[gmx-users] Re: cannot read frames out from trr files

[Justin Lemkul]

On 4/3/13 11:03 PM, mu xiaojia wrote:
> Thanks Justin,
>
> It never happens to my simulations before, is there anyway to figure out
> what caused this corruption?  the log file and screen outputs didn't
> mentioned any error message. I am wandering if it was caused by the
> unstable super-cluster I am currently using, because on another
> super-clusters, it never happened before.
>

I wouldn't necessarily call it an unstable cluster, but file system blips can 
happen.  I've seen it a few times on different clusters.  Judicious use of 
gmxcheck and running jobs in intervals (while saving checkpoints) is the only 
foolproof way to ensure the integrity of your output files.

-Justin

>
> On Wed, Apr 3, 2013 at 7:45 PM, Justin Lemkul <jalemkul at vt.edu> wrote:
>
>>
>>
>> On 4/3/13 8:42 PM, mu xiaojia wrote:
>>
>>> Sorry for the typo, the last evidence"4) I also compared the generated
>>> files' sizes, A = B in trr, edr , gro, cpt, if B crashes before reaching
>>> 100ns, the size would **NOT*** be the same as A."
>>>
>>>
>> You probably have a corrupted frame.  The simulation continued and the
>> .trr file continued being written, so you see comparable file sizes.
>>
>> -Justin
>>
>>
>>> On Wed, Apr 3, 2013 at 7:25 PM, mu xiaojia <muxiaojia2010 at gmail.com>
>>> wrote:
>>>
>>>   Dear GMX users,
>>>>
>>>> I am not sure if someone has similar problems before. I cannot read half
>>>> of the frames from trr file after md simulation, and I believe my
>>>> simulation has already completed.
>>>>
>>>> I have finished simulations A and B with both gmx4.5.4 and gmx4.5.5. the
>>>> only difference between A and B is they are npt simulations under
>>>> different
>>>> temperature (in mdp file, "ref_t" part )and such difference would not
>>>> cause
>>>> crash of the simulation. I have successfully finished A, everything looks
>>>> perfect (frames can be successfully extracted), but for B, I could not
>>>> extract last 70ns' frames from trr.
>>>>
>>>> 1) My command is like: mdrun_mpi -s -deffnm MD_A -cpi -append, because I
>>>> am using super computer so the job will be restarted for multiple times.
>>>>
>>>> 2) both simulation are 100ns, both log files indicate that they have
>>>> finished 100ns simulations. I checked my screen out put for B, whenever
>>>> it
>>>> restarted, the output on screen is correct and continuously ( something
>>>> like: 100000000 steps, 100000.0 ps (continuing from step 95963500,
>>>>   95963.5
>>>> ps).
>>>>
>>>> 3) With gmxcheck, it looks that my B simulation only reached 29,7ns(which
>>>> means only the first two restart simulation were recorded into trr, which
>>>> log files are not saying in this way.)
>>>> *
>>>> *
>>>> *For system A (the successful one), it shows*
>>>>
>>>> Reading frame       0 time    0.000
>>>> # Atoms  72166
>>>> Last frame     2000 time 100000.000
>>>>
>>>>
>>>> Item        #frames Timestep (ps)
>>>> Step           2001    50
>>>> Time           2001    50
>>>> Lambda       2001    50
>>>> Coords         2001    50
>>>> Velocities     2001    50
>>>> Forces           0
>>>> Box            2001    50
>>>>
>>>> *for system B (the failed one)*
>>>>
>>>> Reading frame       0 time    0.000
>>>> # Atoms  72166
>>>> Reading frame     500 time 25000.000
>>>>
>>>>
>>>> Item        #frames Timestep (ps)
>>>> Step           596    50
>>>> Time           596    50
>>>> Lambda         596    50
>>>> Coords         596    50
>>>> Velocities     596    50
>>>> Forces           0
>>>> Box            596    50
>>>>
>>>> 4) I also compared the generated files' sizes, A = B in trr, edr , gro,
>>>> cpt, if B crashes before reaching 100ns, the size would be the same as A.
>>>>
>>>> I am wondering what was wrong with my trr file of B . I appreciate for
>>>> any
>>>> suggestions!
>>>>
>>>> Sincerely
>>>>
>>>> Xiaojia
>>>>
>>>>
>>>>
>> --
>> ==============================**==========
>>
>> Justin A. Lemkul, Ph.D.
>> Research Scientist
>> Department of Biochemistry
>> Virginia Tech
>> Blacksburg, VA
>> jalemkul[at]vt.edu | (540) 231-9080
>> http://www.bevanlab.biochem.**vt.edu/Pages/Personal/justin<http://www.bevanlab.biochem.vt.edu/Pages/Personal/justin>
>>
>> ==============================**==========
>> --
>> gmx-users mailing list    gmx-users at gromacs.org
>> http://lists.gromacs.org/**mailman/listinfo/gmx-users<http://lists.gromacs.org/mailman/listinfo/gmx-users>
>> * Please search the archive at http://www.gromacs.org/**
>> Support/Mailing_Lists/Search<http://www.gromacs.org/Support/Mailing_Lists/Search>before posting!
>> * Please don't post (un)subscribe requests to the list. Use the www
>> interface or send it to gmx-users-request at gromacs.org.
>> * Can't post? Read http://www.gromacs.org/**Support/Mailing_Lists<http://www.gromacs.org/Support/Mailing_Lists>
>>

-- 
========================================

Justin A. Lemkul, Ph.D.
Research Scientist
Department of Biochemistry
Virginia Tech
Blacksburg, VA
jalemkul[at]vt.edu | (540) 231-9080
http://www.bevanlab.biochem.vt.edu/Pages/Personal/justin

========================================