Basic Statistics Post Processor

[Barthle, Jonathan]

Good morning,

I have developed a RAVEN file that runs different cases of a RELAP5-3D model. I have a BasicStatistics post processing step that is supposed to calculate time dependent sensitivity coefficients throughout the simulation. I have gotten the whole process to work with a shortened test run; however, when I try the full simulation, my BasicStatistics step is failing. It seems like RAVEN is killing the run, but I do not see any error messages. The following message is displayed in the terminal output:

/home/bartjona/Bitterroot/raven/raven_framework: line 80: 510983 Killed                  $PYTHON_COMMAND -O $RAVEN_DIR/raven_framework.py "${ARGS[@]}"

My RELAP models are running to completion and the respective variables and FOMs are being output correctly. I was wondering what might be happening. Attached is a simplified version of my RAVEN file. Let me know if there are any additional questions.

Thank you

Jonathan

[Paul W. Talbot]

Hi Jonathan,

 

Are you running this in a high performance computing environment? It looks like you may be on Bitterroot. This is the kind of message we generally see when you hit wall time or run out of allocated memory, or something similar.

 

If you’re collecting a lot  of data from the RELAP runs, or using a lot of memory per run, you may need to increase your requested memory allocation.

 

Thanks,

 

Paul Talbot (pahl TAL’-bit), Ph.D.

Software Architect   |   Integrated Energy Systems Program

Group Lead   |   B842 Statistical Simulation and Optimization

paul....@inl.gov   |   208-526-5425

Idaho National Laboratory   |   Idaho Falls, ID   |   83415

--
You received this message because you are subscribed to the Google Groups "INL RAVEN Users Group" group.
To unsubscribe from this group and stop receiving emails from it, send an email to inl-raven-use...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/inl-raven-users/CAMsK15%2B2mWs%3DaOO7NVbksTQKg8jT2sBt95OSUb8NzFdM0_Fj-w%40mail.gmail.com.

[Barthle, Jonathan]

Yes, I have been running these initial cases on an interactive node on Bitterroot, which should have 2 GB per core. I created a separate RAVEN file that just ran the post processing steps after reading in the needed run data and it was being killed only after about 10 seconds of calculations. There is plenty of walltime left, so it could perhaps be a memory issue. Although, I would not have expected memory to be an issue since these output HistorySets are only 20 kB, and I only have 5 cases at the moment.

Thank you

Jonathan

[Paul W. Talbot]

You may be able to reach out to HPC folks to find out why the job is being killed. We unfortunately don’t have access to run details outside of what’s in the terminal printout.

 

---

Thanks!

 

Paul Talbot

[Barthle, Jonathan]

Sounds good, I'll try a few more things on my end and then reach out to HPC support if needed. Thank you for pointing me in the right direction.

Jonathan