EuroHPC PL

The project was the Polish stage of the construction of a European exascale supercomputer system, created by the EuroHPC Joint Undertaking (EuroHPC JU). It enabled large-scale calculations across national borders and complex simulations, big data analysis, advanced visualizations, and the creation of solutions using artificial intelligence. 

The state-of-the-art infrastructure created as part of EuroHPC PL consists of high-performance computers, quantum and neuromorphic accelerators, memory resources, software, and expert knowledge. The integrated large-scale HPC (High Performance Computing) environment allows for the creation of advanced high-performance computing services. The project involved the construction of supercomputer systems with a total computing power of at least 32 petaflops. 

Detailed project objectives:

• creating infrastructure for supercomputers, specialized accelerators, and laboratory services for the needs of science, the economy, and society;
• improving research and development work;
• implementing modern methods in personalized medicine, including personalized radiotherapy;
• greater focus of R&D work on market needs,
• technological development of enterprises and innovation in areas of national smart specialization,
• wider access for enterprises to modern research infrastructure resources,
• comprehensive support for scientists and entrepreneurs in scaling software and adapting it to the requirements of exascale supercomputers.

The project work was carried out in four laboratories. The Wrocław Supercomputing and Networking Center was responsible for selected tasks in each of them and was also the leader of the fourth laboratory.     

1. The laboratory for modeling and parallel data processing in a pre-exascale environment focused on the design and implementation of advanced large-scale computing systems (at Cyfronet and WCSS) and smaller machines for the development of computing technologies and applications for next-generation systems (at Cyfronet and PCSS). It developed tools for effective system management in a production environment and for the integration of the EuroHPC PL infrastructure with other Polish and European e-infrastructures, as well as tools enabling the use of supercomputers for urgent calculations. The task of this Laboratory was to build:

• a supercomputing platform for scientific research,
• a platform of specialized accelerators,
• a platform for urgent calculations (WCSS is the leader of this task),
• a platform for the integration of e-infrastructures (EuroHPC environment, quantum accelerator environment, and neuromorphic accelerator environment).

2. The Hybrid Computing Applications Laboratory focused on the application of quantum computers and their implementation in practical use. To this end, it was necessary to develop methods and techniques that would allow quantum computing tools to be used to solve existing problems or to formulate the problem in such a way that quantum techniques could be applied. The laboratory built:

• a platform for adapting scientific problems to EuroHPC, quantum, and neuromorphic accelerator technologies,
• an e-platform for quantum machine learning,
• a platform for supporting quantum computing with classical algorithms,
• a platform for designing multi-stage computations using quantum accelerators,
• a platform for certification and error mitigation on quantum computers,
• a platform for quantum operations research and discrete optimization in hybrid computations.

 3. The Laboratory for the Application of Supercomputers in Medicine conducted research on the use of computing power to create individualized therapies tailored to specific patients. The tasks performed within this Laboratory focused on the practical application of high computing power in medical applications. As part of the project, it built:

• a modeling platform for personalized medicine,
• a simulation platform for radiotherapy,
• a platform for quantum simulations and medical imaging for positron emission tomography scanners.

4. The HPC software energy and computational efficiency laboratory worked on developing techniques to reduce the operating costs of computing machines, such as energy consumption and cooling. Some of the tasks within this laboratory concerned the optimization of software running in a supercomputer environment. The task of this laboratory was to build:

• a computational energy efficiency platform,
• a computational efficiency platform in an exascale environment,
• a massively parallel software platform,
• an HPC software acceleration platform,
• a platform for planning multi-stage computations using quantum accelerators,
• a platform facilitating and visualizing quantum computations,
• a pilot platform for efficient domain applications.

The tasks of this Laboratory also included:

• developing a software library for exascale computing,
• developing a methodology for analyzing HPC codes for exascale computing,
• developing a methodology for producing software for exascale computing,
• adapting selected scientific software packages to the requirements of exascale computing,
• developing a methodology for code analysis in the context of a given scientific problem.

The project was carried out by a consortium consisting of:

• AGH University of Science and Technology – Cyfronet AGH Academic Computer Center (project leader)
• Institute of Bioorganic Chemistry of the Polish Academy of Sciences – Poznań Supercomputing and Networking Center
• Gdańsk University of Technology – IT Center of the Tri-City Academic Computer Network
• Wrocław University of Technology – Wrocław Centre for Networking and Supercomputing
• National Center for Nuclear Research – Świerk Computing Center
• Institute of Theoretical and Applied Computer Science of the Polish Academy of Sciences
• Center for Theoretical Physics of the Polish Academy of Sciences.