open TELEMAC-MASCARET The mathematically superior suite of solvers

My understanding of these two arrays is below:
IPOBO - its size is the total number of mesh nodes. Each array element value is either 1 or 0, where 1 indicates a boundary point, and 0 not. The array should be set when generating an input Serafin file (e.g. geometry).

KNOLG - its size depends on how the entire mesh is split for parallel run. This array is mainly used by the TELEMAC code internally for writing the output Serafin file in the parallel run. For example, an input mesh with 10 nodes is split to two pieces with 5 nodes each. So for the first piece, KNOLG array size will be 5, and KNOLG(2)=8 means the local node ID 2 corresponds to global node ID 8. It is not related to how an input mesh should be prepared.

Could someone kindly help to confirm if this is correct? Your input is much appreciated!

Hailiang

Hi,

Yes this is almost it you can find a more detail description of the format in the developer guide:
http://ot-svn-public:telemac1*@svn.opentelemac.org/svn/opentelemac/tags/v8p1r1/documentation/Misc/developer_guide/developer_guide_v8p1.pdf

The definition for IPOBO is 0 or boundary node number and not 1. But it should work within Telemac as it is only counting the number of non zero values.

Also a Python class is available to read/write SERAFIN files (TelemacFile in scripts/python3/data_manip/extraction/telemac_file.py). It is a wrapper on the Hermes (Telemac I/O Fortran Library).

Yugi, thank you for the speedy response! Do you mean in IPOBO array a boundary node value can be any non-zero integer? I see page 139 (comment section) in the developer manual says 1 for a boundary point.

Have all python2 scripts been ported to python3?

What you are looking at is what the hermes library is returning.
But IF you look at page 184 you have the description of the format.

Yes all the scripts have been ported since v8p0 (I think).
And in v8p2 only python3 will be available.