一种新颖的软件可控Cache优化方法

A Novel Method on Software-Controlled Cache Optimization

由于Cache污染问题,传统的仅由硬件控制的Cache替换策略不能得到令人满意的Cache利用率。随着软件可控Cache机制的出现,编译器开始可以直接控制Cache替换,改善Cache行为。本文证明了一个Cache提示优化定理,并依该定理提出了一个由编译器辅助控制的Cache替换策略:最优Cache划分(OCP)。OCPCache替换策略简化了Cache行为和Cache失效分析方法。实验结果表明OCPCache替换策略能有效地降低Cache失效率。

Caches have an increasing impact on overall performance because of the growing gap between CPU cycle times and memory access times.Therefore,improving the cache behavior is essential to obtain good execution speeds.With the advent of software-controlled cache,such as cache hints in EPIC architectures,the compiler has the means to steer the cache replacement directly and to improve the cache behavior.Cache hints are used to specify the cache level at which the data is stored after accessing it.This paper presents and proves a Cache Hints Optimization theorem that results in many interesting conclusions about cache hints optimization,Based on this theorem,this paper presents a compiler-assisted cache replacement policy,Optimum Cache Partition(OCP),which can be carried out through cache hints and LRU replacement policy.OCP replacement policy has been implemented in our compiler test-bed and evaluated on a set of scientific computing benchmarks.Initial results show that our approach is effective on reducing the cache miss rate.