open TELEMAC-MASCARET The mathematically superior suite of solvers

Hi,

I'm trying to read non-erodable depth (ZR) from a selafin file. I have tried to program NOEROD.f to read (a) the BINARY DATA FILE 1, (b) the TELEMAC2D GEOMETRY FILE, or (c) the SISYPHE GEOMETRY FILE, using FIND_IN_SEL subroutine, and I have obtained always the same error:

forrtl: severe (157): Program Exception - access violation
Image PC Routine Line Source
out2684_win32.exe 00405D6B Unknown Unknown Unknown
out2684_win32.exe 004018C7 Unknown Unknown Unknown
out2684_win32.exe 00401308 Unknown Unknown Unknown
out2684_win32.exe 005101CA Unknown Unknown Unknown
out2684_win32.exe 0048EFD1 Unknown Unknown Unknown
out2684_win32.exe 00474CFB Unknown Unknown Unknown
out2684_win32.exe 0046E493 Unknown Unknown Unknown
out2684_win32.exe 0044EE72 Unknown Unknown Unknown
kernel32.dll 773DD309 Unknown Unknown Unknown
ntdll.dll 77701603 Unknown Unknown Unknown
ntdll.dll 777015D6 Unknown Unknown Unknown
Duration of job : 1 seconds ( 0:0:1 ) (system=0.032 sec)


I'm attaching all the files in case someone has some time to look at it.

Some additional comments:

(1) Compilation seems to work fine only if defining the FORTRAN FILE in TELEMAC2D parameters file. If FORTRAN FILE is included in both, TELEMAC2D and SISYPHE parameter files, it doesn't compile, and if defined only in SISYPHE file it is not used at all.

(2) The problem is always the same and is related with FIND_IN_SEL subroutine (this was checked including some PRINT commands in the FORTRAN file). I have tried Intel Compiler 11 and 2011, Windows Vista and 7, and parallel and non-parallel configuration, and the problem is always the same.

(3) I’ve been looking FONSTR.f as a reference, and tried to copy the way it reads BOTTOM, but still didn’t works.

Any advice will be very welcome.

Regards,
Sebastián

I don't have time to look at your fortran file now, so won't comment on that one.

However, have your tried to add the rigid bed layer to the geometry file and use that as the previous sedimentological computation file
with
computation continued = yes?

That seems to work best for me.

Regarding remark (2): you can only provide one fortran file. This should be named in the TELEMAC steering file. If it contains SISYPHE subroutines, these will be used by SISYPHE.

Hope this helps,

Michiel

Hello,

At least one reason is that your BIEF_OBJ structure ROCA is declared but not allocated, see how to do it with BIEF_ALLVEC in the guide for programming or inspire from subroutine point_sisyphe.f, copy what is done for ZF for example. However why don't you use directly ZR ? this would save allocating ROCA...

With best regards,

Jean-Michel Hervouet

Dear Michiel and Jean-Michel,

Thank you very much; both of your suggestions have worked.

I’m attaching a new version of the FORTRAN file, including the allocation of the variable and reading the non-erodable level from the BINARY DATA FILE 1, just in case someone having the same problem wants to use it.

Regards,
Sebastián