Optimizing Java-Specific Overheads: Java at the Speed of C?

Veldema, Ronald; Kielmann, Thilo; Bal, Henri E.

doi:10.1007/3-540-48228-8_78

Cited by 4 publications

(5 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Nevertheless, most real-world programs do not use dynamic class loading: they consist of a set of classes which is completely defined at compile time. Such programs comply with the closed world assumption: all classes are available at compile time [12]. Our static algorithm works correctly under the closed world assumption.…”

Section: Devirtualization and Dynamic Loadingmentioning

confidence: 90%

“…A different, static approach requires that the programmer annotates as final all methods that will not be overridden. But this choice limits an incremental design of the class hierarchy [12]. Nevertheless, most real-world programs do not use dynamic class loading: they consist of a set of classes which is completely defined at compile time.…”

Section: Devirtualization and Dynamic Loadingmentioning

confidence: 99%

“…The features of portability, robustness, security and ease of development have made the Java language [1] popular in the software engineering community. While early implementations of Java Virtual Machines (JVMs) were based on an unsophisticated interpretation of bytecodes, recent developments have led to the diffusion of a family of high-performance execution environments including Just In Time (JIT) compilers [2], optimizing runtime compilers [3][4][5] and native compilers [6][7][8][9][10][11][12].…”

Section: Introductionmentioning

confidence: 99%

“…Recent studies have shown that Java is competitive with traditional languages such as C, C++ and Fortran for some applications [13]. By evaluating the Manta native compiler on three applications kernels, Veldema et al [12] have shown that the Manta-generated code in two cases executes faster than the corresponding gcc-compiled C versions. Comparing Java, C and Fortran implementations of a subset of the Java Grande Benchmarks [14], Bull et al [15] have measured a limited execution time penalty for Java.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Evaluation and optimization of method calls in Java

2004

View full text Add to dashboard Cite

Recent advances in Virtual Machine's technology have led to the diffusion of Java execution environments which provide performance levels comparable to those of traditional languages such as C/C++ and Fortran.This paper's focus is on method calls, a well researched area of compilers' optimization. In particular, we evaluate the impact on the performance of the Java virtual call mechanism and the advantages which can be obtained through the inlining optimization. Moreover, we propose a new heuristic methodology which allows us to consider whole sequences of methods for the inlining decision and not only simple couples of caller-callee methods.The proposed methodology operates at bytecode level and is applicable only to the calls that can be resolved at compile time. Thus, we execute a devirtualization algorithm before the inlining decision, in order to enlarge the basis for the optimization.We made empirical measures on three major platforms (Unix, Windows and Linux) by using the SPEC JVM98 benchmark suite. Results reveal an average performance improvement of 1.9% with devirtualization and 5.5% with inlining. We discuss the different behaviour of single benchmarks and virtual machines, arguing that dynamic compilers are not yet able to fully exploit the wider scope for intraprocedural optimization offered by eliminating the calls. Finally, we show that taking into consideration whole sequences of calls-instead of single calls-is effective only on a few benchmarks where hotspots of execution involve cycles of nested methods.

show abstract

Section: Devirtualization and Dynamic Loadingmentioning

confidence: 90%

Section: Devirtualization and Dynamic Loadingmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Evaluation and optimization of method calls in Java

2004

View full text Add to dashboard Cite

show abstract

“…Works are being developed to make Java become a real time language [25,26,27,28,29,30]. The creation of a real time Java virtual machine will certainly be an important step to improve application performances.…”

Section: Results and Conclusionmentioning

confidence: 99%

An implementation of a distributed algorithm for detection of local knots and cycles in directed graphs based on the CSP model and Java

Souza

Pfitscher

Proceedings. Sixth IEEE International Workshop on Distributed Simulation and Real-Time Applications

View full text Add to dashboard Cite

Cycles and knots in directed graphs are problems that can be associated with deadlocks in database and communication systems. Many algorithms to detect cycles and knots in directed graphs were proposed. Boukerche and Tropper have proposed a distributed algorithm that solve the problem in a efficient away. Their algorithm has a message complexity of 2m vs. (at least) 4m for the Chandy and Misra algorithm, where m is the number of links in the graph, and requires O (n log n) bits of memory, where n is the number of nodes. We have implemented Boukerche and Tropper's algorithm according to the construction of processes of the CSP model. Our implementation was done using JCSP, an implementation of CSP for Java, and the results are presented.

show abstract