Using scratchpad to exploit object locality in Java

Lebsack, Carl S.; Chang, J. Morris

doi:10.1109/iccd.2005.111

Cited by 6 publications

(4 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Both systems had equivalent on-chip storage capacity. Applications that had the highest access density for small short-lived objects saw 17 to 20 percent decrease in traffic to memory when using scratchpad [3]. Applications that had highest access density for large longlived objects saw an increase in memory traffic when using scratchpad.…”

Section: Verifying Object Localitymentioning

confidence: 96%

System level perspective on object locality

Lebsack

Chang

2005

Companion to the 20th Annual ACM SIGPLAN Conference on Object-Oriented Programming, Systems, Languages, and Applications

View full text Add to dashboard Cite

Performance of modern computers is tied closely to the effective use of cache because of the continually increasing speed discrepancy between processors and main memory. Optimum system performance is achieved when software and hardware work symbiotically to increase performance. This work focuses on identifying locality in object-oriented systems and developing techniques for approaching optimum performance with respect to the memory hierarchy.

show abstract

Section: Verifying Object Localitymentioning

confidence: 96%

System level perspective on object locality

Lebsack

Chang

2005

Companion to the 20th Annual ACM SIGPLAN Conference on Object-Oriented Programming, Systems, Languages, and Applications

View full text Add to dashboard Cite

show abstract

“…Other related work such as [23] demonstrates the advantages of locality of small objects in Java applications by employing a garbage collector in a design with caches and SPM. In this method, small long-lived heap objects are mapped to SPM to reduce the memory traffic between the cache system and the main memory.…”

Section: Spm Utilization In Interpreted Environmentmentioning

confidence: 99%

Scratch-pad memory allocation without compiler support for java applications

Nguyen

Domínguez

Barua

2007

Proceedings of the 2007 International Conference on Compilers, Architecture, and Synthesis for Embedded Systems

View full text Add to dashboard Cite

This paper presents the first scratch-pad memory allocation scheme that requires no compiler support for interpreted-language based applications. A scratch-pad memory (SPM) is a fast compilermanaged SRAM that replaces the hardware-managed cache. Its uses are motivated by its better real-time guarantees as compared to cache and by its significantly lower overheads in energy consumption, area and access time. Interpreted languages are languages such as Java that are interpreted by a run-time environment instead of being executed directly on hardware.All existing memory allocation schemes for SPM require compiler analysis to develop the allocation strategy. Specifically, existing allocation schemes for Java-based applications determine the allocations at compile-time. They then annotate the Java bytecodes with these allocation decisions for the Java Virtual Machine (JVM) to implement the actual allocation at runtime. These existing allocation schemes tie the resulting bytecode to specific SPM sizes, therefore preventing the applications from being portable to different SPM sizes. Further, existing methods do not work for unmodified third-party bytecodes produced by compilers other than their specialized compilers.In this paper, we propose the first ever SPM allocation scheme that is completely implemented inside the JVM. Our method requires no compiler support and works for unmodified bytecodes from any source. Moreover, unlike existing methods, it preserves the portability of bytecode to any SPM size. We investigate our method on the Sun Hotspot JVM on which we achieve a 27.8% improvement on runtime and 21.8% on energy saving versus not using the SPM -the only existing alternative for unmodified bytecodes.

show abstract

“…This division is supported by our previous simulation works involved in isolating the nursery region in on-chip resources through physical separation (49) and logical separation through cache replacement biasing (50).…”

Section: Java Environmentmentioning

confidence: 85%

“…The technique allows for partitioning of one sixteenth (no updates, leaving lines in LRU), one eighth, one fourth and one half of the cache. Although it has also been reported that converting half of the cache resources to scratchpad for allocation can reduce memory traffic (49), our technique offers the flexibility of using a smaller percentage of the cache resources for allocation. One eighth most closely matches the life-span of Java objects and this is the size for which we report results.…”

Section: Cache Replacement Policymentioning

confidence: 99%

Performance analysis and optimization of the Java memory system

Chang

Lebsack

View full text Add to dashboard Cite

I would first like to dedicate this thesis to God by whom all things were made and in whom they are held together. Only by His grace and mercy through the gift of His Son, Jesus Christ, was I afforded the opportunity to live and work and play. I would also like to dedicate this thesis to my parents, Stephen and Debra Lebsack, whose support was pivotal in the completion of this work. I want to thank my siblings, Mary and her husband, Aaron Hudlemeyer, David and Laurel, and the rest of my family for their unwavering belief in my ability to complete PhD study and their continual love and support during my unbelief. I also thank Mary for her help in proofing and editing the thesis text. I, am, terrible, with, commas. Laurel deserves an extra thanks for enduring the front lines of battle while living with me during her own transition into college. I would also like to thank my fiancee, Tegwin Taylor, who, although she came along near the end of my study, was extremely uplifting and provided significant motivation to finish at long last. Thanks also goes out to the rest of my friends and family for their continued encouragement

show abstract

Using scratchpad to exploit object locality in Java

Abstract: Abstract

Cited by 6 publications

References 20 publications

System level perspective on object locality

System level perspective on object locality

Scratch-pad memory allocation without compiler support for java applications

Performance analysis and optimization of the Java memory system

Contact Info

Product

Resources

About