A scalable technique for characterizing the usage of temporaries in framework-intensive Java applications

Dufour, Bruno; Ryder, Barbara G.; Sevitsky, Gary

doi:10.1145/1453101.1453111

Cited by 83 publications

(65 citation statements)

References 29 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Each of these tools detects one specific type of performance bugs, such as inefficient nested loops [37], under-utilized data structures [46], and temporary object bloat [12,44,45], through static or dynamic program analysis. They are not designed to cover a wide variety of performance bugs.…”

Section: Motivationmentioning

confidence: 99%

Statistical debugging for real-world performance problems

Song

2014

Proceedings of the 2014 ACM International Conference on Object Oriented Programming Systems Languages &Amp; Applications

View full text Add to dashboard Cite

Design and implementation defects that lead to inefficient computation widely exist in software. These defects are difficult to avoid and discover. They lead to severe performance degradation and energy waste during production runs, and are becoming increasingly critical with the meager increase of single-core hardware performance and the increasing concerns about energy constraints. Effective tools that diagnose performance problems and point out the inefficiency root cause are sorely needed.The state of the art of performance diagnosis is preliminary. Profiling can identify the functions that consume the most computation resources, but can neither identify the ones that waste the most resources nor explain why. Performance-bug detectors can identify specific type of inefficient computation, but are not suited for diagnosing general performance problems. Effective failure diagnosis techniques, such as statistical debugging, have been proposed for functional bugs. However, whether they work for performance problems is still an open question.In this paper, we first conduct an empirical study to understand how performance problems are observed and reported by real-world users. Our study shows that statistical debugging is a natural fit for diagnosing performance problems, which are often observed through comparison-based approaches and reported together with both good and bad inputs. We then thoroughly investigate different design points in statistical debugging, including three different predicates and two different types of statistical models, to understand which design point works the best for performance diagnosis. Finally, we study how some unique nature of performance bugs allows sampling techniques to lower the over- * Shan is now with University of Chicago.

show abstract

Section: Motivationmentioning

confidence: 99%

Statistical debugging for real-world performance problems

Song

2014

Proceedings of the 2014 ACM International Conference on Object Oriented Programming Systems Languages &Amp; Applications

View full text Add to dashboard Cite

show abstract

“…Each of these tools detects one specific type of performance bugs, such as inefficient nested loops [30], under-utilized data structures [38], and temporary object bloat [11,36,37], through static or dynamic program analysis. They are not designed to cover a wide variety of performance bugs.…”

Section: Motivationmentioning

confidence: 99%

“…Some tools [11,36,37] detect runtime bloat, a common performance problem in object-oriented applications. Xu et al [38] targets low-utility data structures with unbalanced costs and benefits.…”

Section: Performance Bug Detectionmentioning

confidence: 99%

Statistical debugging for real-world performance problems

SongLinhai

2014

SIGPLAN Not.

View full text Add to dashboard Cite

Design and implementation defects that lead to inefficient computation widely exist in software. These defects are difficult to avoid and discover. They lead to severe performance degradation and energy waste during production runs, and are becoming increasingly critical with the meager increase of single-core hardware performance and the increasing concerns about energy constraints. Effective tools that diagnose performance problems and point out the inefficiency root cause is sorely needed.The state of the art of performance diagnosis is preliminary. Profiling can tell where computation resources are spent, but not where and why the resources are wasted. Performance-bug detectors can identify specific type of inefficient computation, but are not suited for diagnosing general performance problems. Effective failure diagnosis techniques, such as statistical debugging, have been proposed for functional bugs. However, whether they work for performance problems is still an open question.In this paper, we first conduct an empirical study to understand how performance problems are observed and reported by real-world users. Our study shows that statistical debugging is a natural fit for diagnosing performance problems, which are often observed through comparison-based approaches and reported together with both good and bad inputs. We then thoroughly investigate different design points in statistical debugging, including three different predicates and two different types of statistical models, to understand which design point works the best for performance diagnosis. Finally, we study how some unique nature of performance bugs allows sampling techniques to lower the overhead of run-time performance diagnosis without extending the diagnosis latency.

show abstract

“…Software bloat analysis [19,4,20,1,21,22,3,2,23,24] attempts to identify and remove performance problems due to run-time inefficiencies in the code execution and the use of memory. Mitchell et al [5] propose a manual approach that detects bloat by structuring behavior according to the flow of information, and their later work [4] introduces a way to find data structures that consume excessive amounts of memory.…”

Section: Related Workmentioning

confidence: 99%

CoCo: Sound and Adaptive Replacement of Java Collections

2013

ECOOP 2013 – Object-Oriented Programming

View full text Add to dashboard Cite

Abstract. Inefficient use of Java containers is an important source of run-time inefficiencies in large applications. This paper presents an application-level dynamic optimization technique called CoCo, that exploits algorithmic advantages of Java collections to improve performance. CoCo dynamically identifies optimal Java collection objects and safely performs run-time collection replacement, both using pure Java code. At the heart of this technique is a framework that abstracts container elements to achieve efficiency and that concretizes abstractions to achieve soundness. We have implemented part of the Java collection framework as instances of this framework, and developed a static CoCo compiler to generate Java code that performs optimizations. This work is the first step towards achieving the ultimate goal of automatically optimizing away semantic inefficiencies.

show abstract

A scalable technique for characterizing the usage of temporaries in framework-intensive Java applications

Cited by 83 publications

References 29 publications

Statistical debugging for real-world performance problems

Statistical debugging for real-world performance problems

Statistical debugging for real-world performance problems

CoCo: Sound and Adaptive Replacement of Java Collections

Contact Info

Product

Resources

About