Welcome, Guest
Username: Password: Remember me
  • Page:
  • 1
  • 2

TOPIC: Parallel Mode - Ubuntu

Parallel Mode - Ubuntu 12 years 9 months ago #3732

  • mat.deiber
  • mat.deiber's Avatar
Dear all,

I am trying to run Telemac-2d in parallel mode with Ubuntu. Once I launch the simulation I obtain the following information on the scream :




Loading Options and Configurations
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

... parsing configuration file: /home/mathieu/opentelemac/v6p1/config/systel.cfg


Running test.cas with telemac2d under /home/mathieu/VOR
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

... reading module dictionary
/home/mathieu/VOR/test.cas
... running in English
copying: test.cas
copying: test.cli
copying: test.geo
copying: telemac2dv6p1.dico
copying: dec272010_0015NADD.prn
partitioning: T2DGEO
duplicating: T2DIMP
/usr/bin/mpiexec -wdir /home/mathieu/VOR/test.cas_2012-02-13-08h23min43s -n 2 /home/mathieu/VOR/test.cas_2012-02-13-08h23min43s/out_telemac2dv6p1
mathieu@Modelling-Ub:~/opentelemac/v6p1/pytel$






The program is doing a few things (like dividing the mesh....), but it doesn't run the hydrodynamic model...and I can't find any error message telling me why it suddenly stops. Does anyone encounter this problem before or get an idea about how to solve this problem?

Thanks in advance for your help,

Mathieu
The administrator has disabled public write access.

Re: Parallel Mode - Ubuntu 12 years 9 months ago #3734

  • c.coulet
  • c.coulet's Avatar
  • OFFLINE
  • Moderator
  • Posts: 3722
  • Thank you received: 1031
How could you say the model doesn't run?
Maybe the program is running on background?
Have you seen if some process runs?
Otherwise, maybe a communication problem?
Christophe
The administrator has disabled public write access.

Re: Parallel Mode - Ubuntu 12 years 9 months ago #3737

  • mat.deiber
  • mat.deiber's Avatar
Dear Christophe,

Once I execute the command, I check what is created in the text.cas_2012-02-13**** folder, the file T2DGEO00001-00001 T2DGEO00001-00000,T2DCLI00001-00001..... are present, the log file end up with the comment "+---- PARTEL: NORMAL TERMINATION ----+"

However I can't find the file T2DRES which is always present when I run telemac2d on Windows in scalar mode....

I check the cpu usage and nothing particular is running. Is it any test I can do to check where the problem comes from?

Thanks in advance for your help
The administrator has disabled public write access.

Re: Parallel Mode - Ubuntu 12 years 9 months ago #3738

  • c.coulet
  • c.coulet's Avatar
  • OFFLINE
  • Moderator
  • Posts: 3722
  • Thank you received: 1031
Hi

As i understand, everything is OK in the temp directory.
Maybe you could try to run the simulation manually in the temps directory.
Try with mpiexec -n2 out...

Hope this helps
Christophe
The administrator has disabled public write access.

Re: Parallel Mode - Ubuntu 12 years 9 months ago #3803

  • mat.deiber
  • mat.deiber's Avatar
Hi Christophe,

I try your suggestion by starting it manually but it seems that there is still something wrong :


mathieu@Modelling-Ub:~/opentelemac/v6p1/pytel$ mpiexec -n2 /home/mathieu/VOR/test.cas_2012-02-16-11h02min01s/out_telemac2dv6p1
mpiexec was unable to launch the specified application as it could not find an executable:

Executable: -2
Node: Modelling-Ub

while attempting to start process rank 0.
-


I wonder if there is a way to check if OPENmpi is well install and linked properly to Telemac...?
The administrator has disabled public write access.

Re: Parallel Mode - Ubuntu 12 years 9 months ago #3806

  • c.coulet
  • c.coulet's Avatar
  • OFFLINE
  • Moderator
  • Posts: 3722
  • Thank you received: 1031
Hi

Please go into the temp directory and just execute mpiexec -n2 out_...

For the check of openmpi, I don't know...
Christophe
The administrator has disabled public write access.

Re: Parallel Mode - Ubuntu 12 years 9 months ago #3808

  • UHM
  • UHM's Avatar
Dear Mathieu,

maybe it is a problem with the new perl script in v6p1 that makes the .bat file.
We had a similar problem last week on Ubuntu and I changed the runcode.pl

It was only "./" missing for 2 commands.

I attached our file.

Greetings, Uwe
The administrator has disabled public write access.

Re: Parallel Mode - Ubuntu 12 years 9 months ago #3810

  • mat.deiber
  • mat.deiber's Avatar
Dear UHM,

Thanks a lot for your answer.
I forgot to mention that I am running the code with the Python installation so the runcode.pl will probably not work for me. But I might install Telemac using the pearl installation if I can't solve my problem...so I am still interested to now where the 2 "./" missing commands are (I can't find an attached file).
The administrator has disabled public write access.

Re: Parallel Mode - Ubuntu 12 years 9 months ago #3811

  • UHM
  • UHM's Avatar
In general the parallel installation of v6p1 has some curios aspects.
I installed the new version on 5 computers,

1xUbuntu 11
2xLinux Mint 9 = Ubuntu 10.04
1xSuse 10.3
1xSuse+SGIMPI

The installation worked out of the box for Linux Mint and the SGI Machine, but not for Suse Desktop and not for Ubuntu 11 because of incompatibilities between installed libs and the libs demanded by telemac. Try to use exactly the same versions for metis (4.0.3), mpich2 or openmpi as in the "Perl installation guide". Newer versions might need different flags or different folders or whatever ...

We changed the following:
- Fixing the Perl script runtel.pl
- changing the /bin/sh link from dash to bash
- uninstalling either mpich2 or openmpi, if both are present, there is a strange illegal mixing of the links.

Good luck, Uwe


P.S.:

now I got the following combinations running stable:
Mint 9 / 64bit gfortran 4.4.3 Mpich2 1.2.1p1

Mint 9 / 64bit gfortran 4.4.3 Mpich2 1.2.1p2
Suse /kernel 2.6.27 ifort 11.1 SGI Mpi 2.01 ???

Ubuntu 10.04 gfortran 4.4.5 Mpich2 1.4.1

Suse ifort 11.1
openmpi ???
The administrator has disabled public write access.

Re: Parallel Mode - Ubuntu 12 years 9 months ago #3809

  • mat.deiber
  • mat.deiber's Avatar
Dear Christophe,

I tried to execute mpiexec -n2 out_ but it directly stop running :


mathieu@Modelling-Ub:~/VOR/test.cas_2012-02-16-12h54min00s$ /usr/local/bin/mpiexec -n2 out_telemac2dv6p1
mpiexec was unable to launch the specified application as it could not find an executable:

Executable: -2
Node: Modelling-Ub

while attempting to start process rank 0.
The administrator has disabled public write access.
  • Page:
  • 1
  • 2
Moderators: borisb

The open TELEMAC-MASCARET template for Joomla!2.5, the HTML 4 version.