Engineering Efficient Paging Algorithms

Abstract

In the field of online algorithms paging is a well studied problem. LRU is a simple paging algorithm which incurs few cache misses and supports efficient implementations. Algorithms outperforming LRU in terms of cache misses exist, but are in general more complex and thus not automatically better, since their increased runtime might annihilate the gains in cache misses. In this paper we focus on efficient implementations for the OnOPT class described in [13], particularly on an algorithm in this class, denoted RDM, that was shown to typically incur fewer misses than LRU. We provide experimental evidence on a wide range of cache traces showing that our implementation of RDM is competitive to LRU with respect to runtime. In a scenario incurring realistic time penalties for cache misses, we show that our implementation consistently outperforms LRU, even if the runtime of LRU is set to zero.

Partially supported by the DFG grants ME 3250/1-3 and MO 2057/1-1, and by MADALGO (Center for Massive Data Algorithmics, a Center of the Danish National Research Foundation). A full version of this paper containing all experimental results is available online at www.ae.cs.uni-frankfurt.de/sea12/ .