open TELEMAC-MASCARET The mathematically superior suite of solvers

Hello

I try to launch a simulation with artemis, but there is a problem with linking. I work with V6P1, intel fortran and widows7. here is the problem :


=========================================================
Telemac System 5.6 to 6.1 - Perl scripts version 6.1
=========================================================
starting...

HOSTTYPE : wintel32s
PROJECT : C:\opentelemac\wintels
BASE DIRECTORY : C:\simulations\BP\agitation
LAUNCH DIRECTORY : C:\simulations\BP\agitation
WORK DIRECTORY : C:\simulations\BP\agitation\para_2905648_tmp
PARAMETER FILE : para_290


*** Fichier de configuration SPECIFIQUE :
C:\opentelemac\wintels\config\systel.ini ***



*** ARTEMIS sur STATION ***


*** Lancement en interactif ***


*** VERSION v6p1 ***

________________________________________________________
Fichier des parametres : para_290
________________________________________________________

________________________________________________________
Starting execution: artemis.bat
________________________________________________________
- FICHIER FORTRAN : borh_290.f
______________________________________________________________________________
*** COMPILATION ***

ifort.exe /c /O2 /iface:cref /iface:nomixed_str_len_arg /nologo /names:upperc
ase /convert:big_endian /extend_source:132 /include:C:\opentelemac\wintels\art
emis\arte_v6p1\wintel32s /include:C:\opentelemac\wintels\bief\bief_v6p1\wintel32
s /include:C:\opentelemac\wintels\damocles\damo_v6p1\wintel32s /include:C:\opent
elemac\wintels\paravoid\paravoid_v6p1\wintel32s /include:C:\opentelemac\wintels\
special\special_v6p1\wintel32s artfort.f
______________________________________________________________________________
*** BIBLIOTHEQUES ***

- C:\opentelemac\wintels\artemis\arte_v6p1\wintel32s\artemisv6p1.lib
- C:\opentelemac\wintels\bief\bief_v6p1\wintel32s\biefv6p1.lib
- C:\opentelemac\wintels\damocles\damo_v6p1\wintel32s\damov6p1.lib
- C:\opentelemac\wintels\paravoid\paravoid_v6p1\wintel32s\paravoidv6p1.lib
- C:\opentelemac\wintels\special\special_v6p1\wintel32s\specialv6p1.lib

*** EDITION DE LIENS ***

xilink: executing 'link'
Microsoft (R) Incremental Linker Version 10.00.30319.01
Copyright (C) Microsoft Corporation. All rights reserved.

/stack:536870912
/subsystem:console
/incremental:no
/stack:536870912
/nodefaultlib:libc.lib
/out:out5648_wintel32s.exe
artfort.obj
C:\opentelemac\wintels\artemis\arte_v6p1\wintel32s\artemisv6p1.lib
C:\opentelemac\wintels\bief\bief_v6p1\wintel32s\biefv6p1.lib
C:\opentelemac\wintels\damocles\damo_v6p1\wintel32s\damov6p1.lib
C:\opentelemac\wintels\paravoid\paravoid_v6p1\wintel32s\paravoidv6p1.lib
C:\opentelemac\wintels\special\special_v6p1\wintel32s\specialv6p1.lib
biefv6p1.lib(solve.obj) : error LNK2019: symbole externe non rÚsolu _PRE4_MUMPS
rÚfÚrencÚ dans la fonction _SOLVE
out5648_wintel32s.exe : fatal error LNK1120: 1 externes non rÚsolus
ERREUR A L'EDITION DE LIEN

________________________________________________________
Execution terminee: artemis.bat
________________________________________________________




Thanks a lot for your help !!

Hi
This is a configuration problem
The mumpsvoid library isn't declare in the library list
Have a look in the systel.ini file (if you use perl) and update the variable libmumpspar and libmumpsseq.

Hope this helps

hello

thanks a lot, I done this before but there was a declaration missing. Now it works

hello,

in fact, it doesn't work. Whan I lauch artemis, this message appears : 'out420_wintel32s.exe a cessé de fonctionner'. I don't know why...


In the command board there is :

Thanks a lot for your help

Hi
This time it's not linked to the source but to your computation.
I don't know artemis well but as it's indicate in the listing you've got a O or negative value in the matrix, and the direction is negative.
Maybe you could change the direction from -33 to 327.

Hope it's help

hello,

In telemac v6p0, there were the message " CNTPRE (ARTEMIS) : PRECONDITIONNEMENT DIAGONAL NON APPLIQUE (UN ELEMENT DIAGONA
L DE LA MATRICE EST NEGATIF OU NUL)" too but the simulation lauched.

may it be an error in the dico file ?

thank you

Hi
You're right! CNTPRE just check the matrix and eventually avoid the diagonal scalling but it doesn't stop the computation.

It seems there is a problem durind the resolution (Solve).
Maybe you could add the DEBUGGER=1 keyword, in order to add some information in the listing.

It's strange taht you only have a message from windows that indicate the program interruption without any other information in telemac listing.

With the debugger option, you will see the run and check in which subroutine the program was when it stop.

Hope it's help

Hello,

I don't find any "debbuger" keyword in artemis...

Can you have a look to my steering file ? It may be a problem with my computer...

Thank you !

Hi
I made a test on my computer and i also have a problem
Tha access violation means a memory problem but it's not simple to find where in the code ... but
I check the model works without fortran file so i suppose the problem is in it. Maybe you give a wrong index in a vector and this exceed the size of this vector...
Or maybe it's linked to the message boundary with incident wave. When i omit the fortran file Ihave 1 boundary with incident wave.

Hope it's helps

hello !

you were right ! I had an error in my fortran file.

Thank you very much !