cherezoff - stock.adobe.com
Efforts to upgrade the storage and compute capabilities of the Leibniz Supercomputing Centre’s (LRZ) SuperMUC-NG high-performance computing (HPC) setup are set to begin later this year, with the Lenovo Infrastructure Solutions Group and Intel to continue their involvement in the project.
The SuperMUC-NG is currently ranked the 15th most powerful supercomputing entity in the world, and is set to receive a number of enhancements to bolster its ability to serve the LRZ user community so that it can handle artificial intelligence and complex data analytics workloads more easily.
Lenovo and Intel were brought in to work on the project back in December 2017, confirming at the time that they had joined forces with LRZ to support its bid to create a “next-generation” supercomputer for use by the Bavarian Academy of Sciences in Munich, Germany.
Pegged for completion in 2018, the SuperMUC-NG was initially configured to deliver 26.7 petaflops of compute capacity powered by nearly 6,500 Lenovo server nodes and Intel’s Xeon Platinum processor technology.
Since completing the first phase of the project, the water-cooled SuperMUC-NG has been used to carry out a range of tasks, including traditional simulation and modelling workloads, as well as automating image and pattern recognition in planet observations and processing climate data from satellites.
“Given the successful utilisation of SuperMUC-NG in these projects, the demand for high-performance data analytics, machine learning and fast memory performance has further increased,” said Lenovo in a statement.
To meet this demand, the capabilities of the SuperMUC-NG computer are to be expanded through a second-phase buildout of the supercomputer, which is being funded by the Free State of Bavaria and the German Federal Ministries of Education and Research.
This work will see the SuperMUC-NG equipped with Intel’s next-generation Xeon Scalable processors and the chipmaker’s upcoming HPC graphics processing unit (GPU) integrated with Lenovo’s ThinkSystem platform technology. This is expected to happen in the second quarter of 2022.
Read more about HPC and supercomputing environments
- Gamers are being urged to donate their idle PC compute power to a US-based startup that has set its sights on creating a distributed computing network that it hopes will go on to become the world’s largest supercomputer.
- The Swedish National Infrastructure for Computing has commissioned Hewlett Packard Enterprise to build a replacement supercomputer for the KTH Royal Institute of Technology in Stockholm, Sweden.
As a result of these changes, it is predicted that its compute nodes will deliver four times the performance per watt than the phase one setup is currently capable of.
Other phase two enhancements will include the introduction during the final quarter of 2021 of a one petabyte asynchronous object storage system into the setup so it can accommodate a faster throughput of large data volumes.
The company has also confirmed that all the work completed during phase two of the project will have a sustainability slant to it, as the objective is to bolster the supercomputer’s performance in a environmentally friendly way.
As such, it will continue to make use of Lenovo’s liquid-cooling system, in the form of its Neptune direct water-cooling technology, which reportedly removes about 90% of the heat generated by the compute system.
“Delivering resources and services that empower researchers to accelerate their projects is at the heart of everything we do at LRZ,” said Dieter Kranzlmüller, LRZ director and professor.
“With Lenovo’s warm-water cooling technology, we are able to deliver these enhancements in a way that is as sustainable and energy-efficient as possible.”