Chalmers University of Technology is utilizing Lenovo and Nvidia’s expertise infrastructure to energy its large-scale pc useful resource, or supercomputer, Alvis.
The challenge has seen the supply and implementation of a clustered computing system for synthetic intelligence and machine studying analysis, in what’s Lenovo’s largest high-performance computing cluster for AI and ML within the Europe, Middle East and Africa area.
Alvis – outdated Norse which means “all-wise” or “all-knowing” – is a national supercomputer useful resource throughout the Swedish National Infrastructure for Computing (SNIC).
It started initially in 2020 and has since developed to maintain a capability that solves bigger analysis duties on a broader scale. Financed by the Knut and Alice Wallenberg Foundation, the pc system is equipped by Lenovo and positioned at Chalmers University of Technology in Gothenburg, house to the EU’s largest analysis initiative, Graphene Flagship.
The collaborative challenge permits any Swedish researcher who wants to enhance their mathematical calculations and fashions to benefit from Alvis’ providers by means of SNIC’s software system, whatever the analysis area.
This helps researchers who’re already utilising machine studying to analyse complicated issues, and those that are investigating the usage of machine studying to clear up points inside their respective area, with the potential to lead to ground-breaking educational analysis in fields corresponding to quantum computing and data-driven analysis for healthcare and science.
Noam Rosen, EMEA director, HPC and AI at Lenovo Infrastructure Solutions Group, says: “The Alvis challenge is a primary instance of the position of supercomputing in serving to to clear up humanity’s best challenges, and Lenovo is each proud and excited to be chosen as a part of it.
“Supported by Lenovo’s efficiency main expertise, Alvis will energy analysis and use machine studying throughout many numerous areas with a serious influence on societal growth, together with environmental analysis and the event of prescribed drugs.
“This computing useful resource is actually distinctive, constructed on the premise of architectures for various AI and machine studying workloads with sustainability in thoughts, serving to to save power and cut back carbon emissions through the use of our pioneering heat water-cooling expertise.”
Sverker Holmgren, director of Chalmers e-Infrastructure Commons, internet hosting the Alvis system, says: “The first pilot useful resource for Alvis has already been utilized by greater than 150 analysis initiatives throughout key Swedish universities.
“By making it a lot bigger, and totally opening the Alvis Systems to all Swedish researchers, Chalmers and Lenovo are taking part in an necessary position in offering a national HPC ecosystem for future analysis.”
Chalmers has chosen to implement a scalable cluster with quite a lot of Lenovo ThinkSystem servers to ship the right combination of NVIDIA GPUs to its customers in a means that prioritises power financial savings and workload steadiness.
This consists of the Lenovo ThinkSystem SD650-N V2 to ship the facility of Nvidia A100 Tensor Core GPUs, and the Nvidia-Certified ThinkSystem SR670 V2 for Nvidia A40 and T4 GPUs.
Rod Evans, EMEA director of high-performance computing at Nvidia, says: “The work we’re doing with Chalmers University and its Alvis national supercomputer will give researchers the facility they want to simulate and predict our world.
“Together, we’re giving the scientific neighborhood instruments to clear up the world’s best supercomputing challenges – from forecasting climate to drug discovery.”
The storage structure delivers a brand new Ceph resolution with 7.8 petabytes, to be built-in into the prevailing storage setting at Chalmers.
Nvidia Quantum 200 Gb/s InfiniBand offers the system with low- latency, excessive knowledge throughput networking and good in-computing acceleration engines.
With these high-speed infrastructure capabilities, customers have virtually 1000 GPUs, primarily Nvidia A100 Tensor Core, together with over 260,000 processing cores and over 800 TFLOPS of compute energy to drive a sooner time to reply of their analysis.
In addition, Alvis leverages Lenovo’s Neptune liquid-cooling applied sciences to ship unparalleled compute effectivity. Initially, full air cooling was proposed for the challenge, however Chalmers as a substitute determined to deploy Lenovo Neptune heat water-cooling capabilities to cut back long-term operational prices and end in a “greener” AI infrastructure system.
As a outcome, the college anticipates that there might be important power financial savings thanks to efficiencies by means of water cooling.
The HPC resolution has been in manufacturing since February 2022 and might be totally operational by Summer 2022.
You may also like…
https://roboticsandautomationnews.com/2022/06/19/chalmers-university-selects-lenovo-and-nvidia-to-build-swedish-national-supercomputer/51579/
