Understanding the connectivity of heap objects

HirzelMartin,; HenkelJohannes,; DiwanAmer,; HindMichael,

doi:10.1145/773039.512435

Cited by 8 publications

(14 citation statements)

References 35 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Recent work on connectivity-based garbage collection [21,22] shows that connected objects often die together, which bolsters our hypothesis. However, their collector organization completely eliminates write barriers with a static analysis that allocates objects into static partitions that contain connected objects, forming a hierarchical directed acyclic graph of partitions.…”

Section: Related Worksupporting

confidence: 70%

“…Colocation makes a unique use of static and dynamic information, and should play well with other optimizations to further improve performance. Previous work suggests heap organizations that segregate objects by connectivity, but with the restriction that the objects must never install cross region pointers [21,22]. The success of colocation instead suggests collector organizations that group connected objects into separately collected regions where a write barrier handles cross region pointers.…”

Section: Resultsmentioning

confidence: 99%

“…Since collectors use connectivity to determine survivors, the lifetimes of connected objects are correlated [22], and placing them in the same space can improve collector efficiency. Colocation produces this effect dynamically by determining in which space the source of a pointer resides, and then allocating the target of the pointer in the same space.…”

Section: Dynamic Object Colocationmentioning

confidence: 99%

“…Since collectors must assume that pointers between separately collected regions are live, they will retain this data structure, perhaps needlessly. Furthermore, previous work shows that connected objects usually die together [21,22]. The goal of colocation is to allocate a new object directly into the same region as an object that will reference it.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Finding your cronies

GuyerSamuel

McKinleyKathryn

2004

SIGPLAN Not.

View full text Add to dashboard Cite

This paper introduces dynamic object colocation, an optimization to reduce copying costs in generational and other incremental garbage collectors by allocating connected objects together in the same space. Previous work indicates that connected objects belong together because they often have similar lifetimes. Generational collectors, however, allocate all new objects in a nursery space. If these objects are connected to data structures residing in the mature space, the collector must copy them. Our solution is a cooperative optimization that exploits compiler analysis to make runtime allocation decisions. The compiler analysis discovers potential object connectivity for newly allocated objects. It then replaces these allocations with calls to coalloc, which takes an extra parameter called the colocator object. At runtime, coalloc determines the location of the colocator and allocates the new object together with it in either the nursery or mature space. Unlike pretenuring, colocation makes precise per-object allocation decisions and does not require lifetime analysis or allocation site homogeneity. Experimental results for SPEC Java benchmarks using Jikes RVM show colocation can reduce garbage collection time by 50% to 75%, and total performance by up to 10%.

show abstract

Section: Related Worksupporting

confidence: 70%

Section: Resultsmentioning

confidence: 99%

Section: Dynamic Object Colocationmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Finding your cronies

GuyerSamuel

McKinleyKathryn

2004

SIGPLAN Not.

View full text Add to dashboard Cite

show abstract

“…GC traces are used not only for simulations, however, but have also been used to gain a deeper understanding of the issues affecting object lifetimes [Hirzel et al 2002a[Hirzel et al , 2002bShaham et al 2000] and to measure the effects of GC optimizations [Shaham et al 2002]. Because of the speedup in trace generation achieved by the Merlin algorithm, it is now feasible to consider generating traces at granularities finer than each allocation.…”

Section: Using Merlin For Granulated Tracesmentioning

confidence: 99%

Generating object lifetime traces with Merlin

Hertz

Blackburn

Moss

et al. 2006

ACM Trans. Program. Lang. Syst.

View full text Add to dashboard Cite

Programmers are writing a rapidly growing number of programs in object-oriented languages, such as Java and C#, that require garbage collection. Garbage collection traces and simulation speed up research by enabling deeper understandings of object lifetime behavior and quick exploration and design of new garbage collection algorithms. When generating perfect traces, the brute-force method of computing object lifetimes requires a whole-heap garbage collection at every potential collection point in the program. Because this process is prohibitively expensive, researchers often use granulated traces by collecting only periodically, for example, every 32 KB of allocation.We extend the state of the art for simulating garbage collection algorithms in two ways. First, we develop a systematic methodology for simulation studies of copying garbage collection and present results showing the effects of trace granularity on these simulations. We show that trace granularity often distorts simulated garbage collection results compared with perfect traces. Second, we present and measure the performance of a new algorithm called Merlin for computing object lifetimes. • 477Merlin timestamps objects and later uses the timestamps of dead objects to reconstruct when they died. The Merlin algorithm piggybacks on garbage collections performed by the base system. Experimental results show that Merlin can generate traces over two orders of magnitude faster than the brute-force method which collects after every object allocation. We also use Merlin to produce visualizations of heap behavior that expose new object lifetime behaviors.

show abstract

Semi-Automatic Region-Based Memory Management for Real-Time Java Embedded Systems

Salagnac

Rippert

Verimag

2007

13th IEEE International Conference on Embedded and Real-Time Computing Systems and Applications (RTCSA 2007)

View full text Add to dashboard Cite

In this paper we address the problem of dynamic memory management in real-time Java embedded systems. Our work aims at suppressing the need for garbage collection in order to avoid unpredictable pause times. For that we use a simple static analysis algorithm coupled with region-based memory management as presented in [15]. To overcome the well-known limitations of region inference, we propose in this paper to involve the developer in the analysis process by providing feedback on programming constructs likely to produce memory leaks. Experiments show that for most programming patterns, our system behaves as efficiently as a garbage collector in terms of memory consumption. Our analysis tool is furthermore able to provide useful feedback to the programmer to pinpoint problematic constructs.

show abstract

Understanding the connectivity of heap objects

Cited by 8 publications

References 35 publications

Finding your cronies

Finding your cronies

Generating object lifetime traces with Merlin

Semi-Automatic Region-Based Memory Management for Real-Time Java Embedded Systems

Contact Info

Product

Resources

About