On Design and Implementation of Adaptive Data Classification Scheme for DSM Systems

Abstract

Distributed Shared Memory (DSM) environment is built by using specific softwares, to combine a number of computer hardware resources into one computing environment. Such environment not only provides an easy way to execute parallel applications, but also combines resources to speedup execution of these applications. DSM systems need to maintain data consistency in memory, what usually leads to communication overhead. Therefore, there exist a number of strategies that can be used to overcome this overhead and improve overall performance. Prefetching strategies have been proven to show great performance in DSM systems, since they can reduce data access communication latencies from remote nodes. However, these strategies also transfer unnecessary prefetching pages to remote nodes. In this research paper, we focus on the analysis of data access pattern during execution of parallel applications. We propose an Adaptive Data Classification scheme to improve prefetching strategy, with the goal to improve overall performance. Adaptive Data Classification scheme classifies data according to the access behavior of pages, so that home node uses past history access patterns of remote nodes to decide whether it needs to transfer related pages to remote nodes. From experimental results, our method can improve the performance of prefetching strategies in DSM systems.