Bright Cluster Manager has been selected as the enterprise-grade cluster management platform, which is certified to run on DGX systems as part of the NVIDIA DGX-Ready Software program.
AMSTERDAM (PRWEB) May 04, 2021
Bright Computing, a global leader in automation and management software for edge-to-core-to-cloud high-performance computing, today announced that Bright Cluster Manager has been selected to support the new artificial intelligence supercomputer to be installed at National Supercomputer Centre (NSC) at Linköping University.
Named Berzelius after Swedish scientist Jacob Berzelius, NSC’s supercomputer is set to deliver 300 petaflops of AI performance and will be based on the NVIDIA DGX SuperPOD™ computing infrastructure. Berzelius will consist of 60 NVIDIA DGX™ A100 systems combined with NVIDIA HDR 200Gb/s InfiniBand networking and 1.5 petabytes of storage from DDN. The supercomputer will be supported by Atos Codex AI Suite, enabling NSC’s researchers to use machine learning and analytics to speed up processing times on their complex data and unlock faster insights.
Bright Cluster Manager has been selected as the enterprise-grade cluster management platform, which is certified to run on DGX systems as part of the NVIDIA DGX-Ready Software program. This will allow NSC and Atos to quickly build and deploy Berzelius from bare metal with full support from Bright’s highly experienced engineering team. Previously, NSC developed and operated its own cluster management software stack inhouse.
Bright software will bring several benefits to NSC:
- Simplified administration of the NVIDIA DGX SuperPOD system, freeing system administrators from day-to-day tasks, ensuring that the system stays up to date and that all upgrades are carried out seamlessly and with no impact to service.
- It will reduce risk. Berzelius is a high profile project, funded by the Knut and Alice Wallenberg Foundation, so it’s of great importance that the platform runs seamlessly and manages multiple AI projects simultaneously and with ease.
- It will enable NSC to introduce new best practices to optimize the supercomputer environment and provide a system that is both flexible and reliable for researchers.
Niclas Andersson, Technical Manager at NSC, explained: “We have been evaluating Bright for some time, so it’s great that the new supercomputer has given us the opportunity to try it out. We believe that Bright will ensure the stability and efficiency of Berzelius, allowing our researchers to carry out important AI research projects.”
Terry Rush, Sales Director at Bright Computing, added: “This is an important project for NSC and Sweden as a whole. Berzelius will be at the heart of Swedish AI and machine learning groundbreaking research, and Bright is delighted to be providing the technology that underpins this exciting supercomputer.”
Berzelius will be the largest supercomputer in Sweden that focuses on AI and machine learning research. It will be used by a broad range of researchers, in areas such as computational fluid dynamics, materials science, physics and chemistry, life sciences, and meteorology/climatology.
For more information about Bright Computing and the NSC at Linköping University, please email email@example.com.
About Bright Computing
Bright Computing software helps organizations manage complex, distributed computing environments that combine on-premises systems with public cloud and edge computing to unlock the potential of real-time machine learning for their business. Designed with platform independence in mind, Bright Computing’s automation software supports modern GPU-accelerated systems and reduced-power Arm-based architectures in conjunction with traditional x86-based systems, maximizing flexibility and eliminating complexity. Trusted by industry leaders of the Fortune 500 and leading data center equipment manufacturers to automate building and managing distributed high-performance clusters, Bright Computing enables enterprises to quickly and efficiently establish AI-ready and IoT-ready infrastructure to support their digital transformation.