片上网络通信性能分析建模与缓存分配优化算法

王坚; 李玉柏; 蒋勇男

doi:10.3724/SP.J.1146.2008.00491

片上网络通信性能分析建模与缓存分配优化算法

doi: 10.3724/SP.J.1146.2008.00491

基金项目:

国家自然科学基金(60575031)资助课题

计量
- 文章访问数: 3103
- HTML全文浏览量: 99
- PDF下载量: 938
- 被引次数: 0
出版历程
- 收稿日期: 2008-04-24
- 修回日期: 2008-09-29
- 刊出日期: 2009-05-19

Communication Performance Analytical Model and Buffer Allocation Optimizing Algorithm for Network-on-Chip

摘要

摘要: 该文建立了一种面向应用设计的片上网络的性能分析模型，并在此基础上提出了片上缓存优化策略和分配算法。在硬件实现平台上的仿真表明，该文建立的片上网络分析模型能很好地分析片上网络通信时延和路由节点各方向的阻塞概率，以此进行片上网络的缓存资源优化，能在同等缓存资源的情况下降低数据通过网络的平均时延，使片上网络的性能得到改善。
- 片上网络;分析模型;缓存优化;分配算法;性能优化
Abstract: For application aimed NoC design, this paper proposes an analytical model of communication performance and designs a buffer optimizing strategy and allocation algorithm. The hardware simulation results show that the model can analyze the average delay of NoC and the blocking probability of each port of routers, and the algorithm can reduce the average delay of NoC using the same amount of resources, which improves the NoC performance.

HTML全文

参考文献(1)

Benini L and De Micheli G. Networks on chips: A new SoCparadigm [J]. Computer, 2002, 35(1): 70-78.[2]Murali S, Meloni P, and Angiolini F, et al.. Designingapplication-specific networks on chips with floorplaninformation [C]. ICCAD '06. IEEE/ACM InternationalConference on Computer-Aided Design, San Jose, CA, USA,Nov. 2006: 355-362.[3]Jeang Yuan-Long, Hung Chung-Wei, and ChiangChuen-Muh. A methodology based on maximal-profitspanning tree for designing application specific networks onchip [C][J].ICICIC '06. First International Conference onInnovative Computing, Information and Control, Beijing,Aug.2006, 2:18-21[4]Andreas H, Maarten W, and Arno M, et al.. Applyingdataflow analysis to dimension buffers for guaranteedperformance in networks on chip [C]. NoCS 2008. SecondACM/IEEE International Symposium on Networks-on-Chip,Newcastle University, UK, 7-10 April 2008: 211-212.[5]Hu J and Marculescu R. DyADsmart routing for networkson-chip [C]. Proc. DAC, San Diego, Jun 2004: 260-263.[6]Saastamoinen I and Alho J N M. Buffer implementation forproteo networks-on-chip [C]. Proc Int Symp Circuits andSyst, 2003: 113-116.[7]Dally W J and Towles B. Route packets, not wires: On-chipinterconnection networks [C]. Proceedings of DAC, LasVegas, Nevada, USA, June 18-22, 2001: 683-689.[8]Guerrier P and Greiner A. A generic architecture for on chippacket-switched interconnections [C]. Proceedings ofDATE,Paris, France, March 27-30, 2000: 250-256.[9]Wu Ning, Ge Fen, and Wang Qi. Simulation andperformance analysis of network on chip architectures usingOPNET [C]. ASICON '07. 7th International Conference onASIC, Guilin, China, 22-25 Oct 2007: 1285-1288.[10]Adve V S and Vernon M K. Performance analysis of meshinterconnection networks with deterministic routing [J].IEEE Trans. on Parallel Distrib. Syst.1994, 5(3):225-246[11]Dally W J. Performance analysis of k-ary n-cubeinterconnection networks [J]. Computer, 1990, 39(6):775-785.[12]Kermani P and Kleinrock L. Virtual cut-through: A newcomputer communication switching technique [J]. ComputerNetworks, 1979, 3: 267-289.[13]Kleinrock L. Queueing Systems [M]. Vol.I: Theroy. WileyInterscience, New York, 1975: 17-85.[14]Hu Jingcao, Ogras U Y, and Marculescu R. Applicationspecificbuffer space allocation for networks-on-chip routerdesign [J].IEEE Trans. on Computer-Aided Design ofIntegrated Circuits and Systems.2006, 25(12):2919-2933

施引文献

资源附件(0)

访问统计