An algorithm for optimally exploiting spatial and temporal locality in upper memory levels

摘要

In this study, we present an extension of Belady's MIN algorithm that optimally and simultaneously exploits spatial and temporal locality. Thus, this algorithm provides a performance upper bound of upper memory levels. The purpose of this algorithm is to assess current memory optimizations and to evaluate the potential benefits of future optimizations. We formally prove the optimality of this new algorithm with respect to minimizing misses and we show experimentally that the algorithm produces nearly minimum memory traffic on the SPEC95 benchmarks.