Overview
In an exclusive design, if data is in the L1 cache, it is removed from the L2 cache. This maximizes the total amount of unique data that can be stored in the hierarchy.
Pros and Cons
- Pros: More efficient use of total cache capacity.
- Cons: More complex to manage. When data is evicted from L1, it must be written back to L2 (often using a 'Victim Cache' approach).
Usage
Commonly used in AMD processors (e.g., Zen architecture).