espressomd-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [ESPResSo-devel] About "blockfile read auto" with multiple nb intera


From: Axel Arnold
Subject: Re: [ESPResSo-devel] About "blockfile read auto" with multiple nb interaction
Date: Wed, 2 Feb 2011 11:27:16 +0100
User-agent: KMail/1.13.5 (Linux/2.6.34.7-0.7-desktop; KDE/4.4.4; x86_64; ; )

On Wednesday 02 February 2011 05:05:24 Mingyang Hu wrote:
> Dear all,
> 
> I came across this situation recently:
> 
> In a simulation system, there is more than one non-bonded interactions
> between particle type 1 and 2, e.g. BOTH Lennard Jones and LJ-Cos2. So when
> a checkpoint file is written, the interaction block has a line of the
> following form:
> 
> { 1 1 lennard-jones *...some_lj_parameters...* lj-cos2 *
> ...some_ljcos2_parameters...* }
> 
> However, after the checkpoint is read in by some later simulations, only
> the first nb_interaction, i.e. the lennard-jones, is set. The lj-cos2 is
> ignored.
> 
> After finding this, I went into the scripts and found such a paragraph in
> file ESPRESSO/scripts/blockfile_support.tcl
> 
> proc blockfile_read_auto_interactions {channel read auto} {
>     set data [blockfile $channel read toend]
>     foreach d $data { eval "inter $d" }
>     return "interactions"
> }
> 
> Does this mean when the above interactions are read in, ESPResSo actually
> runs the following line?
> 
> inter 1 1 lennard-jones ...some_lj_parameters... lj-cos2
> ...some_ljcos2_parameters...
> 
> If so, it is apparent why only the first part of lennard-jones interaction
> is set and why the second part of lj-cos2 is ignored. More importantly, it
> this explanation happens to be the case, then I guess some revision is
> necessary on how interactions are stored into and extracted from checkpoint
> files.

Hi!

Actually, this is a rather recent bug in the LJ interaction, which simply eats 
up all trailing parameters. If it wouldn't do that,

inter 1 1 lennard-jones ...some_lj_parameters... lj-cos2
 ...some_ljcos2_parameters...

would work perfectly fine. You can try that by combining. e.g. lj-cos2 and the 
hertzian potential.

Olaf will fix LJ in a couple of days.

Axel

-- 
JP Dr. Axel Arnold
ICP, Universität Stuttgart
Pfaffenwaldring 27
70569 Stuttgart, Germany
Email: address@hidden
Tel: +49 711 685 67609



reply via email to

[Prev in Thread] Current Thread [Next in Thread]