摘要
文章从C64x+内核DSP的RAM存储器使用效率、Cache工作机制和缓存一致性3个方面出发,分析了数据处在不同级别的存储器对运算性能的影响,指出制约算法效率的因素,给出提高算法效率的方法。H.264的数据类型包括原始帧、重建帧、参考帧、内插数据、预测数据等,在不同级别的存储器上安排这些数据对运算速度有不同程度的影响,该文针对H.264视频编码算法的特点,优化配置数据储存器,并进行了实验验证,实验结果表明文中的优化配置方法能有效提高编码速度。
In view of the utilization efficiency of RAM with C64x＋DSP,the working mechanism and coherence of cache,the effect of memory with different levels of data on the computational efficiency is analyzed,the factors that restrict the algorithm efficiency are pointed out,and an improving method is proposed.H.264 video coding standard contains complicated data types,including original frames,reconstructed frames,reference frames,interpolated data and predictive,data and the data arranged to different levels of memory influence the computational speed in varying degrees.In this paper,the configuration of data memory is optimized according to the features of H.264 standard.The simulation results show that the proposed optimized configuration method can speed up the coding process effectively.
出处
《合肥工业大学学报（自然科学版）》
CAS
CSCD
北大核心
2011年第4期501-505,共5页
Journal of Hefei University of Technology：Natural Science
基金
教育部重点实验室开放课题基金资助项目(2008HGXJ0350)
合肥工业大学博士学位专项基金资助项目(GDBJ2009-067)
关键词
视频编码
H.264标准
数字信号处理
缓存优化
video coding
H.264 standard
digital signal processor（DSP）
cache optimization
