HPC

Toodle-oo Raijin and g'day Gadi, you beauty! Australia's fastest super 'puter will bench 38 PFLOPS later this year

HPC beaut built by Fujitsu to debut in November


Australia's most powerful supercomputer is - we are told - about to get 10 times faster in November, thanks to an AU$70m (£39.4m) cash injection from the government.

The new machine, called Gadi – meaning "to search for" in the language of the Ngunnawal people of Canberra - will replace Raijin, which entered operation in 2013.

Gadi will comprise 3,200 nodes based on second-generation Intel Xeon Scalable processors and featuring V100 Nvidia GPUs on machine learning duties. The nodes will be interconnected using 200Gbps Mellanox HDR InfiniBand fabric, and storage will be provided by NetApp, backed by DDN's distribution of Lustre software.

Intel's Optane DC persistent memory is also a part of the design, and Altair will provide scheduling and workload management software, while Schneider Electric's APC will presumably handle power protection.

The supercomputer will depend on liquid cooling, using warm water-based, direct-to chip cooling systems from both Fujitsu and Lenovo.

"Gadi is funded through the Morrison Government's $2.2bn Research Infrastructure Investment Plan and will help Australia break into the top 30 nations for its high-performance computing capacity,” said Dan Tehan, Australia’s Minister for Education.

No exact numbers on performance have been provided by the machine’s owner, Australia's National Computational Infrastructure (NCI) agency, but since it has stated Gadi will be about ten times faster than Rajin, which has a theoretical peak performance of 3,801.4 TFLOPS, we can estimate that the new hardware will bench a tad over 38 PFLOPS.

Just like its predecessor, Gadi will be delivered by Fujitsu, and its resources will be shared with the industry and academia.

Raijin was once considered to be the most powerful supercomputer in the southern hemisphere, equipped with 57,472 Sandy Bridge cores and 160 TB of memory at launch. The hardware was upgraded multiple times, most recently in 2017, when it received a dollop of IBM’s Power8-based nodes.

But the old design couldn’t keep up with the youngsters: Raijin was rated as the 76th fastest system in the world in 2017; by July this year, it had slid down to 173rd place.

Gadi will live at NCI’s high-performance computing and data services facility, located on the campus of the Australian National University (ANU) in Canberra.

NCI's infrastructure is used by 40 universities, five national science agencies, three medical research institutes, and industrial partners.

Professor Brian Schmidt, the astroboffin who won a Nobel for the accelerating universe expansion discovery, and vice-chancellor of the ANU, said: "Gadi will give researchers the tools to unlock the mysteries of the universe, predict and manage natural disasters, advance cancer research and design new materials for future technologies.

"This new machine will keep Australian research and the 5,000 researchers who use it at the cutting edge. It will help us get smarter with our big data. It will add even more brawn to the considerable brains already tapping into NCI."

Gadi might be new and expensive, but it's not very impressive by international standards: the fastest computer in operation today, Oak Ridge National Labs' Summit, has a theoretical peak performance of around 200 PFLOPS – more than five times that of the new Australian flagship.

China, the US, Japan and the EU are currently locked in a race to produce the first exascale machine – a supercomputer with peak performance of 1,000 PFLOPS. ®

Broader topics


Other stories you might like

  • DuckDuckGo tries to explain why its browsers won't block some Microsoft web trackers
    Meanwhile, Tails 5.0 users told to stop what they're doing over Firefox flaw

    DuckDuckGo promises privacy to users of its Android, iOS browsers, and macOS browsers – yet it allows certain data to flow from third-party websites to Microsoft-owned services.

    Security researcher Zach Edwards recently conducted an audit of DuckDuckGo's mobile browsers and found that, contrary to expectations, they do not block Meta's Workplace domain, for example, from sending information to Microsoft's Bing and LinkedIn domains.

    Specifically, DuckDuckGo's software didn't stop Microsoft's trackers on the Workplace page from blabbing information about the user to Bing and LinkedIn for tailored advertising purposes. Other trackers, such as Google's, are blocked.

    Continue reading
  • Despite 'key' partnership with AWS, Meta taps up Microsoft Azure for AI work
    Someone got Zuck'd

    Meta’s AI business unit set up shop in Microsoft Azure this week and announced a strategic partnership it says will advance PyTorch development on the public cloud.

    The deal [PDF] will see Mark Zuckerberg’s umbrella company deploy machine-learning workloads on thousands of Nvidia GPUs running in Azure. While a win for Microsoft, the partnership calls in to question just how strong Meta’s commitment to Amazon Web Services (AWS) really is.

    Back in those long-gone days of December, Meta named AWS as its “key long-term strategic cloud provider." As part of that, Meta promised that if it bought any companies that used AWS, it would continue to support their use of Amazon's cloud, rather than force them off into its own private datacenters. The pact also included a vow to expand Meta’s consumption of Amazon’s cloud-based compute, storage, database, and security services.

    Continue reading
  • Atos pushes out HPC cloud services based on Nimbix tech
    Moore's Law got you down? Throw everything at the problem! Quantum, AI, cloud...

    IT services biz Atos has introduced a suite of cloud-based high-performance computing (HPC) services, based around technology gained from its purchase of cloud provider Nimbix last year.

    The Nimbix Supercomputing Suite is described by Atos as a set of flexible and secure HPC solutions available as a service. It includes access to HPC, AI, and quantum computing resources, according to the services company.

    In addition to the existing Nimbix HPC products, the updated portfolio includes a new federated supercomputing-as-a-service platform and a dedicated bare-metal service based on Atos BullSequana supercomputer hardware.

    Continue reading

Biting the hand that feeds IT © 1998–2022