Description

Modern multicore and many-core platforms can offer impressive performance for various applications. However, achieving this performance is far from trivial: programming these processors has proved to be much more difficult than programming traditional parallel machines, due to memory size and bandwidth constraints, multiple parallelism layers, heterogeneous components, and so on. This so-called multicore programmability gap calls for new ways of thinking about parallel programming and algorithms that can properly exploit these platforms. Therefore, this topic focuses on novel solutions for efficient programming of multicore and many-core processors in both general-purpose and embedded systems.