forrtl: error (73): floating divide by zero in a coupled model

[yaozhixiong@foxmail_com]

Hi, I am running coupled models in CESM1.2.0. I run a coupled model with composet of BC5 and resolution of f19g16. No setting in the model are changed. I submitted the job after build successfully. But the job could not work with an error message "process killed (SIGTERM)" . Then I set  "DEBUG" to "TRUE". Now the error message turned to be "forrtl: error (73): floating divide by zero[/b]".  But, I run the same case successfully some months ago. I also check cases with other resolution, such as with composet BC5 but with resolution of f09g16, it can work. I feel so confused. I think it maybe something wrong with the inputdada. Some log files are attached.Thanks so much. Yao   

 

[jedwards]

What you describe is typical of a system change - was the compiler updated since the last time you ran?   

 

[jedwards]

What you describe is typical of a system change - was the compiler updated since the last time you ran?   

 

[yaozhixiong@foxmail_com]

Thank you for your prompt reply. I have checked that with the cluster administrator. He told me that none is updated.  Four cases are run for comparison. Still nothing are changed   composet   resolution1)   B            f19g162)   B            f09g163)   BC5        f19g164)   BC5        f09g16 All of them run successfully except the third one.   

 

[yaozhixiong@foxmail_com]

Thank you for your prompt reply. I have checked that with the cluster administrator. He told me that none is updated.  Four cases are run for comparison. Still nothing are changed   composet   resolution1)   B            f19g162)   B            f09g163)   BC5        f19g164)   BC5        f09g16 All of them run successfully except the third one.   

 

[yaozhixiong@foxmail_com]

I do not know which inputdata is wrong, so I reupload all the inputdata. And then all cases run successfully.  Maybe some data was modified but I did not know. 

 

[yaozhixiong@foxmail_com]

I do not know which inputdata is wrong, so I reupload all the inputdata. And then all cases run successfully.  Maybe some data was modified but I did not know.