open TELEMAC-MASCARET The mathematically superior suite of solvers

Hello,

I programmed some output to the format output file (T3DFRI) using a user defined version of the subroutine UTIMP. This works fine when I run it in serial. However, when I run in it in parallel, I find the output files (one per processor) in the subdirectory cas_duinkerken_present_condition.txt_2012-08-07-15h44min46s

and I get the following error message:
... did not create outfile /mnt/iscsi/projects/12091/Telemac/Present/calib04/DuinKerkenPresentTimeSeries.txt ( T3DRFO )

Is there any way I can prevent this error from occuring?

Kind regards,

Alexander

Hello,

This file is declared as parallel in the dictionary (then it will be duplicated), but it depends on what the user will do. If you want one file only, change in the dictionary the line starting with SUBMIT that contains T3DRFO : at the end change PARAL into SCAL.

Then you have to decide what you will write and who will write it. For example you can decide that processor 0 will do the job, so in UTIMP put all under a test :

IF(IPID.EQ.0) THEN
WRITE(T3D_FILES(T3DRFO)%LU,*) 'example'
ENDIF

but this requires that this processor has all the information, which depends on what you write.

With best regards,

Jean-Michel Hervouet

Hello,

Thanks for the quick reply, and the interesting information. It will have a look whether it is possible to use a single output file. However, for my purposes, I think it would be sufficient if all the generated files are copied by the python script one folder (and it would continue such that all the other partial files are deleted. Is there any way in which one can do that?

Kind regards,

Alexander