Professor Paolo Bientinesi's was born in Italy and was installed as professor at Umeå University in 2019. His research focuses on mathematical computations that arise in disciplines such as biology, chemistry, data science and materials science, and aims at achieving simultaneously both human productivity and computer efficiency.
“I am thrilled to join HPC2N, one of the six Swedish HPC centers within SNIC, which offers both large-scale computing and large-scale storage to the entire nation. In the immediate future, I want to increase the user base to include scientific disciplines, such as computational medicine, forestry, and atmospheric research, which can benefit tremendously from high-performance resources and algorithms,” says Paolo Bientinesi.
“For more than 20 years, HPC2N has provided Swedish scientists with both hardware and human support. Our expertise covers all the paradigms and tools for parallel programming, and extends into computational areas such as machine learning, materials science, and computational chemistry. Our aim is to expand our computing and storage solutions to enable Swedish science at unprecedented scales,” says Paolo Bientinesi.
25 years of high-performance computing
Professor Bo Kågström has been the Director of HPC2N since the inauguration of the center 25 years ago (1996). As one of six national centers funded by the Swedish National Infrastructure for Computing (SNIC), the center in Umeå has contributed to an increasing development of high-performance computing in Sweden.
“It has been a challenging, formative and very fruitful journey to be responsible for the long-term development of HPC2N. Of course, this has only been possible with important contributions from our very competent staff at HPC2N and collaboration with our "Parallel and Scientific Computing" research group at the Department of Computing Science. This development has also created a strong network at national, Nordic and international level,” says Bo Kågström.
“I also want to gratefully acknowledge for the strong support we have had during these years from the leadership of Umeå University and our faculty, as well as the generous funding of computing resources at HPC2N by The Kempe Foundations and the Knut and Alice Wallenberg Foundation and not the least the grants for large scale HPC resources from SNIC, which we hope and anticipate will continue,” Bo Kågström continues.
“I will continue as Senior Professor at the Department of Computing Science focusing on our research group and my engagement in eSSENCE, a strategic research programme and collaboration in e-Science between Lund, Uppsala and Umeå universities. Besides, I will still be involved in HPC2N as an advisor for future endeavors,” says Bo Kågström.
Raising the national level of high-performance computing
Users of HPC2N cover virtually all major computational disciplines, including chemistry, physics, life sciences, as well as social sciences and humanities. HPC2N also have an increasing number of users in Artificial Intelligence and Machine Learning applications that also require large storage resources. The applications are as diverse as the design of nanomaterials, market simulations, analysis of decision-making processes, deep learning for protein prediction, and weather forecasting.
The primary objective of HPC2N is to raise the national level of competence in high-performance computing (HPC) and to transfer HPC knowledge and technology to new users in academia and industry. Today, the use of HPC includes compute-intensive as well as data- and communication-intensive applications.
One important task of HPC2N is to supply adequate scalable HPC and VR/visualization resources to meet different and growing needs in the Swedish scientific community, especially in Northern Sweden. The center takes a lead in developing efficient computational methods, application parallel software and user-friendly environments for state-of-the-art scalable computer systems. A goal is also to provide and further develop an international research environment for scientific and parallel computing in various scientific disciplines and application areas.
Paolo Bientinesi Paolo Bientinesi was born in 1973 in Livorno, Italy. He studied computer science at the University of Pisa, and at The University of Texas at Austin, where he received his Ph.D. in 2006. In 2007-2008, he was a research associate at Duke University, and between 2008 and 2018, he was professor at RWTH Aachen University (Germany).
His research interests are centered on computer automation, and range from numerical linear algebra to computer music. He was the recipient of the Karl Arnold Prize from the North Rhine-Westphalian Academy of Sciences and Humanities.
Bo Kågström Bo Kågström was appointed Full Professor of Numerical Analysis and Parallel Computing at Umeå University in 1989 and Director of Super Computer Center North (SDCN). Since 1996 he has been the Director of High Performance Computing Center North (HPC2N). He was also Head of the Department of Computing Science (former Information Processing) during 15 years (1983 – 1998). Appointments include Member of the Royal Swedish Academy of Engineering Sciences (IVA), Division of Information Technology since 1992 (Vice Chair 2003-05), Fellow of the Society of Industrial and Applied Mathematics (SIAM, 2016). Moreover, expert in the European Exascale Software Initiative (EESI), panel chair and coordinator for international evaluation of INRIA Theme: Distributed and High Performance Computing. Assignments also encompass periods as Visiting Scientist at MIT, Stanford and Cornell. Awards and prizes include Nordbanken Scientific Prize 2000, SIAM / SIAG Linear Algebra Prize 2000, and prestigious funding from EC Horizon 2020 Future Emerging Technology Program for the project “Parallel Numerical Linear Algebra for Future Extreme Scale Systems” (www.nlafet.eu).
Bo Kågström has over 150 scientific publications, and he has been invited speaker at several international conferences, workshops and universities worldwide. Principal advisor and supervisor of 14 PhD Theses, 16 Licentiate Theses and several Master Theses. Organizer and co-organizer of more than ten international conferences.
High Performance Computing Center North (HPC2N) High Performance Computing Center North (HPC2N, www.hpc2n.umu.se) is a national center for Scientific and Parallel Computing. It is located in Umeå, North of Sweden, and was founded in 1996 as one of the national supercomputing centers in Sweden.
HPC2N is mainly funded by Swedish National Infrastructure for Computing (SNIC) and the partners Umeå University, Luleå University of Technology, The Swedish Institute of Space Physics (IRF) in Kiruna, Swedish University of Agricultural Sciences (SLU) in Umeå, and Mid-Sweden University in Sundsvall, Härnösand, Örnsköldsvik and Östersund. The Kempe Foundations and the Knut and Alice Wallenberg Foundation have generously funded computing resources at HPC2N.
Kebnekaise is the latest supercomputer at HPC2N. It is named after the massif of the same name, which has some of Sweden's highest mountain peaks (Sydtoppen and Nordtoppen). Just as the massif, the supercomputer Kebnekaise is a system with many faces.
Kebnekaise was delivered by Lenovo and installed in three phases during 2016 - 2018. The first phase of Kebnekaise was opened up for general availability on November 7, 2016. In the third phase, 2018, Kebnekaise was extended with 52 Intel Xeon Gold 6132 (Skylake) nodes, as well as 10 NVidia V100 (Volta) GPU nodes. Kebnekaise is a heterogeneous resource with standard nodes, GPUs and nodes with a lot of memory, to meet the divergent needs of many different research groups. At installation, Kebnekaise was the most powerful GPU system in Sweden. After being expanded twice, Kebnekaise will be ready for replacement within the next two years.
Kebnekaise in numbers 602 nodes 15 racks 19288 cores (of which 2448 cores are KNL-cores) 18840 available for users (the rest are for managing the cluster) 501760 CUDA cores (80 * 4992 cores/K80 + 20 * 5120 cores/V100) 12800 Tensor cores (20 * 640 cores/V100) More than 136 TB memory (20*3TB + (432 + 36) * 128GB + (52 + 10) * 192 GB + 36 * 192GB) 71 switches (Infiniband, Access and Managment networks) 984 TFlops/s Peak performance 791 TFlops/s HPL HPL: 80% of Peak performance