open TELEMAC-MASCARET The mathematically superior suite of solvers

Hello,

Please find attached the modified subroutines, so far the equivalent subroutines works properly for telemac2d in scalar mode.
Please let me know your thought about it.
[attachment:1]C:\fakepath\Test_par.rar[/attachment]

Many thanks,

Mathieu

Hello,

I understand that you did a file with space and time varying free surface elevations of a number of points, that you read into SL3, with relevant modifications because your functions are called ETA(NG), NG being the global number of the point. I may be wrong but it seems that NG in your implementation comes from the call of SL3 in bord3d.f, and it is the local number IPOIN2, which changes if you change the number of sub-domains. In fact you should send MESH2D%KNOLG%I(IPOIN2) or as well MESH3D%KNOLG%I(IPOIN2) instead of IPOIN2 in the call to SL3 to get the original number of this point.
Tell me if this is the problem (in this case we could actually do the change for everybody in the next version, as it is less probable that someone needs the local number IPOIN2 in SL3).

With best regards,

Jean-Michel Hervouet

Hello,

I made the changes you suggested :

I replace
HBOR%R(K) = SL3(ICOT,AT,IPOIN2,INFOGR)-ZF%R(IPOIN2)
by
HBOR%R(K) = SL3(ICOT,AT,MESH2D%KNOLG%I(IPOIN2),INFOGR)-ZF%R(IPOIN2)

Everything works fine so far, thanks a lot for your help again.

Kind Regards,

Mathieu

Hello,

OK, I'll probably put this in next version. Beware that what you do may not work in scalar mode, something like:

IF(NCSIZE.GT.1) THEN
HBOR%R(K)=SL3(ICOT,AT,MESH2D%KNOLG%I(IPOIN2),INFOGR)-ZF%R(IPOIN2)
ELSE
HBOR%R(K)=SL3(ICOT,AT,IPOIN2,INFOGR)-ZF%R(IPOIN2)
ENDIF

would be safer.

JMH