Re: RE: NAMD-l: energy jump between minimization and simulation

From: Eva Gonzalez Noya (evanoya_at_uchicago.edu)
Date: Wed Feb 16 2011 - 16:37:09 CST

Yes, I have set to read the coordinates as binary but I am doing
everything at a time with a single conf file that has a minimize and
then a run command. Even doing it like this is reading again data
between the minimization and simulation run from the restart files?
Are you suggesting that perhaps binary files are what are causing the
problem?
I send you an example of the log file where it can be seen that
temperature (or what is the same velocities) does not seem to be the
problem, however bond, electrostatic and other contribitions to the
potential energy suffer a jump in the step sitching from minimization to
simulation.
Thanks again
Eva

ETITLE: TS BOND ANGLE DIHED
IMPRP ELECT
     VDW BOUNDARY MISC KINETIC
TOTAL TEMP POTENTIAL
         TOTAL3 TEMPAVG PRESSURE
GPRESSURE VOLUME PRESSAVG GPR
ESSAVG
ENERGY: 20000 9841.1304 6733.7863 1235.5120
56.2450 -306542.8519 33195.3210 1709.6964
0.0000 0.0000 -253771.1608 0.0000
-253771.1608 -253771.1608 0.0000 -4846.8643
-4846.8716 639048.6136 -4846.8643 -4846.8716
ENERGY: 20000 1330.6651 1236.8253 1235.3343
56.1684 -278139.0332 32570.9172 1709.9092
0.0000 0.0000 -239999.2137 0.0000
-239999.2137 -240002.4128 0.0000 -2133.9392
-2070.3058 639048.6136 -2133.9392 -2070.3058

On 2/16/11 1:48 PM, Bennion, Brian wrote:
> Do you read in the coordinates as binary as well?
> Can you repost the log files again?
> Brian
>
> -----Original Message-----
> From: Eva Gonzalez Noya [mailto:evanoya_at_uchicago.edu]
> Sent: Wednesday, February 16, 2011 11:15 AM
> To: Bennion, Brian
> Cc: namd-l_at_ks.uiuc.edu
> Subject: Re: namd-l: energy jump between minimization and simulation
>
> Actually I do not know. I am a new NAMD user and not sure how it works.
>
> But the way I was doing these calculations was by setting up the
> minimization and run commands in the same conf file, so I do not know
> what the program does. ]Anyway I get the same result when in read the
> velocities from the binary restart file.
>
> Also I checked in the log file and temperature is low (about 1 K), which
> is in keeping with what I would expect. I am starting the MD from a
> configuration close to the minimum and set the temperature close to
> zero, so some movement appears because most likely I am not exactly at
> the minimum and forces are not exactly zero, so that particles move a
> bit and the temperature increases a little. What I do not understand is
> why the potential energy (I think alll contributions to the potential
> energy) change.
>
> Thanks for your help. It is very much appreciated.
>
> Eva
>
> On 2/16/11 10:58 AM, Bennion, Brian wrote:
>> Are you using binary restart files created from the minimization when you start your dynamics?
>>
>> brian
>>
>>
>> ________________________________________
>> From: owner-namd-l_at_ks.uiuc.edu [owner-namd-l_at_ks.uiuc.edu] On Behalf Of Eva Gonzalez Noya [evanoya_at_uchicago.edu]
>> Sent: Wednesday, February 16, 2011 8:14 AM
>> To: Ajasja Ljubetič
>> Cc: namd-l_at_ks.uiuc.edu
>> Subject: Re: namd-l: energy jump between minimization and simulation
>>
>> Thank you very much for your answer.
>> Actually I tried many things already. I was doing first an NpT (using lagevin pistons) setting the initial velocity close to zero, and also tried NVE but the problem persists
>> Regads,
>> Eva
>>
>> On 2/16/11 3:27 AM, Ajasja Ljubetič wrote:
>> Hi,
>> perhaps the energy jumps due to temperature and/or pressure control?
>> How are you heating the system?
>>
>> Regards,
>> Ajasja
>>
>>
>> On Wed, Feb 16, 2011 at 01:06, Eva Gonzalez Noya<evanoya_at_uchicago.edu<mailto:evanoya_at_uchicago.edu>> wrote:
>> Hi:
>>
>> I found that when running a MD simulation after a minimization the total energy suffers a jump when switching from the minimization to the simulation run (the conf file I am using is copied at the end of this message). In addition there two lines labeled with the same step number (the last step of minimization and the first step of the simulation) and the energy does not much for those neither (I am writing energies every step).
>> I am not rescaling velocities after the minimization. I am using a constraining harmonic potential on the alpha carbons of the protein both in the minimization and simulation runs (which I perform using a single conf file).
>> Anyone can help with this issue?
>> Thanks
>>
>>
>>
>>
>>
>

This archive was generated by hypermail 2.1.6 : Mon Dec 31 2012 - 23:19:50 CST