Is it possible to reduce the storage used during the optimization process?

[cfddinner]

Dear Professor,
Hi! I'm encountering an issue when running a steady-state case with 1.7 million meshes using a container deployed on a single node with 512GB memory. Around the 50th iteration, an error always occurs: "mpirun noticed that process rank 15 with PID 2123105 on node node547 exited on signal 9 (Killed)". I would like to ask if it's possible to optimize for more iterations by not storing intermediate optimization results.
Additionally, during my optimization process, inf_du never reaches 10e-5, but I'm quite satisfied with the optimization results. Can these be considered good optimization results?
Thank you very much for taking the time to answer my questions. Wish you a pleasant workday!

[friedenhe]

I am not sure why your case ended at the 50th iteration. But I think it should not be caused by saving too many intermediate opt results. If you have a decent amount of performance improvement and are happy with the optimized design, go ahead to use it. I believe there is no need to always converge the optimality to 1e-5.

[cfddinner]

Thank you so much for your reply and advice!