Learning Forum

[H.Shirzad2]

I have a 12900 CPU with 16 cores and 24 threads, but I'm having trouble finding the best way to allocate tasks to the efficient cores for my simulations. I feel like I'm not using the full potential of my i9 processor, and the simulations are running slower than I expected. Do you have any suggestions to improve this? Maybe an article or link? I’m asking because, on two occasions, during parametric analysis and under HPC and analysis options, the lecturer in " getting to Know Maxwell" adjusted the core/task combinations, and it seemed to help.

[Timos]

According to Ansys documentation, you should consider the total number of compute jobs that will be run simultaneously on your machine. Each solver or distribution is considered one task. The total cores specify the number of cores that will be used for the given machine, which is how you specify multiprocessing. For instance, if you want to run two threads for each task, you would specify Total Cores as 2 times the Number of Tasks. It's also important to note that the number of cores must always be greater than or equal to the number of tasks. If the number of cores is not an exact multiple of the number of tasks, some tasks will use more cores than others.

For further information please visit Help documentation:

Maxwell Help > Running Simulations > Distributed Analysis > Editing Distributed Machine Configurations