The Optimization And Implementation Of The Embedded Javascript Engine

With the development of the3G mobile Internet technology, Browser operatingSystem, and dynamic webpage technology, browser is becoming the platform in whichusers use those rising technologies. The performance of browser plays a very importantrole in users’ experience. So high-performance embedded browser technology has beendrawing great public attention.There are many factors have something to do with browser’s performances, andJavaScript engine’s performance is one of those most important factors. With thedevelopment of JavaScript engine, many classic engines arise.Such as Google’s V8engine, Webkit’s squirrelfish engine, and Mozilla FireFox’s JavaScript engine.Thoseengines put forward many effective optimization methods.The JavaScript engine weresearching bases on SquirrelFish Extreme, it compiles all the scripts into native codethen execute, it improves engine’s performance effective. In some resources limitedplatform, its compilation delay becomes serious bottleneck.So hotspot is adopted to optimize our engine, and a new kind of engine namedHotCode is constructed. HotCode engine mainly learns the hot spot idea from JavaHotspot virtual machine, referring to the traits of Embedded Browser. It focuses on hotspot which plays a very important role in engine’s execution efficiencies and analysesits varieties, distribution, general features associating with JavaScript language. Then ittakes a deeper look into hot spot by constructing its general feature models. Basing onthose models, it further constructs a lot of hot spot exploration models realized bydifferent resolutions.Hot spot exploration could divide into preliminary exploration and monitorexploration by different exploration opportunity.The former could adjust executioncodes before engine executes them. The latter could dig out key regions hinder engine’sefficiencies during executing. Our engine uses not only the traditional exploration byhot spot region idea, but also exploration by hot tracing. We put forward a new kind ofmixed-way, combining the both advantages of the two ways. Taking quick response of embedded browser into account, we separate some regions named less-hotspot fromhotspot. The less-hot spot’s compilation is a delayed procedure, so it relieves situationmaking by hardware shortage of embedded platform. It accelerates engine’s responsespeed, and finds a way to make balanced use of system resources moreover.Our HotCode engine has different solutions towards the same issue. Thosedifferent solutions become a strategy warehouse. It is rational called by IntelligentAnalysis and Decision model according to engine’s performance data. In this way,engine will always cope with the mutative execution demand dynamically by choosingthe best solution. At last, we optimize architecture of engine by using multithreadingidea; this will be an exploration to engine’s parallel processing ability.