Finland acquires a new national data management and computing environment that will ensure competitive resources for Finnish researchers. The environment purchased consists of a new supercomputer, data management system and cloud computing capacity. The CSC – IT Center for Science and IT company Eviden sign contract to deliver the new environment.
The new supercomputer called Roihu is a major investment in the national research infrastructure. Roihu will triple Finland’s national supercomputers’ resources and enable future scientific breakthroughs. At the same time, CSC will renew its data management system Allas and acquire more cloud computing capacity.
The value of the whole purchase is almost 30 million euros including maintenance costs. Investing in high-performance computing has been very profitable, as according to a study by Taloustutkimus, an investment of one euro in CSC’s high-performance computing services has generated a benefit to society of 25–37 euros during 2018–2023.
For the needs of Finnish researchers
The environment acquired is designed and built specifically for the needs of researchers at Finnish higher education institutions and research institutes. Supercomputer Roihu provides the majority of Finnish researchers with an efficient and versatile computing environment and ensures the competitiveness of Finnish research.
“Finland has invested in its own national computing environment on a long-term basis. We need to be able to serve all researchers who need data management and computing resources. Computational grand challenges that require very large computing resources will still be solved with the pan-European EuroHPC LUMI supercomputer. Roihu’s mission as a national supercomputer is to complement LUMI and to offer competitive resources to the majority of Finnish researchers,” said Director Pekka Lehtovuori at CSC.
Data becomes increasingly important
The role of data-driven research and high-performance computing is becoming increasingly important. To meet this demand, Roihu’s performance and storage capacity are crucial. Roihu offers significantly more GPU capacity than CSC’s current supercomputers, Mahti and Puhti and its high-performance storage system enables efficient processing of large datasets. The theoretical peak power of Roihu is 49 Pflops (49 quadrillion operations per second).
GPUs are ideal for analyzing large volumes of data and supporting artificial intelligence applications. Most of researchers using CSC AI-related services train small to medium-sized models, for which Roihu is well suited. Roihu also enables machine learning with sensitive or confidential data.
In addition to supporting AI applications, Roihu is highly versatile platform for researchers across various fields. Researchers can use it to analyze audio and video materials, compute atomistic molecular dynamic simulations, screen potential drug compounds from extensive libraries, simulate glaciers and calculate climate scenarios. The CPU capacity of Roihu provides an easy-to-use platform for all disciplines. The educational use of HPC in higher education institutions is a new and growing field. While it doesn’t demand extensive computing resources, it involves a large number of users. This approach helps cultivate the next generation of researchers across all disciplines and prepares skilled experts for industry.
The data management system Allas services the entire environment and its storage capacity will be extended. Allas is intended for project-time storage. To support diverse computational research, cloud computing service Pouta is being modernized. There will be significantly more GPU capacity suitable for the use of ready-made AI services.
“The data management and computing research infrastructures available to everyone doing research in Finland are a great example of long-term investment in research and development. The know-how gained over the decades and modern national research infrastructures have made CSC an important and reliable player, which has enabled the hosting of top European equipment and a strong international position as a research infrastructure for data management and high-performance computing,” said Director Erja Heikkinen at Ministry of Education and Culture.
Modern supercomputer
Roihu is a modern platform that enables more advanced and versatile services and improved information security for researchers. Roihu is built on BullSequana XH3000 hybrid system by Eviden and it replaces CSC’s current supercomputers Mahti and Puhti. Roihu will be located in CSC’s Kajaani data center, and it will be in researchers’ use by the end of the year 2025.
“Roihu is a significant investment in Finland’s national research infrastructure, set to triple the current supercomputing resources and become a major accelerator for scientific breakthroughs. CSC – IT Center for Science is a world-renowned research center, equipped with some of the most powerful systems globally. It is an honor to support them and Finland on their large-scale simulation journey. With our BullSequana DLC technologies, we will support Finnish research while ensuring reduced environmental impacts. Together, we are advancing sustainable HPC AI computation, ensuring Finland remains at the forefront of innovation to leverage AI workloads,” said Bruno Lecointe, VP, Global head of HPC, AI and Quantum Computing, Eviden, Atos Group.
Roihu technical details
Roihu’s total theoretical peak power is 49 Pflops (Peta floating point operations per second) which means 49 quadrillion operations per second
- CPU: 486 nodes, 2*192-core AMD Turin processors each, altogether 186624 cores
- GPU: 132 nodes, 528 Nvidia GH200 GPUs
- Special GPU nodes for visualization and nodes with extended memory.
This article was first published on 19 November by CSC.
Receive the Funding Newswire each Tuesday, our Policy Bulletin each Thursday, and news about bridging Europe’s east-west innovation gap twice a month in The Widening. 
A unique international forum for public research organisations and companies to connect their external engagement with strategic interests around their R&D system.
