HPC failure - Fluent
Hello,
I have a fluent case file with an overset grid. the grid is 5.6 million cells with an 8 million cell overset. I have previously ran high grid cases (20 million cells successfully with standard RAM issuance before) However to initialize and run this case I need 16Gb of RAM per processor as opposed to 4Gb before. I am currently running on 4 nodes with 20 PPN (320 Gb of RAM per node). This seems extremly high and means the simulations take over the HPC I am using. Is there ways I can reduce the RAM requirements of my simulation.
I don't have a stong HPC background so some of this is new to me. Any help would be appreciated.
Note: If I have lower memory usage I get this output:
*** glibc detected *** /opt/aci/sw/ansys/2019R3/v195/fluent/fluent19.5.0/lnamd64/3ddp_node/fluent_mpi.19.5.0: free(): corrupted unsorted chunks: 0x0000000005df0a00 ***
Thanks,
Pierce
Comments
I attached the entire output message file in this comment.
I am running: Steady, compressible (subsonic), SA. No mesh adaption as has been suggested in other posts.
Hello,
What is the operating system that is being used? Please check if the platform you are working on is supported : https://www.ansys.com/-/media/ansys/corporate/files/pdf/solutions/it-professionals/platform-support/platform-support-by-application-product-2020-r2.pdf?la=en-gb&hash=2589D6B982A5E5D86091E2005B5B57E10FA612F6
Hi Rahnkumar, the mesh failed due to the overset conditions, nothing to do with the OS, i've managed to run several cases without overset and with (in 2D) with no issues on the HPC clusters i'm using.