Modern research in chemistry, biomedicine and materials science has long relied on complex computations and the analysis of vast datasets. In late October, Kyiv hosted a science communication event dedicated to promising interdisciplinary projects and the official launch of a new computing cluster at Taras Shevchenko National University of Kyiv (KNU). The powerful computing system was successfully established and deployed in its full configuration thanks to grant funding from the National Research Foundation of Ukraine (NRFU). The project, “Addressing Contemporary Issues in Chemistry, Biomedicine, Physics, and Materials Science Using High-Performance Computing and Machine Learning”, was selected for funding under the NRFU Call for Proposals “Infrastructures for Advanced Research”.
How exactly does the new cluster help address complex research challenges? Who will be able to access it, and how? And what are the first findings achieved within the framework of the project? These were among the questions discussed by Ukrainian and international researchers at a two-day conference entitled “Prospects for International Interdisciplinary Research in Chemistry, Biomedicine, Physics and Materials Science”.
Building Computing Capacity for the Future
The project’s PI is Igor Komarov, Deputy Director of the Institute of High Technologies of Taras Shevchenko National University of Kyiv. As Professor Ganna Tolstanova, who initiated the project, noted, Igor Komarov brought together experts from different university departments. His talent and respect for people helped ensure the successful implementation of the initiative and the creation of the most advanced computing cluster in Ukraine.
Ganna Tolstanova recalls that when the NRFU announced its call for proposals for funding research infrastructure projects in 2023, KNU researchers had a clear goal: to establish a modern research infrastructure and deliver tangible practical results. With a budget of UAH 30 million, this Call for proposals was highly competitive for higher education institutions across the country.
The research team focused on high-performance computing and artificial intelligence. The project built on the infrastructure that had been developed at the University since 1999 and brought together chemists, physicists, mathematicians, biologists, IT specialists and nuclear physicists. The cluster was conceived as a shared resource for researchers across a range of scientific disciplines, as well as for start-ups and knowledge-intensive businesses.
Energy efficiency posed a separate challenge. Faced with rising electricity costs and the risk of power outages, the research team began looking for alternative solutions. Alongside the development of the computing centre, they secured funding from the Ministry of Education and Science of Ukraine to install a solar power plant.
The new cluster was deployed at the Information and Computing Centre (ICC) of KNU under the direction of Yuriy Boyko, an associate professor in the Faculty of Radiophysics, Electronics and Computer Systems. The equipment was selected to meet the specific needs of researchers, with some components custom-built. The system is designed to meet the needs of university departments – from physicists to astronomers – as well as to provide computing resources to external users.
From the First Desktops to a Supercomputer: How the KNU’s New Cluster Works
The history of the university’s computing cluster began back in 1999 in response to a practical need among researchers. As Yuriy Boyko explains, a request from the Department of Nuclear Physics prompted the research team to connect ordinary workstations into a single network. The Centre continued to develop, providing researchers across different fields with remote access to computing resources.
New cluster, developed with the grant support of the NRFU, significantly expands the Centre’s capabilities. The technical upgrade required unconventional solutions: much of the equipment was custom-built to meet the researchers’ specific requirements and research needs. Computing power and disk array storage capacity have increased, while modern high-speed data transfer protocols enable more efficient processing of large datasets.
However, ensuring a stable power supply is essential for the system to operate reliably. As Yuriy Boyko pointed out, batteries alone, even high-capacity ones, are insufficient for extended autonomous operation, so the research team is developing a comprehensive solution. A solar power system with an inverter was purchased with funding from the Ministry of Education and Science of Ukraine, while the KNU provided funding for the installation of a diesel generator.
“We can operate on solar power for several hours, which saves money and staff time. We are already using a solar panel in combination with a diesel generator to ensure uninterrupted operation of the KNU’s internal internet network. I hope we will be able to implement the same technology for the computing cluster”, emphasises Yuriy Boyko.
The cluster’s equipment has been installed and is fully operational, as confirmed during a test run. However, the full-scale launch depends on completing the work needed to protect the power supply. Under unstable voltage conditions, fluctuations can damage expensive equipment, in particular sensitive network switching equipment. Therefore, the installation of backup power systems must be completed before the cluster can operate at full capacity.
Another equally important issue is data storage and the use of cloud services. At the beginning of the full-scale invasion, critical university resources were temporarily migrated to the cloud. This helped safeguard them during power outages and internet disruptions. However, storing large volumes of data on foreign platforms on a permanent basis is costly and creates dependence on external service providers.
“Moving to the cloud is easy, getting out is difficult”, Yuriy Boyko noted. Without their own infrastructure and specialists, migrating data and services back to local resources can be challenging. Moreover, hardware and network equipment needs to be located where it is used. The University now has backup equipment distributed across secure locations, while ensuring the energy resilience of the computing centre remains one of its priorities.
The financial aspects of operating the cluster were also considered. According to the specialists’ estimates, running the generator costs around €10 per hour, meaning that 100 hours of autonomous operation would cost approximately €1,000. Electricity and computing service costs can be included in grant budgets. This makes it possible to plan the use of computing resources as an integral part of research projects.
A Supercomputer in the Classroom
Computing infrastructure is important not only for research but also for training future specialists. Ievgen Sliusar, Associate Professor and Head of the Centre for the Collective Use of Scientific Equipment “The High Performance Computing and Machine Learning Centre” at Taras Shevchenko National University of Kyiv, spoke about how this infrastructure is integrated into the KNU’s educational process.
Undergraduate and postgraduate students are given remote access to the cluster, where they learn to distribute tasks, manage parallel computations and work with scientific data. This enables them to gain practical experience with machine-learning algorithms, develop mathematical models and carry out tasks that require substantial computing resources.
This approach helps bridge the gap between theory and practice: students can contribute to research projects while still studying and acquire skills needed both in academic research and in high-tech businesses.
From Cheminformatics to Astrophysics
The second day of the event focused on international interdisciplinary collaboration and the potential of high-performance computing across various fields of science. Participants in the conference discussed how computing power and artificial intelligence methods help address complex research challenges – from discovering new chemical compounds and modelling chemical reactions to developing herbicides, investigating the properties of atomic nuclei and exploring astrophysical phenomena.
Professor Igor Komarov (KNU) shared his experience of international collaboration involving Ukrainian chemists and cheminformaticians. Professor Alexandre Varnek, University of Strasbourg (France), devoted his presentation to the role of chemistry in the development of artificial intelligence, while Olga Tarkhanova, Ph.D., CEO of Chemspace, spoke about methods for exploring colossal chemical spaces. Other presentations covered quantum chemical calculations, bioinformatics approaches to herbicide development, research into the dineutron, the use of computing clusters in astrophysics and the training of large language models.
The establishment of the cluster marks an important step in the development of Ukraine’s research infrastructure. The additional computing power expands researchers’ capabilities, creates opportunities for interdisciplinary research and supports collaboration between Ukrainian and international research teams.
Interviewed by Svitlana Galata