CFD Online Discussion Forums

CFD Online Discussion Forums (https://www.cfd-online.com/Forums/)
-   OpenFOAM (https://www.cfd-online.com/Forums/openfoam/)
-   -   User defined classes error (https://www.cfd-online.com/Forums/openfoam/82743-user-defined-classes-error.html)

tonyuprm December 4, 2010 02:36

User defined classes error
 
Hi all,

I created a class which imposes body forces on my domain and inserted it into the main code for the solver. The code compiles an runs without that class being included. When I do include it, it compiles, but when I try to run it I get the following error:

*** glibc detected *** ABLPisoFoam.1.2.w: corrupted double-linked list: 0x000000001c01b240 ***

"ABLPisoFoam.1.2.w" is my solver

what does this message mean? Thanks

tkrks December 4, 2010 09:35

http://tinyurl.com/3699b76

Openfoam has errors itself. Do we also have to explain user defined errors?

Tino

tonyuprm December 4, 2010 15:10

Please avoid this kind of comments on this forum Tino.




The code will not show this error when running in parallel. Maybe this is a memory issue?

Thanks,

Tony

wyldckat December 4, 2010 16:21

Greetings to all!

Tony, although Tino did make a harsh comment, he does have a point: with such limited information, we would have to do the same you should Tony, i.e., Google it ;)
And if it were something related to OpenFOAM's code itself, a few hints on which tutorial case it breaks with, could also help us help you, because we would have access to the very same code.

By following Tino's link ("Let Me Google That For You" is really one of those helpfully hilarious sites :D), the very first hit indicates two things:
  1. "Corrupted double-linked list" is commonly related to pointers that point to freed memory or memory out of bounds of the allocated space;
  2. Valgrind is your friend! Or at least, can be your friend, if you learn how to use it.
Personally, I've never used Valgrind, so it's not my friend yet either, but since Eclipse+gdb or "old school isolate and conquer" debugging usually get me to the bug itself, I've never felt the need to resort to it.

Working in parallel mode but not in serial, it's pretty much as likely that the reverse being true with some other simulation case. Just try changing your mesh to have 1 more cell in each direction and the bug might hide itself entirely or reveal itself on both runs!

You might also want to check out these wiki pages:
Best regards and good luck!
Bruno


All times are GMT -4. The time now is 17:46.