近年來多媒體應用有越來越複雜的趨勢，如3D遊戲及影像壓縮…等，若將之實現在嵌入式系統卻一件不簡單的事，因為必須考慮到應用程式的複雜度、高效能的晶片設計及產品及時上市的種種問題，而這些問題在單處理器上似乎越來越難解決，因此高效能的內建網路通訊的晶片(Network-on-Chip)的研究與發展便油然而生，此NoC平台允許設計者能在系統晶片(System-on-Chip)上快速整合眾多的矽智財(SIP)，而我們亦稱此系統平台為多處理器系統單晶片(Multi-Processor SoC)，而此MPSoC平台能在同一時間即時提供大量的處理量。
在此論文中，一個分散式JPEG編碼器將應用在NoC平台上的多處理器系統，不同以往JPEG編碼器應用在單處理器上，我們將JPEG編碼器切割給數個區塊，再將各個區塊分別分派給NOC平台上的不同處理器去做平行處理，而且處理效能會比單處理器來得更好，尤其是處理越大的圖片，效果越顯著。而在此次的測驗中我們利用多處理系統上的4個處理器來做JPEG編碼的動作，最多處理效能可提升180%。

Recently, there is a trend towards complex multimedia applications, such as 3D games and video compression, implemented by embedded devices. The implementation of these applications on small and fixed devices is a very difficult task, due to the divergence and the complexity of embedded applications, the higher chip design cost, and the shorter time-to-market. To meet these requirements, single core (or single processor) systems are not capable of providing the required computational power for such complex applications. Thus a high performance platform-based design approach called Network-on-Chip (NoC) should be applied. The platform-based NoC paradigm allows the designers to fast integrate (and reuse) tens of Silicon Intellectual Property (SIP) cores in a single System-on-Chip (SoC), and a NoC-based Multi- Processor SoC (MPSoC) can provide the computational concurrency required to handle concurrent events in real- time.
In the thesis, a distributed JPEG encoder on a NoC-based MPSoC platform is presented. Unlike the traditional JPEG encoder designed with a single core, first this new JPEG encoder is partitioned into some parallel threads, which are then mapped to the processor cores of the scalable NoC-based MPSoC platform respectively. Therefore, it can be executed concurrently and distributedly, and the performance is better than that of the single-core design, especially as processing larger pictures. The results show that the encoding performance with 4 processor cores indeed increases up to 180 %.

[1] Davide Bertozzi and Luca Benini, “Xpipes : A Network- on-chip Architecture for Gigascale Systems-on-Chip” IEEE Circuits and Systems Magizine, Vol. 4, Issue 5, pp 18-31, 2004.
[2] G.M. Chiu, “The Odd-Even Turn Model for Adaptive Routing,” IEEE Trans. Parallel and Distributed Systems, vol.11, no.7, pp.729-737, 2000.
[3] Jingcao Hu and Radu Marculescu, “DyAD – smart routing for networks-on-chip”, Proc. of DAC., pp.260-263, June 2004.
[4] K. Goossens, J. Dielissen, and A. Radulescu, “ AEthereal Networks on Chip: Concepts, Architectures, and Implementations”, IEEE Design & Test of Computers, Vol. 22, No. 5, pp. 414-421, 2005.
[5] K. Lee, S.-J. Lee & H.-J. Yoo, “Low-Power Network-on-Chip for High- Performance SoC Design”, IEEE Trans. on VLSI, Vol.14, No.2, 2006.
[6] L. Benini and G. D. Micheli, “Networks on Chip: A New Paradigm for Systems on Chip Design,” Proc. of DATE, pp. 418-419, 2002.
[7] P. Lieverse, et al., “A Methodology for Architecture Exploration of Heterogeneous Signal Processing Systems”, Journal of VLSI Signal Processing, vol. 29, pp. 197–206, Kluwer Academic Publishers, 2001.
[8] R. Saleh, S. Wilton, et.al, “System-on-chip: Reuse and Integration”, Proc. of IEEE, Vol.94, No.6, pp:1050–1069, 2006.
[9] S.S. Bhattacharyya, P.K. Murthy, and E.A. Lee, Software Synthesis from Dataflow Graphs, Kluwer Academic Press, ISBN 0-7923- 9722-3, 1996.
[10] W. J. Dally and C. Seitz, “Deadlock-Free Message Routing in Multiprocessor Interconnection Networks,” IEEE Trans. on Comp, 1987.
[11] W.Dally "A VLSI Architecture for Concurrent Data Structures", Kluwer 1987
[12] Available：http://www.es.ele.tue.nl/~mininoc/
[13] OpenMP, Available：http://www.openmp.org/
[14] The MPI standard, Available：http://wwwunix.mcs.anl.gov/mpi/
[15] Open SystemC Iniative, OSCI, SystemC documentation. Available：http://www.systemc.org , 2004.
[16] 戴顯權, 「資料壓縮」. 出版商:紳藍. 出版日期: 2002.
[17] 戴顯權, 「多媒體通訊」. 出版商:紳藍. 出版日期: 2002.