Research On Java Virtual Machine Optimization Of Mobile Communication Terminal

In recent years, Java in PC workstation and server performance hasbeen greatly improved, consistent in the level of the performance of Cprograms, because in the Java virtual machine, there are multipleperformance optimization techniques are applied. But in the field ofmobile communication, the calculation of mobile communicationequipment, power supply and storage capacity is limited, in current thecomplex optimization technology can hardly applied in communicationequipment.This paper analyzes several performance optimization techniqueused in a Java virtual machine, according to the mobile communicationequipment hardware features, find the technology facing the problem inthe embedded Java virtual machine.Some optimization methods adaptedto specific places, under certain circumstances, it can achieve goodresults, but to not adapt occasion, it will show obvious drawbacks. On thebasis of that, this paper designs a Java virtual machine overall optimization program which is suitable for mobile communicationequipment. This is a dynamic compiler optimization scheme based on theinterpretation of the implementation, Mainly used in embedded Javavirtual machine performance optimization.This scheme includes threeaspects the optimization,that is threaded directly interrupt optimization,optimization, dynamic compilation optimization for hot spot detection.The technical ideas of this optimization scheme is basicd on thread directinterrupt, through design pretreatment of Java byte code, LRU tablemanagement link code, simplified hot looking for operation, set theappropriate threshold for dynamic compilation to part of thecomputationally intensive, Realizing the improvement of the embeddedJava virtual machine implementation on the mobile communicationequipment.In order to verify this algorithm, the paper introduces test platformof the CLDC1.1and MIDP2.0, use FeiLing OK6410development boardand WinCE embedded operating system, through the operational testbinarytrees and input/output test OutPutTest, Respectively Test operation optimization effect and input output optimization effect,obtained a general conclusion of this optimization algorithm: he moresimilar code, can generate more heat,the optimization scheme of thispaper can play a greater role. For the compute-intensive programs, canget greater improvement in performance, to the input/output intensiveprogram, performance is likely to be a small drop.