| 摘要: |
| 以异构系统的过程间相关性分析为基础,研究分析异构系统硬件故障在软件之中的传播行为,指导优化基于异构系统的应用级checkpointing 检查点保存问题,并通过实验验证其可行性及性能,对异构系统的容错优化研究具有重大意义. |
| 关键词: 通用GPU 异构系统 过程间相关性分析 传播行为 容错优化 |
| DOI：10.3724/SP.J.1001.2012.04057 |
| 分类号: |
| 基金项目:国家自然科学基金(60921062, 61003087) |
|
| Propagation Behavior Analysis and Fault Tolerance Optimization of Hardware Fault in Heterogeneous Systems |
|
JIA Jia, YANG Xue-Jun
|
|
National Key Laboratory for Parallel and Distributed Processing, College of Computer, National University of Defense Technology, Changsha 410073, China
|
| Abstract: |
| Based on the inter-procedural dependence analysis, this paper studies the propagation behavior in software of hardware fault in heterogeneous systems. This research can be used for optimizing application-level checkpointing techniques. Experimental results demonstrate that this method is viable and can be very helpful for the research of fault tolerance optimization of heterogeneous systems. |
| Key words: general purpose GPU heterogeneous system inter-procedural dependence analysis propagation behavior fault tolerance optimization |
