|
Recently, the paper "High-performance IPv6 Forwarding Algorithm for Multi-core and Multi-threaded Network Processors" done by the graduate Hu Xianghui of USTC Computer Department, his tutor associate professor Hua Bei and USTC schoolfellow Dr. Tang Xinan is issued on ACM SIGPLAN Symposium on Principles and Practice of Parallel Programming (PPoPP 2006), the international parallel process conference. Dr. Tang Xinan and Hu Xianghui announced the paper at the conference held on 31st, March. It is the first time that conference has accepted paper from China mainland since 1988.
ACM PPoPP is one of the top-leveled conferences in the computer system field. The topics of the conference are parallel programming principles, parallel programming tools and parallel programming applications. Given this international conference is the top one which is concerned about computing parallel processes, the papers acceptation competition is extraordinarily intense and full of challenges. From 1988 to 2005, except America, the academic papers from other countries and areas, which are issued on PPoPP, are less than 6, 2 papers from Taiwan of China. This paper representing the universities and institutions of China mainland breaks the zero. It indicates USTC has advanced to the international research pioneer line of parallelism process.
The paper is the fruit of labor from three years of study by Dr. Tang Xinan and Associate professor Hua Bei. Dr. Tang Xinan firstly introduced the multiprocessor and multithreaded parallel system, which represents the future development tendency of network processor. Then, the research team led by Dr. Tang Xinan, far-sightedly choose "Research on Network -processor-based Core Algorithms for High-performance Routers",which needs to deal with multi-subjects, including parallel process, network, systematic CMOS chip design, compiling principles, embedded system. After their three-year unremitting efforts, they finally gain the achievements on high-speed route search algorithm on IPv6 and parallel of network processors. The experiments demonstrate the TrieC algorithm which they put forward can support the liner speed above 10Gbps on Intel IXP2800 network processors. This outstanding achievement bases their entering the internal top conference. |