摘要
为了进一步提高JPEG编码效率,对JPEG压缩算法进行研究,分析得出JPEG核心步骤可以并行化处理。因此,实现平台宜采用以并行计算为优势的GPU,而不是以串行计算为主的CPU。NVIDIA新推出的CUDA(计算统一设备架构)为此实现提供了软硬件环境。CUDA是基于GPU进行通用计算的开发平台,非常适合大规模的并行数据计算。在GPU流处理器架构下用CUDA技术实现编码并行化,并针对流处理器架构特点进行内存读写等方面的优化,提高了JPEG编码的速度。实验结果表明了CUDA技术在并行处理方面的优越性,JPEG编码效率得到了极大提高。
To further improve the coding efficiency of JPEG,the JPEG compression algorithm is researched. And the core steps in JPEG can be drawn parallel processing,which should be implemented on GPU,which has an advantage in parallel computing,but not on CPU,which has an advantage in sequential computing. CUDA （compute unified device architecture） launched by NVIDIA provides a hardware and software environment to accomplish the aim. CUDA is based on the GPU for general-purpose computing development platform and suitable for large-scale parallel data computing. Under the stream processor architecture on GPU,Using CUDA to implement the parallel processing of the JPEG compression algorithm,And according to the features of stream processor architecture,doing opti-mization on aspects of reading and writing memory,to improve the speed. The result of experiments indicates that the CUDA technology has the advantages of parallel processing and the efficiency of JPEG encoding is improved greatly.
出处
《计算机工程与设计》
CSCD
北大核心
2010年第14期3302-3304,3308,共4页
Computer Engineering and Design
基金
上海市教委创新基金项目(08ZZ76)
