위상 결합을 기반으로 한 연결 망 설계 및 시뮬레이션

Design and Simulation of Interconnection Network Based on Topological Combination

  • 장창수 (여수대학교 컴퓨터공학과) ;
  • 최창훈 (상주대학교 소프트웨어공학과)
  • 발행 : 2004.06.01

초록

본 논문에서는 정적 네트워크 위상과 동적 위상을 결합한 새로운 부류의 MIN(Multistage Interconnection Network)인 Combine MIN을 제안한다. Combine MIN은 단일 경로 성질을 갖는 MIN보다도 적은 하드웨어 비용을 가지면서도 다중 경로를 제공한다. 또한 Combine MIN은 빈번한 통신을 갖는 프로세서-메모리에 짧은 경로의 지름길 경로 및 다중 경로를 제공함으로써 지역화된 통신에 적합하게 구성할 수 있게 설계되었다. 성능 평가를 위한 시뮬레이션 결과에 따르면 Combine MIN은 높은 지역화된 통신에서 같은 네트워크 크기를 갖는 기존의 MIN보다 우수한 성능을 보였다. 따라서 Combine MIN은 공유 메모리 다중 프로세서 시스템에서 지역화된 통신구조를 갖는 병렬 응용 분야에서 효율적으로 활용될 수 있을 것이다.

In this paper, we propose a new class of MIN(Multistage Interconnection Network) called Combine MIN which combines static network topology and apimic network topology. Combine U provides multiple paths at a hardware cost lower than that of MIN with unique path property. Combine MIN can be constructed suitable for localized communication by providing the shortcut path and multiple paths inside the processor-memory cluster which has frequent data communications. According to the results of analysis and simulation for performance evaluation, Combine MIN shows higher performance than MINs of the same network size in the highly localized communication Therefore, Combine MIN can be used as an attractive interconnection network for parallel applications with a localized communication pattern in shared-memory multiprocessor systems.

키워드

참고문헌

  1. Proc. Int'l Conf on Parallel Processing A Communication Model for Optimizing Hierarchical Multiprocessor System S.G.Abraham;E.S.Davidson
  2. IEEE, Compt. A Survey and Comparison of Fault Tolerant Multistage Interconnection Network G.B.Adams Ⅲ;D.P.Agrawal;H.J.Siegel
  3. IEEE Trans. Compt. v.C-37 Partitioning Techniques for Large-Grained Parallelism R.Agrawal;H.V.Jagadish
  4. IEEE Compt. Performance of Multiprocessor Interconnection Networks L.N.Bhuyan;D.P.Agrawal
  5. The Technology of Parallel Processing : Parallel Processing Architectures and VLSI hardware v.Ⅰ A.L.Decegama
  6. Cache and Interconnect Architectures in Multiprocessors M.Dubois;S.S.Thakkar
  7. Proc. 10th Symp., Computer Arch. Using Cache Memory to Reduce Processor-Memory Traffic J.R.Goodman
  8. The Journal of Supercomputing v.18 no.2 The Multi-Level Communication:Efficient Roting for Interconnection Networks H.B.Hadin
  9. Advanced Computer Architecture: Parallelism Scalability Programmability K.Hwang
  10. Computer Architecture and Parallel Processing K.Hwang;F.A.Briggs
  11. Advances in Compt. v.26 Multistage Interconnection Networks for Multiprocessor Systems S.C.Kothari
  12. ACM SIG COMM. D-ARM : A New Proposal for Multi-dimensional Interconnection Network L.L.Ling;A.Fiho
  13. Proc. of the Seventh Asia-Pacific Conference on Computer Architecture v.6 Metacube: A New Interconnection Network for Large Scale Parallel System Y.Li;S.Peng;W.Chu
  14. Shared Memory Multiprocessing N.Suzuki
  15. Computer Architecture A Quantitative Approach D.A.Patterson;J.L.Hennessy
  16. Fault-Tolerant Computer System Design D.K.Pradhan
  17. SES/workbench (Rel. 3.0)
  18. Interconnection Networks for Large-scale Parallel Processing H.J.Siegel
  19. Int'l Journal of Computer Applications in Technology v.13 CPMBIC: An improved Cluster-based interconnection network H.Wand