I am using parallel simulation for simulating a model with a model reference hierarchy. I precompiled my models and saved all generated cache files in a cache folder. I assumed that in the sequence of my parallel simulation, these models would not need to be recompiled again. However, it seemed not to be this case. In every run, the referenced models seemed to be recompiled as it took way more time than not using model referencing.
My question is, how are the models updated in every run in the sequence of parallel simulation?
Can I avoid updating the model in every run to save time?