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

TOPIC: mpi problem when running telemac in parallel

mpi problem when running telemac in parallel 11 years 4 days ago #11035

  • Willvan
  • Willvan's Avatar
hello,
i have an issue with the installation of telemac with the python install path. i have it working in sequential mode and teh examples work like a charm but when i set it up to run in parallel mode it tells me :
I don't know how to run MPI, please help.

although the mpiexec is in my PATH variable and the config looks allright.
can anyone help?
Attachments:
The administrator has disabled public write access.

mpi problem when running telemac in parallel 11 years 4 days ago #11036

  • Willvan
  • Willvan's Avatar
i am already a step further:

i changed the config file :

configs: susgfortrans susgfortransdbg susgfopenmpi

which was wrong of course
The administrator has disabled public write access.

mpi problem when running telemac in parallel 11 years 4 days ago #11042

  • Willvan
  • Willvan's Avatar
i have a following issue with running in parallel:

LECDON_TELEMAC2D: ERROR FOR FILE NUMBER
I= 1 NAME=
THIS FILE SHOULD HAVE A STRING SUBMIT
IN DICTIONARY

PLANTE: PROGRAM STOPPED AFTER AN ERROR


this prevents running in parallel mode.

can someone help?
The administrator has disabled public write access.

mpi problem when running telemac in parallel 11 years 4 days ago #11043

  • jmhervouet
  • jmhervouet's Avatar
Hello,

If you are not in coupling mode, file 1 is the geometry file, check that its name is duly given in the steering file. If yes and if it works in scalar mode I do not know, check with one of the test-cases provided to see if it is linked to your case only.

Regards,

JMH
The administrator has disabled public write access.

mpi problem when running telemac in parallel 11 years 3 days ago #11053

  • Willvan
  • Willvan's Avatar
hi, i don't really know what you mean with coupling mode but i see in the working directory file appearing like:
T2DGEO
PARAL
PARTEL.PAR
partel_T2DGEO.log
T2DCAS
T2DCLI
T2DDICO

in the partel.TD2GEO.log there still is the line :
ERROR: TRY TO RUN PARTEL WITH A SERIAL CONFIGURATION

which i don't understand because it is compiled to run in parallel mode.
The administrator has disabled public write access.

mpi problem when running telemac in parallel 11 years 3 days ago #11054

  • jmhervouet
  • jmhervouet's Avatar
Hello,

OK, so the mesh has not been decomposed by partel, the files are not in the folder.
You get the message: ERROR: TRY TO RUN PARTEL WITH A SERIAL CONFIGURATION

when you do not add -DHAVE_MPI in the compiler directives. This is to link with MPI. SO I'm afraid you will have to recompile.

Regards,

Jean-Michel Hervouet (with help by Yoann Audouin)
The administrator has disabled public write access.

mpi problem when running telemac in parallel 11 years 3 days ago #11056

  • Willvan
  • Willvan's Avatar
i have added the -DHAVE_MPI in the config file and recompiled but to no avail.

i have added the new version of the config file.
Attachments:
The administrator has disabled public write access.

mpi problem when running telemac in parallel 11 years 3 days ago #11060

  • Willvan
  • Willvan's Avatar
after the compile : with modules : clean system (start from scratch)
i now have the problem that it crashes with a segmentation fault.

Program received signal SIGSEGV: Segmentation fault - invalid memory reference.

Backtrace for this error:
#0 0x2AAAAACE5E87
#1 0x2AAAAACE6454
#2 0x2AAAAC9B491F
#3 0x452977 in METIS_PartMeshDual
#4 0x4068E5 in partitioner_
#5 0x41882E in MAIN__ at partel.F:0
sh: line 1: 23422 Segmentation fault (core dumped) /home/user/telemac/v6p3r1/builds/susgfopenmpi/bin/partel < PARTEL.PAR >> partel_T2DGEO.log
The administrator has disabled public write access.

mpi problem when running telemac in parallel 11 years 3 days ago #11061

  • yugi
  • yugi's Avatar
  • OFFLINE
  • openTELEMAC Guru
  • Posts: 851
  • Thank you received: 244
Hi,

You get this error if you are using a earlier version of metis.
You need metis 5.x.x since v6p2.

Hope it helps.
There are 10 types of people in the world: those who understand binary, and those who don't.
The administrator has disabled public write access.

mpi problem when running telemac in parallel 11 years 3 days ago #11065

  • Willvan
  • Willvan's Avatar
ok, installed metis-5.xxx and recompiled the whole lot, so thanks.
It doesn't give any crashes anymore but it looks like i am back at square one.

First when i tried to use parallel processor =2 it gave me an error that the nc parameter didn't match the mpi nc parameter which is at 1. So I put the parallel processors back to 1, and ran it again.

and guess what, another error :

LECDON_TELEMAC2D: ERREUR POUR LE FICHIER
I= 1 NOM=
IL MANQUE UNE CHAINE SUBMIT DANS LE
DICTIONNAIRE

this is giving me a headache.
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.