High Performance Computing (HPC) refers to the development, operation, and use of large scale computers to solve the most challenging problems in Computational Science. Some examples are the simulation of the full Earth system to understand and predict climate change, designing environmentally friendlier materials, simulating the human brain under healthy and diseased conditions, or replying fundamental questions in particle physics, just to name some of the applications that very strongly rely on HPC infrastructures.
Achieving the highest computational performance (typically measured in floating operations per second, FLOP/s) at an affordable energy consumption and cost requires analyzing a highly multi-dimensional design space and taking trade-off decisions from chip- to system design level, including hardware and software constrains, and taking application and user requirements into account.
This group touches upon all these layers in the development of novel system level architectures aiming at the integration of heterogeneous computer resources in the most efficient manner, both for the system's operation and its use. Some of the research aspects that are touched in this context are listed below: