High Performance Computing (HPC) at RWTH
The RWTH Compute Cluster managed by the IT Center bundles all High Performance Computing (HPC) ressources at RWTH. It offers a comprehensible point of access for all RWTH institutions and staff who want to use these ressources for research and teaching. Since the ressources are limited, there is an allocation process for computing time exceeding a basic quota. Methodological support for using the HPC components as well as workshops and trainings are, for example, offered by the IT Center or the Jülich Super Computing Center (JSC)
The HPC system at RWTH is called CLAIX (Cluster Aix-la-Chapelle) and consists of three parts installed between 2016 and 2018. They offer different levels of availability (Tier-2 and Tier-3), and some of them can be used only for research, while others are also available for teaching and thesis projects. To manage this heterogeneous system landscape and provide access to various user groups with different demands, the RWTH IT Center introduce the "1-Cluster-Concept". This operational strategy makes all HPC ressources available using one interface and thereby enables optimal usability as well as scalability for all processes.
All RWTH students and employees are granted a basic computing time quota for research purposes within this cluster. For projects exceeding this quota, there is an allocation process for the limited computing time. Depending on the scope, there are different project categories (see illustration). The IT Center Help offers an overview of the respective application processes for each category. For smaller projects such as theses, the process is very lean. Within each project category, both CPU and GPU ressources can be requested.
National HPC for Computational Engineering Science - NHR4CES
Within the NHR4CES, RWTH Aachen University and TU Darmstadt bundle their competences and experience in the field of High Performance Computing (HPC). The NHR4CES belongs to a coordinated network of national HPC centers, the NHR, which was initiated by the Joint Science Conference of the federal and state governments in 2018 to support HPC at German universities. The aim of NHR4CES is to create an ecosystem that combines best practices in HPC and research data management. In addition to providing an HPC infrastructure, this also includes supporting the development of scalable algorithms and software in various application disciplines. For this purpose, the NHR4CES scientists have joined together in topic-specific "Simulation and Data Laboratories" and "Cross-Sectional Groups". More information can be found on the NHR4CES website.
JARA Center for Simulation and Data Science (CSD)
Within the Jülich-Aachen Research Alliance (JARA), the CSD is a competence center for simulation, data analysis and HPC technologies. The CSD’s goal is to contribute to today’s research landscape by making the use of data analysis systems and HPC systems available to a broad scientific user spectrum and at the same time encouraging top-level research, that would not be possible without the joint activities and resources within the Center. In order to enable the computational science communities to solve problems with high scientific impact through efficient use of HPC, several Simulation and Data Laboratories and Cross-Sectional Groups were established since 2009 in the context of JARA. An overview can be found on the JARA-CSD website.