[gmx-developers] how to extend a run using tpbconv?

Berk Hess hessb at mpip-mainz.mpg.de
Wed Sep 17 15:26:18 CEST 2008


Hi,

The problem was that the number of pme nodes, which is read from the 
checkpoint file,
was only set on the master node and not on all the other nodes.
I have fixed it.

Thanks for reporting this bug,

Berk

Berk Hess wrote:
> Hi,
>
> I have fixed problems like this a few weeks ago.
> Is this with the newest CVS code?
>
> If not could you try it with the newest code.
> If it still you can mail me your new tpr file and your cpt file.
>
> Berk
>
> spitaleri.andrea at hsr.it wrote:
>> Hi Berk,
>> thanks for your reply. I followed your suggestion but I did not work.
>> Basically:
>> 1. tpbconv -s old -extend (or until) -o new -> it does not work (it ask
>> for trr)
>>
>> 2. tpbconv -s old -nsteps -o new 2a. mdrun -s new -cpi file.cpt it 
>> crashes with this error:
>>
>> Program mdrun, VERSION 4.0_beta1
>> Source code file: domdec.c, line: 5627
>>
>> Fatal error:
>> The size of the domain decomposition grid (12) does not match the number
>> of nodes (24). The total number of nodes is
>>  24
>>
>> The only way to get to work it is in the old fashion:
>> tpbconv -s -extend -o new -f -e
>> mdrun -s new (without file.cpt)
>>
>> I am just wondering wheter the last procedure is still correct.
>>
>> Thanks in advance
>>
>> andrea
>>
>> Andrea Spitaleri PhD
>> Dulbecco Telethon Institute c/o DIBIT Scientific Institute
>> Biomolecular NMR, 1B4
>> Via Olgettina 58
>> 20132 Milano (Italy)
>> Tel: 0039-0226434348/5622/3497/4922
>> Fax: 0039-0226434153
>>
>> ----- Original Message -----
>> From: Berk Hess <hessb at mpip-mainz.mpg.de>
>> Date: Tuesday, September 16, 2008 5:00 pm
>> Subject: Re: [gmx-developers] how to extend a run using tpbconv?
>>
>>  
>>> Hi,
>>>
>>> Extending is only required if you want more steps then you initially 
>>> set in your mdp file.
>>>
>>> The procedure is to first modify only nsteps in your tpr file and 
>>> then continue from the checkpoint file:
>>> 1 tpbconv -s old -extend (or -until or -nsteps) -o new
>>> 2 mdrun -s new -cpi file.cpt
>>>
>>> Berk
>>>
>>> andrea spitaleri wrote:
>>>    
>>>> Hi then,
>>>> I see that now gromacs can read checkpoint file. Which is now       
>>> the    
>>>> right procedure to extend a simulation using tpbconv?
>>>>
>>>> 1. tpbconv -s file.tpr -f file.cpt -extend 1000 -o new
>>>> or
>>>> 1. tpbconv -s file.tpr -f file.trr -extend 1000 -o new
>>>>
>>>> 2. mdrun -s new
>>>> or
>>>> 2. mdrun -s new -cpi file.cpt
>>>>
>>>> Then, do I need to keep the same number of cpus of the previous       
>>> run or    
>>>> I can change?
>>>>
>>>> thanks in advance
>>>>
>>>> Regards,
>>>>
>>>> andrea
>>>>
>>>>       
>>> _______________________________________________
>>> gmx-developers mailing list
>>> gmx-developers at gromacs.org
>>> http://www.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.
>>>
>>>     _______________________________________________
>>> gmx-developers mailing list
>>> gmx-developers at gromacs.org
>>> http://www.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.
>>> ------------------------------------------------------------------------ 
>>>
>>>
>>> This email was Anti Virus checked by Astaro Security Gateway. 
>>> http://www.astaro.com
>
> _______________________________________________
> gmx-developers mailing list
> gmx-developers at gromacs.org
> http://www.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.




More information about the gromacs.org_gmx-developers mailing list