【摘要】：隨著移動互聯(lián)網(wǎng)技術(shù)和嵌入式技術(shù)的飛速發(fā)展,大量消費類多核嵌入式設(shè)備,如智能手機、平板電腦、智能電視,已應(yīng)用到人們生活的方方面面。人們?yōu)g覽網(wǎng)頁的習(xí)慣已開始從桌面電腦向這些移動終端轉(zhuǎn)變。然而,現(xiàn)有主流嵌入式瀏覽器基本采用原先單核處理器上的設(shè)計架構(gòu)——串行架構(gòu),直接運行在多核嵌入式設(shè)備中并未充分發(fā)揮多核處理器并行處理能力,使得瀏覽器性能仍未得到多大提升,并不能滿足用戶快速瀏覽網(wǎng)頁的需求。為此針對多核平臺并行執(zhí)行能力,本文重構(gòu)了瀏覽器串行架構(gòu),提出了適應(yīng)多核平臺的并行化架構(gòu)。本文首先在研究傳統(tǒng)串行瀏覽器架構(gòu)和網(wǎng)頁處理流程的基礎(chǔ)上,采用并行化思想將瀏覽器重構(gòu)成五個子系統(tǒng):用戶界面子系統(tǒng)、資源管理子系統(tǒng)、DOM子系統(tǒng)、CSS子系統(tǒng)、和Rendering子系統(tǒng)。各子系統(tǒng)之間通過消息進行通信,眾消息由消息管理器統(tǒng)一管理,采用優(yōu)先級調(diào)度策略進行分發(fā)處理。在各子系統(tǒng)內(nèi)部也采用一些并行化設(shè)計,如多線程圖片解碼、DOM并行樣式化和并行化渲染。對于圖片解碼,采用線程池解碼策略,每加載一個圖片就從線程池中分配一個線程負(fù)責(zé)解碼,進而降低了因圖片解碼引起的延時;對于DOM樣式化,將節(jié)點的樣式化分成兩個任務(wù)——規(guī)則匹配任務(wù)和樣式應(yīng)用任務(wù),各節(jié)點規(guī)則匹配任務(wù)的并行執(zhí)行加快了DOM樹的樣式化過程;對于渲染,將Render樹劃分成若干子樹,給每個子樹分配一個渲染線程,各子樹可同時渲染降低了整顆Render樹渲染時間。最后,在特定平臺上實現(xiàn)和測試該并行化設(shè)計并與原始串行瀏覽器做了性能對比與分析。
[Abstract]:With the rapid development of mobile Internet technology and embedded technology, a large number of consumer multi-core embedded devices, such as smart phones, tablets, intelligent television, have been applied to all aspects of people's lives. The habit of browsing the web has begun to shift from desktop computers to these mobile terminals. However, the existing mainstream embedded browsers basically adopt the design architecture of the original single-core processor-serial architecture, and run directly in multi-core embedded devices without giving full play to the parallel processing ability of multi-core processors. So that the browser performance has not been greatly improved, and can not meet the needs of users to quickly browse the web. Aiming at the parallel execution ability of multi-core platform, this paper reconstructs the browser serial architecture and proposes a parallel architecture suitable for multi-core platform. On the basis of studying the traditional serial browser architecture and web page processing flow, this paper reconstructs the browser into five subsystems: user interface subsystem, resource management subsystem, DOM subsystem and CSS subsystem. And Rendering subsystem. Each subsystem communicates through messages, and the message is managed uniformly by the message manager, and the priority scheduling policy is used to distribute the messages. Some parallel designs are also used in each subsystem, such as multithread picture decoding, DOM parallel styling and parallel rendering. For picture decoding, a thread pool decoding strategy is adopted, and a thread is allocated from the thread pool to decode each picture, which reduces the delay caused by picture decoding. For DOM styling, the node styling is divided into two tasks: rule matching task and style application task. The parallel execution of each node rule matching task speeds up the styling process of DOM tree. For rendering, the Render tree is divided into several subtrees, each subtree is assigned a rendering thread, and each subtree can be rendered simultaneously to reduce the rendering time of the whole Render tree. Finally, the parallel design is implemented and tested on a specific platform, and the performance of the parallel design is compared and analyzed with the original serial browser.
【學(xué)位授予單位】：電子科技大學(xué)
【學(xué)位級別】：碩士
【學(xué)位授予年份】：2014
【分類號】：TP393.092

【參考文獻(xiàn)】

相關(guān)碩士學(xué)位論文 前1條

1 孫瑋;嵌入式瀏覽器解析與排版布局引擎的研究優(yōu)化[D];電子科技大學(xué);2012年

，

本文編號：2352311