From: Jérôme Hénin (jhenin_at_ifr88.cnrs-mrs.fr)
Date: Wed May 23 2012 - 13:21:26 CDT
Hi Niklaus and Jeremy,
This was indeed a bug due to improper handling of the singularity in
the distance gradient when two atoms are exactly superimposed.
The source file attached corrects this. We'll try to get the fix into
NAMD CVS as soon as possible.
Thanks for the report,
Jerome
On 19 May 2012 22:11, Jérôme Hénin <jhenin_at_ifr88.cnrs-mrs.fr> wrote:
> Hi Niklaus,
>
> This is very intriguing. Can you please send me off-list the minimum
> files needed to reproduce the issue?
>
> Thanks,
> Jerome
>
>
> On 18 May 2012 16:13, Niklaus Johner <nij2003_at_med.cornell.edu> wrote:
>> Hi all,
>>
>> I'm having the same issue with an FEP calculation of the protonation of a
>> sidechain.
>> My FEP calculations run fine as long as I don't use colvar to add a
>> restraint. If I turn colvar on and add a restraint between an atom from my
>> vanishing sidechain and any atom in my system but the appearing sidechain,
>> it also works fine. But if I add a restraint between atoms from the
>> appearing and vanishing sidechain, then it crashes at the first time step
>> (even if I use forceConstant=0).
>>
>> It seems that adding a restraint between the vanishing and appearing atoms
>> somehow makes them see each other and clash. Maybe it breaks the automatic
>> generation of the non-bonded exclusion list?
>>
>> I tried this on namd2.7, namd2.8 and namd2.9.
>> I haven't tried to use alchemify to create the exclusion list myself yet--20cf300513142c297804c0b835ce--
This archive was generated by hypermail 2.1.6 : Mon Dec 31 2012 - 23:21:32 CST