"forrtl: severe (174): SIGSEGV, segmentation fault occurred" when running wrf.exe with FNL datas

[wenli]

Dear Users,
I am running WRF Elec V4.5 using FNL input data.
While running wrf.exe I am getting the error: "forrtl: severe (174): SIGSEGV, segmentation fault occurred"
I have already tried the solutions showed in this url: Error while running wrf.exe forrtl: severe (174): SIGSEGV, segmentation fault occurred
Besides, I have also tried to reduce or increase the time-step and narrow down the region area.
But it still encounter same problem.
The namelist, rsl files are attached.
Thank you for your help!

 

[Ming Chen]

This case crashed immediately, indicating that it is either a data issue or a memory issue.
Since it is a triply-nested case, I am suspicious that you may not have enough memory to run this case.

Please clarify how you compile WRF, and how many processors did you use to run this case.

Also, please rerun this case with more processors and see whether it works.

 

[wenli]

This case crashed immediately, indicating that it is either a data issue or a memory issue.
Since it is a triply-nested case, I am suspicious that you may not have enough memory to run this case.

Please clarify how you compile WRF, and how many processors did you use to run this case.

Also, please rerun this case with more processors and see whether it works.

Dear Ming Chen, thank you for your response. I used one node and one processor to run this case. I also tried to use 4 nodes and 64 processors to run it, but it still encounter the same problem.
After trying more processors, I changed it to doubly-nested and reran, then it completed with no error.
Does this mean that I have enough memory to run a triply-nested case?
If I need to run a triply-nested case, is narrowing the model region a workable way?
Thank you for your help!

 

[Ming Chen]

Hi,
Thank you for the update.
If this case is able to run over 2 nested domains, it indicates that your input data is correct. I believe the segmentation fault you saw previously is caused by insufficient memory. In this case, you will need to enlarge your memory for running this case.
You may try with more number of processors, which will give you larger memory.
Or you can reduce the number of grids and make this a smaller case, then try again.