[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [ESPResSo-users] help
From: |
Axel Arnold |
Subject: |
Re: [ESPResSo-users] help |
Date: |
Mon, 18 Jun 2012 19:22:46 +0200 |
User-agent: |
KMail/1.13.6 (Linux/2.6.38-15-generic; KDE/4.6.5; x86_64; ; ) |
On Monday 18 June 2012, mehdi ezatabadi wrote:
> Dear Dr.Arnold
>
>
> In the past step of my project I used the espresso command to make the
> neighbor list of particles to see if they made special structures. The
> command was:
>
> set nbhood [analyze nbhood idParticle cutOffRadious]
This command won't help you. It uses the analysis data structures which are
not up-to-date during the integration, and are not parallel.
> In condensed material simulations I figured it out that lots of
> complementary bases would interact which a single base that is different
> from the physics of base pair(A_T C_G) interactions. Now I dont know
> how to include it in Integration by Espresso. Usually in the tcl program we
> define the interactions among the types of particles. But now I should look
> at the particles at every step and then decide which particles interact
> with each other.
Yes, and that requires a bit of coding, as I said. You might want to take a
_good_ look e.g. at the technical slides from the Espresso workshop last year
(http://www.cecam.org/workshop-5-532.html).
The steps are roughly, plus a lot of debugging
- make a new interaction, e.g. copy the Lennard-Jones code, and adjust all the
places the LJ code appears (#ifdef LENNARD_JONES).
- make the global maximal cutoff twice the one of your A-T interaction, in
order to get all T-particles surrounding one A-particle onto one processor.
(recalc_maximal_cutoff_nonbonded in interaction_data.c)
- to loop over the neighbors of a particle, calculate its cell position in the
domain decomposition (compare dd_position_to_cell in domain_decomposition.c),
then loop over all surrounding cells and their particles (compare
calc_link_cell in domain_decomposition.c, but of course restrict to the
relevant cells).
- alternatively, you can directly loop over all A-particles and their T-
neighbors, might be faster. That you wold add as a separate calc_link_cell-
like loop to calc_forces in forces.c.
If you look for an easier route, good luck, I am not aware of one. And I wrote
most of the short ranged force calculation.
Axel
--
JP Dr. Axel Arnold Tel: +49 711 685 67609
ICP, Universität Stuttgart Email: address@hidden
Pfaffenwaldring 27
70569 Stuttgart, Germany