SHF: Small: Software Testing Cognizant of Just-in-time Compilers

University Of Texas At Austin, Austin TX

Investigators

Abstract

Just-in-time (JIT) compilers are integral to many popular programming languages, such as Java, C#, or JavaScript. These languages use managed runtime environments, where the execution of a program starts in an interpreted mode, but as code gets executed more times, a JIT compiler recompiles code blocks into an optimized native form, ensuring faster and more optimized versions of code. JIT compilers are instrumental in ensuring that software written in the aforementioned languages is highly performant. Despite the immense importance of JIT compilers for optimizing deployed software, there is little research focused on JIT compilation, particularly in the area of software testing. The goal of this project is to develop techniques to support both JIT compiler developers to better test their compilers and general software developers to improve their software testing process by taking advantage of the underlying JIT compiler with which they run. To accomplish this goal, this project will (1) develop test generation techniques that allow JIT compiler developers to bring their domain knowledge to better test JIT compilers, (2) accelerate software testing via compiler-cognizant analyses, and (3) integrate existing testing techniques closer with the compilers. This proposal has the potential to substantially reduce the cost of software testing and software development, as well as increase the quality of JIT compilers and the code they produce. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.