Program received signal SIGSEGV: Segmentation fault - invali

Clearly defined bug reports and their fixes
Post Reply
pagrumiaux
Posts: 3
Joined: 21 Jul 2016, 12:42
Antispam: Yes

Program received signal SIGSEGV: Segmentation fault - invali

Post by pagrumiaux »

Hi,

I'm new to Elmer and I have a lot of pain to debug with it.

I'm working on Ubuntu 16.04 LTS with Elmer downloaded and installed from https://www.csc.fi/web/elmer/sources-and-compilation.
I'm trying to run the tutorial on eigenvalues analysis (from http://www.nic.funet.fi/index/elmer/doc ... nonGUI.pdf) with ElmerSolver, but I get this error :

Code: Select all

Program received signal SIGSEGV: Segmentation fault - invalid memory reference.

Backtrace for this error:
#0  0x7F1DC9EB1E08
#1  0x7F1DC9EB0F90
#2  0x7F1DC9B0249F
#3  0x7F1DCA410987
#4  0x7F1DCA5CB238
#5  0x7F1DCA5BF0BC
#6  0x40218E
#7  0x4024B0
#8  0x7F1DC9AED82F
#9  0x401E18
#10  0xFFFFFFFFFFFFFFFF
Erreur de segmentation (core dumped)
I've managed to run another example which worked perfectly though.
I don't understand what the error comes from as I just followed the tutorial ...

If you can help thank you :)
I attached the sif :
case.sif
(1.37 KiB) Downloaded 426 times
Cheers,

Pierre-Amaury
pagrumiaux
Posts: 3
Joined: 21 Jul 2016, 12:42
Antispam: Yes

Re: Program received signal SIGSEGV: Segmentation fault - invali

Post by pagrumiaux »

Hi,

For those who have the same error, I have solved my issue.

This was simple (but not easy to discover ...) : my mesh was not fine enough, so the solver could not be able to run a solution.
Be careful to have a nice mesh (by setting some parametr in the configure menu in ElmerGUI)

Bye
raback
Site Admin
Posts: 4812
Joined: 22 Aug 2009, 11:57
Antispam: Yes
Location: Espoo, Finland
Contact:

Re: Program received signal SIGSEGV: Segmentation fault - invali

Post by raback »

Hi

The problem was solved but just wanted to discuss what the problem probably was.

Now the eigenvalue solver assumes a orthogonal basis that has perhaps 2*N vectors, where N is the number of eigenmodes requested. It could be that if the mesh was very sparse the number of vectors was larger than the number of possible degrees of freedom. This would mean that the solution is exhausted before all vectors in the basis are used. Hence the algorithm probably was used in a way that was not foreseen resulting to the crash.

-Peter
Post Reply