为提高并行绘制系统中的帧率，从数据压缩的角度出发，搭建了基于 InfiniBand的并行绘制系统。首先对并行绘制系统中数据进行分类，根据当前帧的数据量大小设置临界值，对达到临界值的几何数据进行无损压缩。本文研究并对比了 LZ77,LZW,LZO三种数据压缩算法对系统的影响。实验表明， LZ77,LZW和LZO算法均对系统的帧率有所提高，而 LZO算法对系统的帧率提高最明显且稳定性最好，使系统帧率提高近 60%。无损数据压缩对提高并行绘制系统性能具有很明显的效果。

From the perspective of data compression, parallel rendering system based on InfiniBand is set up in order to improve the frame rate. Data classification in parallel rendering system is performed; and the thresholds are set according to the data of the current frame; the lossless compression is applied to the geometric data which reach up to the critical value. The effects of three types of data compression algorithms, LZ77, Lempel-Ziv-Welch(LZW)，Lempel -Ziv-Oberhumer(LZO), are researched and compared. The experiments results show that，LZ77,LZW and LZO algorithms can improve the frame rate of system; LZO algorithm increases the frame rate of system most significantly with the best stability, and the system frame rate is increased by nearly 60%. Lossless data compression can improve the performance of parallel rendering system significantly.