CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   ANSYS Meshing & Geometry (http://www.cfd-online.com/Forums/ansys-meshing/)
-   -   [ICEM] "segmentation violation" error (http://www.cfd-online.com/Forums/ansys-meshing/63005-segmentation-violation-error.html)

Jasmine March 25, 2009 14:56

"segmentation violation" error
 
Hi,

Got this error meassage "segmentation violation" many times. For example, when I tried to generate a find grid in ICEM CFD by chaning the maximum element size from 30 to 24. or even when I check a coarse grid quanlity, the error meassage show up!
I actually just added 2GB memory which is not the same brand as the old 2G meomry. Will that be a problem?

What might be the reason?And how should I deal with it? Thanks in advance.

fauka March 25, 2009 23:49

Hi!!

Most of the cases of "segmentation violation" error are because of memory problems. You can check it with the task manager if there is any such problem. Other than that, I think you need to work on your geometry for holes or any such defects.

Jasmine March 26, 2009 17:29

Thanks!
Well, I changed 2 new memory which is the same brand as my old one, and it works. That kind of error message does not show up so far.

hashem_1064 October 29, 2009 05:49

yes, this error is strongly related to the lack of memory.

rwryne October 29, 2009 08:03

Quote:

Originally Posted by hashem_1064 (Post 234466)
yes, this error is strongly related to the lack of memory.

I do not think that is the sole case. A coworker and I were getting this message often with ICEM 11. We would basically have to restart the project fresh from the CAD geometry when it would happen. Would use exact same parameters (aka exact same memory usage) and it would work sometimes, others not.

Have not seen this message at all since switching to ICEM 12.

PSYMN November 11, 2009 23:00

Clear the memory.
 
2 Attachment(s)
Hey Ryne,

It may have been that some processes were still running, or at least still hogging memory. For instance, during the tetra process, elements are swapped around or removed. The matrix keeps the memory slots for these cells for a while and may have an issue clearing the memory in some situations. Other times, a process like Prism.exe may continue to run and be hogging memory even after you think it is done.

The undo buffer may also hog a lot of memory, for instance, if you did a global mesh edit operation operation like converting linear tets to quadratic (or tet to hex), your memory would hold both the old and new tets so that you could "undo" the operation.

Many of these processes, and UNDO its self, were significantly improved at 12.0, which gave us memory improvements, etc. Actually, add mid-side nodes still needs this fix :rolleyes:, but at least you can clear the undo buffer easily now.

When you were quitting and restarting 11.0, you were clearing your memory in a rough but effective way.

I am attaching an image of the memory improvements for Octree Tetra at 12.1 (due out later this month).

rwryne November 12, 2009 09:12

Quote:

Originally Posted by PSYMN (Post 235922)
Hey Ryne,

It may have been that some processes were still running, or at least still hogging memory. For instance, during the tetra process, elements are swapped around or removed. The matrix keeps the memory slots for these cells for a while and may have an issue clearing the memory in some situations. Other times, a process like Prism.exe may continue to run and be hogging memory even after you think it is done.

The undo buffer may also hog a lot of memory, for instance, if you did a global mesh edit operation operation like converting linear tets to quadratic (or tet to hex), your memory would hold both the old and new tets so that you could "undo" the operation.

Many of these processes, and UNDO its self, were significantly improved at 12.0, which gave us memory improvements, etc. Actually, add mid-side nodes still needs this fix :rolleyes:, but at least you can clear the undo buffer easily now.

When you were quitting and restarting 11.0, you were clearing your memory in a rough but effective way.

I am attaching an image of the memory improvements for Octree Tetra at 12.1 (due out later this month).

I am excited for 12.1 then!

v12.0 has been great, haven't gotten the segmentation violation error with it. However, I am getting a large memory spike on the "disconnecting orphan cells" step of mesh generation. This often crashes ICEM with no warning (i.e. the process just disappears). I submitted a help ticket to the techsupport who confirmed this spike, and they recommended using batch mode until v12.1 comes out to avoid this. This seems to be working, but I often forget to check the box ;)

Thanks for the reply,

Ryne

alinik December 20, 2013 18:06

I am getting this error in cfx-solve. Any idea how I can resolve this issue?
I have increased the memory allocation factor upto 10 and still getting the same error. I am running CFX remotely on a linux based supercomputer.

Thanks

vishwanath January 16, 2015 06:00

how to allocate memory for segemental violation error
 
hi every one,

I am finding error as segemental error. how to resolve it? ...
can you explain the steps to avoid it.:confused:


All times are GMT -4. The time now is 02:04.