基于DCT并行加速算法图像渲染平台系统设计

Accelerating Application of Rendering Platform Based on DCT Parallel Algorithm

摘要:

此次主要研究了基于GPU的集群渲染系统平台设计;为了提高平台的工作效率、增强集群渲染系统平台的数据传输能力,提出了一种采用DCT变换的方法来加速图像渲染速度;该方法利用DCT变换算法加速图像的实时压缩,加入CPU监控器和任务分配器模块,让GPU和CPU共同承担了绘图和渲染的目的,这样有效地降低处理流程对CPU的占用,实现了三维绘图和特效渲染的加速;为了验证平台的有效性以及图像压缩处理的效果,做了相应的功能验证;对640×480的RCB图像使用上述压缩方法和JPEG标准库在不同压缩设置下进行实验;仿真实验结果表明所提方案具有更高的压缩效率。

Abstract:

This paper mainly researches the design of cluster rendering system platform based on GPU. In order to improve work efficiency, enhance the platform data transmission capability of cluster rendering system platform, this paper presents a DCT transform to accelerate the rendering speed. The method using DCT transform algorithm to accelerate the image real-time compression, joined the CPU monitor and the task distributor module, GPU and CPU shared the drawing and the purpose of rendering, so reduce the occupancy of the CPU treatment process, the accelerated 3D graphics and special effects rendering. Finally, with the JPEG Library in different compression settings were compared, the simulation results show that the proposed scheme has higher compression efficiency.