| 引用本文: | 陈莉丽,姚益平,蔡潞.多核环境下负载均衡的并行离散事件全局调度机制.[J].国防科技大学学报,2012,34(4):108-113.[点击复制] |
| CHEN Lili,YAO Yiping,CAI Lu.A global schedule mechanism for PDES on multi-core environments[J].Journal of National University of Defense Technology,2012,34(4):108-113[点击复制] |
|
| |
|
|
| 本文已被：浏览 7996次 下载 6521次 |
| 多核环境下负载均衡的并行离散事件全局调度机制 |
| 陈莉丽1,2, 姚益平3, 蔡潞4 |
|
(1.北京系统工程研究所 复杂系统仿真总体重点实验室,北京 100101;2.2.国防科技大学 计算机学院，湖南 长沙 410073;3.2.国防科技大学 计算机学院，湖南 长沙 410073;4.3.长安大学 信息工程学院，陕西 西安 710064)
|
| 摘要: |
| 分析了多核环境下传统的离散事件时间弯曲并行系统的性能，针对其事件调度开销小和负载均衡能力强难以兼得的问题，提出了一种基于分布式队列的全局调度机制，设计了相应的数据结构和调度算法，大大减少了锁开销。通过大量实验对多核环境下几种典型离散事件系统并行策略的性能分析表明，本文提出的全局调度策略不仅事件调度开销小，而且回滚率大大降低，有效克服了传统策略回滚量较大或难以实现动态负载平衡的情况，并具备良好的可扩展性。 |
| 关键词: 并行离散事件仿真 时间弯曲协议 并发 多核 多线程 |
| DOI： |
| 投稿日期：2011-07-07 |
| 基金项目:国家自然科学基金资助项目（61170048） |
|
| A global schedule mechanism for PDES on multi-core environments |
| CHEN Lili1,2, YAO Yiping3, CAI Lu4 |
|
(1.Science and Technology on Complex System Simulation Laboratory, Beijing 100101, China;2.2.College of Computer, National University of Defense Technology, Changsha 410073, China;3.2.College of Computer, National University of Defense Technology, Changsha 410073, China;4.3.School of Information Management and Information Systems, Chang’an University, Xi’an 710064, China)
|
| Abstract: |
| The current trend in processor architecture design adopts the integration of multiple cores on a single processor. The tightly integrated processing cores in one chip with communication latencies substantially lower than those present in conventional clusters provide potential performance improvement especially for the fine-grained PDES. Thus, in the PDES domain, one of the research focuses is on modifying software platforms to efficiently utilize the computation resources of multi-core processors. The current dynamic load balancing technologies for PDES cannot reach the twin goals of good balance and low event-scheduling overhead. By taking advantage of multi-core architecture with shared memory address space and low communication, a global schedule mechanism based on a distributed event queue is proposed. Its specially designed data structures and algorithms reduced the cost of lock operations much. In comparison with the distributed event queue local schedule mechanism, the experiment results show that the distributed queue global schedule mechanism can effectively reduce the rollback rate and balance the workloads at a low event scheduling cost for Time Warp system on multi-core platforms. |
| Keywords: PDES Time Warp Parallelism Multi-core Multi-thread |
|
|
|
|
|
