基于视频场景的带宽感知Cache模型和划分策略-计算机系统结构专业论文.docxVIP

华中 科技 大学 华 中 科 技 大 学 硕 士 学 位 论 文 ———————————————————————————————————————————————— ———————————————————————————————————————————————— II Abstract Multimedia stream computing, such as video encoding and decoding, is not only with high computational complexity, but also with intensive data access in computing process. Multi-level Cache architecture could diminish the speed-gap between processor and memory. But the memory space of on-chip Cache is very limited usually, and the energy consumption is high relatively. So, allocating the Cache memory according to the run-time requirement features of multimedia stream computing is necessary. It could make full use of Cache, ensuring the performance and decreasing energy consumption of system. Cache modeling and partitioning strategy are investigated based on video scenarios for video decoding, which is a typical multimedia stream computing application. The bandwidth-aware Cache modeling and partitioning strategy are proposed based on video scenarios in this paper. It is designed in the premise of guaranteeing the soft real-time of video decoding, combing two features of video decoding. The one is the data volume should be accessed during video decoding is different while video scenario is changed. The other one is the energy consumption is different when decoding under the configuration of different Cache size. Firstly, after analyzing the variation characteristics of performance and energy consumption when decoding different video sequences under the configuration of different Cache size, the Cache model which makes decoding meet the delay requirements and reduces the energy consumption is established. Then, the best Cache sizes of multigroup video sequences are calculated through the Cache model. These Cache sizes are divided into several groups through a simple clustering algorithm and the characteristics of the corresponding video scenarios are analyzed in each Cache group. To perceive video scenarios switch from the compl