News: 1708397890

  ARM Give a man a fire and he's warm for a day, but set fire to him and he's warm for the rest of his life (Terry Pratchett, Jingo)

Australia has no next-gen HPC investment plan and clouds can't fill the gap

(2024/02/20)


Australia needs an exascale computer system, and a refresh of its current HPC fleet, but lacks a plan or the budget for either – and can't expect cloud providers or quantum computers to offer a suitable substitute for sovereign capacity.

That's the thrust of a [1]brief [PDF] published today by the Australian Academy of Science.

Announced at the SuperComputing Asia conference today in Sydney by the Academy's president, professor Chennupati Jagadish, the brief opens with a bang: "Australia has no national strategy to acquire and sustain state-of-the-art high-performance computing and data (HPCD) – putting the country's future prosperity and security at risk."

[2]

The document, titled "The future computing needs of the Australian science sector," describes Australia's current HPCD capability as "moderate" and suggests existing hardware will need to be replaced by the end of the decade.

[3]

[4]

It also suggests that Australia will need "at least one" exascale capacity, "to secure Australia's sovereign capability and enable science and research to meet national and regional priorities."

The brief dismisses cloud providers as a potential source of resources.

[5]

"Cloud computing is useful for smaller scale research projects but is unsuitable for large scale research or those using sensitive data," the brief states, adding: "Commercial cloud services are not financially viable for researchers with large datasets or complex requirements. Using commercial services also has risks for securing intellectual property."

[6]A visa to fill Australia's empty tech jobs is getting more expensive, but maybe better value

[7]Australian Tax Office probed 150 staff over social media refund scam

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

[9]Supercomputer pinpoints exact origin of 'Black Beauty' meteorite from Mars

Quantum computing, a field in which Australia punches above its weight, is also dismissed as an alternative to more big iron. The tech is described as having "potential as a future complement to HPCD, offering unique capabilities for certain types of complex problems."

"However, its practical development, integration and full realization depend on utilizing HPCD infrastructure. Currently, quantum computing cannot replace all functions of traditional HPCD and requires further development and integration efforts before it can be considered a viable technology for Australia's computing capability."

The document also warns that without an investment in GPU-based systems, "Australia's AI capability will be entirely dependent on other nations."

The brief notes that the US, Japan, China, and the UK have allocated funds to build exascale systems, and that some European Union members "are already planning for post-exascale computing."

[10]

Australia, the document argues, needs to make investment in HPCD a priority – and may be able to do so by pitching itself as a regional hub for such infrastructure.

But failure to do something will, the Academicians warn, harm Australia's science community and its economy. ®

Get our [11]Tech Resources



[1] https://www.science.org.au/files/userfiles/support/evidence/2024/brief-1-hpcd-in-australia.pdf

[2] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_onprem/hpc&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=2&c=2ZdQx980SVtuT7XcQwnWgGwAAAQo&t=ct%3Dns%26unitnum%3D2%26raptor%3Dcondor%26pos%3Dtop%26test%3D0

[3] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_onprem/hpc&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=4&c=44ZdQx980SVtuT7XcQwnWgGwAAAQo&t=ct%3Dns%26unitnum%3D4%26raptor%3Dfalcon%26pos%3Dmid%26test%3D0

[4] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_onprem/hpc&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=3&c=33ZdQx980SVtuT7XcQwnWgGwAAAQo&t=ct%3Dns%26unitnum%3D3%26raptor%3Deagle%26pos%3Dmid%26test%3D0

[5] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_onprem/hpc&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=4&c=44ZdQx980SVtuT7XcQwnWgGwAAAQo&t=ct%3Dns%26unitnum%3D4%26raptor%3Dfalcon%26pos%3Dmid%26test%3D0

[6] https://www.theregister.com/2024/02/19/australia_tech_visa_job_visa_application/

[7] https://www.theregister.com/2024/02/14/ato_operation_protego_tax_scam/

[8] https://www.theregister.com/2020/10/19/new_pawsey_supercomputer/

[9] https://www.theregister.com/2022/07/14/black_beauty_mars_rock_origin/

[10] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_onprem/hpc&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=3&c=33ZdQx980SVtuT7XcQwnWgGwAAAQo&t=ct%3Dns%26unitnum%3D3%26raptor%3Deagle%26pos%3Dmid%26test%3D0

[11] https://whitepapers.theregister.com/



Easy decision

HuBo

NCI and CSIRO could definitely benefit from access to Exaflopping HPC capacity. The US ORNL's Frontier's been out for 2 years already and that it remains the only offical Exaflopper on Top500, for this long(!), is quite unacceptable IMHO. Australia's Pawsey Centre Setonix is like a mini-Frontier (3rd gen EPYC 64C + MI250x), but 27 PF/s is 40x too doggone slow. Canada's even worse off though, with the Robert and Underhill twins at 8 PF/s (need at least 100x speedup).

At $10 per person ($300 million) Exaflopping machines are just about commodity hardware today for governing bodies, and any country with 30 million people should either already have one, or have plans to get one within the next 2 years in my opinion. AMD's MI300A will make this an even easier decision (or Nvidia's GraceHoppers if you insist). Falling further behind Russia's Yandex should not be an option!

The wise shepherd never trusts his flock to a smiling wolf.