open TELEMAC-MASCARET The mathematically superior suite of solvers

Good morning everyone,
I'm try to run a simulation but I get this error:

"Program received signal SIGSEGV: Segmentation fault - invalid memory reference.
"
I have already read some topics about this problem.. but I don't find the manner to fix mine..
I enclose the complete error message and the .cas files.

I think it could be a problem of mesh size...By the way, if that is the problem how can I deal with it?

I have windows 8.1 with 4 Gb RAM, intel core3.
Telemac is run with a virtual machine with Ubuntu for which I set RAM= 3014MB and 2 processors.







Thank you in advance

Daniele

Hi
You could add the keyword debugger = 1 to precisely see where the program crash (which subroutine)
Maybe not realted but as the type of selafin file is not given in the title, I think telemac automatically consider this file as a SELAFIN file (simple precision). If you build your geometry file in double precision, this could create a misunderstanding. You could precise the geometry file format in the steering file.

Hope this helps

thanks for your reply.
I think the "DEBUGGER=1" is already selected, I can see it in the .cas file.

I have already tried to set in "GEOMETRY FILE FORMAT" the value "SERAFIND" and now the error is different:

LIT: READ ERROR
ONE INTENDED TO READ
A RECORD OF ****** VALUES
OF TYPE: R8
ON LOGICAL UNIT : 1
PLANTE: PROGRAM STOPPED AFTER AN ERROR



Any idea about how to solve this problem??

Thank you in advance

Daniele

Hello

Are you sure that your mesh contains more that 5 million elements ?
this could not be handled by a 4GB RAM computer (I think !)
verify also tha tdebugger =1 because we can not see its effects on the listing

kind regards

Riadh

Thank you riadh for your reply!

I've verified that the debugger = 1; it is also in the cas file (you can check), but as you said and I don't know why, it doesn't give any effects !

By the way, I am also (quite) sure that the mesh contains 5 milion elements.. the geometry file is about 98MB... I created it in Blue Kenue, starting from a DEM 1x1m..
I attach the mesh analysis results for my geometry file.




what do you suggest?
if the problem is the RAM, what do you think I could do? which is the amount of RAM that you think is able to deal with my project?
I would like to have a very defined mesh because I need to simulate a urban flooding..
thank you very much for your time.

Daniele

Hi

Hard to say about the RAM but it's probably the reason of the crash.
Even you need to simulate a urban flooding, 1m is really a small size.
Maybe you could try to build a mesh with 2m or 5m and check if the simulation run.

Hope this helps

problem solved increasing the size of the mesh (now it is 2 meters)
now the elements are about 1 milion but the simulation runs :)

Thank you all for your help

Daniele