From: Jim Phillips (jim_at_ks.uiuc.edu)
Date: Mon Jul 18 2011 - 17:00:02 CDT
On Wed, 13 Jul 2011, Austin B. Yongye wrote:
> Dear NAMD users, I ran into the "Low global exclusion count errors"
> after about 20 ns of production dynamics.
> 1). I increased the pairlist distance and set "COMmotion no" after the
> first instance, but still encountered the same error around ~20-23 ns.
> Temperature, potential and kinetic energies are all stable throughout.
Neither of those should be necessary. All excluded pairs of atoms should
be well within the cutoff distance and net system motion shouldn't matter.
> 2). After reading the section on "Bad global exclusion count errors" on
> this link:
> I included the "splitPatch position" option set "rigidbonds none" and
> "COMmotion no." Now the temperature is no longer maintained at 310K.
If the issue is happening after many ns this isn't the issue either.
Is that 20ns of continuous running or are there restarts? If you restart
from the last checkpoint before the error does it run for another 20ns?
The NAMD version, platform, and number of cores might be useful to know,
particularly if errors only happen at certain core counts.
This archive was generated by hypermail 2.1.6 : Wed Feb 29 2012 - 15:57:28 CST