open TELEMAC-MASCARET The mathematically superior suite of solvers

Here is the content of the TEMPDIR at the end of the run :

-rwxr-xr-x 1 * * 4.3M Mar 23 15:18 out_telemac2d
-rw-r--r-- 1 * * 172K Mar 24 10:57 T2DDICO
-rw-r--r-- 1 * * 1.8K Mar 24 10:57 T2DCAS
-rw-r--r-- 1 * * 41K Mar 24 10:57 T2DCLI
-rw-r--r-- 1 * * 1.5M Mar 24 10:57 T2DGEO
-rw-r--r-- 1 * * 33K Mar 24 10:57 T2DIMP00003-00003
-rw-r--r-- 1 * * 33K Mar 24 10:57 T2DIMP00003-00002
-rw-r--r-- 1 * * 33K Mar 24 10:57 T2DIMP00003-00001
-rw-r--r-- 1 * * 33K Mar 24 10:57 T2DIMP00003-00000
-rw-r--r-- 1 * * 33K Mar 24 10:57 T2DIMP
-rw-r--r-- 1 * * 9.6M Mar 24 10:57 T2DPRE
-rw-r--r-- 1 * * 4 Mar 24 10:57 CONFIG
-rw-r--r-- 1 * * 27 Mar 24 10:57 MPI_HOSTFILE
-rw-r--r-- 1 * * 56 Mar 24 10:57 PARAL
-rw-r--r-- 1 * * 361K Mar 24 10:57 T2DGEO00003-00000
-rw-r--r-- 1 * * 368K Mar 24 10:57 T2DGEO00003-00001
-rw-r--r-- 1 * * 359K Mar 24 10:57 T2DGEO00003-00002
-rw-r--r-- 1 * * 363K Mar 24 10:57 T2DGEO00003-00003
-rw-r--r-- 1 * * 3.8K Mar 24 10:57 partel_T2DGEO.log
-rw-r--r-- 1 * * 24 Mar 24 10:57 PARTEL.PAR
-rw-r--r-- 1 * * 57K Mar 24 10:57 T2DCLI00003-00000
-rw-r--r-- 1 * * 51K Mar 24 10:57 T2DCLI00003-00001
-rw-r--r-- 1 * * 51K Mar 24 10:57 T2DCLI00003-00002
-rw-r--r-- 1 * * 48K Mar 24 10:57 T2DCLI00003-00003
-rw-r--r-- 1 * * 2.4M Mar 24 10:57 T2DPRE00003-00000
-rw-r--r-- 1 * * 2.5M Mar 24 10:57 T2DPRE00003-00001
-rw-r--r-- 1 * * 2.4M Mar 24 10:57 T2DPRE00003-00002
-rw-r--r-- 1 * * 2.4M Mar 24 10:57 T2DPRE00003-00003
-rw-r--r-- 1 * * 4.5K Mar 24 10:57 partel_T2DPRE.log
-rw-r--r-- 1 * * 516K Mar 24 11:09 TelemacBayonne.cas_2016-03-24-10h57min03s.sortie
-rw-r--r-- 1 * * 15 Mar 24 11:09 gretel_T2DRES.par
-rw-r--r-- 1 * * 367 Mar 24 11:09 gretel_T2DRES.log

And still no T2DRESxxx and no PExxx.log.

The thing is that during the run there is only the T2DRES file as a result file, even though the executable is runned on several processors. At the end of the run this file is copied as "TelemacBayonne_Resultats.slf" and I checked with BlueKenue and this result file looks normal.

Hi
this probably means you have a problem with the parallel run.
Even your model was split it seems to me that you run it in serial mode...

You could look into the sortie file to check the parallel message Barrier Passed which exist in //

regards

You mean that there should be a "Barrier Passed" message in my sortie file ? Hum, this is not the case. The command "grep -ir 'Barrier Passed'" launched in my Bayonne file returns nothing.

What should be the origin of the problem ? A compilation problem ?

I'm not sure of the exact message but I know there is some message which indicate the parallel run in the beginning of the sortie file.
There is a mention of master and slave processors and the call to P_INIT

As the listing is not so huge, maybe you could join it (with a txt extension)

About the origin, I think it's more a // configuration problem rather than a compilation problem. -> join config file

regards

OK, I cleaned the folder where I run TELEMAC, I compiled TELEMAC with susgortranopenmpi again, and now it looks like the results are normal... I'm sorry I don't know what was the problem and what made things right. Nonetheless I must thank you for your help and your attention