[Top][All Lists]

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

Re: [ESPResSo-users] Fwd: latest version of Espresso

From: Konrad Breitsprecher
Subject: Re: [ESPResSo-users] Fwd: latest version of Espresso
Date: Mon, 24 Mar 2014 16:02:10 +0100

Hello Xikai,
I tested your script and altered it at some points so that it works fine for me. See the comments in the tcl file for more explanations.
I think the main problem was that you called iccp3m without setting up p3m and that you used individual temperatures for the ions.
Note that you will get NaNs for the coulomb (and also total) energy. I think this happens because the wall charges are very close nearby which causes the shortrange part of the coulomb energy to explode. There might be a fix in the future, for now you will have to calculate the coulomb energy yourself (if you need it anyways).
About the compilation, one error was related to a cuda library, maybe try configure --without-cuda.
No clue about the other error (conftest.cpp(11): error #2379: cannot open source file "ac_nonexistent.h"). Maybe someone can help here...

2014-03-24 2:40 GMT+01:00 Xikai Jiang <address@hidden>:
Dear Konrad:

Thank you very much for providing your code. I have downloaded the code from your branch, but I've no luck compiling them, there are errors with MPI (the config.log file is attached) :

undefined reference to 'ompi_mpi_double'
undefined reference to 'MPI_Bcast'

For the code from ESPResSoMD Github, I can compile without problems. I'm not sure where the problem is, but I found a way around. I copied the newest files (see below)

iccp3m_tcl.cpp/hpp, forces.cpp/hpp, iccp3m.cpp/hpp, thermostat.cpp/hpp, subt_elec.cpp/hpp, drude.cpp/hpp, interaction_data.cpp/hpp, energy.hpp, pressure.hpp, mmm1d.cpp/hpp, constraint.cpp/hpp

from your branch to the lastest code from ESPResSoMD Github, then I can compile.

After this, I started a test drive. My system contains two parallel walls, which held at zero electric potential, and there is Lennard-Jones fluid confined between them. First I use forcecapping to warm up my system, but the message pops up:

"ICCP3M found zero electric field on a charge. This must never happen

I also tried to put iccp3m command after the simulation ran for a while with zero-charged wall, but it still gave me the above message. My code is attached.

This doesn't appear when I used 3.2.0 version. The walls were built using fixed atoms, will this lead to the problem? Do you have any clues or suggestions for this type of error?

Thank you for you time.



Subject: Re: [ESPResSo-users] Fwd: latest version of Espresso
From: address@hidden
Date: Fri, 21 Mar 2014 22:08:30 -0400
To: address@hidden

Hello Konrad:

Thank you very much for your information. I'll give the code a test run.



在 Mar 21, 2014,16:46,"Konrad Breitsprecher" <address@hidden> 写道:

Hello Jimmy,
you can try to use my branch where the langevin thermostat+iccp3m is working. Be aware that I'm still running some final tests to check the validity of icc, so if you encounter sth strange related to the method, please contact me.
Konrad Breitsprecher

2014-03-17 23:57 GMT+01:00 Stefan Kesselheim <address@hidden>:

Begin forwarded message:

> From: Stefan Kesselheim <address@hidden>
> Subject: Re: latest version of Espresso
> Date: March 17, 2014 11:51:36 PM GMT+01:00
> To: Xikai Jiang <address@hidden>
> Dear Jimmy,
> On Mar 15, 2014, at 1:37 AM, Xikai Jiang <address@hidden> wrote:
>> Hello Stefan:
>> Last time you mentioned that the thermostat is not enabled using ICCP3M in Espresso 3.2.0. Could you help me to find where I can download bug-fixed version of Espresso that enables thermostat during ICCP3M?
> That is true. Konrad Breitsprecher has a working patch, and we'll try to upload it in the next couple of days. We however first have to check it actually does what it should :-).
> Cheers
> Stefan

Attachment: test_constant_potential_parallel_plate_nvt.tcl
Description: Tcl script

reply via email to

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