Homeless and Home-based Lazy Release Consistency Protocols on Distributed Shared Memory

Yu, B.-H., Cranefield, S., Huang, Z. and Purvis, M.

    This paper describes the comparison between homeless and home-based Lazy Release Consistency (LRC) protocols which are used to implement Distributed Shared Memory (DSM) in cluster computing. We present a performance evaluation of parallel applications running on homeless and home-based LRC protocols. We compared the performance between Tread- Marks, which uses homeless LRC protocol, and our home-based DSM system. We found that the home-based DSM system has shown better scalability than TreadMarks in parallel applications we tested. This poor scalability in the homeless protocol is caused by a hot spot and garbage collection, but we have shown that these factors do not affect the scalability of the home-based protocol.
Cite as: Yu, B.-H., Cranefield, S., Huang, Z. and Purvis, M. (2004). Homeless and Home-based Lazy Release Consistency Protocols on Distributed Shared Memory. In Proc. Twenty-Seventh Australasian Computer Science Conference (ACSC2004), Dunedin, New Zealand. CRPIT, 26. Estivill-Castro, V., Ed. ACS. 117-123.
pdf (from crpit.com) pdf (local if available) BibTeX EndNote GS