Making the Most of SMT in HPC
Making the Most of SMT in HPC
复制标题
DOI：
10.1145/2687651
复制
发表时间：
2015-01
期刊：
影响因子：
--
通讯作者：
Leo Porter;M. Laurenzano;Ananta Tiwari;Adam Jundt;W. A. Ward;R. Campbell;L. Carrington
中科院分区：
文献类型：
--
作者：
Leo Porter;M. Laurenzano;Ananta Tiwari;Adam Jundt;W. A. Ward;R. Campbell;L. Carrington
这项工作提出了一个端到端的方法，量化的性能和功耗的好处，同时多线程（SMT）的HPC中心，并将此方法应用于生产系统和工作负载。最终，SMT在整个系统中的价值取决于用户是否在应用程序级别有效地使用SMT。然而，预测SMT对HPC应用程序的好处是具有挑战性的;通过将线程数量增加一倍，应用程序的特性可能会发生变化。这项工作提出了统计建模技术来预测加速SMT赋予HPC应用程序。这种方法的精确度在8%以内，只使用在应用程序单次运行期间收集的轻量级、透明的性能监视器。
This work presents an end-to-end methodology for quantifying the performance and power benefits of simultaneous multithreading (SMT) for HPC centers and applies this methodology to a production system and workload. Ultimately, SMT’s value system-wide depends on whether users effectively employ SMT at the application level. However, predicting SMT’s benefit for HPC applications is challenging; by doubling the number of threads, the application’s characteristics may change. This work proposes statistical modeling techniques to predict the speedup SMT confers to HPC applications. This approach, accurate to within 8%, uses only lightweight, transparent performance monitors collected during a single run of the application.