E. Suarez, Jülich Supercomputing Centre (DE)
Abstract: Reaching Exascale compute performances at an affordable budget requires increasingly heterogeneous HPC systems, which combine general purpose processing units (CPUs) with acceleration devices (e.g. graphic cards (GPUs) or many-core processors). The Modular Supercomputing Architecture developed within the EU-funded DEEP project series orchestrates all these resources at system-level, organizing them in compute modules. The goal is to provide cost-effective computing at extreme performance scale fitting the needs of a wide range of Computational Sciences.
In a modular supercomputer each application can dynamically decide which kinds and how many nodes to use, mapping its intrinsic requirements and concurrency patterns onto the hardware. Codes that perform multi-physics or multi-scale simulations can run across compute modules thanks to a global system-software and programming environment. Application workflows that execute different actions after (or in parallel) to each other can also be distributed in order to run each workflow-component on the best suited hardware, and exchange data either directly (via message-passing communication) or via the file-system. This talk described the Modular Supercomputing Architecture – which constitutes the central element in the JSC’s roadmap to Exascale –, including its history, its hardware and software elements, and its current and upcoming implementations.