The University of Innsbruck secured ongoing access to the Leonardo supercomputer, current no. 7 of the fastest supercomputers worldwide, with two HPC access proposals accepted from Computer Science and Theoretical Chemistry. Located at the CINECA facility in Bologna, Italy, Leonardo is equipped with 13824 NVidia A100 GPUs, making it an ideal site for cutting-edge research in modelling and simulation.
The first project, lead by Peter Thoman, focuses on multi-GPU multi-node scheduling optimization of scientific workloads using the Celerity programming model and runtime system. Celerity is built on the SYCL standard, allowing seamless integration with existing C++ applications. It introduces a high-level abstraction that simplifies the development of distributed applications, enabling developers to focus on algorithm design rather than the complexities of managing data distribution and synchronization across GPU clusters.
The second project, lead by Klaus Liedl, focuses on molecular dynamics in theoretical chemistry, specifically the role of antigen-binding mechanisms and involved fragments in the function of antibodies in the adaptive immune system. Using Amber, a state-of-the-art molecular dynamics simulation software, the project will use the power of GPUs to analyze the structure and capture the effects of the antigen-binding framents.