Why two processes took extreme high RAM when running MFiX in DMP?

I am running a DEM case with 8 threads DMP. This probelm always harras me. Some threads seem cannot release their RAM correctly, and resulting in higher and higher RAM being taken up. Finally the program will crash, or the system stuck.

What is more weird is that two process showed “0.022t” in the column where it should be the RAM in kb, and their %MEM are abnormally much higher than other threads…

How many particles are you trying to simulate?

If there is lots of ram usage on two of the process, then I suspect all of your particles are sitting on those two processes. You might want to think about decomposing your domain differently.