When I run my program on Setonix, it occasionally gets killed, although sometimes it completes successfully. The logs indicate that the termination is due to an out-of-memory error. Could you provide information on the memory limitations for our program? The memory usage is primarily for the two matrices, totaling approximately 20 gigabytes—10 gigabytes for each matrix. I am considering alternatives such as splitting the matrices into smaller pieces or avoiding the creation of the original large matrices. However, I'm uncertain if these solutions are appropriate. My primary concern is understanding why my program is killed sporadically, as it only happens intermittently. Additionally, I have a question regarding execution. Do I need to use the lab script to run my program, or is it sufficient to compile it with gcc and execute it directly using ./program? In my view, both methods are acceptable since they both utilize OpenMP. Nevertheless, I am seeking clarification on this matter.

You can always set the memory limit in the parameters, like `--mem=60g`.