期刊文献+

动态时间片缩放:一种优化Linux任务公平性的方法

Dynamic Time-Slice Scaling:A Scheme to Improve Fairness of Tasks in Linux
下载PDF
导出
摘要 处理器动态电压频率调节技术,对Linux系统中并发任务的性能产生不同程度的变化,从而影响并发任务计算资源分配的公平性.提出了一种利用动态时间片缩放来优化任务公平性的方法,并基于Linux操作系统任务调度程序,加入动态时间片缩放模块,该模块通过读取CPU性能监控计数器,在线计算时间片缩放系数,并利用该系数对任务时间片长度进行动态缩放.实验表明,这种方法以较小的系统开销为价,极大地提高了Linux中并发任务计算资源分配的公平性. The processor dynamic voltage and frequency scaling(DVFS)leads to unequal performance loss for concurrently running tasks in Linux, affecting the fairness of CPU sharing. This paper proposes a dynamic time-slice scaling scheme to improve the fairness of CPU sharing for concurrently running tasks. The method which is implemented based on Linux task scheduler dynamically scales the length of time-slice for each task using a time-slice scaling factor which is calculated based on the processor performance monitoring unit statistics. Experiment results show that dynamic time-slice sealing scheme can significantly improve the fairness of the CPU sharing with low overhead compared with the conventional Linux scheduler algorithm.
作者 高旭宏 李曦
出处 《计算机系统应用》 2012年第10期208-213,共6页 Computer Systems & Applications
基金 江苏省产学研前瞻性联合研究项目(BY2009128)
关键词 动态电压频率调节 Linux任务调度程序 并发任务 公平性 动态时间片缩放 dynamic voltage and frequency scaling Linux task scheduler concurrently running tasks fairness dynamictime-slice scaling
  • 相关文献

参考文献20

  • 1Donald J, Martonosi M. Techniques for multicore thermal management:classification and new exploration. Proc. of 33rd International Symposium on Computer Architecture. 2006, 78 -88.
  • 2Benini L, De Micheli G System-Level power optimization: Techniques and tools. ACM Trans. on Design Automation of Electronic Systems, 2000: 115-192.
  • 3Jejurikar R,Gupta R. Dynamic slack reclamation with procra- stination scheduling in real-time embedded systems Proc. of the 42nd annual Design Automation Conference. New York, 2005. 111-116.
  • 4Zanini F, Jones CN. Multicore thermal management using approximate explicit Model Predictive Control. Proc. of IEEE International Symposium on Circuits and Systems. 2010. 3321-3324.
  • 5Cai Q, Gonzalez J, Magklis G, Chaparro P, Gonzalez A. Thread shuffling: Combining DVFS and thread migration to reduce energy consumptions for multi-core systems. Proc. of 2011 International Symposium on Low Power Electronics and Design. Aug. 2011. 379-384.
  • 6PaUipadi V, Starikovskiy A.The ondemand governor, Proc, of the Linux Symposium, 2006, 2:223-238.
  • 7Hughes C, Li T. Optimizing throughput/power tradc-offs in hardware transactional memory using DVFS and intelligent scheduling. Proc. of the International Conference on Su- pcrcomput/ng, 2011.141-150.
  • 8Dhiman G; Rosing TS. Dynamic voltage frequency scaling for multi-tasking systems using online learning. Proc. of the 2007 International Symposium on Low Power Electronics and Design. New York, 2007. 207-212.
  • 9Spiliopoulos V, Keramidas G, Kaxiras S, Efstathiou K. Poster: DVFS management in real-processors. Proc. of the International Conference on Supercomputing, 2011.
  • 10Kolpe T, Zhai SA, et al. Enabling improved power management in multicore processors through clustered DVFS. Proc. of Design, Automation & Test in Europe Con- ference & Exhibition. 2011.1-6.

相关作者

内容加载中请稍等...

相关机构

内容加载中请稍等...

相关主题

内容加载中请稍等...

浏览历史

内容加载中请稍等...
;
使用帮助 返回顶部