Hey guys,
Thanks for the discussion. I think I understand better what is going on.
I'm going to use the binary format as suggested. Peace.
Sid
On Tue, Dec 15, 2009 at 7:45 AM, Daniel Roe <droe.nist.gov> wrote:
> If you want higher precision trajectories you can use the netcdf trajectory
> format (ioutfm=1). The higher precision will likely not change your results,
> but netcdf files are more compact and faster to process than formatted
> trajectories so it's a good idea to use them anyway.
>
> -Dan
>
>
> Jason Swails wrote:
>
>> The problem is not with the mdcrd precision, it is the fact that the
>> time step of the restart file is one greater than the time step of the
>> final frame in the mdcrd file. The trajectory file should have 3
>> decimal places in each coordinate (unless maybe your coordinates have
>> absolute values > 1000? This would most likely not happen if you use
>> iwrap=1 for your MD simulations) which should provide plenty of
>> resolution for its intended usage (see Professor Case's description).
>> Properties derived from the mdcrd will not be that sensitive to
>> changes in only the 4th decimal place for each coordinate, so the
>> existing precision is fine for that use. Restart files have a greater
>> precision since they're used to restart simulations.
>>
>> Thus, I'd suggest that you leave the mdcrd format intact, though if
>> you were really intent on changing this it would have to be done in
>> the source code (you'd have to find the appropriate format statement
>> that the mdcrd file uses for its output and change it to include more
>> decimal places).
>>
>> However, the existing tools used to analyze trajectories (ptraj, or
>> sander if you're using imin=5) may expect to see the default format,
>> so you would have to either change these programs to read your new
>> format, or write your own analysis scripts/programs. In any case, it
>> seems a daunting task for so little gain (if any gain is to be made,
>> which I'm doubtful of).
>>
>> Good luck!
>> Jason
>>
>> On Tue, Dec 15, 2009 at 3:21 AM, Sidney Elmer <paulymer.gmail.com> wrote:
>>
>>
>>> I realized that the problem was with the precision of the mdcrd
>>> coordinates,
>>> which is 2 decimal places. Is there any way to increase the precision of
>>> formatted trajectory files? Two decimal places is too low resolution.
>>> Thanks.
>>>
>>> Sid
>>>
>>>
>>> On Mon, Dec 14, 2009 at 7:28 PM, case <case.biomaps.rutgers.edu> wrote:
>>>
>>>
>>>
>>>> On Mon, Dec 14, 2009, Sidney Elmer wrote:
>>>>
>>>>
>>>>> 3. thus, after execution finishes, the coordinates in "restrt" file
>>>>>
>>>>>
>>>> should
>>>>
>>>>
>>>>> match coordinates from frame 500, correct?
>>>>>
>>>>>
>>>> No. The coordinates stored in the restart file are one time step ahead
>>>> of the last snapshot in the trajectory file. Restart files should be
>>>> used
>>>> for exactly that: restarting an existing trajectory. Trajectory files
>>>> should
>>>> be used for sampling the coordinates traversed by the simulation.
>>>>
>>>> ...dac
>>>>
>>>>
>>>> _______________________________________________
>>>> 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
>>>
>>>
>>>
>>
>>
>>
>>
>>
>
> --
> Daniel R. Roe, Ph.D.
> Research Chemist
> National Institute of Standards and Technology
> 100 Bureau Drive, Stop 8443
> Gaithersburg, MD 20899-8443
> (301) 975-8741
>
>
>
> _______________________________________________
> 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
Received on Tue Dec 15 2009 - 12:30:02 PST
