What is the Effective Access Time of a Cache
Also know as the average time per access, EAT helps you figure out how performant a cache memory system is on average.
Parallel (overlapping) vs Sequence (non-overlapping)
When we access memory we have two options that could increase the EAT, depending on the hit ratio.
- Parallel access or overlapping access means we access both the cache and the main memory in parallel so that we potentially decrease the time it takes to access the memory if there is a cache miss.
- Sequential access or non-overlapping access occurs when we access the cache before we access the main memory. This is beneficial if we have an exceptionally high cache hit ratio because we would avoid unnecessarily accessing the main memory.
Calculating the Effective Access Time
$$ \text{EAT} = (\text{hit probability} \times\text{hit access time}) + (\text{miss probability}\times\text{miss access time}) $$