| 引用本文: | 马安国,成玉,唐遇星,等.GPU异构系统中的存储层次和负载均衡策略研究.[J].国防科技大学学报,2009,31(5):38-43.[点击复制] |
| MA Anguo,CHENG Yu,TANG Yuxing,et al.Research on Memory Hierarchy and Load Balance Strategy in Heterogeneous System Based on GPU[J].Journal of National University of Defense Technology,2009,31(5):38-43[点击复制] |
|
| |
|
|
| 本文已被：浏览 7685次 下载 6452次 |
| GPU异构系统中的存储层次和负载均衡策略研究 |
| 马安国, 成玉, 唐遇星, 邢座程 |
|
(国防科技大学 计算机学院，湖南 长沙 410073)
|
| 摘要: |
| GPU体系结构的革新和相应开发平台的发展使得GPU广泛地应用于科学计算领域。通过深入地分析GPU体系结构和存储层次的优缺点以及GPU上的关键性能特征，阐明了GPU体系结构、编程模型和存储层次之间的关系。针对GPU异构系统上的应用映射提出三种基本负载均衡优化策略：预取、流化、任务划分。试验结果揭示了不同的优化因子与优化效率之间的具体关联。 |
| 关键词: GPGPU 存储层次 负载均衡策略 流计算 任务划分 |
| DOI： |
| 投稿日期：2009-07-03 |
| 基金项目:国家自然科学基金资助项目(60873016)；国家863计划资助项目(2009AA01Z102)；教育部“高性能微处理器技术”创新团队资助项目(IRT0614) |
|
| Research on Memory Hierarchy and Load Balance Strategy in Heterogeneous System Based on GPU |
| MA Anguo, CHENG Yu, TANG Yuxing, XING Zuocheng |
|
(College of Computer, National Univ. of Defense Technology， Changsha 410073，China)
|
| Abstract: |
| Owing to the revolution of GPU architecture and improvement of developing platforms, GPU is widely used in scientific computing nowadays. Relationships among GPU architecture, programming model and memory hierarchy are illustrated by analyzing memory hierarchy and exploring key performance features of GPU. Three basic load balance strategies on mapping applications onto GPU are presented: Prefetch, stream computing, task division. The effective relationships among different factors and optimization efficiency are tested and exposed by experiments. |
| Keywords: GPGPU memory hierarchy load balance strategy stream computing task division |
|
|
