Learning Forum

[dkmenend]

Hello,

I am running a transient case with mesh adaptation to capture a detonation wave. I am running simulations on a high power computing cluser and I am getting Out of Mem Errors. I am running 1.4 million cells on 48 cores. What is the potential causes of the OOM error? I have noticed that they tend to crash after running for ~24 hours. Thanks for your help.

The slurm out file says "slurmstepd: error: Detected 1 oom_kill event in StepId=34836945.batch. Some of the step tasks have been OOM Killed."

 

[Rob]

If you're repartitioning and saving data/images it's possible the solver hasn't freed RAM up as well as it might. I'm assuming you can resume from the (auto)saved case & data?