[gmx-developers] H-REMD restart

[Mark Abraham]

On 24/01/2011 8:22 PM, Luca Bellucci wrote:
> Dear Mark,
> I have reproduced the run with two replica.
> If you want, i can send you the results with a private mail for a deepest
> check. It is about 8/9M.

I only want to see the .log file, stderr and stdout up until the restart 
simulation start doing the MD steps, to see the messages they provide 
about what sense they've made of the checkpoint files. That should be 
only a few hundred KB at most. Please email me off-list.

Mark

> Thanks
> Luca
>
>> On 01/24/11, Mark Abraham<mark.abraham at anu.edu.au>  wrote:
>>> On 01/24/11, Luca Bellucci<luca.bellucci_s3 at unimore.it>  wrote:
>>>> hanks Mark for quick replay.
>>>> it is 4.5.3
>>>> This is the log file at restart point.
>>>>
>>>> -----------------------------------------------------------
>>>> Restarting from checkpoint, appending to previous log file.
>>>>
>>>> Log file opened on Sun Jan 23 23:40:07 2011
>>>> Host: ionode8  pid: 2916  nodeid: 0  nnodes:  64
>>>> The Gromacs distribution was built Thu Nov 18 14:35:49 CET 2010 by
>>>> cin0644a at loginfn (Linux 2.6.16.60-0.42.10-ppc64 ppc64)
>>>> [..]
>>>> Initializing Replica Exchange
>>>> Repl  There are 64 replicas:
>>>> Multi-checking the number of atoms ... OK
>>>> Multi-checking the integrator ... OK
>>>> Multi-checking init_step+nsteps ... OK
>>>> Multi-checking first exchange step: init_step/-replex ... OK
>>>> Multi-checking the temperature coupling ... OK
>>>> Multi-checking the number of temperature coupling groups ... OK
>>>> Multi-checking the pressure coupling ... OK
>>>> Multi-checking free energy ... OK
>>>> [..]
>>> Seeing the whole logfile up to the first energy output would be more
>>> helpful.
>> Also, stderr and stdout can have useful diagnostic information here.
>>
>> Mark
>>
>>> Mark
>>>
>>>> Started mdrun on node 0 Sun Jan 23 23:40:07 2011
>>>>
>>>>> On 24/01/2011 8:57 AM, Luca Bellucci wrote:
>>>>>> Dear all,
>>>>>> I known that in repl_ex.c is implemented the
>>>>>> H-REMD.  It seems to work, however
>>>>>> H-REMD is not documented, thus I am testing it.
>>>>>> I performed a test  in a BlueGene cluster:
>>>>>>
>>>>>> mpirun -mode VN -np 4096 mdrun_mpi_bg  -s md.tpr -np 16 -multi 128
>>>>>> -replex 2000
>>>>>>
>>>>>> it works fine.
>>>>>>
>>>>>> I restarted with this command:
>>>>>>
>>>>>> mpirun -mode VN -np 4096 mdrun_mpi_bg   -s md.tpr -cpi state.cpt
>>>>>> -append -np 16 -multi 128 -replex 2000
>>>>>>
>>>>>> I followed the GROMACS  "Doing Restarts" section, however it does
>>>>>> not work as expected; the trajecory and energy files are
>>>>>> reinitialized and I lost most of the MD informations.
>>>>>> Only log files were updated as expected.
>>>>> Is this with the latest GROMACS version? If not, please try that.
>>>>>
>>>>> Mark
>>>> --
>>>> gmx-developers mailing list
>>>> gmx-developers at gromacs.org
>>>> http://lists.gromacs.org/mailman/listinfo/gmx-developers
>>>> Please don't post (un)subscribe requests to the list. Use the
>>>> www interface or send it to gmx-developers-request at gromacs.org.
>