New Zealand weather and research bodies are to get three new Cray supercomputers under a six-year $31.7m contract announced today by science and innovation minister Paul Goldsmith.
New Zealand eScience Infrastructure (NeSI), a collaborative partnership between the University of Auckland, National Institute of Weather and Atmosphere (NIWA), the University of Otago, and Landcare Research, said the new supercomputers would replace its national compute and data platform, extend its scope and performance to meet changing needs. The new supercomputers will be commissioned by early 2018.
NeSI director, Nick Jones, detailed plans for the upgrade in the organisation’s March newsletter, saying: “Early in 2016 NeSI started preparations to replace both the NIWA IBM P575/POWER6 and Auckland IBM iDataPlex x86 platforms, selecting a single procurement process with NIWA as the lead agent working alongside NeSI’s other collaborators.”
Jones said funding would come from the NeSI Investors: MBIE, NIWA, University of Auckland, University of Otago and Landcare Research. NeSI describes itself as “a virtual organisation developing advanced computation and analytics capabilities across the science sector and delivering underpinning computing and data infrastructure.”
The largest supercomputer, a Cray XC50 will be situated in the High Performance Computing Facility at NIWA’s Greta Point campus in Wellington. It has a theoretical peak performance equivalent to 1.4 trillion calculations per second. It will replace the IBM P575/POWER6, which, NIWA said, has reached the end of its operating life and cannot keep up with demand.
NIWA said it would invest $18m in the facility and a further $4.8m would be spent on behalf of the Universities of Auckland and Otago and Landcare Research for a Cray CS400 to replace the IBM iDataPlex x86s. No details have been provided on where the balance of the $31.7m figure cited in NeSI’s announcement will come from.
The third supercomputer (model not specified) will be housed at the University of Auckland’s Tamaki data centre and will act as a back-up system for NIWA.
NIWA said the three supercomputers together would have up to 13 times the computing capability and more than six times the storage capacity of their predecessors. The CX50 has a theoretical peak performance of 1.425 PetaFLOPS—equivalent to 1.425 trillion calculations per second.
“Altogether, the new supercomputers will provide more than 33,500 compute cores—equivalent to about 16,000 laptops—and enough storage to hold 3 million DVDs online and 8 million DVDs offline,” NIWA said. Together the will deliver up to 12 times the computing capability and more than four times the storage capacity of their predecessors, and have a combined weight of 23 tonnes.
NIWA said it had signed the contract with Cray on behalf of the other NeSI partners. NIWA CEO, John Morgan, described it as “One of the most significant investments in science in this country,” saying it would enable NIWA to provide better information on hugely important issues such as how climate change will affect New Zealand.
“The ability of the new supercomputers to process vast amounts of data in very short spaces of time will also enable us to build more precise forecasting tools to help farmers and environmental managers make more informed decisions using the best information available,” he said.
“Industries that are weather-sensitive—such as the energy sector, farming, horticulture and tourism—will benefit directly from NIWA’s ability to make more accurate and more specific forecasts. The new supercomputers will also allow NIWA to improve early warnings of the effects of severe events, such as flooding and storm surge.”
NeSI director Nick Jones said: “We are propelling our research into a new world where big compute and big data are going to impact many aspects of our lives and through this investment into a high-performance computing platform, New Zealand research will benefit across the board and allow us to compete with the rest of the world.
“This investment will allow our researchers to approach larger scale and more complex questions through the power of computational and analytical techniques and that is vital in being able to respond to some of our most difficult research challenges.”