題目：Static Energy Management in Supercomputer Interconnection Networks Using Topology-Aware Partitioning

時(shí)間：2018年6月25日上午10:30-11:30

地點(diǎn)：明理樓B306

主辦單位：西南石油大學(xué)科研處計(jì)算機(jī)科學(xué)學(xué)院

主講人：陳娟（Juan Chen）

Abstract：With the parallel systems being scaled-up, the static energy consumed by their interconnection networks has been increasing substantially. The key to reducing static energy in supercomputers is switching off their unused components. Routers are the major components of a supercomputer. Whether routers can be effectively switched off or not has become the key to static energy management for supercomputers. For many typical applications, the routers in a supercomputer exhibit low utilization. However, it is very difficult to switch the routers off when they are idle. By analyzing the router occupancy in time and space, we present a routing-policy guided topology partitioning methodology to solve this problem. We propose topology partitioning methods for three kinds of commonly used topologies (mesh, torus and fat-tree) equipped with the three most popular routing policies (deterministic routing, directionally adaptive routing and fully adaptive routing). Based on the above methods, we propose the key techniques required in this topology partitioning based static energy management in supercomputer interconnection networks to switch off unused routers in both time and space dimensions. Three topology-aware resource allocation algorithms have been developed to handle effectively different job-mixes running on a supercomputer. We validate the effectiveness of our methodology by using Tianhe-2 and a simulator for the aforementioned topologies and routing policies.

主講人簡介：陳娟，博士，國防科技大學(xué)計(jì)算機(jī)學(xué)院副教授、碩士生導(dǎo)師。國家公派加州大學(xué)河濱分校訪問學(xué)者。2007年6月于國防科技大學(xué)獲得計(jì)算機(jī)科學(xué)與技術(shù)專業(yè)博士學(xué)位。ACM/IEEE會員、中國計(jì)算機(jī)學(xué)會（CCF）高級會員、中國計(jì)算機(jī)學(xué)會理論計(jì)算機(jī)專委會委員、ACM SIGCSE China副秘書長、常務(wù)理事。曾作為核心技術(shù)骨干參與天河超級計(jì)算機(jī)系統(tǒng)研制，在高性能編譯優(yōu)化、高效異構(gòu)協(xié)同計(jì)算、多核多線程并行優(yōu)化、軟件低功耗優(yōu)化技術(shù)、能量有效性優(yōu)化技術(shù)等方面開展了長期深入的研究。獲省部級科技進(jìn)步一等獎(jiǎng)1項(xiàng)、省部級科技進(jìn)步二等獎(jiǎng)2項(xiàng)、校教學(xué)成果二等獎(jiǎng)1項(xiàng)。主持和參與了國家自然科學(xué)基金青年基金項(xiàng)目、國家自然科學(xué)基金面上項(xiàng)目、國家重點(diǎn)實(shí)驗(yàn)室基金項(xiàng)目、核高基重大專項(xiàng)、國家863重點(diǎn)項(xiàng)目等10余項(xiàng)課題。在IEEE Transactions on Computers、Parallel Computing、FCS、JCST、RSC Advances、Cluster、SIGCSE等國際著名期刊和會議上發(fā)表論文多篇。累計(jì)共發(fā)表學(xué)術(shù)論文60余篇。獲Cluster 2010最佳論文獎(jiǎng)。獲得國家發(fā)明專利9項(xiàng)。

擔(dān)任2018中國圖靈大會SIGCSE程序委員會共同主席。擔(dān)任SIGCSE '17、SIGCSE '18、ITiCSE '17、ACM TURC (SIGCSE China) '17 - '18、ICESS '14 - '16、HPCC '08等國際會議程序委員會委員。擔(dān)任IEEE TPDS、Journal of Supercomputing、Frontiers of Computer Science in China、IEEE Systems Journal等多個(gè)重要學(xué)術(shù)期刊的審稿人。擔(dān)任《清華大學(xué)學(xué)報(bào)自然科學(xué)版（英文版）》高性能計(jì)算專刊編委。

研究方向：大規(guī)模并行計(jì)算機(jī)系統(tǒng)軟件低功耗優(yōu)化技術(shù)、能量有效性優(yōu)化技術(shù)、功耗感知的并行算法、高性能互連的能量優(yōu)化方法、大規(guī)模科學(xué)計(jì)算應(yīng)用優(yōu)化及可擴(kuò)展性研究、基于GPU/Intel MICs的性能優(yōu)化、基于GPU的能量有效性優(yōu)化、CPU-GPU異構(gòu)系統(tǒng)的低功耗優(yōu)化、能量建模與預(yù)測技術(shù)、基于機(jī)器學(xué)習(xí)的能量有效的任務(wù)調(diào)度方法等。