open TELEMAC-MASCARET The mathematically superior suite of solvers

I am running a T3D model with MPI, with 3 liquid boundaries defined in a timeseries data file. The MPI processes are spread over 10 cores.

An SL() boundary update info print only outputs to the LOG file for the processor that handles it. While the other Q() boundaries output on all processors LOG files including the one with the SL() boundary! This is despite the fact that those boundaries are handled by just 2 other independent processors.

This was very confusing while trying to track down a boundaries configuration error.

Does this suggests some strange behaviour in the MPI setups?

Hello,

Is it some print that you added or something we did ?

The thing to know is that :

PRINT*,'Hello World'

will be printed by all processors, while :

WRITE(LU,*) 'Hello World'

will be redirected to log files for all processors except number 0 which appears on the listing.

With best regards,

Jean-Michel Hervouet