Parallel Processing Using Data Localization for MPEG2 Encoding on OSCAR Chip Multiprocessor

Overview

Need for efficient processing of multimedia applications on PCs, mobile phones, games and so on have been increasing. Especially, low cost, low power consumption and high performance processors for multi-media applications have been expected. To satisfy the demands, chip multiprocessor architectures which allow giving scalability using multigrain parallelism are attracting much attention. However, to get performance of chip multiprocessor architectures, data locality optimization for target applications is also required. This paper describes a parallel processing scheme for MPEG2 encoding using data localization technique which improves execution efficiency by using global data locality optimization among different loops with coarse grain task parallel processing, and evaluates the performance of the proposal scheme on OSCAR chip multiprocessor architecture.