HPE called in to overhaul Swedish university’s supercomputing setup
The Swedish National Infrastructure for Computing has commissioned HPE to build a supercomputer for one of Sweden’s largest technical universities
The Swedish National Infrastructure for Computing (SNIC) has commissioned Hewlett Packard Enterprise (HPE) to build a replacement supercomputer for the KTH Royal Institute of Technology in Stockholm, Sweden.
The Institute is one Sweden’s largest technical universities, and the supercomputer will be used to support the science and engineering research work it does to advance the development of new innovations in the life sciences, automotive, transportation and energy industries.
The supercomputer will have a theoretical peak performance of 13.5 petaflops, and will be housed within the PDC Center for High Performance Computing at the KTH campus in Stockholm.
It will also go by the name of Dardel, which is the surname of noted Swedish novelist Thora Dardel and her first husband, Nils Dardel, who was a post-impressionist painter.
The supercomputer will consist of a large central processing unit (CPU) partition that will be used by academic researchers to run computational applications, as well as another partition housing Graphics Processing Units (GPU) to support compute-heavy workloads too.
“In recent times, we have seen a dramatic increase in the extent to which researchers need to use accelerators, mainly in the form of GPUs. We will soon be able to meet that demand through the accelerator partition in Dardel,” said SNIC director Professor Hans Karlsson. “Dardel will significantly increase the Swedish capacity for research that requires access to large-scale computational resources.”
The deployment will be based on the HPE Cray EX supercomputer technology and come equipped with AMD EPYC processors, AMD Instinct graphics processing unit (GPU) accelerators and will replace the KTH’s existing system, named Beskow.
Read more about HPC
- With the evolution of HPC over the past 30 years, the huge amounts of information inherent in living systems can be treated with the same industrial efficiency as enjoyed by the engineering, financial and media fields.
- Nissan Motor Company has committed to moving its on-premise HPC-based performance and engineering workloads to the Oracle Cloud Infrastructure.
HPE said it expects the first phase of the supercomputer’s deployment, which will include the buildout of the CPU partition, to be ready for use by July 2021. The second phase, meanwhile, should be installed later this year in anticipation of a go-live date in January 2022.
“The HPE Cray EX supercomputer will include HPE Slingshot for purpose-built HPC networking to address demands for higher speed and congestion control for data-intensive workloads,” said HPE in a statement.
“It will also feature next generation AMD EPYC processors and AMD Instinct GPU accelerators to improve efficiency and achieve the performance required to process and harness insights from computationally complex data.”
The SNIC is investing 100m SEK to bring the project to fruition and has also committed to covering the cost of its operations for at least the next five years.
Peter Ungaro, senior vice president and general manager for HPC and mission critical solutions at HPC, said the supercomputer’s deployment will bolster the ability of Swedish firms to innovate.
“High performance computing (HPC) technologies are broadly used in R&D to advance the technologies that go into the products and services we use each day,” said Ungaro. “For decades, the PDC Center for High Performance Computing at the KTH Royal Institute of Technology has empowered a community of researchers to make breakthroughs across a range of industries using HPC.”
“We are honoured to have been selected by the KTH Royal Institute of Technology to deliver the latest HPE Cray EX supercomputer with next-generation AMD technologies and provide an even greater level of performance to improve and expand Swedish academic research and drive innovation for Swedish industrial companies.”