Memory Management And Optimization Strategy Of Embedded JVM

The development of wireless internet and the internet of things technology promote the wide application of embedded devices. because the smart mobile phone and home appliance, there will be a higher requirement for the performance of embedded devices. with portability, security and compatibility of Java, brought huge boost to the production of J2ME.Due to the limitations of the mobile equipment,make us have to consider performance optimization to improve operation efficiency.Garbage collection mechanism of the Java virtual machine has a lot of room to improve now,this causes a lot of different recovery algorithm is put forword.In this article, we are to do same analysis with the basic principle of embedded Java virtual machine, at the same time, we discuss the use and development trend of it. This paper will do some research in the KVM garbage collection mechanism of the J2ME.And research the method of original garbage collection, such as reference-counting, mark-sweep, coping, generational collection algorithm. We analysis and research the compiled technology on the HotSpot virtual machine.Meanwhile Some optimization strategy is proposed by the generational collection,in the new and the old generation.for garbage collection of memory,it adopted the mark-clear and replication algorithm respectively.It puts forward the idea of dynamic partitioning for the new and old generation.Dynamic classification is used with top pointer two-way movement prediction,the new generation of long survival time object can quickly turn into old age and improving the efficiency of garbage collection,number and pause time.In last,it compares the dynamic generational collection and tag-collectiong,and inpromove the dynamic generational collection.