open TELEMAC-MASCARET The mathematically superior suite of solvers

Hello,

I am facing an error when lauching partitionning for parallel computations.
The node mentionned in the error changes when I try to change the number of parallel processors. The mesh looks clean (no overconstrained elements)..

Here is the error:

runPARTEL: Could not split your file T3DGEO (runcode=1) with the error as follows:
|
|... The following command failed for the reason above (or below)
|C:\opentelemac-mascaret\v8p0\builds\wing64mpi\bin\partel.exe < PARTEL.PAR >> partel_T3DGEO.log
|
| You may have forgotten to compile PARTEL with the appropriate compiler directive
| (add -DHAVE_MPI to your cmd_obj in your configuration file).




AND BELOW



BEGIN PARTITIONING WITH METIS
|
| RUNTIME OF METIS 0.00000000 SECONDS
|
| THE MESH PARTITIONING STEP HAS FINISHED
|
| ERROR IN BOUNDARIES J= 174
|
| WHILE NPTFR_P(I)= 200
|
| WHILE L= 200




enclosed my config file.

Does anybody faced this problem before?
Thanks,

Thomas

Hi
In such case, your systel file is not important. What we need to try to understand your problem are the project file (geo, cli, cas)

Regards

Hi,

Of course here are the files.
Thanks,

Regards

Thomas

Hi,

Your mesh seems ok, but I think it's a precision issue in your case.
I just shifted your coordinates and your case runs.

Xshift=-1000000
Yshift=-6260000

But now, your tide settings (commented in your steering file) and source point location shouldn't be good anymore..
For the source point, you just have to shift the coordinates also. And for tide, you can re-add the shifted coordinates before recomposing the tidal signal for instance (in tidal_model_t3d.f : see XSHIFT and YSHIFT).

Maybe using a double precision geometry file can also be a solution.

Regards,
Laurent

Hi,

Thank you very much.
Best regards,

Thomas

Hello all,

I have the same problem with shifted coordinates (v8p4), my geometry file has been shifted using run_telfile.py alter --orig 437682 6922105 geo.slf geo_shifted.slf as recommended in opentelemac.org/index.php/kunena/blue-ke...oblem?start=10#43758.

Can parallel run can be used on data shifted this way ?
For tide, I use TPXO. Attached is my steering file, and my mesh.

Thank you very much,
Coline

Hi
I just opened quickly your mesh in BK.
Strange there is no IPOBO value read by BK...
This could be the reason of your problem...

Regards