Implementation of cache coherence protocol for COMA multiprocessor systems based on the scalable coherent interface

作者: M. AL-Rousan , S. Ahmed

DOI: 10.1016/J.CSI.2004.03.010

关键词: Distributed computingCommunications protocolMESIF protocolShared memoryCache coherenceScalable Coherent InterfaceMESI protocolParallel computingMemory architectureBus sniffingComputer science

摘要: Abstract The Scalable Coherent Interface (SCI) provides high bandwidth, low latency communication in systems with point-to-point links. SCI can also support cache coherence shared memory multiprocessors. Existing Cache-Only Memory Architecture (COMA) are all based on and protocols other than SCI. Hierarchical COMA generally suffer from message latencies. In this paper, we examine the implementation of multiprocessors using protocol. Implementing protocol hierarchical reduces latencies its will reduce directory sizes. work, a system was modified so that it fits addition, new replacement policy is proposed which replaced lines swap storage locations ones fetched into attraction memory. Similar to systems, still gives performance when part left unallocated. However, slightly influenced by initial distribution data across system.

参考文章(16)
Ross Evan Johnson, Extending the scalable coherent interface for large-scale shared-memory multiprocessors University of Wisconsin at Madison. ,(1993)
A. Grujic, M. Tomasevic, V. Milutinovic, A simulation study of hardware-oriented DSM approaches IEEE Parallel & Distributed Technology: Systems & Applications. ,vol. 4, pp. 74- 83 ,(1996) , 10.1109/88.481689
Tom Lovett, Russell Clapp, STiNG: A CC-NUMA Computer System for the Commercial Marketplace international symposium on computer architecture. ,vol. 24, pp. 308- 317 ,(1996) , 10.1145/232973.233006
M Al-Rousan, J.K Archibald, L Bearnson, Evaluating the impact of locality on the performance of large-scale SCI multiprocessors Performance Evaluation. ,vol. 46, pp. 275- 302 ,(2001) , 10.1016/S0166-5316(01)00055-4
Sangyeun Cho, Jinseok Kong, Gyungho Lee, Coherence and Replacement Protocol of DICE—A Bus-Based COMA Multiprocessor Journal of Parallel and Distributed Computing. ,vol. 57, pp. 14- 32 ,(1999) , 10.1006/JPDC.1998.1524
Robert W. Todd, Matthew C. Chidester, Alan D. George, Comparative performance analysis of directed flow control for real-time SCI Computer Networks. ,vol. 37, pp. 391- 406 ,(2001) , 10.1016/S1389-1286(01)00221-3
M.A Sarwar, A.D George, Simulative performance analysis of distributed switching fabrics for SCI-based systems Microprocessors and Microsystems. ,vol. 24, pp. 1- 11 ,(2000) , 10.1016/S0141-9331(00)00063-6
Fredrik Dahlgren, P Stenström, Mårten Björkman, Reducing the Read-miss Penalty for Flat COMA Protocols The Computer Journal. ,vol. 40, pp. 208- 219 ,(1997) , 10.1093/COMJNL/40.4.208
Damian M Gonzalez, Alan D George, Matthew C Chidester, Performance modeling and evaluation of topologies for low-latency SCI systems Microprocessors and Microsystems. ,vol. 25, pp. 343- 356 ,(2001) , 10.1016/S0141-9331(01)00128-4
Z. Radovic, E. Hagersten, Hierarchical backoff locks for nonuniform communication architectures high-performance computer architecture. pp. 241- 252 ,(2003) , 10.1109/HPCA.2003.1183542