1. Introduction
Ultra High Definition Television (UHDTV) has been popularized in the recent period. by supporting up to 4K or 8K high resolution, UHDTV can provide better visual experience compared to previous Full HD specification. However, higher resolution means to higher data throughput requirement, which leads to several challenges on real-time design. Firstly, massive data volume means that increasing bit rate is required to compress the huge video content. In order to deal with compression problem, new High Efficiency Video Coding has been standardized by JCT-VC in January 2013. As the successor to H.264, HEVC can achieve 50% video compression ratio while guarantee the equivalent compression quality [1]. Secondly, UHDTV increases the design complexity as more hardware resources are needed to ensure higher throughput. Meanwhile, HEVC also adds extra burden on complexity for hardware implementation because of new coding tools introduced for better compression performance, such as flexible hierarchy structure for coding unit (CU) and so on [2]. Thirdly, memory traffic problem becomes more serious for UHDTV. The demand for off-chip memory bandwidth is proportionally increased to the video pixel rate. Bandwidth requirement may exceed more than 10GB/s for UHDTV. In total, new VLSI hardware architecture is required for HEVC-based UHDTV decoder.