Next: Synchronous Vs. Asynchronous Prefetching
Up: Sequential Prefetching
Previous: Sequential Prefetching
The cost of caching an entity is equal to the size of the entity
multiplied by the amount of time for which it is present in the cache.
The benefit of caching an entity, on the other hand, is the number of hits it
produces.
We define two self-evident rules that any prefetching algorithm should follow:
- Avoid Wastage Rule: Do not prefetch any page that will be evicted before it is requested by the workload.
- Avoid Cache Pollution Rule: Do not prefetch any page that will evict other
prefetched pages without providing any net gain in performance.
root
2006-12-19