DescriptionOpenMP is the leading node-level programming model for high performance computing (HPC) systems. This ubiquitous interface supports a wide range of high-level parallel programming paradigms, including task-based parallelism as well as execution on devices such as GPUs. The language also provides many low-level capabilities that enable users to control details of code generation. This range of features combines with its ubiquity to make OpenMP an ideal choice for cross-platform portability. This talk will discuss the unifying concepts that guide OpenMP's design and on-going evolution. It will explore recent extensions and potential future directions with a particular focus on directions that enhance its support for portable and performant programming.