Energy-efficient localised rollback via data flow analysis and frequency scaling

Energy-efficient localised rollback via data flow analysis and frequency scaling
复制标题

通过数据流分析和频率缩放实现节能的局部回滚

DOI:
10.1145/3236367.3236379
复制
发表时间:
2018
期刊:
--
影响因子:
--
通讯作者:
Dichev K
Dichev K
中科院分区:
--
文献类型:
--
作者:
Dichev K

文献摘要

参考文献

被引文献

相似文献

Exascale系统每小时都会发生故障。HPC程序员主要依赖应用程序级检查点和全局回滚来恢复。近年来,减少回滚进程数量的技术已经通过消息日志记录实现。然而,基于日志的方法有弱点,例如依赖于MPI实现中的复杂修改,以及在一般情况下可能需要完全重启的事实。为了解决所有基于日志的机制的局限性,我们返回到仅检查点的机制,但提倡数据流回滚(DFR),一个根本不同的方法依赖于分析迭代代码的数据流,以及众所周知的数据流图的概念。我们展示了DFR的MPI模板代码的本地化回滚的好处,然后通过频率缩放将空闲节点上的能耗降低10-12%。我们还提供了大规模的估计DFR的节能相比,全球回滚,其中模板代码增加为n2的过程计数n。
Exascale systems will suffer failures hourly. HPC programmers rely mostly on application-level checkpoint and a global rollback to recover. In recent years, techniques reducing the number of rolling back processes have been implemented via message logging. However, the log-based approaches have weaknesses, such as being dependent on complex modifications within an MPI implementation, and the fact that a full restart may be required in the general case. To address the limitations of all log-based mechanisms, we return to checkpoint-only mechanisms, but advocate data flow rollback (DFR), a fundamentally different approach relying on analysis of the data flow of iterative codes, and the well-known concept of data flow graphs. We demonstrate the benefits of DFR for an MPI stencil code by localising rollback, and then reduce energy consumption by 10-12% on idling nodes via frequency scaling. We also provide large-scale estimates for the energy savings of DFR compared to global rollback, which for stencil codes increase as n2 for a process count n.
DOI: 10.1177/1094342010391989
发表时间: 2011-02-01
影响因子: 3.1
作者:
Dongarra, Jack;Beckman, Pete;Yelick, Kathy
通讯作者: Yelick, Kathy
通过域分解应用程序的协调检查点/回滚协议优化恢复
DOI: --
发表时间: 2008
期刊: Modelling, Computation and Optimization in Information Systems and Management Sciences
影响因子: --
作者:
Xavier Besseron;T. Gautier
通讯作者: T. Gautier
DOI: --
发表时间: 2013
期刊: PMBS@SC
影响因子: --
作者:
Scott Levy;Bryan Topp;Kurt B. Ferreira;D. Arnold;T. Hoefler;Patrick M. Widener
通讯作者: Patrick M. Widener
嵌套 Fork-Join 程序的局部故障恢复
DOI: --
发表时间: 2017
期刊: IEEE International Parallel and Distributed Processing Symposium
影响因子: --
作者:
Gokcen Kestor;S. Krishnamoorthy;Wenjing Ma
通讯作者: Wenjing Ma
DOI: --
发表时间: 2013
期刊: The international journal of high performance computing applications
影响因子: --
作者:
Wesley Bland;A. Bouteiller;T. Hérault;G. Bosilca;J. Dongarra
通讯作者: J. Dongarra