open TELEMAC-MASCARET The mathematically superior suite of solvers

Hello,

I have multiple open boundaries, forced with tidal constituents read in using 'FORMATTED DATA FILE 1' (channel 26).
My bord.f looks something like this:

REWIND 26
DO K=1,NPTFRL
READ(26,*) ...
ENDDO

DO 5 K=1,NPTFR
IF(NCSIZE.GT.1) THEN ! PARALLEL MODE
NBC = BOUNDARY_COLOUR%I(K)
ELSE ! NON-PARALLEL MODE
NBC = K
ENDIF

IF(LIHBOR(NBC).EQ.KENT.AND.NCOTE.NE.0) THEN
IF(NCOTE.GE.0) THEN
H1 = ...
HBOR(K) = H1 - ZF(NBOR(K))
HBOR(K) = MAX(0.D0,HBOR(K))
H%R(NBOR(K)) = HBOR(K)
...

When I run in parallel, some boundaries don't read the correct data. So I either have a problem reading the FORMATTED DATA FILE, or I have a problem switching from local boundary nodes to global boundary nodes, or both!

So, I have 2 questions:
1. Is this the best way to read in data (i.e. use channel 26), in parallel?
2. Is this the best way (i.e. use BOUNDARY_COLOUR%I) to assign boundary nodes, in parallel?

Thank you very much!
Peter

Hello,

It seems to me that the test on LIHBOR(NBC) should be in fact on LIHBOR(K) because LIHBOR has a local numbering.

Regards,

Jean-Michel Hervouet

Thank you, Jean-Michel, that sorted it.
Peter