Re: Managing Log files

From: bo liu (liubo.njuer_at_gmail.com)
Date: Sat Sep 27 2008 - 03:22:21 CDT

Hi Navdeep,

In my view, minimization of 10000 steps for a time is unnecessary in NAMD.
I'd like to recommand you check the .log file first. maybe after hundreds of
steps for minimization, the TOTAL energy for the system changes little. In
this case, continuing minimization of tens of thousand steps is not economic
for burning CPUs.

you may want to alternate "short minimization" and "short equilibration
(i.e. real MD run)" for several times to drive the system to the thermally
accessible state that is ready for a production run.

OK, let me explain a little bit why this strategy:
First, you get an initial structure of the simulation system, it has bad
contacts and needs minimization and equilibration.
Then, you do minimization to drive the system into a *local* minimum. (from
the peak of the mountain to a nearby lowland.)
Note that owning to the algorithm of minimization, it may be hard for
programs to escape the local minimum and find the *global* minimum. So,
continuing minimization for a *very long* timesteps merely means the system
lingers around in that local lowland. In this case, you may need to do
equilibration to propel the system out of that local miminum. Several times
of alternating minimization and equilibration would be able to drive your
system from mountains to valleys.

please let me know if you have any questions.

Cheers!

LIU BO

2008/9/27 Nd S <navdeep79_at_gmail.com>

> Hi Liubo
>
> I am usually going for approximately 10000 steps of minimization. For large
> systems I am facing a problem as it takes hell lot of time. I have to do
> this as usually my starting geometry is in pretty bad shape. Can you please
> explain to me the strategy you mentioned.
>
> Thanks
> Navdeep
>
> 2008/9/26 bo liu <liubo.njuer_at_gmail.com>
>
> Hi Navdeep,
>>
>> How many timesteps do you set for the minimization?
>> minimization does output information every step, but this shouldn't bother
>> you because in most cases, a thousand steps of minimization is enough. I
>> think this wouldn't contribute to a very long .log file... In other cases,
>> more sophisticated protocol of minimization should be used, such as
>> alternating MD runs and Minimizations, rather than *too many* steps for
>> minimization at a time.
>>
>> Regards,
>>
>> 2008/9/25 Nd S <navdeep79_at_gmail.com>
>>
>> Hi Alexander
>>>
>>> Is there any way out to reduce output information during minimization.
>>>
>>> Thanks
>>> Navdeep
>>>
>>>
>>> On Wed, Sep 24, 2008 at 1:01 PM, Alexander A. Vakhrushev <
>>> makaveli.lcf_at_gmail.com> wrote:
>>>
>>>> Hi Navdeep!
>>>> Perhaps your running minimization. In that case NAMD will still output
>>>> information in log file each step.
>>>>
>>>> 2008/9/24, Joshua Adelman <jadelman_at_berkeley.edu>:
>>>> > Hi Navdeep,
>>>> >
>>>> > That should work, and I've never seen it fail to. Is the problem that
>>>> the
>>>> > output files are still too long or is it that NAMD is really ignoring
>>>> the
>>>> > outputEnergies parameter?
>>>> >
>>>> > Josh
>>>> >
>>>> >
>>>> >
>>>> >
>>>> > On Sep 24, 2008, at 9:52 AM, Nd S wrote:
>>>> >
>>>> > Hi all
>>>> >
>>>> > While running a MD simulation in NAMD, the log file created is very
>>>> large as
>>>> > the energy values are stored at every step. I added the command
>>>> > "outputEnergies 5000" to output energy into stdout after every 5000
>>>> steps as
>>>> > per the NAMD manual, but it doesnt work. What's the other way out.
>>>> >
>>>> > Thanks
>>>> > Navdeep
>>>> >
>>>> >
>>>> >
>>>> ------------------------------------------------------------------------------------------------------
>>>> > Joshua L. Adelman
>>>> > Biophysics Graduate Group Lab: 510.643.2159
>>>> > 218 Wellman Hall Fax: 510.642.7428
>>>> > University of California, Berkeley
>>>> > http://nature.berkeley.edu/~jadelman
>>>> > Berkeley, CA 94720 USA jadelman_at_berkeley.edu
>>>> >
>>>> ------------------------------------------------------------------------------------------------------
>>>> >
>>>> >
>>>>
>>>>
>>>> --
>>>> Best regards,
>>>>
>>>> Dr. Alexander Vakhrushev
>>>> Institute of Applied Mechanics
>>>> Dep. of Mech. and Phys.-Chem.
>>>> of heterogeneous media
>>>> UB of Russian Academy of Sciences
>>>> 34 T. Baramzinoy St.
>>>> Izhevsk, Russia 426067
>>>> ----------------------------------
>>>> ӧѧا֧ߧڧ֧,
>>>> ѧ֧ ݧ֧ܧѧߧէ ݧ֧ܧѧߧէӧڧ
>>>>
>>>> ..-.., ...
>>>> ߧڧ ڧܧݧѧէߧ ާ֧ѧߧڧܧ
>>>> ѧݧܧԧ է֧ݧ֧ߧڧ
>>>> ڧۧܧ ܧѧէ֧ާڧ ѧ
>>>> 426067, . ا֧ӧ
>>>> . . ѧѧާ٧ڧߧ 34
>>>> +7(3412)21-45-83
>>>>
>>>
>>>
>>>
>>> --
>>> --------------------------------------------
>>> Gig'em
>>> Navdeep Singh
>>>
>>
>>
>>
>> --
>> -Liu bo
>> -----------------------------------------------------------
>> Computational Biochemistry&Biophysics, Nano-bio systems: MD method;
>> College of Chemistry and Chemical Engineering, Graduate University of
>> Chinese Academy of Sciences, Beijing P.R. China
>> Office: (86)-010-88233187
>> Home: (86)-010-88259765
>> Cell: 13426057875
>>
>
>
>
> --
> --------------------------------------------
> Gig'em
> Navdeep Singh
>

-- 
-Liu bo
-----------------------------------------------------------
Computational Biochemistry&Biophysics, Nano-bio systems: MD method;
College of Chemistry and Chemical Engineering, Graduate University of
Chinese Academy of Sciences, Beijing P.R. China
Office: (86)-010-88233187
Home: (86)-010-88259765
Cell: 13426057875

This archive was generated by hypermail 2.1.6 : Wed Feb 29 2012 - 15:49:53 CST