This paper presents a new combined pointer and escape analysis for multithreaded programs. The algorithm uses a new abstraction called parallel interaction graphs to analyze the interactions between threads and extract precise points-to, escape, and action ordering information for objects accessed by multiple threads. The analysis is compositional, analyzing each method or thread once to extract a parameterized analysis result that can be specialized for use in any context. It is also capable of analyzing programs that use the unstructured form of multithreading present in languages such as Java and standard threads packages such as POSIX threads.We have implemented the analysis in the MIT Flex compiler for Java and used the extracted information to 1) verify that programs correctly use region-based allocation constructs, 2) eliminate dynamic checks associated with the use of regions, and 3) eliminate unnecessary synchronization. Our experimental results show that analyzing the interactions between threads significantly increases the effectiveness of the region analysis and region check elimination, but has little effect for synchronization elimination.

We present a new classification system for aspect-oriented programs. This system characterizes the interactions between aspects and methods and identifies classes of interactions that enable modular reasoning about the crosscut program. We argue that this system can help developers structure their understanding of aspect-oriented programs and promotes their ability to reason productively about the consequences of crosscutting a program with a given aspect.We have designed and implemented a program analysis system that automatically classifies interactions between aspects and methods and have applied this analysis to a set of benchmark programs. We found that our analysis is able to 1) identify interactions with desirable properties (such as lack of interference), 2) identify potentially problematic interactions (such as interference caused by the aspect and the method both writing the same field), and 3) direct the developer's attention to the causes of such interactions.

We present an interprocedural and compositional algorithm for finding pairs of compatible allocation sites, which have the property that no object allocated at one site is live at the same time as any object allocated at the other site. If an allocation site is compatible with itself, it is said to be unitary : at most one object allocated at that site is live at any given point in the, execution of the program. We use the results of the analysis to statically preallocate memory space for the objects allocated at unitary sites, thus simplifying the computation of an upper bound on the amount of memory required to execute the program. We also use the analysis to enable objects allocated at several compatible allocation sites to share the same preallocated memory. Our experimental results show that, for our set of Java benchmark programs, 60% of the allocation sites are unitary and can be statically preallocated. Moreover, allowing compatible unitary allocation sites to share the same preallocated memory leads to a 95% reduction in the amount of memory preallocated for these sites.