Storage mode analysis is a region-based program analysis that aims at determining whether a memory region can be reset before allocating in the region. Storage mode analysis is an integral part of MLKit, a region-based Standard ML compiler, and serves as an important mechanism for improving the overall memory usage for programs that run without a reference tracing garbage collector and for decreasing the pressure on the garbage collector when region-based reclamation of memory is integrated with reference tracing garbage collection. Originally, storage mode analysis was presented as a global program analysis that suffered from very conservative properties when used in concert with separate compilation. We present a modular storage mode analysis that is effective also for use across compilation unit boundaries. The modular storage mode analysis has been adopted for use in the MLKit, which now benefits from the supplied modular reasoning principles.

Professor in the Programming Languages and Theory of Computation (PLTC) section at Department of Computer Science, University of Copenhagen (DIKU). Conducts research in the design and implementation of programming languages, including compilation techniques for functional languages, parallelism, memory management, and program optimisation.