open TELEMAC-MASCARET The mathematically superior suite of solvers

Hi everyone

I am currently working with Telemac and Tomawac on a triangular mesh of the North sea. To save calculation time I had to adjust the network. I deleted a part of it in BlueKenue and made new boundary conditions (instead of 1 open and 1 closed boundary there are now 2 open and 2 closed boundaries). I can run the simulation with a single processor but no longer using multiple processors (I had no problem with this in the old, unadjusted network). I believe it has something to do with the partitioning of the mesh but I am not sure how to solve this since I am new to Telemac. Any ideas to what I have forgotten? The error is shown in the attachments.

Thanks in advance.

Hello

All my best wishes for this new year !

This looks like a problem with your boundary condition file. You have to generate it again properly with BK. If the mesh has been changed, you need absolutely to generate once again the boudnary condition file.

I hope this helps

Riadh

Dear Riadh

Thank you for this information. I made a new boundary conditions file (.cli) in BK an saved it in the correct map. You can see the conditions file in attachment. It has however not solved the problem I am afraid. In the bottom right corner is where I cut of part of the previous network.


Thanks for your help so far.

Joren

Hi
If you started from a selafin file and just cut a part of the mesh and save the new selafin mesh, regenerate the BC file is not sufficient because with such manipulation you let in you selafin file a wrong ipobo array...
Sometimes this could create such problems.
Try to create a new selafin file and drag and drop the bottom in it to ensure the generation of a compatible ipobo array.

Hope this helps

Best wishes for 2018

Thank you Coulet!

This helped (a bit). I am now able to go further in the calculations but still get an error regarding the MPI "received suspend command for a pmi context that doesn't exist" any ideas on how to solve this? In single processor mode it works just fine but parallel doesn't work. Other parts in the file directory are identical as before the adjustments to the network (fortran file, meteo file, etc)

Thanks for your help

Hi
Could you try to activate the debugger mode to try to see in which part of the code the problem appear?
DEBUGGER = 1 in your steering file

Saying this, I suppose the message about point 1 is normal in your case...

regards

The debugger is set to 1 in both steering files. The error given in previous attachment is the one I get with activated debugger mode.

Kind regards

I read it might have something to do with the boundary file (however it does work with a single processor which is odd). I refer to the .cli file inside the steering file, is this ok? Do I have to do anything with the more extensive boundary.bc2 file?

Hi
Too hard to find the reason of the crash without your files.
If you could attach them (in a zip) this will allow us to try to reproduce the problem and possibly understand why

Regards

You can find the files in attachment.

Thanks for the help