General Mechanical

General Mechanical

Topics related to Mechanical Enterprise, Motion, Additive Print and more.

direct solver in-core vs. out-of-core mode and RAM usage

    • Fabian Potschul
      Subscriber

      Hello,

      i wanted to run a a thermomechanical simulation of the lpbf-process which consits of transient thermal and a linked static-structural system. The model consists of 808797 nodes. To solve the static-structural system i wanted to use the direct solver but i received an error message:

      My problem is that i do not unterstand why the direct solver does not change to the out-of-core memory mode. Earlier in the solution output file the solver is set to run in the in-core memory mode:

      Why does the solver choose to run in the in-core mode but then later fails to solve, because there is not enough physical memory available?

      Another question of mine is regarding the numbers at then end of the solution output file:

      What is the meaning of the following numbers and why is the number for allocated memory larger than the number for physical memory available? Is "sum of memory used on all processes" the amount of RAM which was used?

      I would be very grateful for any help.

    • peteroznewman
      Subscriber

      Open Ansys Help, select Mechanical APDL, Performance Guide, and look at section 4.2 on Understanding Memory… you can read the definintion of all the terms in the solution output file and see this illustration:

      The solver can start off using in-core, and as the memory required increases as the solution progresses, the solver automatically switches to out-of-core.  Then as the solution continues, at some point even more memory is required but there is no more.  That is what the error message is telling you.

    • Fabian Potschul
      Subscriber

      Thank you for your reply. I will look it up.

      When the solver switches from in-core to out-of-core mode would there be a note about it in the solution output file (?), because in this case ther is no such note. That is why I am confused about the solver not having enough memory to proceed using the out-of-core mode. Additionally there should be enough free disk storage in my working directory for the direct solver to continue solving, even when using the out-of-core mode but I will definitley check that. Thank you and have a nice weekend!

Viewing 2 reply threads
  • The topic ‘direct solver in-core vs. out-of-core mode and RAM usage’ is closed to new replies.