In hierarchical access when miss in L1 cache , go to L2 cache (means higher level cache) and if not getting in L2, go to L3 , and if not found then access main memory.
In hierarchical access, Tavg= H1*T1+ (1- H1) * (T1+T2) by considering L1 access time T1 and L2 access time is T2.
Simultaneous access is like a parallel access. Means when accessing from L2, no need the access time of L1 because we can access L1,L2 at the same time, and when accessing the main memory ,then no need to access time to other cache block because at the same time cache and memory can be accessed. So, simultaneous access is parallel .
Tavg= H1* T1+ H2* T2 , in case of simultaneous access.
Why hierarchical is default access ?
Because it avoids unnecessary memory requests whenever there is a hit in cache and this usually results in better overall performance though in problems, we get better answer for simultaneous cache ( means less time in simultaneous access ).
A good read :
https://gateoverflow.in/108780/simultaneous-vs-hierarchical-memory
