I am looking to model cache for multicore processors, including cache coherence. Do such PROMELA implementations already exist. I tried to search for it, but couldn't find any. Secondly, if I have to implement it myself, is it feasible in PROMELA to declare very large arrays as in to represent cache structures?
I personally don't know of such existing Promela models. Moreover, large array structures sounds like a serious state blow-up.
Depending on what properties you want to show, I would suggest to abstract from reality as much as possible. Modeling things with a high precision compared to the real world is typically nothing one should do in Promela.
Two alternative suggestions: