AMD scores 200,000 cores worth of secret silicon at new Australian supercomputer

Chips will power $34m machine doing Square Kilometre Array work and more

Australia’s Pawsey Supercomputing Centre has bet on unannounced AMD Epyc silicon to power a new 50-petaFLOP supercomputer.

The centre refused to reveal clock speeds or other details of the silicon to us but did say that it will comprise over 1,600 dual-socket nodes and utilise over 200,000 cores. Those figures yield a per-machine core count of 125. We're willing to round that up to 128 and suggest that the centre is betting on the forthcoming Milan edition of AMD’s Epyc processors, which is widely expected to debut late in 2020. Today's Rome generation already reaches 64 cores per CPU, so if Milan goes there, too, it won’t be a surprise.

Pawsey executive director Mark Stickells told The Register the silicon is “an AMD architecture that is not yet available but is the same as being installed in other exascale labs.”

The new ‘super will use HPE’s Cray EX architecture, pack over 750 GPUs, 548TB of RAM and put HPE’s Slingshot interconnect tech to work to pump at least 200Gb/sec of data into compute nodes and 800Gb/sec into GPU-equipped nodes.

The machine will supplant Pawsey’s current supers, Magnus and Galaxy, yet have a smaller footprint: it will occupy eight cabinets compared to the 11 housing its predecessors.

Square Kilometre Array precursor shrinks 5TB of data to 22MB – every second!


Installation will take place in two phases, with two racks worth to be installed in Q3 2021 before the rest follows by Q2 2022.

The Pawsey Centre is located in Western Australia, a state that has currently closed its borders to both the rest of Australia and the world due to the coronavirus pandemic. HPE told The Register that even if international borders are closed, it has local staff who can install the new super.

Among the tasks the new machine will perform is data ingestion for the Australian Square Kilometre Array Pathfinder, which is also located in Western Australia.

Pawsey executive director Stickells told us the new machine will be welcomed by local boffins, as Magnus has been more than doubly oversubscribed in 2020 and Australian scientists are as keen as any others to do some heavy data crunching.

The new machine has been budgeted at AU$48m (US$34m). ®

Similar topics

Narrower topics

Other stories you might like

  • VMware claims 'bare-metal' performance from virtualized Nvidia GPUs
    Is... is that why Broadcom wants to buy it?

    The future of high-performance computing will be virtualized, VMware's Uday Kurkure has told The Register.

    Kurkure, the lead engineer for VMware's performance engineering team, has spent the past five years working on ways to virtualize machine-learning workloads running on accelerators. Earlier this month his team reported "near or better than bare-metal performance" for Bidirectional Encoder Representations from Transformers (BERT) and Mask R-CNN — two popular machine-learning workloads — running on virtualized GPUs (vGPU) connected using Nvidia's NVLink interconnect.

    NVLink enables compute and memory resources to be shared across up to four GPUs over a high-bandwidth mesh fabric operating at 6.25GB/s per lane compared to PCIe 4.0's 2.5GB/s. The interconnect enabled Kurkure's team to pool 160GB of GPU memory from the Dell PowerEdge system's four 40GB Nvidia A100 SXM GPUs.

    Continue reading
  • Nvidia promises annual datacenter product updates across CPU, GPU, and DPU
    Arm one year, x86 the next, and always faster than a certain chip shop that still can't ship even one standalone GPU

    Computex Nvidia's push deeper into enterprise computing will see its practice of introducing a new GPU architecture every two years brought to its CPUs and data processing units (DPUs, aka SmartNICs).

    Speaking on the company's pre-recorded keynote released to coincide with the Computex exhibition in Taiwan this week, senior vice president for hardware engineering Brian Kelleher spoke of the company's "reputation for unmatched execution on silicon." That's language that needs to be considered in the context of Intel, an Nvidia rival, again delaying a planned entry to the discrete GPU market.

    "We will extend our execution excellence and give each of our chip architectures a two-year rhythm," Kelleher added.

    Continue reading
  • Now Amazon puts 'creepy' AI cameras in UK delivery vans
    Big Bezos is watching you

    Amazon is reportedly installing AI-powered cameras in delivery vans to keep tabs on its drivers in the UK.

    The technology was first deployed, with numerous errors that reportedly denied drivers' bonuses after malfunctions, in the US. Last year, the internet giant produced a corporate video detailing how the cameras monitor drivers' driving behavior for safety reasons. The same system is now apparently being rolled out to vehicles in the UK. 

    Multiple camera lenses are placed under the front mirror. One is directed at the person behind the wheel, one is facing the road, and two are located on either side to provide a wider view. The cameras are monitored by software built by Netradyne, a computer-vision startup focused on driver safety. This code uses machine-learning algorithms to figure out what's going on in and around the vehicle.

    Continue reading

Biting the hand that feeds IT © 1998–2022