Many of the University’s staff have refocused their research to help stop the spread of COVID-19, with some of these projects relying on the resources of a particular facility on site. The University’s High-Performance Computing (HPC) facility has significantly contributed towards the fight against the pandemic since mid-March 2020. The “supercomputer” and its team, led by Prof. Pascal Bouvry and Dr. Sébastien Varrette, have supported University researchers and external partners in more than seven projects with its computational resources.
HPC delivers high performance in order to solve large problems faster. Tasks which would typically require several years to be computed on a typical desktop computer may only require a couple of hours, days or weeks on a HPC system. Accelerating time-to-solution is a critical criterion to efficiently fight the spread of the pandemic, which makes this resource all the more valuable for our researchers on the COVID-19 Task Force.
The supercomputer’s vast computing power and storage capacities were used for enabling and accelerating COVID-19 research in the areas of biomedical and life sciences, ICT and material sciences. Among others, it facilitates machine learning-based lung bioavailability estimations; business ecosystem modelling and simulation techniques to inform economic policy-makers in Luxembourg and abroad; while allowing for computing future predictions of the visibility of the coronavirus on surfaces.
Among the projects the HPC has supported are four University-led projects funded by the FNR COVID-19 Fast Track Call, one project from the Research Luxembourg COVID-19 Task Force and one collaboration between the University’s Luxembourg Centre for Systems Biomedicine, TU Munich and the Flatiron Institute.
The HPC facility is one element of the extensive digital research infrastructure and expertise developed by the University over the last few years. It also supports the University’s ambitious digital strategy, in particular the creation of a Facility for Data and HPC Sciences. This facility aims to provide a world-class user-driven digital infrastructure and services for fostering the development of collaborative activities related to frontier research and teaching in the fields of computational and data sciences, including high performance computing, data analytics, big data applications, artificial intelligence and machine learning.
More than 1030 jobs were scheduled on the dedicated reservations set by the HPC team (the longest job running for 58 days). An overview of the associated load usage in the most critical period of the pandemic is depicted in Figure 1. The high utilisation rate of the resources during this critical period shows the strong involvement and collaboration of all University partners to fight the pandemic.
The University’s HPC remains determined to provide resources and guidance to current and future COVID-19 related projects.