Memory Performance Optimizations For Real-Time Software HDTV Decoding

Pure software HDTV video decoding is still a challenging task on entry-level to mid-range desktop and notebook PCs, even with today’s microprocessors frequency measured in GHz. This paper shows that the performance bottleneck in a software MPEG-2 decoder has been shifted to memory operations, as microprocessor technologies including multimedia instruction extensions have been improving at a fast rate during the past years.Our study exploits concurrencies at macroblock level to alleviate the performance bottleneck in a software MPEG-2 decoder. First, the paper introduces an interleaved block-order data layout to improve CPU cache performance. Second, the paper describes an algorithm to explicitly prefetch macroblocks for motion compensation. Finally, the paper presents an algorithm to schedule interleaved decoding and output at macroblock level. Our implementation and experiments show that these methods can effectively hide the latency of memory and frame buffer. The optimizations improve the performance of a multimedia-instruction-optimized software MPEG-2 decoder by a factor of about two. On a PC with a 933 MHz Pentium III CPU, the decoder can decode and display 1280 × 720-resolution HDTV streams at over 62 frames per second.This work was done while the author was a Ph.D. candidate in the Computer ScienceHan Chen is a research staff member in IBM T.J. Watson Research Center. His research interests include distributed computing systems, scalable display system, and multimedia. He received his Ph.D. degree in 2003 and his M.A. degree in 1999 from Princeton University. He received his B.S. degree from Tsinghua University of Beijing, China in 1997.Kai Li is a Charles Fitzmorris professor at the Computer Science Department of Princeton University. His research interests include operating systems, computer architecture, distributed systems, and scalable display systems. He received his Ph.D. degree from Yale University in 1986. Prior to that, he received his M.S. degree from University of Science and Technology of China, Academy of Sciences of China in 1981 and a B.S. degree from Jilin University in China in 1977. He was a visiting faculty member at University of Toronto in 1988 and a visiting professor at Stanford University during his sabbaticals in 1996 and 2000. He has served on dozens of program committees and served as chair or vice chair several times. He has been elected as an ACM fellow in 1998.Bin Wei received a Ph.D. in Computer Science from Princeton University in 1998 and joined the research community at AT&T Shannon Laboratories since then. His research interests are in the areas of high-performance computer systems, multimedia, and service platforms for mobile users. He received a BS in Computer Science from Tianjin University, China in 1983 and an MS in Computer Science from the Institute of Computing Technology, Chinese Academy of Sciences, in 1989.