【摘要】：高性能計(jì)算機(jī)在能源、生物、氣象、科研、地質(zhì)勘探等計(jì)算密集型應(yīng)用中得到長(zhǎng)足發(fā)展，是一個(gè)國(guó)家科技綜合實(shí)力的重要象征，在現(xiàn)代社會(huì)中發(fā)揮著重要的作用，然而它的發(fā)展現(xiàn)在卻面臨著功耗墻的重大挑戰(zhàn)。 隨著半導(dǎo)體工藝的持續(xù)發(fā)展和芯片集成度的顯著提高，微處理器的性能在得到大幅度提升的同時(shí)也造成了功耗密度的急劇增大，而且伴隨著計(jì)算機(jī)結(jié)點(diǎn)規(guī)模的擴(kuò)大，高性能計(jì)算機(jī)的功耗急劇增高，目前千萬(wàn)億次的計(jì)算機(jī)功耗基本以兆瓦（MW）計(jì)，其中K Computer功耗更是達(dá)到了驚人的9.898MW。高功耗同時(shí)帶來(lái)高發(fā)熱，消耗了更多的冷卻成本，造成巨大的電力壓力，大大限制了高性能計(jì)算機(jī)的發(fā)展和應(yīng)用。 本文以高性能計(jì)算機(jī)功耗管理為研究?jī)?nèi)容，基于對(duì)系統(tǒng)負(fù)載、能耗分布以及硬件特征等總體把握的基礎(chǔ)上，設(shè)計(jì)功耗管理系統(tǒng)，該系統(tǒng)集監(jiān)控、策略制定、功耗調(diào)整、設(shè)備控制于一體，結(jié)合資源管理、作業(yè)管理、低功耗編譯、動(dòng)態(tài)電源管理、動(dòng)態(tài)電壓調(diào)整以及外圍設(shè)備控制等方法技術(shù)，力求在保證系統(tǒng)性能的同時(shí)降低系統(tǒng)的整體功耗。 在功耗管理系統(tǒng)的研究中，自主開(kāi)發(fā)了模擬器Simschedule對(duì)ParallelWorkloads Archive標(biāo)準(zhǔn)負(fù)載進(jìn)行分析，，總結(jié)出高性能計(jì)算機(jī)作業(yè)負(fù)載的特點(diǎn)。建立結(jié)點(diǎn)能耗模型，結(jié)合作業(yè)負(fù)載特點(diǎn)，提出了三種基于關(guān)閉空閑結(jié)點(diǎn)的低功耗調(diào)整算法，三種算法分別為關(guān)閉間隔時(shí)間調(diào)整算法、反比關(guān)系時(shí)間調(diào)整算法、歷史記錄調(diào)整算法，通過(guò)模擬驗(yàn)證的方式證明了三種算法的有效性。在滿(mǎn)足用戶(hù)切換頻率限定約束的情況下，三種算法都大大降低了系統(tǒng)的功耗。 本文以天河-1A超級(jí)計(jì)算機(jī)為實(shí)驗(yàn)平臺(tái)對(duì)外圍設(shè)備的功耗管理進(jìn)行了研究。經(jīng)過(guò)分析，基于CPU負(fù)載、CPU溫度以及風(fēng)機(jī)轉(zhuǎn)速之間存在的相互關(guān)系，提出了基于感知的風(fēng)機(jī)智能調(diào)控策略，在達(dá)到精確制冷的同時(shí)，能夠降低風(fēng)機(jī)22.9%的功耗；對(duì)電源單元提出了一種按需分配的供電策略，在保證電源單元時(shí)時(shí)處于高效率工作的前提下，達(dá)到節(jié)省功耗17.5%的效果。 采用功耗管理系統(tǒng)較好地實(shí)現(xiàn)了對(duì)高性能計(jì)算機(jī)系統(tǒng)級(jí)功耗管理和對(duì)外圍設(shè)備的調(diào)控，大大降低了整個(gè)計(jì)算機(jī)系統(tǒng)的整體功耗，節(jié)約了經(jīng)濟(jì)成本，同時(shí)具有良好的社會(huì)效應(yīng)。
[Abstract]:High-performance computer has made great progress in computing intensive applications such as energy, biology, meteorology, scientific research, geological exploration and so on. It is an important symbol of the comprehensive strength of science and technology in a country and plays an important role in modern society. However, its development is now facing major challenges from the power-consuming wall. With the continuous development of semiconductor technology and the remarkable improvement of chip integration, the performance of microprocessor not only improves greatly, but also results in a sharp increase of power density, and accompanied by the expansion of computer node size. The power consumption of high-performance computers has increased dramatically. At present, the power consumption of thousands of terabytes of computers is basically megawatt (MW), and the K-Computer power consumption has reached an astonishing 9.898MW. High power consumption at the same time leads to high heat consumption, consuming more cooling costs, resulting in huge power pressure, which greatly limits the development and application of high-performance computers. Based on the overall grasp of system load, energy consumption distribution and hardware characteristics, this paper designs a power management system based on high-performance computer power management. The system includes monitoring, policy formulation, power consumption adjustment, and so on. Combined with resource management, job management, low-power compilation, dynamic power management, dynamic voltage regulation and peripheral equipment control, equipment control can ensure the performance of the system and reduce the overall power consumption of the system. In the research of power management system, the simulator Simschedule is developed independently to analyze the ParallelWorkloads Archive standard load, and the characteristics of high performance computer job load are summarized. According to the characteristics of job load, three low-power adjustment algorithms based on closing idle nodes are proposed. The three algorithms are closed interval time adjustment algorithm and inverse ratio relation time adjustment algorithm, respectively. The validity of the three algorithms is verified by simulation. The three algorithms greatly reduce the power consumption of the system when the limit of user switching frequency is satisfied. In this paper, the power management of peripheral devices is studied with Tianhe-1A supercomputer as the experimental platform. Based on the analysis of the relationship among CPU load, CPU temperature and fan speed, an intelligent fan control strategy based on perception is proposed, which can reduce the power consumption of the fan by 22.9% while achieving accurate refrigeration. This paper presents a power supply strategy based on demand distribution for the power supply unit, which can save 17.5% of power consumption on the premise that the power supply unit is in high efficiency all the time, and that the power consumption of the power supply unit can be reduced by 17.5%. The power management system is adopted to realize the power management of high performance computer system and the control of peripheral devices, which greatly reduces the overall power consumption of the whole computer system, saves the economic cost, and has good social effects at the same time.
【學(xué)位授予單位】：國(guó)防科學(xué)技術(shù)大學(xué)
【學(xué)位級(jí)別】：碩士
【學(xué)位授予年份】：2012
【分類(lèi)號(hào)】：TP38

【參考文獻(xiàn)】

相關(guān)期刊論文 前7條

1 卜愛(ài)國(guó),劉昊,李杰;PBALT動(dòng)態(tài)電源管理策略[J];電路與系統(tǒng)學(xué)報(bào);2005年04期

2 趙霞;陳向群;郭耀;楊芙清;;操作系統(tǒng)電源管理研究進(jìn)展[J];計(jì)算機(jī)研究與發(fā)展;2008年05期

3 曹仰杰;錢(qián)德沛;伍衛(wèi)國(guó);董小社;;一種面向多處理器系統(tǒng)的在線低功耗調(diào)度算法[J];西安交通大學(xué)學(xué)報(bào);2010年08期

4 李順;曾超;李軍;;一種自適應(yīng)的動(dòng)態(tài)功耗管理預(yù)測(cè)策略[J];信息與電子工程;2010年04期

5 卜愛(ài)國(guó),劉昊,胡晨,王超;基于AR模型的PCF動(dòng)態(tài)電源管理預(yù)測(cè)策略[J];應(yīng)用科學(xué)學(xué)報(bào);2005年05期

6 黃少珉;戚隆寧;楊軍;胡晨;;能耗期望的動(dòng)態(tài)功耗管理預(yù)測(cè)性策略[J];應(yīng)用科學(xué)學(xué)報(bào);2009年03期

7 谷立靜;周伏秋;孟輝;;我國(guó)數(shù)據(jù)中心能耗及能效水平研究[J];中國(guó)能源;2010年11期

本文編號(hào)：2442813