[gmx-developers] Doubts to use Gromacs as Library

Sun Sep 7 03:21:32 CEST 2014

Hi, Rodrigo-

We've been talking and planning about better integrator frameworks and a MC framework for a while, though it's gotten hung up on lack of time for people involved.  The really hard part is getting the data structures and logic right, especially in conjunction with parallel processing -- all the algorithms themselves are pretty simple.  There are a few postings on redmine, and that might be the best place to post comments and continue the discussion.  I'm guessing it's not really going to be until October or November before that effort restarts significantly.

Best,
~~~~~~~~~~~~
Michael Shirts
Associate Professor
Department of Chemical Engineering
University of Virginia
michael.shirts at virginia.edu
(434)-243-1821

From: Rodrigo Faccioli <rodrigo.faccioli at gmail.com<mailto:rodrigo.faccioli at gmail.com>>
Reply-To: "gmx-developers at gromacs.org<mailto:gmx-developers at gromacs.org>" <gmx-developers at gromacs.org<mailto:gmx-developers at gromacs.org>>
Date: Saturday, September 6, 2014 at 10:41 AM
To: Discussion list for GROMACS development <gmx-developers at gromacs.org<mailto:gmx-developers at gromacs.org>>
Subject: Re: [gmx-developers] Doubts to use Gromacs as Library

Hi Ronald,

Thanks for your feedback.

I will be happy and I have interested in contribute to Gromacs project.

> I have read about Gromacs 5 as library But, I am not sure what is the best way to implement it in my project. In [5] is an image of my data >structure.  In solution_t, representation field  stores a protein conformation (protein_t) that is computed by Gromacs (here Gromacs is >called by simple system commands) the potencial energy of this conformation. This energy value is stored at obj_values field.
> We don't have a proper API for anything other than the writing analysis tools yet. We do have the other headers exported but it isn't a real >API. Do you want to keep your MC a separate project or is it an option to contribute it to Gromacs? An API for integrator extensions is still >a while in the future. We are of course always happy about contributions, but it might be quite a project to define a good API for that.

Are you interest in develop it or similar work? Can we work together in this project and start to develop an API for integrator extensions? I am interested.

> Based on Gromacs API, is it a good practice modify my pdb_atom_t and top_global_t data structures for their respective  Gromacs structures?
>Your copying the data into your structures? Without knowing the details that seems a bit inefficient (might not matter) but helps having a >better interface. Seems reasonable.

Here I thought to add structures from Gromacs topology in my top_global_t struture. So, my project (2PG) call Gromacs function to build whole topology and I referenced it by pointer to avoid coping data. However, based on your email I believe that we can implement  more efcient way. So, I guess an API or wrapper for it is a good way.

> My second interest is Modularization of the program pdb2gmx. I would like to know about developers that have been working with it since I would like to participate/attend to this project idea.
> I'm not aware of anyone having started anything.

I would like to start it. Let me know if there is some discussion/presentation/objective about this project among developers gromacs or do I start?I have read about it what was written in [1].

[1] http://www.gromacs.org/Project_ideas#Modularize_pdb2gmx_(and_friends)

Best regards,

--
Rodrigo Antonio Faccioli, Ph.D
Development Software for Structural Biology
Barao de Maua University
University of Sao Paulo
Lindedin - br.linkedin.com/pub/rodrigo-antonio-faccioli/7/589/a5/<http://br.linkedin.com/pub/rodrigo-antonio-faccioli/7/589/a5/>
Curriculum Lattes - http://lattes.cnpq.br/1025157978990218

On Fri, Sep 5, 2014 at 5:57 PM, Roland Schulz <roland at utk.edu<mailto:roland at utk.edu>> wrote:

On Fri, Sep 5, 2014 at 3:25 PM, Rodrigo Faccioli <rodrigo.faccioli at gmail.com<mailto:rodrigo.faccioli at gmail.com>> wrote:
>
> Hi Gromacs developers,
>
>
> Based on Gromacs project Ideas [1] and feature wishlist [2], I am interested in two projects:
>
>
> 1. Monte Carlo;
>
> 2. Modularization of the program pdb2gmx.
>
>
> In Monte Carlo, I would like to work with Gromacs 5 API integrated with my Monte Carlo implementation. Nowadays, my implementation uses Gromacs 4.6. I have already commented here about the first version of my project [3]. However, It was updated totally. Its new version can be found in [4].
>
>
> I have read about Gromacs 5 as library But, I am not sure what is the best way to implement it in my project. In [5] is an image of my data structure.  In solution_t, representation field  stores a protein conformation (protein_t) that is computed by Gromacs (here Gromacs is called by simple system commands) the potencial energy of this conformation. This energy value is stored at obj_values field.

We don't have a proper API for anything other than the writing analysis tools yet. We do have the other headers exported but it isn't a real API. Do you want to keep your MC a separate project or is it an option to contribute it to Gromacs? An API for integrator extensions is still a while in the future. We are of course always happy about contributions, but it might be quite a project to define a good API for that.

> More specific, I would like to get help in two parts:
>
> How can I remove Gromacs system calls using Gromacs API?

Not sure what you mean with that.

> Based on Gromacs API, is it a good practice modify my pdb_atom_t and top_global_t data structures for their respective  Gromacs structures?

Your copying the data into your structures? Without knowing the details that seems a bit inefficient (might not matter) but helps having a better interface. Seems reasonable.

>
> It is important to emphasize that all implemented algorithms in my project, they  use solution_t data structure to represent a solution that is a protein conformation or polymer (next release). For example, in [6] is the data structure of NSGA-II algorithm (it is multi-objective evolutionary algorithm). In this case ea_nsga2_t means an individual. Therefore, in obj_values field is stored more than one value of protein conformation such as Potential, Gyrate, number of Hydrogen Bond and others. All values are calculated by Gromacs.
>
>
> My second interest is Modularization of the program pdb2gmx. I would like to know about developers that have been working with it since I would like to participate/attend to this project idea.

I'm not aware of anyone having started anything.

Roland

>
>
> I've seen that Mone Carlo there is 1137 issue [7]. Nonetheless, I haven't found for Modularization of the program pdb2gmx.
>
>
>
> [1] http://www.gromacs.org/Project_ideas
>
> [2] http://www.gromacs.org/Developer_Zone/Roadmap/Feature_wishlist
> [3] https://mailman-1.sys.kth.se/pipermail/gromacs.org_gmx-developers/2013-September/007089.html
> [4] http://lcrserver.icmc.usp.br/projects/2pg_cartesian
> [5] https://www.dropbox.com/s/z3z86h6jma1okid/all_typedef.png?dl=0
> [6] https://www.dropbox.com/s/red11o3fr6h4lyo/nsga2_typedef.png?dl=0
> [7] http://redmine.gromacs.org/issues/1137
>
>
> Best regards,
>
> --
> Rodrigo Antonio Faccioli, Ph.D
> Development Software for Structural Biology
> Barao de Maua University
> University of Sao Paulo
> Lindedin - br.linkedin.com/pub/rodrigo-antonio-faccioli/7/589/a5/<http://br.linkedin.com/pub/rodrigo-antonio-faccioli/7/589/a5/>
> Curriculum Lattes - http://lattes.cnpq.br/1025157978990218

--
ORNL/UT Center for Molecular Biophysics cmb.ornl.gov<http://cmb.ornl.gov>
865-241-1537<tel:865-241-1537>, ORNL PO BOX 2008 MS6309

--
Gromacs Developers mailing list

* Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/GMX-developers_List before posting!

* Can't post? Read http://www.gromacs.org/Support/Mailing_Lists

* For (un)subscribe requests visit
https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-developers or send a mail to gmx-developers-request at gromacs.org<mailto:gmx-developers-request at gromacs.org>.

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://maillist.sys.kth.se/pipermail/gromacs.org_gmx-developers/attachments/20140907/b934ed1f/attachment-0001.html>