open TELEMAC-MASCARET The mathematically superior suite of solvers

Hello friends!
I'm working with the coupling of the telemac3d module with the tomawac module. Both function in parallel when set separately. When I do the coupling it starts the simulation with the number of processors selected but it does not do so. In the result it does not divide the mesh during the simulation and its processing speed is the same when it works in a scalar simulation. Someone came across this problem.
Grateful for the attention

Hello Pablo

There is an example of a coupling between telemac3d and tomawac in the examples : examples/tomawac/3Dcoupling, mays be that you can check you did the same thing.

I just tested this case and controlled that geometry is divided using the command

telemac3d.py --ncsize=4 -w=repertoire T3D_littoral.cas

to check what is done during simulation in the directory 'repertoire'

Do not forget that the command (--ncsize ) has the priority over the steering file.
Or may be that you are using an old version.

Hope this helps

Hello!

Hello! The simulation divides the mesh but I'm encountering another problem. Wind data is having problems. When I open the simulation in BlueKenue it shows the message:

WARNING: 69244 bad values found in object WIND ALONG XY

The wind does not present this problem when the simulation is in scalar. Any suggestions to solve this problem?

Grateful for the attention

Hello,

Are you introducing wind in both codes or only in tomawac ?
Which options do you use to read wind data ?
Are your wind outputs right if you run the code alone in parallel mode ?

Regards,
Laurent

Hello! I use a wind file for the Telemac3d and another wind file for the Tomawac. At first there was a conflict between the two wind data but this has already been resolved.
In tomawac I use SUBROUTINE ANAVEN and my data is in an array where the first column is the wind x and the second column the wind y. I computed the wind data and are correct no wrong value. The error appears when I run the tomawac simulation (I'm testing it separately because it has the same error when coupling with telemac3d) in parallel. When I run with a core the wind does not present problem.
Grateful for the attention

Hi,

So the bad values also appears when you run tomawac alone in parallel mode ?
In parallel mode, each subdomain has a local numbering. Did you pay attention to that in anavent ?

Regards,
Laurent

Hi!
Yes when I run the tomawac alone in parallel mode it also appears. But the wave parameter data from WW3 behaves well in the split mesh declared in SUBROUTINE LIMWAC, the problem is only with the wind. I could not identify in SUBROUTINE ANAVEN where I change that.

Grateful for the help

Attached is the fortran file of simulation

Hello Pablo

I have looked at your fortran file and if i understood well it is normal that in parallel it is not working. Your wind file is a formated file so partel does not know how to cut it like it cuts your mesh.
So i would see 2 solutions

- you can use a binary (SERAFIN) file for your wind

- you have to do the job by yourself, telling to the code to which point what you read have to be set, using the tab knolg of your mesh which gives you the global number of your local node.
There is a sisyphe example in of the use in
littoral/user_fortran/bord.f

thank you so much! It worked