Production of parallel processing of surface data

[weixinyu]

Dear all
Because it may be that I don't have enough memory on my computer, causing me to use this command. /mkmapdata.sh -r $GRIDNAME -f $GRIDFILE -t regional When I make a map of the surface data, it breaks at the last data, and I looked up related posts on the internet, and it said that parallel compilation can solve this problem.
So I tried parallel compilation, but I don't understand how parallel compilation should work, so I just modified the input command with
mpirun -np 10 . /mkmapdata.sh -r $GRIDNAME -f $GRIDFILE -t regional
I also modified the mkmapdata.sh file to include active=“NO”
It shows this problem, is it because I still don't have enough RAM? Am I using 50g of RAM and then 10 in parallel, or do I need to manipulate the rest that
Can you tell me exactly what I should do if I need to parallelize the processing when making surface data, thank you very much, looking forward to a reply!

 

[slevis]

If you have the option of running a newer version (e.g. ctsm5.2.0), then you can use the new mksurfdata_esmf tool that is designed to run parallel.

Otherwise, I will let others chime in with suggestions.

 

[weixinyu]

If you have the option of running a newer version (e.g. ctsm5.2.0), then you can use the new mksurfdata_esmf tool that is designed to run parallel.

Otherwise, I will let others chime in with suggestions.

Dear slevis
I'm afraid I'm stuck with this version of ctsm, can I trouble you for some viable suggestions for using this version?

 

[slevis]

I will move this thread to "porting" in case you are able to find help there.

 

[jedwards]

Try running mkmapdata.sh with the -verbose flag and post more of the output log. Check your ulimits are adequate, talk to your system support staff if you have one.

 

[weixinyu]

尝试使用 -verbose 标志运行 mkmapdata.sh 并发布更多输出日志。检查您的 ulimits 是否足够，如果有，请咨询您的系统支持人员。

非常感谢您的回复，我现在已经解决了这个问题。
我设法制作了表面数据，但是当我检查日志文件时，它在某些数组的顶部显示了所有 0，我犯了错误吗？还是它就是这样。

 

[jedwards]

Google translate says: Thank you very much for your reply, I have solved the problem now.I managed to make the surface data, but when I checked the log file it showed all 0s at the top of some arrays, did I make a mistake? Or is it just like that.

@slevis it's back in your court.

 

[slevis]

Thank you @jedwards
@weixinyu you are the best person to judge whether the results in the log file make sense to you. You may want to also visualize the contents of the fsurdat file. This may help you identify problems that you suspect with the file.