Improving Trace Precision for Concurrent Garbage Collection on Multicore Platform

Abstract

Garbage Collection removes the programmer's responsibility for managing heap object deal location, and is now in widespread use in programming languages such as Java, C#, Python and Perl. On the other hand garbage collection is the most difficult component to be parallel and concurrent in runtime system. This paper proposed a high precision concurrent tracing garbage collector based on tracing reference count. The concurrent collector is modeled, by using the reference count in recycling process, most of the modifications by the application thread are monitored. The barriers are classified as either store protection or deletion protection. In this paper, the approach can effectively track the modifications of the object references, reducing the number of objects in the process of rescanning process, while effectively reducing the of floating garbage during the concurrent collection.