbb262a0481ff4f>
However, replacing csg_stat is no-go, mainly because csg_stat can
calculate IMC correlation matrices, too and it is parallel.
My guess is that once you have written a mapping file like the one
generated by the script I attached csg_stat will work.
In case of lammps, which has multiple unit system, we only support the
native one, meaning we assume everything is in nm and that will work
fine, because all the scaling is linear. E.g. VOTCA reads a LAMMPS
trajectory in any units assuming it is nm and then calculates the
potentials in nm, which are basically the original units. It only
becomes problematic if you pdb or xyz files because these are defined
to be in Angstrom and hence VOTCA scales them by 10.

Christoph

[Lee]

Hi frank, 

        I'm doing the IBI process for dgeba. After added the pressure correction, the distribution and density are all close to the full-atomic system. But when I use these final potential tables to do the tensile simulation for CG system, the system is unstable. And the stress-strain curve is extrmely terrible. I saw you get a extremely good stress-strain curve from your attched file. I want to konw how could I get a good stress-strain curve with these potential tables? Or there still need some post process for these tables?

I would really appreciate your advise. Thanks in advance!

Lee

在 2015年10月18日星期日 UTC+8下午11:15:50，frank Zack写道：