CFD Online Discussion Forums

CFD Online Discussion Forums (https://www.cfd-online.com/Forums/)
-   CONVERGE (https://www.cfd-online.com/Forums/converge/)
-   -   UDF segmentation fault (https://www.cfd-online.com/Forums/converge/255233-udf-segmentation-fault.html)

dnwnlife0109 March 27, 2024 00:52

UDF segmentation fault
 
Hi.



I'm currently trying to use udf on convergeCFD.



I'd followed the instruction in udf manual but finally the program print segmentation fault (11). Could somebody explain this thing? There was no error or warning on building process.



Currently I'm suspecting the version of building tools

This is the versions of program that I'd used.



cmake 3.15.4

make 3.8.2 (minimum version)

gcc 8.3.0


The only thing I did different from the manual is that i'd invoked cmake and make to use GCC version 8.3.0. Which is 4.8.5 default on the cluster.



C udf file was flux_limiter.c from example folder. No code modification.



Thank you.

ksrivast March 29, 2024 14:54

Hello Samuel,

Were there any messages in the converge.log file when the case crashed, apart from the Signal 11 segmentation fault error?
Does the converge.log show the simulation crashed while loading the UDF libraries?
What version of CONVERGE solver are you using? And on which platform (Windows/Linux)?
Did you try to run one of our example cases with the UDFs?

Hopefully your answers to the above will help us narrow down the issue. We ran one of our example cases (Pipe CHT) using our latest v3.1.11 solver with the default flux_limiter.c compiled UDFs and the case seems to run fine.

Sincerely,

dnwnlife0109 April 1, 2024 05:45

Thank you for your response.



The test results was as follows.



1&2. . No nothing specific. Final message was just "cell_count in rank 123: is ~~~~"

3. Currently using 3.10 with linux cluster

4. Solver crashes with segmentation 9, 11. (On pipe-CHT case) No warning was in converge.log. I'd used user limiter only on momentum limiter.

ksrivast April 1, 2024 08:37

Hello Samuel,

Given that we have been unable to reproduce the issue on our end, we recommend you reach out to us at support@convergecfd.com. Please use your official email for all correspondence with Convergent Science. Please mention your issue, attach your case setup along with your compiled UDFs and add the cfd-online thread, as reference.

We will check whether there were any issues in your compiled UDFs and also forward you with compiled UDFs we created to see if the case runs fine.

Sincerely,


All times are GMT -4. The time now is 15:05.