|
From: | Olaf Lenz |
Subject: | Re: [ESPResSo-users] error message about min_local_box_l |
Date: | Fri, 29 Nov 2013 10:52:56 +0100 |
Hi!First of all, please keep the discussion on-list, so that the mailing gets archived and others can profit from it as well.2013/11/29 guozhicheng222 <address@hidden>Dear Olaf:set warm_steps 10000
Thanks for your reply. Based on your suggestion, I increaed the warm-up integrate simulation time( set warm_steps 10000, 1000 as before) to ensure eliminate the large force due to overlap generated from place the particles randomly. In addition, the time step is chosed 0.001. It is still running. When I am running on the single core, it's ok and no error message generated. It seems like that error occured in the parallel simulation only. Is it right?
So, you modified all at once and ask me now what has happened? This is hard to answer.For the warmup, I would not set a fixed number of steps, but instead you should check whether the large forces have gone. For example, you could simply measure the force on each particle and check whether the maximal force is below a certain threshold If that is the case, end the warmup.You say that it runs now, does that mean that the error has disappeared? Maybe you should visualize the system (using VMD) to see what is happening.Olaf--Dr. rer. nat. Olaf LenzInstitut für Computerphysik, Allmandring 3, D-70569 StuttgartPhone: +49-711-685-63607
[Prev in Thread] | Current Thread | [Next in Thread] |