Tile Based Wavefront Parallelism in HEVC

2020 15th International Workshop on Semantic and Social Media Adaptation and Personalization (SMA(2020)

引用 2|浏览39
暂无评分
摘要
High Efficiency Video Coding (HEVC) was developed to tackle the 4K resolution era, achieving up to 50% more compression efficiency (for the same video quality) compared to its predecessor H.264/AVC. However, this performance improvement comes at the expense of increased coding time. In order to ameliorate the situation, HEVC provides three coarse grained parallelization primitives namely: slices, tiles and wave front. Due to the higher overhead incurred by slice parallelism, most related studies in the literature focused on exploring tile and wave front parallelism as standalone mechanisms, or in combination with other finer grained approaches such as SIMD instructions. Nevertheless, the possibility of combining wave front and tile parallelism in order to speed up the encoding process was typically overlooked. In this paper we propose and evaluate a multi-threaded approach for coarse grained parallel video encoding, that is based on tile partitioning and per tile wave front parallelism. Through experimental evaluation with common test sequences, the proposed Wave Front per Tile Parallelism (WTP) is shown to provide a valid trade-off between speedup and video coding efficiency, effectively combining the best of two worlds, i.e., tile and wave front parallelism.
更多
查看译文
关键词
video coding,tile parallelism,wave front parallelism,HEVC,WTP
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络
Chat Paper
正在生成论文摘要