[Top][All Lists]

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

Re: [ESPResSo-users] Problem With ghost cell communication

From: Axel Arnold
Subject: Re: [ESPResSo-users] Problem With ghost cell communication
Date: Wed, 02 Jul 2014 21:38:44 +0200
User-agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10.9; rv:24.0) Gecko/20100101 Thunderbird/24.6.0


That sounds a bit confuse. Note, that ghost particles only carry forces if they have interactions with locally stored particles, so most ghost particles will have zero forces. The ghost forces are initialized to 0 and just serve to collect the contributions that would need to cross a processor border. The ghost forces are then sent back to the processors physically carrying the particle.

So, you can only reliably access the forces of locally stored particles (that are in cells from local_cells) after the force computation is finished. You should not look at the forces of the ghost particles.


On 02.07.14 16:41, Mohammad Abdulkhaliq wrote:
I have created a new bonded interaction and it works fine with single processor runs however I found that when running multiple processors the ghost cells in calc_particle_lattice_ia in lb.cpp are not set with the forces given to the particles in forces.hpp I printed out the forces on the local and ghost cells and  found only a small number of them set with a force the rest were zeroes.

By the way this was not a problem on the release version 3.1.1 this tests are done on the current development version.

Can someone help me with the proper way to give the forces to all the ghost and local particles ?

Best Regards,
             Mohammad Abdulkhaliq

JP Dr. Axel Arnold
ICP, Universität Stuttgart
Allmandring 3
70569 Stuttgart, Germany
Email: address@hidden
Tel: +49 711 685 67609

reply via email to

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