本文選題：片上網(wǎng)絡(luò)　切入點(diǎn)：服務(wù)質(zhì)量　出處：《吉林大學(xué)》2015年博士論文

【摘要】：隨著半導(dǎo)體工藝進(jìn)入納米時代，單芯片上集成數(shù)十億支晶體管已經(jīng)成為現(xiàn)實(shí)，從2015年開始的兩年內(nèi)，集成電路的工藝尺寸有望縮小到10nm。集成電路制作工藝的飛速發(fā)展推動了“片上系統(tǒng)”(System-on-Chip，SoC)從多核到眾核(100-1000)、從總線到網(wǎng)絡(luò)、從集中式存儲到分布式存儲、從二維到三維方向上的發(fā)展，片上多核間的復(fù)雜通信使得基于總線的傳統(tǒng)片上通信結(jié)構(gòu)成為主要的性能瓶頸。為了適應(yīng)通信復(fù)雜度的需求，片上網(wǎng)絡(luò)(Network-on-Chip，NoC)已經(jīng)成為當(dāng)前片上多核間標(biāo)準(zhǔn)的通信架構(gòu)。面向不同的業(yè)務(wù)需求，NoC應(yīng)提供不同的性能保證，比如：像電子郵件、文件傳輸?shù)葟椥詷I(yè)務(wù)需要系統(tǒng)提供足夠的通信帶寬，采取盡力服務(wù)的方式，提高網(wǎng)絡(luò)資源的利用率，改善網(wǎng)絡(luò)的平均性能；而對那些諸如多媒體業(yè)務(wù)等對時間敏感度高的剛性需求，則需要NoC提供確定性的端到端的延時保證。 本文在保證網(wǎng)絡(luò)服務(wù)質(zhì)量(Quality of Service，QoS)的基礎(chǔ)上，建立網(wǎng)絡(luò)模型，評價網(wǎng)絡(luò)通信延時，并采取相關(guān)技術(shù)，優(yōu)化網(wǎng)絡(luò)性能。主要針對以下四方面展開研究： 第一，在以往的片上網(wǎng)絡(luò)性能評價中，經(jīng)常假設(shè)存儲器在網(wǎng)絡(luò)節(jié)點(diǎn)中均勻分布，且讀取響應(yīng)具有固定延時，沒有考慮網(wǎng)絡(luò)節(jié)點(diǎn)訪問片外存儲器的情況，導(dǎo)致訪存執(zhí)行過程過于簡單，得到的評價結(jié)果過于樂觀，不能準(zhǔn)確反映網(wǎng)絡(luò)中存貯器訪問的真實(shí)情況。文中選用Open Core Protocol-International Partnership(OCP-IP)工作組的DRAM模型，采用C++和SystemC語言實(shí)現(xiàn)模型配置與接口，修改片上網(wǎng)絡(luò)仿真平臺Noxim，使其能夠訪問片外存儲器，并運(yùn)行SPLASH-2測試標(biāo)準(zhǔn)，獲得真實(shí)有價值的片上網(wǎng)絡(luò)訪存延時數(shù)據(jù)。 第二，在多核甚至眾核系統(tǒng)中，處于不同路由節(jié)點(diǎn)的存儲器讀寫延時，在整個網(wǎng)絡(luò)延時中占有相當(dāng)比重。本文將訪存延時作為一項重要參數(shù)，對片上網(wǎng)絡(luò)訪存延時均衡性展開研究，建立片上網(wǎng)絡(luò)往返訪存延時預(yù)測模型，提出基于總延時預(yù)測的訪存報文仲裁技術(shù)。首先，依據(jù)訪存報文后續(xù)路徑的擁塞信息預(yù)測訪存報文未來等待延時，并計算出總延時。其次，基于預(yù)測的總延時對競爭同一鏈路的訪存報文進(jìn)行仲裁。本文在2D-mesh片上網(wǎng)絡(luò)路由器中，對該技術(shù)進(jìn)行了設(shè)計和實(shí)現(xiàn)，并對訪存延時做了優(yōu)化，有效地減小了片上訪存的延時，獲得了更好的訪存延時均衡性，提升了系統(tǒng)性能。 第三，減小系統(tǒng)的平均延時，對某些彈性業(yè)務(wù)的性能有一定程度的改善，但是對于一些實(shí)時性要求比較高的剛性業(yè)務(wù)，比如多媒體應(yīng)用技術(shù)，如果網(wǎng)絡(luò)服務(wù)不能保證視頻流的最大延時，會導(dǎo)致某一點(diǎn)上視頻播放效果急劇惡化，影響視頻輸出的整體質(zhì)量，在這種情況下，應(yīng)確保視頻流端到端的有效傳輸。本文以網(wǎng)絡(luò)演算為依據(jù)，首先分析片上網(wǎng)絡(luò)GPS節(jié)點(diǎn)輸入數(shù)據(jù)流的特征，建立緊致的到達(dá)曲線，然后根據(jù)網(wǎng)絡(luò)的繁忙周期，分析每條數(shù)據(jù)流在不同時間段的積壓，最后得到分段線性的GPS節(jié)點(diǎn)的準(zhǔn)確服務(wù)曲線。與傳統(tǒng)的GPS節(jié)點(diǎn)近似服務(wù)曲線相比，文中推導(dǎo)的服務(wù)曲線更緊致，服務(wù)性能更強(qiáng)，確保了數(shù)據(jù)流端到端的QoS。 第四，建立片上網(wǎng)絡(luò)流量控制演算模型，評價網(wǎng)絡(luò)性能。通過在路由節(jié)點(diǎn)添加流量控制器，實(shí)現(xiàn)傳輸路徑上相鄰的下游節(jié)點(diǎn)到當(dāng)前節(jié)點(diǎn)的反饋機(jī)制，根據(jù)反饋信息，控制注入到網(wǎng)絡(luò)中的數(shù)據(jù)量，，從而降低網(wǎng)絡(luò)擁塞，減小網(wǎng)絡(luò)競爭。借助網(wǎng)絡(luò)演算中的到達(dá)曲線和節(jié)點(diǎn)路由器的服務(wù)曲線，分析網(wǎng)絡(luò)中的延時和積壓，推導(dǎo)片上最優(yōu)緩沖區(qū)大小。 綜上所述，本文面向不同業(yè)務(wù)，在確保網(wǎng)絡(luò)QoS的前提下，構(gòu)建片上網(wǎng)絡(luò)通信模型，分析其延時函數(shù)，評價并優(yōu)化系統(tǒng)的性能。
[Abstract]:In order to meet the demands of communication complexity , network - on - chip ( NoC ) has become the main performance bottleneck . In order to meet the needs of communication complexity , the network - on - chip ( NoC ) has become the main performance bottleneck . In order to meet the needs of communication complexity , the NoC should provide different performance guarantees . For example , the system can improve the utilization rate of network resources and improve the average performance of the network . In order to meet the requirements of communication complexity , the network - on - chip ( NoC ) has become the main performance bottleneck .
while there is a need for noc to provide deterministic end - to - end delay guarantees for those rigid requirements such as multimedia services for time sensitivity .

On the basis of ensuring the quality of service ( QoS ) , this paper establishes a network model , evaluates network communication delay , and takes the related technology to optimize the network performance .

First , in the previous network performance evaluation , it is often assumed that the memory is uniformly distributed in the network node , and the read response has a fixed delay , which leads to an excessively simple visit and execution process , which can not accurately reflect the real situation of memory access in the network . In this paper , an Open Core Protocol - International Partnership ( ocp - IP ) working group is adopted to realize the model configuration and interface , and the network simulation platform Noxim on the chip is modified , so that the network simulation platform Noxim can be accessed , and the SPLASH - 2 test standard is run to obtain the real valuable piece - on - chip network access time delay data .

Second , in the multi - core and even public - core system , the memory read - write delay of different routing nodes occupies a significant proportion in the whole network delay . In this paper , the memory delay is used as an important parameter to study the time delay of network visit and memory on the chip , and the total delay is calculated . Secondly , based on the predicted total delay , the time delay of the visiting message is predicted and the total delay is calculated . Secondly , based on the total delay of the forecast , the time delay of the access memory is optimized , the delay of the on - chip visit is optimized , the better memory delay balance is obtained , and the performance of the system is improved .

Third , reduce the average delay of the system , improve the performance of some elastic services to some extent , but for some real - time requirements , such as multimedia application technology , if the network service does not guarantee the maximum delay of the video stream , it can cause the video playing effect to deteriorate sharply at some point . In this case , the accurate service curve of each data stream is analyzed . The service curve derived in this paper is more compact and the service performance is stronger than the traditional GPS node approximate service curve .

fourthly , establishing an on - chip network flow control calculus model to evaluate the network performance , and realizing the feedback mechanism of the adjacent downstream nodes on the transmission path to the current node through adding a flow controller at the routing node , controlling the data quantity injected into the network according to the feedback information , reducing the network congestion and reducing the network competition .

In conclusion , in this paper , the network communication model is constructed on the premise of ensuring the QoS of the network , and the time delay function is analyzed , and the performance of the system is evaluated and optimized .

【學(xué)位授予單位】：吉林大學(xué)
【學(xué)位級別】：博士
【學(xué)位授予年份】：2015
【分類號】：TN47

【參考文獻(xiàn)】

相關(guān)期刊論文 前4條

1 徐寧儀;冷祥綸;周祖成;;基于SystemC的支持異源通信實(shí)體的NoC仿真架構(gòu)[J];半導(dǎo)體技術(shù);2006年04期

2 張連明;陳志剛;黃國盛;;網(wǎng)絡(luò)演算理論及應(yīng)用研究[J];計算機(jī)工程與應(yīng)用;2006年27期

3 劉勝;陳書明;尹亞明;陳勝剛;谷會濤;陳小文;王耀華;;片上網(wǎng)絡(luò)延時差異對存儲系統(tǒng)公平性的影響及對策[J];計算機(jī)學(xué)報;2011年08期

4 林世俊;張凡;金德鵬;曾烈光;;分布式同步的GALS片上網(wǎng)絡(luò)及其接口設(shè)計[J];清華大學(xué)學(xué)報(自然科學(xué)版)網(wǎng)絡(luò).預(yù)覽;2008年01期

相關(guān)博士學(xué)位論文 前3條

1 陳小文;同構(gòu)眾核處理器的片上存儲管理與同步機(jī)制研究[D];國防科學(xué)技術(shù)大學(xué);2011年

2 錢悅;片上網(wǎng)絡(luò)演算模型及性能分析[D];國防科學(xué)技術(shù)大學(xué);2010年

3 蔣紅艷;基于流量監(jiān)控的網(wǎng)絡(luò)性能優(yōu)化關(guān)鍵技術(shù)研究[D];湖南大學(xué);2010年

本文編號：1710309