Abstract: One of key issues in applying computer-based case simulation system for training and test is its response time since the clients often download huge files and mixed files. This paper studies the scheme of implementating multiple threads in Rich Internet Applications（RIA） and presents the optimal approaches to execute multiple threads to download in parallel a mount of audio files, video files and other huge files on the multicore computers. The theoretical analysis and experimental results show that the multi-thread parallel downloading technique can speed up online download of the moduls for the computer-based case simulation system and improve remarkably the execution performance of the system.

Key words: multi-thread downloading, performance optimization, Rich Internet Applications（RIA）, multicore computers, thread-level parallelism

摘要： 在应用计算机模拟病例训练与考试系统的过程中，客户端时常需要在线下载许多大数据文件、音频和视频混合文件，系统响应速度是一个关键问题。研究了在RIA中实现多线程的技术方案，提出在多核计算机上有效实现多线程并行下载大数据文件、音频和视频混合文件的优化方法。算法分析与实验结果表明，提出的多线程并行下载技术能够加速计算机模拟病例系统模块的在线下载，显著优化了系统运行性能。

关键词: 多线程下载, 性能优化, 富互联网应用（RIA）, 多核计算机, 线程级并行性