[Getdp] Antwort: Re: Antwort: Re: zip getdp solution files

m.schoening at e-und-a.ch m.schoening at e-und-a.ch
Wed Dec 21 06:50:20 CET 2016


Thanks again for the answer, I already work with -v2 option.

As workaround, I will try to seperate the solving from the post processing 
to reduce the network traffic.

Nice holidays!

Marc


"getdp" <getdp-bounces at ace20.montefiore.ulg.ac.be> schrieb am 20.12.2016 
15:43:02:

> Von: David Colignon <david.colignon at ulg.ac.be>
> An: m.schoening at e-und-a.ch, getdp at onelab.info
> Datum: 20.12.2016 15:43
> Betreff: Re: [Getdp] Antwort: Re: zip getdp solution files
> Gesendet von: "getdp" <getdp-bounces at ace20.montefiore.ulg.ac.be>
> 
> 
> On 16/12/16 07:03, m.schoening at e-und-a.ch wrote:
> 
> > Hi,
> >
> > thanks for the information and you are right, there is as binary 
> format in getdp. In my tests, it reduces the file size to around 70%
> (perhaps, I did something
> > wrong, I am speaking of element solution files) of the orignal 
> value (it saves 30% traffic) and you are right, gmsh is able to 
> handle this binary files. I would
> > be happy, if the file size can be reduced even more.
> >
> > For example, trying "gzip -6" reduced the file size to around 20% 
> (this saves 80% of the traffic) of the original value and several 
> other languages (e.g.
> > python) are able to handle this file format out of the box, so 
> individual post processing procedures on getdp solution files can be
> adapted with minimal effort.
> >
> > Gmsh is able to read gzipped solution files, but not out of the 
> box. Gmsh extracts the files on the hard disk (not in memory), so 
> you have to clean the working
> > directory manually afterwards.
> 
> Hi,
> 
> you can also try the "-v2" option of getdp:
> 
> Output options:
>    -bin                      create binary output files
>    -v2                       create mesh-based Gmsh output files when 
possible
> 
> that could also reduce the size of your postpro files.
> 
> Regards,
> 
> Dave
> 
> 
> >
> > Best regards,
> >
> > ppa. Marc C. Schöning
> > Dr.-Ing.
> > Leiter Forschung & Entwicklung / Head of R&D
> > Mitglied der Geschäftsleitung / Member of the Management
> > ----------------------------------------------------
> > e+a
> > Elektromaschinen und Antriebe AG
> > Bachstrasse 10
> > 4313 Moehlin
> > SWITZERLAND
> > ----------------------------------------------------
> > Phone +41 61 855 92 92
> > Fax +41 61 855 92 99
> > Mail m.schoening at e-und-a.ch
> >
> > http://www.e-und-a.ch <http://www.e-und-a.ch/>
> > ----------------------------------------------------
> >
> > CONFIDENTIALITY NOTICE
> > This communication and the information it contains is
> > intended for the person(s) or organisation(s) named
> > above and no other person(s) or organisation(s) and
> > may be confidential, legally privileged and protected
> > by law. Unauthorised use, copying or disclosure of any
> > of it may be unlawful. If you have received this
> > communication in error, please contact us immediately
> > by email, telephone or facsimile. We will be happy to
> > accept a reversed charge call (call collect).
> >
> > "getdp" <getdp-bounces at ace20.montefiore.ulg.ac.be> schrieb am 15.
> 12.2016 22:15:54:
> >
> >> Von: David Colignon <david.colignon at ulg.ac.be>
> >> An: m.schoening at e-und-a.ch, getdp at onelab.info
> >> Datum: 15.12.2016 22:16
> >> Betreff: Re: [Getdp] zip getdp solution files
> >> Gesendet von: "getdp" <getdp-bounces at ace20.montefiore.ulg.ac.be>
> >>
> >>
> >> On 15/12/16 14:02, m.schoening at e-und-a.ch wrote:
> >>
> >> > Dear List,
> >> >
> >> > is there a "build in" way to work with zipped solution and mesh
> >> files instead of the "standard" text files? I experience performance 
issues
> >> > running getdp on several cluster nodes, which all write their
> >> results on the same nfs network share.
> >> >
> >> > Additionally, it would be very helpfull, if gmsh could read such
> >> zipped or binary files directly for postprocessing.
> >>
> >>
> >> Hi,
> >>
> >> ace35:~ > ./getdp/BUILD/getdp --help
> >>
> >> GetDP, a General environment for the treatment of Discrete Problems
> >> Copyright (C) 1997-2016 P. Dular and C. Geuzaine, University of Liege
> >> Usage: ./getdp/BUILD/getdp [file] [options]
> >> Processing options:
> >>    -pre 'Resolution'         pre-processing
> >>    -cal                      processing
> >>    -pos 'PostOperation(s)'   post-processing
> >>    -msh file                 read mesh (in msh format) from file
> >>    -gmshread file(s)         read gmsh data (same as GmshRead in 
> resolution)
> >>    -gmshtag tag(s)           tag(s) associated to GmshRead data
> >>    -restart                  resume processing from where it stopped
> >>    -solve 'Resolution'       same as -pre 'Resolution' -cal
> >>    -split                    save processing results in separate 
files
> >>    -res file(s)              load processing results from file(s)
> >>    -name string              use string as generic file name
> >>    -adapt file               read adaptation constraints from file
> >>    -order num                restrict maximum interpolation order
> >>    -cache                    cache network computations to disk
> >> Linear solver options:
> >>    -solver file              specify parameter file (default: 
solver.par)
> >>    -'Parameter' num          override value of solver parameter 
'Parameter'
> >> Output options:
> >>    -bin                      create binary output files
> >> #######################################################
> >>
> >>
> >> And from http://getdp.info/doc/texinfo/getdp.html:
> >>
> >> GetDP can also read more recent versions of the ‘msh’ format (2.0
> >> and above), as well as binary meshes. See the Gmsh documentation for 
more
> >> information about these formats.
> >>
> >> Regards,
> >>
> >> Dave
> >>
> >> --
> >> David Colignon, Ph.D.
> >> 1er logisticien de recherche
> >> Université de Liège
> >> ACE - Applied & Computational Electromagnetics
> >> Sart-Tilman B28
> >> 10, Grande Traverse
> >> 4000 Liège - BELGIQUE
> >> Tél: +32 (0)4 366 37 32
> >> http://www.ulg.ac.be/nic4
> >>
> >>
> >> >
> >> > Best regards,
> >> >
> >> > ppa. Marc C. Schöning
> >> > Dr.-Ing.
> >> > Leiter Forschung & Entwicklung / Head of R&D
> >> > Mitglied der Geschäftsleitung / Member of the Management
> >> > ----------------------------------------------------
> >> > e+a
> >> > Elektromaschinen und Antriebe AG
> >> > Bachstrasse 10
> >> > 4313 Moehlin
> >> > SWITZERLAND
> >> > ----------------------------------------------------
> >> > Phone +41 61 855 92 92
> >> > Fax +41 61 855 92 99
> >> > Mail m.schoening at e-und-a.ch
> >> >
> >> > http://www.e-und-a.ch <http://www.e-und-a.ch/><
http://www.e-und-a.ch/>
> >> > ----------------------------------------------------
> >> >
> >> > CONFIDENTIALITY NOTICE
> >> > This communication and the information it contains is
> >> > intended for the person(s) or organisation(s) named
> >> > above and no other person(s) or organisation(s) and
> >> > may be confidential, legally privileged and protected
> >> > by law. Unauthorised use, copying or disclosure of any
> >> > of it may be unlawful. If you have received this
> >> > communication in error, please contact us immediately
> >> > by email, telephone or facsimile. We will be happy to
> >> > accept a reversed charge call (call collect).
> >> > *
> >> > CONFIDENTIALITY NOTICE*
> >> > This communication and the information it contains is intended for
> >> the person(s) or organisation(s) named above and no other person(s) 
or
> >> > organisation(s) and may be confidential, legally privileged and
> >> protected by law.
> >> > Unauthorised use, copying or disclosure of any of it may be
> >> unlawful. If you have received this communication in error, 
> please contact us
> >> > immediately by email, telephone or facsimile. We will be happy to
> >> accept a reversed charge call (call collect).
> >> >
> >> >
> >> > _______________________________________________
> >> > getdp mailing list
> >> > getdp at onelab.info
> >> > http://onelab.info/mailman/listinfo/getdp
> >> >
> >>
> >> _______________________________________________
> >> getdp mailing list
> >> getdp at onelab.info
> >> http://onelab.info/mailman/listinfo/getdp
> >
> > *
> > CONFIDENTIALITY NOTICE*
> > This communication and the information it contains is intended for
> the person(s) or organisation(s) named above and no other person(s) 
> or organisation(s) and
> > may be confidential, legally privileged and protected by law.
> > Unauthorised use, copying or disclosure of any of it may be 
> unlawful. If you have received this communication in error, please 
> contact us immediately by email,
> > telephone or facsimile. We will be happy to accept a reversed 
> charge call (call collect).
> >
> >
> > _______________________________________________
> > getdp mailing list
> > getdp at onelab.info
> > http://onelab.info/mailman/listinfo/getdp
> >
> 
> 
> _______________________________________________
> getdp mailing list
> getdp at onelab.info
> http://onelab.info/mailman/listinfo/getdp


CONFIDENTIALITY NOTICE
This communication and the information it contains is intended for the 
person(s) or organisation(s) named above and no other person(s) or 
organisation(s) and may be confidential, legally privileged and protected 
by law.
Unauthorised use, copying or disclosure of any of it may be unlawful. If 
you have received this communication in error, please contact us 
immediately by email, telephone or facsimile. We will be happy to accept a 
reversed charge call (call collect).
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://onelab.info/pipermail/getdp/attachments/20161221/a3b89de6/attachment-0001.html>


More information about the getdp mailing list