Thanks for the clarification!

Christoph

2016-07-25 14:34 GMT-06:00 Chang Woon Jang <[email protected]>:
> Dear Christoph,
>
>     It is my misunderstanding during finding ways to solve the domain
> decomposition error on Stampede. Gromacs suggests optimizing the DD grids
> and choosing the proper ranks. When standalone mode is ok, I thought that
> log file might give me some clue about that. Therefore, I looked at the log
> file saying that csg_stat --nt 8 ...
>
> I guessed that this may be a way to solve the error. That is why I asked it.
>
> However, I found the answer after I tried few numbers for "mdrun -nt
> option".
>
> <mdrun>
>     <opts>-nt number</opts>
> <mdrun>
>
>
> Gromacs manual: "With thread-MPI there are additional options -nt, which
> sets the total number of threads."
>
> The IBI is running now.
>
> Thank you for your kind answers.
>
> Best regards,
> Changwoon Jang
>
>
>
>
>
> On Mon, Jul 25, 2016 at 4:09 PM, Christoph Junghans <[email protected]>
> wrote:
>>
>> Can you explain, what the issue with csg_stat was please? Is there an
>> error?
>>
>> 2016-07-25 12:25 GMT-06:00 Chang Woon Jang <[email protected]>:
>> > Dear Christoph,
>> >
>> >    Thank you for your reply. Does it mean that I need to set up the
>> > following like
>> >
>> > <simulation>
>> >    <tasks>--nt number</tasks>
>> > </simulation>
>> >
>> > in the <gromacs> in xml file?
>> Ok, "cg.inverse.simulation.tasks" means
>> <cg>
>>   <inverse>
>>     <simulation>
>>       <tasks>8</tasks>
>>     </simulation>
>>   </inverse>
>> </cg>
>> but make sure to put it in existing blocks!
>>
>> Christoph
>>
>> >
>> > Thank you.
>> >
>> > Best regards,
>> > Changwoon Jang
>> >
>> >
>> >
>> >
>> >
>> > On Mon, Jul 25, 2016 at 1:42 PM, Christoph Junghans <[email protected]>
>> > wrote:
>> >>
>> >> 2016-07-25 9:59 GMT-06:00 Chang Woon Jang <[email protected]>:
>> >> > Dear Votca Users,
>> >> >
>> >> >    I would like to solve the gromacs domain decomposition error on
>> >> > parallel
>> >> > machine. I have asked Gromacs Users about the error. The final
>> >> > conclusion is
>> >> > to choose the proper number of ranks. The option "-nt 27" needs to be
>> >> > included in setting file. I have tried the following options.
>> >> >
>> >> >    <mdrun>
>> >> >         <opts>-nt 27</opts>
>> >> >    </mdrun>
>> >> >
>> >> > However, it does not influence the domain decomposition.
>> >> >
>> >> >
>> >> >
>> >> > I have tested standalone mode in my local machine. It works well. I
>> >> > have
>> >> > looked at inverse.log file. It says "Running critical command
>> >> > 'csg_stat
>> >> > --nt
>> >> > 8 --options /home/tmp/settings.xml --top topol.tpr --trj traj.xtc
>> >> > --begin 10
>> >> > --first-frame 0'.  I think this is the key issue.
>> >> What is exactly is the issue with csg_stat and 8 threads?
>> >> csg_stat should work with all number of threads, otherwise please
>> >> report an issue:
>> >> <https://github.com/votca/csg/issues>
>> >>
>> >> >
>> >> > Therefore, how can I change csg_stat --nt "numbers" via setting.xml
>> >> > file?
>> >> You can limit the number of threads using the xml option
>> >> "cg.inverse.simulation.tasks"!
>> >>
>> >> Christoph
>> >> >
>> >> > Thank you.
>> >> >
>> >> >
>> >> > Best regards,
>> >> > Changwoon Jang,
>> >> >
>> >> > Postdoctoral Research Fellow
>> >> > Department of Chemical & Biological Engineering, Drexel University
>> >> > 3141 Chestnut Street, Philadelphia, PA 19104
>> >> >
>> >> > Voice: (662) 617-2267
>> >> > E-mail: [email protected]
>> >> >
>> >> > --
>> >> > You received this message because you are subscribed to the Google
>> >> > Groups
>> >> > "votca" group.
>> >> > To unsubscribe from this group and stop receiving emails from it,
>> >> > send
>> >> > an
>> >> > email to [email protected].
>> >> > To post to this group, send email to [email protected].
>> >> > Visit this group at https://groups.google.com/group/votca.
>> >> > For more options, visit https://groups.google.com/d/optout.
>> >>
>> >>
>> >>
>> >> --
>> >> Christoph Junghans
>> >> Web: http://www.compphys.de
>> >>
>> >> --
>> >> You received this message because you are subscribed to the Google
>> >> Groups
>> >> "votca" group.
>> >> To unsubscribe from this group and stop receiving emails from it, send
>> >> an
>> >> email to [email protected].
>> >> To post to this group, send email to [email protected].
>> >> Visit this group at https://groups.google.com/group/votca.
>> >> For more options, visit https://groups.google.com/d/optout.
>> >
>> >
>> > --
>> > You received this message because you are subscribed to the Google
>> > Groups
>> > "votca" group.
>> > To unsubscribe from this group and stop receiving emails from it, send
>> > an
>> > email to [email protected].
>> > To post to this group, send email to [email protected].
>> > Visit this group at https://groups.google.com/group/votca.
>> > For more options, visit https://groups.google.com/d/optout.
>>
>>
>>
>> --
>> Christoph Junghans
>> Web: http://www.compphys.de
>>
>> --
>> You received this message because you are subscribed to the Google Groups
>> "votca" group.
>> To unsubscribe from this group and stop receiving emails from it, send an
>> email to [email protected].
>> To post to this group, send email to [email protected].
>> Visit this group at https://groups.google.com/group/votca.
>> For more options, visit https://groups.google.com/d/optout.
>
> --
> You received this message because you are subscribed to the Google Groups
> "votca" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to [email protected].
> To post to this group, send email to [email protected].
> Visit this group at https://groups.google.com/group/votca.
> For more options, visit https://groups.google.com/d/optout.



-- 
Christoph Junghans
Web: http://www.compphys.de

-- 
You received this message because you are subscribed to the Google Groups 
"votca" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
To post to this group, send email to [email protected].
Visit this group at https://groups.google.com/group/votca.
For more options, visit https://groups.google.com/d/optout.

Reply via email to