In: Proc. 16th Performance Engineering Workshop, 24-25 July 2000, Durham, UK, pages 129-140. 2000.
Abstract: Understanding and improving Java performance is an important objective for both applications and tool developers. Current efforts towards developing benchmarks have played a crucial role towards this objective, particularly in identifying general trends that can be observed through such a study. An alternative mechanism for categorizing Java performance is investigated here, based on developing a performance model of Java execution. Two models based on stochastic Petri nets are proposed, one based on a profile of Java source code, the other based on execution semantics of a Java program, aimed at evaluation performance of Java method call.
Keywords: Java, execution semantics, performance analysis, stochastic Petri nets.