The three GCS centres HLRS (High Performance Computing Center Stuttgart), JSC (Jülich Supercomputing Centre) and LRZ (Leibniz Supercomputing Centre, Garching near Munich) have started an initiative which takes the existing close cooperation of the three national sites one additional step forward. Project InHPC-DE aims at uniting the three national HPC centres into one integrated national high-performance computing (HPC) ecosystem, in turn creating the foundation of a homogeneous yet distributed HPC concept of Germany's Tier-1 computing facilities and technologies, its world-class HPC services and support.
At the core of this initiative is a high-speed 100-Gbit/s network to interlink the three centres across the nation, enabling easy and fast, cross-organizational transfer of the massive amounts of data resulting from the large-scale computing runs. This extremely fast network will support distributed workflows, including post processing, and allow for collaborative, remote visualization. Equally important will be the implementation of a distributed yet integrated data management system spread across the three centres. Adding to GCS's most appealing feature—namely, the three centres' complementary system architectures offering its scientific users access to technology best suited to their specific tasks—the centres are revamping their user support structure to include not only HPC support specialists, but also application experts. By having a contact versed in their various science domains, researchers using GCS resources can collaborate more closely with the respective centres and have a liaison available to help solve specific computing challenges in the researchers’ specific areas of study.
Under the umbrella of GCS, the three national centres will continue their status as closely coordinated, highly aligned HPC centres. The technical and conceptual investments involved with project InHPC-DE lay the foundation for Germany's “Smart Scaling” initiative, where an advanced support structure and integrated computing ecosystem will lead to scientists being able to successfully leverage the ever-increasing computing capabilities on the way toward exascale computing.