This paper presents a simple and efficient data flow algorithm for escape analysis of objects in Java programs to determine (i) if an object can be allocated on the stack; (ii) if an object is accessed only by a single thread duriing its lifetime, so that synchronization operations on that object can be removed. We introduce a new program abstraction for escape analysis, the connection graph, that is used to establish reachability rela-tionships between objects and object references. We show that the connection graph can be summarized for each method such that the same summary information may be used effectively in different calling contexts. We present an interprocedural al-gorithm that uses the above property to efficiently compute the connection graph and identify the non-escaping objects for methods and threads. The experimental results, from a proto-type implementation of our framework in the IBM High Per-formance Compiler for Java, are very promising. The percent-age of objects that may be allocated on the stack exceeds 70% of all dynamically created objects in three out of the ten bench-marks (with a median of 19%), 11 % to 92 % of all lock oper-ations are eliminated in those ten programs (with a median of 5 l%), and the overall execution time reduction ranges from 2 % to 23 % (with a median of 7%) on a 333 MHz PowerPC workstation with 128 MB memory. Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advant-age and that copies bear this notice and the full citation on the first page.

Abstract. Java programs perform many synchronization operations on data structures. Some of these synchronizations are unnecessary; in particular, if an object is reachable only by a single thread, concurrent access is impossible and no synchronization is needed. We describe a flow-insensitive, context-sensitive data-flow analysis that finds such situations and a global optimizing transformation that eliminates synchronizations on these objects. For every program in our suite of ten Java benchmarks consisting of SPECjvm98 and others, our system optimizes over 90 % of the alias sets containing at least one synchronized object. As a result, the dynamic frequency of synchronizations is reduced by up to 99%. For two benchmarks that perform synchronizations very frequently, this optimization leads to speedups of 36 % and 20%, respectively. 1.

The Marmot system is a research platform for studying the implementation of high level programming languages. It currently comprises an optimizing native-code compiler, runtime system, and libraries for a large subset of Java. Marmot integrates well-known representation, optimization, code generation, and runtime techniques with a few Java-specific features to achieve competitive performance. This paper contains a description of the Marmot system design, along with highlights of our experience applying and adapting traditional implementation techniques to Java. A detailed performance evaluation assesses both Marmot's overall performance relative to other Java and C++ implementations and the relative costs of various Java language features in Marmot-compiled code. Our experience with Marmot has demonstrated that well-known compilation techniques can produce very good performance for static Java applications---comparable or superior to other Java systems, and approaching that o...

The Java Virtual Machine is now widely available on many architectures and systems, which makes this virtual machine a very appealing target as an execution environment, not only for Java programs but also for Java extensions as well as other high-level languages.

This document describes the areas of research I am interested in, and highlights the aims and scope of work that I intend to pursue. This proposal supercedes the interim statement I wrote in February --- it includes more detail in the discussion of virtual types for genericity, and further detail on lightweight classes. Comments and suggestions are welcome, as are pointers to important literature I have overlooked.

I hereby declare that this thesis has not been submitted, either in the same or different form, to this or any other university for a degree.

From a real-time perspective, the garbage collector (GC) introduces unpredictable pauses that are not tolerated by real-time tasks. Real-time collectors eliminate this problem but introduce a high overhead. Another approach is to use memory regions (MRs) within which allocation and deallocation is customized. This facility is supported by the memory model of the Real-Time Specification for Java (RTSJ). RTSJ imposes strict access and assignment rules to avoid both the dangling inter-region references and the delays of critical tasks of the GC. A dynamic check solution can incur high overhead, which can be reduced by taking advantage of hardware features. This paper provides an in-depth analytical investigation of the overhead introduced by dynamic assignments checks in RTSJ, describing and analysing several solutions to reduce the introduced overhead. Copyright c ○ 2005 John Wiley & Sons, Ltd. KEY WORDS: 1.

Dynamic memory management is a known performance bottleneck of Java applications. The problem arises out of the Java memory model in which all objects (non-primitive type instances) are allocated on the heap and reclaimed by garbage collector when they are no longer needed. This paper presents a simple and fast algorithm for inference of object lifetimes. Given the analysis results, a Java compiler is able to generate faster code, reducing the performance overhead. Besides, the obtained information may be then used by garbage collector to perform more effective resource clean-up. Thus, we consider this technique as "compile-time garbage collection" in Java.