arXiv.itsnick.co

an independent field guide to 23 million papers from arXiv & OpenAlex — not affiliated with arXiv.org

Computer Science · Hardware and Architecture

Victim-Caching for Large Caches and Modern Workloads

Hal Wasserman

See on the map

Abstract

Jouppi [4] has proposed that direct-mapped caches should be assisted by victim caches: small, fully-associative caches which reduce conflict misses by saving the blocks most recently thrown out of the main cache. Jouppi found that victim-caching worked well for 4KB data caches. Here, we continue this work, studying victim-caching for modern workloads (SPEC95) and larger caches (32--64KB). We find that, if victim-cache size is allowed to scale up to 1 4 of main-cache size, victim-caching continues to work well at 32KB, but shows signs of breaking down for larger cache sizes. We conclude that victim-caching is a credible way of reducing miss-rates for modern, medium-sized caches, and that it may be particularly useful for reducing chip-area costs. 1 Introduction The cache misses of a fully-associative cache may be itemized as compulsory misses (the first reference to each memory block) or capacity misses (misses due to a block's having been thrown out of the cache when it overfilled)...

More like this