Why Nobody Should Care About Operating Systems for Exascale Ron Brightwell Reaching the next three orders of magnitude increase in parallel computing performance poses significant new challenges for system software. The landscape of hardware for computing, memory, networking, and storage is changing, which is in turn driving the need for new parallel algorithms and programming models. Power, resiliency, massive on-node parallelism, and heterogeneity are some of the important issues that are driving the need for new and alternative approaches to the design and implementation of scalable system software for high-performance computing. However, these changes are minimizing the importance of the operating system as the lowest-level resource manager, but are rather focusing more attention on the runtime system as the key intermediary between applications and hardware. This talk will discuss some of the important capabilities needed for exascale runtime systems and offer a perspective on addressing some of the fundamental challenges and opportunities of exascale computing.