VQ Compression Algorithms on A Multiprocessor System

A variety of parallel processing technologies have been implemented in a processor, and thus a cutting edge algorithm for multimedia applications should be aware of parallel processing features. We implemented parallel algorithms for VQ compression on two parallel environments and evaluated the effectiveness of the parallel algorithms. On a multiprocessor system with distributed memories, we evaluate two parallel algorithms for the codebook generation of the VQ compression: parallel LBG and aggressive PNN. We measured the speedups and elapsed times of both algorithms on a PC cluster system and find that both algorithms can achieve scalable parallelisms for the case with a large number of training vectors. On the other hand, for a codeword search on a system with a shared memory, the p-dist approach and the c-dist approach with the aggregation of synchronizations are suitable for a small codebook, and the c-dist approach and the p-dist approach with the ADM or the strip-mining method are suitable for a large codebook. However, since the aggregation of synchronizations and the strip-mining method increases the space complexity of the algorithm, the p-dist approach and the c-dist approach are more suitable for a small codebook and for a large codebook, respectively.The original publication is available at JAIST Press http://www.jaist.ac.jp/library/jaist-press/index.htmlIFSR 2005 : Proceedings of the First World Congress of the International Federation for Systems Research : The New Roles of Systems Sciences For a Knowledge-based Society : Nov. 14-17, 2089, Kobe, JapanSymposium 3, Session 3 : Intelligent Information Technology and Applications Networks and Agent