Memory Management Techniques for Embedded System Language Runtimes

点击翻译按钮获取中文摘要

In this research, we studied garbage collection techniques for language runtimes on embedded systems. As a result, we achieved the followings. (1) We developed a partial compaction algorithm that ensures that the heap never gets fragmented more than a degree by compacting only a part of the heap in each garbage collection cycle. (2) We improved a concurrent garbage collection so that garbage collector can run efficiently on embedded systems with multi-core CPUs. (3) We developed a program analysis tool that finds bugs in the Android library that were obstacles of our implementation of the partial compaction. (4) We developed two techniques to improve performance of garbage collections.