Gerald.
If you share your files we can take a look on my group's side.
We have seen this before due to a VERY weird issue. There are cases
where tleap decides to create carbon chains going through the middle of
a phenyl ring for instance. Incredibly enough, some of these are
perfectly stable at 2 fs, even though they are wrong. 4 fs forces the
issues and things blow up, but it has nothing to do with the time step,
it is tleap's fault. I suggest you rerun with 4 fs and ntwx=1, and look
at your trajectory to see what blows up.
Adrian
On 7/15/16 5:14 PM, Gerald Monard wrote:
> Hello,
>
> On 07/15/2016 09:56 PM, Hai Nguyen wrote:
>> quick question: have you used ParmEd to change the mass yet?
>>
> Yes. I forgot to say: I've modified the topology file with the
> HMassRepartition command in parmed and I have tested both the "normal"
> repartioning and the "dowater" option. Same results.
>
> Gerald.
>
>> Hai
>>
>> On Fri, Jul 15, 2016 at 3:54 PM, Gerald Monard <
>> Gerald.Monard.univ-lorraine.fr> wrote:
>>
>>> Hello,
>>>
>>> I have a protein system in water (321,300 atoms including ~100,000 water
>>> molecules). I'm trying to use Hydrogen Mass Repartitioning (HMR) to
>>> benefit a 4 fs timestep. However, it seems to fail. dt = 0.002 is OK,
>>> but with dt = 0.003, the temperature goes bad after a few thousands
>>> steps, and with dt = 0.004, it's even faster. This is for pmemd.cuda
>>> with Amber16 (all last patches applied) as well as pmemd.MPI.
>>>
>>> The mdin file is:
>>> &cntrl
>>> imin = 0, ntx = 5, irest = 1,
>>> dt = 0.004, nstlim = 1000000,
>>> ntb = 1,
>>> ntt = 1, ig = -1, temp0 = 300.,
>>> ntc = 2,
>>> ntpr = 1000, ntwx = 10000, ntwv = 00, ntwe = 10000,
>>> /
>>>
>>> From the mdin of the GPU benchmark, I've also tried to add tol =
>>> 0.000001 as well as ntf = 2. No improvement.
>>>
>>> Any idea?
>>>
>>> Gerald.
>>>
>>> --
>>>
>>> ____________________________________________________________________________
>>>
>>> Prof. Gerald MONARD
>>> SRSMC, Université de Lorraine, CNRS
>>> Boulevard des Aiguillettes B.P. 70239
>>> F-54506 Vandoeuvre-les-Nancy, FRANCE
>>>
>>> e-mail : Gerald.Monard.univ-lorraine.fr
>>> tel. : +33 (0)383.684.381
>>> fax : +33 (0)383.684.371
>>> web : http://www.monard.info
>>>
>>>
>>> ____________________________________________________________________________
>>>
>>>
>>> _______________________________________________
>>> AMBER mailing list
>>> AMBER.ambermd.org
>>> http://lists.ambermd.org/mailman/listinfo/amber
>>>
>> _______________________________________________
>> AMBER mailing list
>> AMBER.ambermd.org
>> http://lists.ambermd.org/mailman/listinfo/amber
>>
--
Dr. Adrian E. Roitberg
University of Florida Research Foundation Professor.
Department of Chemistry
University of Florida
roitberg.ufl.edu
352-392-6972
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Fri Jul 15 2016 - 14:30:03 PDT
