A dynamic analysis to support object-sharing code refactorings

Abstract

Creation of large numbers of co-existing long-lived isomorphic objects increases the memory footprint of applications significantly. In this paper we propose a dynamic-analysis based approach that detects allocation sites that create large numbers of long-lived isomorphic objects, estimates quantitatively the memory savings to be obtained by sharing isomorphic objects created at these sites, and also checks whether certain necessary conditions for safely employing object sharing hold. We have implemented our approach as a tool, and have conducted experiments on several real-life Java benchmarks. The results from our experiments indicate that in real benchmarks a significant amount of heap memory, ranging up to 37% in some benchmarks, can be saved by employing object sharing. We have also validated the precision of estimates from our tool by comparing these with actual savings obtained upon introducing object-sharing at selected sites in the real benchmarks.

Semi-automated Algorithm for Complex Test Data Generation for Interface-based Regression Testing of Software Components

“…Their common idea is that some instances in the memory of a running program are equivalent and all equivalent instances can be replaced by a single instance without affecting the execution of the program [3], [4]. Examples can be found in [20], [21], [22], or [23]. In [20], it is pointed out that a thorough comparison of two objects (in order to determine their equality) requires checking of the graphs of the internal structures of the compared objects for isomorphism.…”

Deep Object Comparison for Interface-based Regression Testing of Software Components

“…In contrast to [12], the technique described in [14] is focused on the long-lived objects, which are candidates for caching. For the speedup of the object comparison, a form of hash value called "fingerprint" of the object is used.…”

“…For the speedup of the object comparison, a form of hash value called "fingerprint" of the object is used. In standard classes from the java.lang package, the equals() method is used [14].…”

Performance Modeling of Maximal Sharing

“…In additional experiments simulating the semantics of related memory profilers, which check for isomorphic heap graphs [18,14], we measured that MASHO uncovers up to 14%, and at median 4%, more unique caching opportunities in the aforementioned benchmarks than the related work can provide -due to the additional abstraction facilities.…”

“…Rama and Komondoor [18] worked on an extension of OEP and introduced a tool, the Object Caching Advisor (OCA), to support introducing hash-consing at the source-code level as a refactoring. The authors reuse a fingerprinting function, introduced by Xu [28], that runs in O(size of object graph) and yields a runtime overhead ranging from 98-2520x.…”